Skip to content

Build your personal knowledge base with Trilium Notes

docs(user): sync

Elian Doranc198706d04ce
diff --git a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Quick search.html b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Quick search.html
index 2f516189db..285c60930c 100644
--- a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Quick search.html
+++ b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Quick search.html
@@ -66,80 +66,86 @@
<li><strong>Match locations</strong>: Bold text indicates where matches occur</li>
</ol>
<h2>Quick Search - Exact Match Operator</h2>
-<p>Quick Search now supports the exact match operator (<code spellcheck="false">=</code>)
- at the beginning of your search query. This allows you to search for notes
- where the title or content exactly matches your search term, rather than
- just containing it.</p>
-<h3>Usage</h3>
-<p>To use exact match in Quick Search:</p>
-<ol>
- <li>Start your search query with the <code spellcheck="false">=</code> operator</li>
- <li>Follow it immediately with your search term (no space after <code spellcheck="false">=</code>)</li>
-</ol>
-<h4>Examples</h4>
-<ul>
- <li><code spellcheck="false">=example</code> - Finds notes with title exactly
- "example" or content exactly "example"</li>
- <li><code spellcheck="false">=Project Plan</code> - Finds notes with title
- exactly "Project Plan" or content exactly "Project Plan"</li>
- <li><code spellcheck="false">='hello world'</code> - Use quotes for multi-word
- exact matches</li>
-</ul>
-<h4>Comparison with Regular Search</h4>
+<p>Quick Search shares the same search engine as the full <a class="reference-link"
+ href="#root/_help_eIg8jdvaoNNd">Search</a>, so the exact match operator (
+ <code
+ spellcheck="false">=</code>) behaves identically in both. Start your query with <code spellcheck="false">=</code> (no
+ space after it) to switch from the default "contains" behavior to exact
+ whole-word or phrase matching.</p>
+<p><strong>What <code spellcheck="false">=</code> actually does:</strong> it
+ finds notes where the title or content contains your term as a <strong>whole word or phrase</strong>,
+ ignoring surrounding punctuation. It does <strong>not</strong> require the
+ whole note to equal your term, and it does <strong>not</strong> do substring
+ or fuzzy matching.</p>
<figure class="table">
<table>
<thead>
<tr>
<th>Query</th>
- <th>Behavior</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ <th>Why</th>
</tr>
</thead>
<tbody>
<tr>
- <td><code spellcheck="false">example</code>
+ <td><code spellcheck="false">sync</code>
+ </td>
+ <td><code spellcheck="false">synchronize the database now</code>
+ </td>
+ <td>Yes</td>
+ <td>default search matches the substring</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">=sync</code>
</td>
- <td>Finds all notes containing "example" anywhere in title or content</td>
+ <td><code spellcheck="false">see (sync) mode</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">=</code> matches the whole word <code spellcheck="false">sync</code>,
+ punctuation ignored</td>
</tr>
<tr>
- <td><code spellcheck="false">=example</code>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">synchronize the database now</code>
</td>
- <td>Finds only notes where the title equals "example" or content equals "example"
- exactly</td>
+ <td>No</td>
+ <td><code spellcheck="false">=</code> never matches substrings</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">="project plan"</code>
+ </td>
+ <td><code spellcheck="false">the (project plan) is ready to share</code>
+ </td>
+ <td>Yes</td>
+ <td>quote a multi-word phrase to match it exactly</td>
</tr>
</tbody>
</table>
</figure>
-<h3>Technical Details</h3>
-<p>When you use the <code spellcheck="false">=</code> operator:</p>
-<ul>
- <li>The search performs an exact match on note titles</li>
- <li>For note content, it looks for exact matches of the entire content</li>
- <li>Partial word matches are excluded</li>
- <li>The search is case-insensitive</li>
-</ul>
+<p>The search is case- and diacritic-insensitive. For the complete explanation
+ of the three matching modes, fuzzy operators and relevance ranking, see
+ <a
+ href="#root/_help_eIg8jdvaoNNd">How search matches your text</a>in the full Search documentation.</p>
<h3>Limitations</h3>
<ul>
<li>The <code spellcheck="false">=</code> operator must be at the very beginning
- of the search query</li>
- <li>Spaces after <code spellcheck="false">=</code> will treat it as a regular
- search</li>
+ of the search query.</li>
+ <li>A space immediately after <code spellcheck="false">=</code> is treated as
+ a regular search.</li>
<li>Multiple <code spellcheck="false">=</code> operators (like <code spellcheck="false">==example</code>)
- are treated as regular text search</li>
-</ul>
-<h3>Use Cases</h3>
-<p>This feature is particularly useful when:</p>
-<ul>
- <li>You know the exact title of a note</li>
- <li>You want to find notes with specific, complete content</li>
- <li>You need to distinguish between notes with similar but not identical titles</li>
- <li>You want to avoid false positives from partial matches</li>
+ are treated as regular text search.</li>
</ul>
<h3>Related Features</h3>
<ul>
- <li>For more complex exact matching queries, use the full <a href="#root/_help_eIg8jdvaoNNd">Search</a> functionality</li>
- <li>For fuzzy matching (finding results despite typos), use the <code spellcheck="false">~=</code> operator
- in the full search</li>
+ <li>For attribute, property, boolean and ordering queries, use the full
+ <a
+ href="#root/_help_eIg8jdvaoNNd">Search</a>functionality.</li>
+ <li>For fuzzy matching (finding results despite typos), use the <code spellcheck="false">~=</code> or
+ <code
+ spellcheck="false">~*</code>operators in the full search.</li>
<li>For partial matches with wildcards, use operators like <code spellcheck="false">*=*</code>,
<code
- spellcheck="false">=*</code>, or <code spellcheck="false">*=</code> in the full search</li>
+ spellcheck="false">=*</code>, or <code spellcheck="false">*=</code> in the full search.</li>
</ul>
\ No newline at end of file
diff --git a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Search.html b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Search.html
index e17d69ac03..b201b2eded 100644
--- a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Search.html
+++ b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Navigation/Search.html
@@ -116,256 +116,816 @@
</ol>
</li>
</ol>
-<h3>Simple Note Search Examples</h3>
+<h2>Viewing search results</h2>
+<p>Results appear below the search pane as a list of <strong>snippet cards</strong>.
+ Each card shows the note title and a short excerpt of the text that matched,
+ with your search terms highlighted.</p>
<ul>
- <li><code spellcheck="false">rings tolkien</code>: Full-text search to find
- notes containing both "rings" and "tolkien".</li>
- <li><code spellcheck="false">"The Lord of the Rings" Tolkien</code>: Full-text
- search where "The Lord of the Rings" must match exactly.</li>
- <li><code spellcheck="false">note.content *=* rings OR note.content *=* tolkien</code>:
- Find notes containing "rings" or "tolkien" in their content.</li>
- <li><code spellcheck="false">towers #book</code>: Combine full-text and attribute
- search to find notes containing "towers" and having the "book" label.</li>
- <li><code spellcheck="false">towers #book or #author</code>: Search for notes
- containing "towers" and having either the "book" or "author" label.</li>
- <li><code spellcheck="false">towers #!book</code>: Search for notes containing
- "towers" and not having the "book" label.</li>
- <li><code spellcheck="false">#book #publicationYear = 1954</code>: Find notes
- with the "book" label and "publicationYear" set to 1954.</li>
- <li><code spellcheck="false">#genre *=* fan</code>: Find notes with the "genre"
- label containing the substring "fan". Additional operators include
- <code
- spellcheck="false">*=*</code>for "contains", <code spellcheck="false">=*</code> for "starts
- with", <code spellcheck="false">*=</code> for "ends with", and <code spellcheck="false">!=</code> for
- "is not equal to".</li>
- <li><code spellcheck="false">#book #publicationYear &gt;= 1950 #publicationYear &lt; 1960</code>:
- Use numeric operators to find all books published in the 1950s.</li>
- <li><code spellcheck="false">#dateNote &gt;= TODAY-30</code>: Find notes with
- the "dateNote" label within the last 30 days. Supported date values include
- NOW +- seconds, TODAY +- days, MONTH +- months, YEAR +- years.</li>
- <li><code spellcheck="false">~author.title *=* Tolkien</code>: Find notes
- related to an author whose title contains "Tolkien".</li>
- <li><code spellcheck="false">#publicationYear %= '19[0-9]{2}'</code>: Use
- the '%=' operator to match a regular expression (regex). This feature has
- been available since Trilium 0.52.</li>
- <li><code spellcheck="false">note.content %= '\\d{2}:\\d{2} (PM|AM)'</code>:
- Find notes that mention a time. Backslashes in a regex must be escaped.</li>
-</ul>
-<h3>Fuzzy Search</h3>
-<p>Trilium supports fuzzy search operators that find results with typos or
- spelling variations:</p>
-<ul>
- <li><code spellcheck="false">#title ~= trilim</code>: Fuzzy exact match -
- finds notes with titles like "Trilium" even if you typed "trilim" (with
- typo)</li>
- <li><code spellcheck="false">#content ~* progra</code>: Fuzzy contains match
- - finds notes containing words like "program", "programmer", "programming"
- even with slight misspellings</li>
- <li><code spellcheck="false">note.content ~* develpment</code>: Will find
- notes containing "development" despite the typo</li>
+ <li>The <strong>total number of results</strong> is always shown, so you can
+ immediately tell how broad a query is.</li>
+ <li>A <strong>page-size selector</strong> lets you choose how many results to
+ display per page. Your choice is remembered and synced across your devices
+ (stored in the <code spellcheck="false">searchResultsPageSize</code> option),
+ so you do not have to reset it on every device.</li>
+ <li><strong>Clicking a result</strong> opens the note and jumps straight to
+ the first match. The in-note find bar opens pre-filled with your search
+ terms, so you can step through the remaining matches with the find controls.</li>
+ <li>If a match is inside a <strong>collapsed section</strong> (for example a
+ folded heading), that section is expanded automatically so the match is
+ visible.</li>
</ul>
-<p><strong>Important notes about fuzzy search:</strong>
-</p>
-<ul>
- <li>Fuzzy search requires at least 3 characters in the search term</li>
- <li>Maximum edit distance is 2 characters (number of character changes needed)</li>
- <li>Diacritics are normalized (e.g., "café" matches "cafe")</li>
- <li>Fuzzy matches work best for finding content with minor typos or spelling
- variations</li>
-</ul>
-<h3>Advanced Use Cases</h3>
-<ul>
- <li><code spellcheck="false">~author.relations.son.title = 'Christopher Tolkien'</code>:
- Search for notes with an "author" relation to a note that has a "son" relation
- to "Christopher Tolkien". This can be modeled with the following note structure:
- <ul>
- <li>Books
- <ul>
- <li>Lord of the Rings
- <ul>
- <li>label: “book”</li>
- <li>relation: “author” points to “J. R. R. Tolkien” note</li>
- </ul>
- </li>
- </ul>
- </li>
- <li>People
- <ul>
- <li>J. R. R. Tolkien
- <ul>
- <li>relation: “son” points to "Christopher Tolkien" note</li>
- <li>Christopher Tolkien</li>
- </ul>
- </li>
- </ul>
- </li>
- </ul>
- </li>
- <li><code spellcheck="false">~author.title *= Tolkien OR (#publicationDate &gt;= 1954 AND #publicationDate &lt;= 1960)</code>:
- Use boolean expressions and parentheses to group expressions. Note that
- expressions starting with a parenthesis need an "expression separator sign"
- (# or ~) prepended.</li>
- <li><code spellcheck="false">note.parents.title = 'Books'</code>: Find notes
- with a parent named "Books".</li>
- <li><code spellcheck="false">note.parents.parents.title = 'Books'</code>:
- Find notes with a grandparent named "Books".</li>
- <li><code spellcheck="false">note.ancestors.title = 'Books'</code>: Find notes
- with an ancestor named "Books".</li>
- <li><code spellcheck="false">note.children.title = 'sub-note'</code>: Find
- notes with a child named "sub-note".</li>
-</ul>
-<h3>Search with Note Properties</h3>
-<p>Notes have properties that can be used in searches, such as <code spellcheck="false">noteId</code>,
- <code
- spellcheck="false">dateModified</code>, <code spellcheck="false">dateCreated</code>,
- <code
- spellcheck="false">isProtected</code>, <code spellcheck="false">type</code>, <code spellcheck="false">title</code>,
- <code
- spellcheck="false">text</code>, <code spellcheck="false">content</code>, <code spellcheck="false">rawContent</code>,
- <code
- spellcheck="false">ownedLabelCount</code>, <code spellcheck="false">labelCount</code>,
+<h2>How search matches your text</h2>
+<p>The single most common source of confusion is that the <code spellcheck="false">=</code> sign
+ means two <em>different</em> things depending on where it appears. Read this
+ section once and the rest of search becomes predictable.</p>
+<p>There are three matching modes:</p>
+<figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Mode</th>
+ <th>How you trigger it</th>
+ <th>What it matches</th>
+ <th>Substring?</th>
+ <th>Fuzzy (typos)?</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><strong>Default</strong>
+ </td>
+ <td>type words with no prefix</td>
+ <td>whole words <em>and</em> substrings, anywhere in title, content or attributes,
+ relevance-ranked</td>
+ <td>Yes</td>
+ <td>Yes</td>
+ </tr>
+ <tr>
+ <td><strong>Exact full-text</strong>
+ </td>
+ <td>a leading <code spellcheck="false">=</code> (e.g. <code spellcheck="false">=sync</code>)</td>
+ <td>the exact whole word or phrase, ignoring surrounding punctuation</td>
+ <td>No</td>
+ <td>No</td>
+ </tr>
+ <tr>
+ <td><strong>Attribute / property equality</strong>
+ </td>
+ <td><code spellcheck="false">=</code> inside a <code spellcheck="false">#label=value</code> or
<code
- spellcheck="false">ownedRelationCount</code>, <code spellcheck="false">relationCount</code>,
- <code
- spellcheck="false">ownedRelationCountIncludingLinks</code>, <code spellcheck="false">relationCountIncludingLinks</code>,
- <code
- spellcheck="false">ownedAttributeCount</code>, <code spellcheck="false">attributeCount</code>,
- <code
- spellcheck="false">targetRelationCount</code>, <code spellcheck="false">targetRelationCountIncludingLinks</code>,
- <code
- spellcheck="false">parentCount</code>, <code spellcheck="false">childrenCount</code>,
- <code
- spellcheck="false">isArchived</code>, <code spellcheck="false">contentSize</code>, <code spellcheck="false">noteSize</code>,
- and <code spellcheck="false">revisionCount</code>.</p>
-<p>These properties can be accessed via the <code spellcheck="false">note.</code> prefix,
- e.g., <code spellcheck="false">note.type = code AND note.mime = 'application/json'</code>.</p>
-<h3>Order by and Limit</h3><pre><code class="language-text-x-trilium-auto">#author=Tolkien orderBy #publicationDate desc, note.title limit 10</code></pre>
-<p>This example will:</p>
-<ol>
- <li>Find notes with the author label "Tolkien".</li>
- <li>Order the results by <code spellcheck="false">publicationDate</code> in
- descending order.</li>
- <li>Use <code spellcheck="false">note.title</code> as a secondary ordering if
- publication dates are equal.</li>
- <li>Limit the results to the first 10 notes.</li>
-</ol>
-<h3>Negation</h3>
-<p>Some queries can only be expressed with negation:</p><pre><code class="language-text-x-trilium-auto">#book AND not(note.ancestor.title = 'Tolkien')</code></pre>
-<p>This query finds all book notes not in the "Tolkien" subtree.</p>
-<h2>Progressive Search Strategy</h2>
-<p>Trilium uses a progressive search strategy that performs exact matching
- first, then adds fuzzy matching when needed.</p>
-<h3>How Progressive Search Works</h3>
-<ol>
- <li><strong>Phase 1 - Exact Matching</strong>: When you search, Trilium first
- looks for exact matches of your search terms. This handles the vast majority
- of searches (90%+) and returns results almost instantly.</li>
- <li><strong>Phase 2 - Fuzzy Fallback</strong>: If Phase 1 doesn't find enough
- high-quality results (fewer than 5 results with good relevance scores),
- Trilium automatically adds fuzzy matching to find results with typos or
- spelling variations.</li>
- <li><strong>Result Ordering</strong>: Exact matches always appear before fuzzy
- matches, regardless of individual scores. This ensures that when you search
- for "project", notes containing the exact word "project" will appear before
- notes containing similar words like "projects" or "projection".</li>
-</ol>
-<h3>Progressive Search Behavior</h3>
-<ul>
- <li><strong>Speed</strong>: Most searches complete using only exact matching</li>
- <li><strong>Ordering</strong>: Exact matches appear before fuzzy matches</li>
- <li><strong>Fallback</strong>: Fuzzy matching activates when exact matches
- return fewer than 5 results</li>
- <li><strong>Identification</strong>: Results indicate whether they are exact
- or fuzzy matches</li>
-</ul>
-<h3>Search Performance</h3>
-<p>Search system specifications:</p>
-<ul>
- <li>Content size limit: 10MB per note (previously 50KB)</li>
- <li>Edit distance calculations for fuzzy matching</li>
- <li>Infinite scrolling in Quick Search</li>
-</ul>
-<h2>Under the Hood</h2>
-<h3>Label and Relation Shortcuts</h3>
-<p>The "full" syntax for searching by labels is:</p><pre><code class="language-text-x-trilium-auto">note.labels.publicationYear = 1954</code></pre>
-<p>For relations:</p><pre><code class="language-text-x-trilium-auto">note.relations.author.title *=* Tolkien</code></pre>
-<p>However, common label and relation searches have shortcut syntax:</p><pre><code class="language-text-x-trilium-auto">#publicationYear = 1954
-~author.title *=* Tolkien</code></pre>
-<h3>Separating Full-Text and Attribute Parts</h3>
-<p>Search syntax allows combining full-text search with attribute-based search.
- For example, <code spellcheck="false">tolkien #book</code> contains:</p>
-<ol>
- <li>Full-text tokens - <code spellcheck="false">tolkien</code>
- </li>
- <li>Attribute expressions - <code spellcheck="false">#book</code>
- </li>
-</ol>
-<p>Trilium detects the separation between full text search and attribute/property
- search by looking for certain special characters or words that denote attributes
- and properties (e.g., #, ~, note.). If you need to include these in full-text
- search, escape them with a backslash so they are processed as regular text:</p><pre><code class="language-text-x-trilium-auto">"note.txt"
-\#hash
-#myLabel = 'Say "Hello World"'</code></pre>
-<h3>Escaping Special Characters</h3>
-<p>Special characters can be enclosed in quotes or escaped with a backslash
- to be used in full-text search:</p><pre><code class="language-text-x-trilium-auto">"note.txt"
-\#hash
-#myLabel = 'Say "Hello World"'</code></pre>
-<p>Three types of quotes are supported: single, double, and backtick.</p>
-<h3>Type Coercion</h3>
-<p>Label values are technically strings but can be coerced for numeric comparisons:</p><pre><code class="language-text-x-trilium-auto">note.dateCreated =* '2019-05'</code></pre>
-<p>This finds notes created in May 2019. Numeric operators like <code spellcheck="false">#publicationYear &gt;= 1960</code> convert
- string values to numbers for comparison.</p>
-<h2>Auto-Trigger Search from URL</h2>
-<p>You can open Trilium and automatically trigger a search by including the
- search <a href="https://meyerweb.com/eric/tools/dencoder/">url encoded</a> string
- in the URL:</p>
-<p><code spellcheck="false">http://localhost:8080/#?searchString=abc</code>
-</p>
-<h2>Search Configuration</h2>
-<h3>Parameters</h3>
+ spellcheck="false">note.property=value</code>clause</td>
+ <td>the <em>entire</em> attribute or property value, exactly</td>
+ <td>No</td>
+ <td>No</td>
+ </tr>
+ </tbody>
+ </table>
+</figure>
+<p>Every example below is backed 1:1 by an automated test, so the documentation
+ cannot drift from the engine.</p>
+<h3>Default matching (no prefix)</h3>
+<p><strong>Rule:</strong> typing words with no prefix finds notes that contain
+ those words anywhere in the title, content or attributes, as whole words
+ or as substrings. The closest matches are ranked first.</p>
<figure class="table">
<table>
<thead>
<tr>
- <th>Parameter</th>
- <th>Value</th>
- <th>Description</th>
+ <th>Query</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ <th>Why</th>
</tr>
</thead>
<tbody>
<tr>
- <td><code spellcheck="false">MIN_FUZZY_TOKEN_LENGTH</code>
+ <td><code spellcheck="false">sync</code>
+ </td>
+ <td><code spellcheck="false">please sync the folders</code>
+ </td>
+ <td>Yes (ranked higher)</td>
+ <td>contains the exact word <code spellcheck="false">sync</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">sync</code>
+ </td>
+ <td><code spellcheck="false">synchronize the database now</code>
+ </td>
+ <td>Yes (ranked lower)</td>
+ <td><code spellcheck="false">sync</code> is a substring of <code spellcheck="false">synchronize</code>
+ </td>
+ </tr>
+ </tbody>
+ </table>
+</figure>
+<h3>Exact match with the <code spellcheck="false">=</code> prefix</h3>
+<p><strong>Rule:</strong> a leading <code spellcheck="false">=</code> switches
+ the full-text search to exact matching. It finds the whole word or phrase
+ only, ignoring surrounding punctuation, with <strong>no</strong> substring
+ and <strong>no</strong> fuzzy matching. Use it when a normal search returns
+ too many near-matches.</p>
+<figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">see (sync) mode</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">(sync)</code> is the whole word <code spellcheck="false">sync</code>;
+ punctuation is ignored</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">in sync, then continue</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">sync,</code> is the whole word <code spellcheck="false">sync</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">he said "sync" out loud</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">"sync"</code> is the whole word <code spellcheck="false">sync</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">synchronize the database now</code>
</td>
- <td>3</td>
- <td>Minimum characters for fuzzy matching</td>
+ <td>No</td>
+ <td><code spellcheck="false">=</code> never matches substrings</td>
</tr>
<tr>
- <td><code spellcheck="false">MAX_EDIT_DISTANCE</code>
+ <td><code spellcheck="false">=sync</code>
+ </td>
+ <td><code spellcheck="false">please send the file</code>
</td>
- <td>2</td>
- <td>Maximum character changes allowed</td>
+ <td>No</td>
+ <td><code spellcheck="false">=</code> never matches typos / fuzzy</td>
</tr>
+ </tbody>
+ </table>
+</figure>
+<p>To match an exact <strong>phrase</strong>, quote it after the <code spellcheck="false">=</code> (single,
+ double or backtick quotes all work). The phrase must appear as consecutive
+ words, but punctuation between or around them is ignored:</p>
+<figure class="table">
+ <table>
+ <thead>
<tr>
- <td><code spellcheck="false">RESULT_SUFFICIENCY_THRESHOLD</code>
+ <th>Query</th>
+ <th>Example note (title or content)</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">="project plan"</code>
+ </td>
+ <td>title <code spellcheck="false">Project Plan</code>
</td>
- <td>5</td>
- <td>Minimum exact results before fuzzy fallback</td>
+ <td>Yes</td>
+ <td>the title is exactly the phrase</td>
</tr>
<tr>
- <td><code spellcheck="false">MAX_CONTENT_SIZE</code>
+ <td><code spellcheck="false">="project plan"</code>
+ </td>
+ <td><code spellcheck="false">the (project plan) is ready to share</code>
</td>
- <td>10MB</td>
- <td>Maximum note content size for search processing</td>
+ <td>Yes</td>
+ <td>the consecutive phrase appears; punctuation is ignored</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">="project plan"</code>
+ </td>
+ <td><code spellcheck="false">the plan for this project is late</code>
+ </td>
+ <td>No</td>
+ <td>the words are present but not consecutive</td>
</tr>
</tbody>
</table>
</figure>
-<h3>Limits</h3>
-<ul>
- <li>Searched note content is limited to 10MB per note to prevent performance
- issues</li>
- <li>Notes exceeding this limit will still be included in title and attribute
- searches</li>
- <li>Fuzzy matching requires tokens of at least 3 characters</li>
-</ul>
\ No newline at end of file
+<h3>Attribute and property equality (<code spellcheck="false">=</code>, <code spellcheck="false">!=</code>)</h3>
+<p><strong>Rule:</strong> when <code spellcheck="false">=</code> compares an
+ attribute or property, as in <code spellcheck="false">#label=value</code> or
+ <code
+ spellcheck="false">note.title=value</code>, it is <strong>strict full-value equality</strong>:
+ the <em>whole</em> value must equal what you typed, ignoring case and diacritics.
+ This is <strong>not</strong> word matching. <code spellcheck="false">!=</code> inverts
+ it.</p>
+<p>The examples below assume four notes: <em>Austria</em> (<code spellcheck="false">#capital=Vienna</code>), <em>Somewhere</em> (
+ <code
+ spellcheck="false">#capital=Vienna Austria</code>), <em>Czech Republic</em> (<code spellcheck="false">#capital=Prague</code>)
+ and <em>Switzerland</em> (<code spellcheck="false">#capital=Zürich</code>).</p>
+<figure
+class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Example label value</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">#capital=Vienna</code>
+ </td>
+ <td><code spellcheck="false">Vienna</code>
+ </td>
+ <td>Yes</td>
+ <td>the whole value equals <code spellcheck="false">Vienna</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#capital=Vienna</code>
+ </td>
+ <td><code spellcheck="false">Vienna Austria</code>
+ </td>
+ <td>No</td>
+ <td>the whole value is <code spellcheck="false">Vienna Austria</code>, not
+ <code
+ spellcheck="false">Vienna</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#capital="Vienna Austria"</code>
+ </td>
+ <td><code spellcheck="false">Vienna Austria</code>
+ </td>
+ <td>Yes</td>
+ <td>quote a multi-word value to match it in full</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#capital=Zurich</code>
+ </td>
+ <td><code spellcheck="false">Zürich</code>
+ </td>
+ <td>Yes</td>
+ <td>equality ignores diacritics</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#capital!=Vienna</code>
+ </td>
+ <td><code spellcheck="false">Prague</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">!=</code> matches every value that is not
+ <code
+ spellcheck="false">Vienna</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#capital!=Vienna</code>
+ </td>
+ <td><code spellcheck="false">Vienna</code>
+ </td>
+ <td>No</td>
+ <td><code spellcheck="false">!=</code> excludes the exact value</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <blockquote>
+ <p><strong>Quick search relaxes this.</strong> The <a href="#root/_help_Ms1nauBra7gq">Quick search</a> bar
+ and autocomplete treat an attribute <code spellcheck="false">=</code> as
+ "contains", so <code spellcheck="false">#capital=Vienna</code> typed there
+ also matches <code spellcheck="false">Vienna Austria</code>. The strict
+ full-value equality described here applies only in the full Search.</p>
+ </blockquote>
+ <h3>Fuzzy operators (<code spellcheck="false">~=</code> and <code spellcheck="false">~*</code>)</h3>
+ <p><strong>Rule:</strong> the fuzzy operators tolerate typos. <code spellcheck="false">~=</code> (fuzzy-equals)
+ matches a value that is a close whole-word variant of your term. <code spellcheck="false">~*</code> (fuzzy-contains)
+ matches when your term appears anywhere inside the value, either as a fragment
+ or as a near-miss. Both work on note properties such as <code spellcheck="false">note.title</code> and
+ <code
+ spellcheck="false">note.content</code>, and on labels (<code spellcheck="false">#label</code>).
+ Fuzzy operators require at least 3 characters.</p>
+ <p>The examples assume a note titled <code spellcheck="false">Books</code> carrying
+ the label <code spellcheck="false">#author=Tolkien</code>, and a note whose
+ content is <code spellcheck="false">learn programming today</code>.</p>
+ <figure
+ class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Example value</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">note.title ~= boks</code>
+ </td>
+ <td>title <code spellcheck="false">Books</code>
+ </td>
+ <td>Yes</td>
+ <td>one edit away from <code spellcheck="false">books</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">#author ~= tolkein</code>
+ </td>
+ <td>author <code spellcheck="false">Tolkien</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">tolkein</code> is a typo of <code spellcheck="false">tolkien</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">note.content ~* progr</code>
+ </td>
+ <td><code spellcheck="false">learn programming today</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">progr</code> is a fragment of <code spellcheck="false">programming</code>
+ </td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">note.content ~* programing</code>
+ </td>
+ <td><code spellcheck="false">learn programming today</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">programing</code> is one edit from <code spellcheck="false">programming</code>
+ </td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Fuzzy tolerance (AUTO)</h3>
+ <p><strong>Rule:</strong> how many typos are tolerated depends on the <strong>length</strong> of
+ your search term (the "AUTO" scheme popularized by Elasticsearch). Short
+ terms must match almost exactly to avoid noise; longer terms tolerate more.</p>
+ <figure
+ class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Term length</th>
+ <th>Edits allowed</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td>1–2 characters</td>
+ <td>0 (exact only)</td>
+ </tr>
+ <tr>
+ <td>3–5 characters</td>
+ <td>1</td>
+ </tr>
+ <tr>
+ <td>6+ characters</td>
+ <td>2</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Term length</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">cat</code>
+ </td>
+ <td>3</td>
+ <td><code spellcheck="false">a bright red car</code>
+ </td>
+ <td>Yes</td>
+ <td>1 edit is within budget for 3–5 character terms</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">ceck</code>
+ </td>
+ <td>4</td>
+ <td><code spellcheck="false">the latest tech trends</code>
+ </td>
+ <td>No</td>
+ <td><code spellcheck="false">ceck</code>→<code spellcheck="false">tech</code> needs
+ 2 edits; only 1 is allowed at this length</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">combinef</code>
+ </td>
+ <td>8</td>
+ <td><code spellcheck="false">the values were combined together</code>
+ </td>
+ <td>Yes</td>
+ <td><code spellcheck="false">combinef</code>→<code spellcheck="false">combined</code> is
+ 1 edit; up to 2 are allowed at 6+ characters</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Relevance ranking</h3>
+ <p><strong>Rule:</strong> results are ordered by <em>how well</em> they match,
+ not merely whether they match. Exact whole-word and phrase matches rank
+ above substring and fuzzy matches, and a note where your words appear as
+ a consecutive phrase outranks one where they are scattered.</p>
+ <figure
+ class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Ranked higher</th>
+ <th>Ranked lower</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">sync</code>
+ </td>
+ <td><code spellcheck="false">please sync the folders</code>
+ </td>
+ <td><code spellcheck="false">synchronize the database now</code>
+ </td>
+ <td>exact word beats substring</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">you and me</code>
+ </td>
+ <td><code spellcheck="false">I like you and me as a phrase</code>
+ </td>
+ <td><code spellcheck="false">the menu is here and you know it</code>
+ </td>
+ <td>consecutive phrase beats scattered words</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>What is searchable</h3>
+ <p>A search looks at more than the visible body text. All of the following
+ are indexed and searchable:</p>
+ <ul>
+ <li>note titles,</li>
+ <li>note body content,</li>
+ <li>labels and relations (<a href="#root/_help_zEY4DaJG4YT5">attributes</a>),</li>
+ <li>link URLs and the titles/descriptions of link previews,</li>
+ <li>the titles of notes that a note reference-links to.</li>
+ </ul>
+ <p>The last point is the least obvious: if a note's only content is a reference
+ link to another note, searching for that other note's <strong>title</strong> still
+ finds the linking note.</p>
+ <figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Setup</th>
+ <th>Matches?</th>
+ <th>Why</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">special topic</code>
+ </td>
+ <td>a note whose only content is a reference link to a note titled <code spellcheck="false">Special Topic</code>
+ </td>
+ <td>Yes, and the linked target ranks first</td>
+ <td>the target's title is indexed into the linking note's searchable text</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">special</code>
+ </td>
+ <td>the same note (only a reference link to <code spellcheck="false">Special Topic</code>)</td>
+ <td>Yes, one word from the target's title is enough</td>
+ <td>the indexed title is normalized just like body text, so even one lowercased
+ word matches</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">zurich</code>
+ </td>
+ <td>a note whose only content is a reference link to a note titled <code spellcheck="false">Zürich</code>
+ </td>
+ <td>Yes</td>
+ <td>the indexed title has its accents normalized, so the plain form matches
+ the accented title</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Diacritics</h3>
+ <p><strong>Rule:</strong> accents are normalized on both sides, so an accented
+ word and its plain form match each other.</p>
+ <figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">ktory</code>
+ </td>
+ <td><code spellcheck="false">slovo ktorý znamena nieco</code>
+ </td>
+ <td>Yes</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">ktorý</code>
+ </td>
+ <td><code spellcheck="false">the word ktory appears here</code>
+ </td>
+ <td>Yes</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Regular expressions (<code spellcheck="false">%=</code>)</h3>
+ <p><strong>Rule:</strong> the <code spellcheck="false">%=</code> operator matches
+ a property or label value against a regular expression.</p>
+ <figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Query</th>
+ <th>Example note content</th>
+ <th>Matches?</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">note.content %= 'colou?r'</code>
+ </td>
+ <td><code spellcheck="false">my favorite color of all</code>
+ </td>
+ <td>Yes</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">note.content %= 'colou?r'</code>
+ </td>
+ <td><code spellcheck="false">my favourite colour of all</code>
+ </td>
+ <td>Yes</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Simple Note Search Examples</h3>
+ <ul>
+ <li><code spellcheck="false">rings tolkien</code>: Full-text search to find
+ notes containing both "rings" and "tolkien".</li>
+ <li><code spellcheck="false">"The Lord of the Rings" Tolkien</code>: Full-text
+ search where "The Lord of the Rings" must match exactly.</li>
+ <li><code spellcheck="false">note.content *=* rings OR note.content *=* tolkien</code>:
+ Find notes containing "rings" or "tolkien" in their content.</li>
+ <li><code spellcheck="false">towers #book</code>: Combine full-text and attribute
+ search to find notes containing "towers" and having the "book" label.</li>
+ <li><code spellcheck="false">towers #book or #author</code>: Search for notes
+ containing "towers" and having either the "book" or "author" label.</li>
+ <li><code spellcheck="false">towers #!book</code>: Search for notes containing
+ "towers" and not having the "book" label.</li>
+ <li><code spellcheck="false">#book #publicationYear = 1954</code>: Find notes
+ with the "book" label and "publicationYear" set to 1954.</li>
+ <li><code spellcheck="false">#genre *=* fan</code>: Find notes with the "genre"
+ label containing the substring "fan". Additional operators include
+ <code
+ spellcheck="false">*=*</code>for "contains", <code spellcheck="false">=*</code> for "starts
+ with", <code spellcheck="false">*=</code> for "ends with", and <code spellcheck="false">!=</code> for
+ "is not equal to".</li>
+ <li><code spellcheck="false">#book #publicationYear &gt;= 1950 #publicationYear &lt; 1960</code>:
+ Use numeric operators to find all books published in the 1950s.</li>
+ <li><code spellcheck="false">#dateNote &gt;= TODAY-30</code>: Find notes with
+ the "dateNote" label within the last 30 days. Supported date values include
+ NOW +- seconds, TODAY +- days, MONTH +- months, YEAR +- years.</li>
+ <li><code spellcheck="false">~author.title *=* Tolkien</code>: Find notes
+ related to an author whose title contains "Tolkien".</li>
+ <li><code spellcheck="false">#publicationYear %= '19[0-9]{2}'</code>: Use
+ the '%=' operator to match a regular expression (regex). This feature has
+ been available since Trilium 0.52.</li>
+ <li><code spellcheck="false">note.content %= '\\d{2}:\\d{2} (PM|AM)'</code>:
+ Find notes that mention a time. Backslashes in a regex must be escaped.</li>
+ </ul>
+ <h3>Advanced Use Cases</h3>
+ <ul>
+ <li><code spellcheck="false">~author.relations.son.title = 'Christopher Tolkien'</code>:
+ Search for notes with an "author" relation to a note that has a "son" relation
+ to "Christopher Tolkien". This can be modeled with the following note structure:
+ <ul>
+ <li>Books
+ <ul>
+ <li>Lord of the Rings
+ <ul>
+ <li>label: “book”</li>
+ <li>relation: “author” points to “J. R. R. Tolkien” note</li>
+ </ul>
+ </li>
+ </ul>
+ </li>
+ <li>People
+ <ul>
+ <li>J. R. R. Tolkien
+ <ul>
+ <li>relation: “son” points to "Christopher Tolkien" note</li>
+ <li>Christopher Tolkien</li>
+ </ul>
+ </li>
+ </ul>
+ </li>
+ </ul>
+ </li>
+ <li><code spellcheck="false">~author.title *= Tolkien OR (#publicationDate &gt;= 1954 AND #publicationDate &lt;= 1960)</code>:
+ Use boolean expressions and parentheses to group expressions. Note that
+ expressions starting with a parenthesis need an "expression separator sign"
+ (# or ~) prepended.</li>
+ <li><code spellcheck="false">note.parents.title = 'Books'</code>: Find notes
+ with a parent named "Books".</li>
+ <li><code spellcheck="false">note.parents.parents.title = 'Books'</code>:
+ Find notes with a grandparent named "Books".</li>
+ <li><code spellcheck="false">note.ancestors.title = 'Books'</code>: Find notes
+ with an ancestor named "Books".</li>
+ <li><code spellcheck="false">note.children.title = 'sub-note'</code>: Find
+ notes with a child named "sub-note".</li>
+ </ul>
+ <h3>Search with Note Properties</h3>
+ <p>Notes have properties that can be used in searches, such as <code spellcheck="false">noteId</code>,
+ <code
+ spellcheck="false">dateModified</code>, <code spellcheck="false">dateCreated</code>,
+ <code
+ spellcheck="false">isProtected</code>, <code spellcheck="false">type</code>, <code spellcheck="false">title</code>,
+ <code
+ spellcheck="false">text</code>, <code spellcheck="false">content</code>, <code spellcheck="false">rawContent</code>,
+ <code
+ spellcheck="false">ownedLabelCount</code>, <code spellcheck="false">labelCount</code>,
+ <code
+ spellcheck="false">ownedRelationCount</code>, <code spellcheck="false">relationCount</code>,
+ <code
+ spellcheck="false">ownedRelationCountIncludingLinks</code>, <code spellcheck="false">relationCountIncludingLinks</code>,
+ <code
+ spellcheck="false">ownedAttributeCount</code>, <code spellcheck="false">attributeCount</code>,
+ <code
+ spellcheck="false">targetRelationCount</code>, <code spellcheck="false">targetRelationCountIncludingLinks</code>,
+ <code
+ spellcheck="false">parentCount</code>, <code spellcheck="false">childrenCount</code>,
+ <code
+ spellcheck="false">isArchived</code>, <code spellcheck="false">contentSize</code>, <code spellcheck="false">noteSize</code>,
+ and <code spellcheck="false">revisionCount</code>.</p>
+ <p>These properties can be accessed via the <code spellcheck="false">note.</code> prefix,
+ e.g., <code spellcheck="false">note.type = code AND note.mime = 'application/json'</code>.</p>
+ <h3>Order by and Limit</h3><pre><code class="language-text-x-trilium-auto">#author=Tolkien orderBy #publicationDate desc, note.title limit 10</code></pre>
+ <p>This example will:</p>
+ <ol>
+ <li>Find notes with the author label "Tolkien".</li>
+ <li>Order the results by <code spellcheck="false">publicationDate</code> in
+ descending order.</li>
+ <li>Use <code spellcheck="false">note.title</code> as a secondary ordering if
+ publication dates are equal.</li>
+ <li>Limit the results to the first 10 notes.</li>
+ </ol>
+ <h3>Negation</h3>
+ <p>Some queries can only be expressed with negation:</p><pre><code class="language-text-x-trilium-auto">#book AND not(note.ancestors.title = 'Tolkien')</code></pre>
+ <p>This query finds all book notes not in the "Tolkien" subtree.</p>
+ <h2>Progressive Search Strategy</h2>
+ <p>Trilium uses a progressive search strategy that performs exact matching
+ first, then adds fuzzy matching when needed.</p>
+ <h3>How Progressive Search Works</h3>
+ <ol>
+ <li><strong>Phase 1 - Exact Matching</strong>: When you search, Trilium first
+ looks for exact matches of your search terms. This handles the vast majority
+ of searches (90%+) and returns results almost instantly.</li>
+ <li><strong>Phase 2 - Fuzzy Fallback</strong>: If Phase 1 doesn't find enough
+ high-quality results (fewer than 5 results with good relevance scores),
+ Trilium automatically adds fuzzy matching to find results with typos or
+ spelling variations.</li>
+ <li><strong>Result Ordering</strong>: Exact matches always appear before fuzzy
+ matches, regardless of individual scores. This ensures that when you search
+ for "project", notes containing the exact word "project" will appear before
+ notes containing similar words like "projects" or "projection".</li>
+ </ol>
+ <h3>Progressive Search Behavior</h3>
+ <ul>
+ <li><strong>Speed</strong>: Most searches complete using only exact matching</li>
+ <li><strong>Ordering</strong>: Exact matches appear before fuzzy matches</li>
+ <li><strong>Fallback</strong>: Fuzzy matching activates when exact matches
+ return fewer than 5 results</li>
+ <li><strong>Identification</strong>: Results indicate whether they are exact
+ or fuzzy matches</li>
+ </ul>
+ <h3>Search Performance</h3>
+ <p>Search system specifications:</p>
+ <ul>
+ <li>Content size limit: 10MB per note (previously 50KB)</li>
+ <li>Edit distance calculations for fuzzy matching</li>
+ <li>Infinite scrolling in Quick Search</li>
+ </ul>
+ <h2>Under the Hood</h2>
+ <h3>Label and Relation Shortcuts</h3>
+ <p>The "full" syntax for searching by labels is:</p><pre><code class="language-text-x-trilium-auto">note.labels.publicationYear = 1954</code></pre>
+ <p>For relations:</p><pre><code class="language-text-x-trilium-auto">note.relations.author.title *=* Tolkien</code></pre>
+ <p>However, common label and relation searches have shortcut syntax:</p><pre><code class="language-text-x-trilium-auto">#publicationYear = 1954
+~author.title *=* Tolkien</code></pre>
+ <h3>Separating Full-Text and Attribute Parts</h3>
+ <p>Search syntax allows combining full-text search with attribute-based search.
+ For example, <code spellcheck="false">tolkien #book</code> contains:</p>
+ <ol>
+ <li>Full-text tokens - <code spellcheck="false">tolkien</code>
+ </li>
+ <li>Attribute expressions - <code spellcheck="false">#book</code>
+ </li>
+ </ol>
+ <p>Trilium detects the separation between full text search and attribute/property
+ search by looking for certain special characters or words that denote attributes
+ and properties (e.g., #, ~, note.). If you need to include these in full-text
+ search, escape them with a backslash so they are processed as regular text:</p><pre><code class="language-text-x-trilium-auto">"note.txt"
+\#hash
+#myLabel = 'Say "Hello World"'</code></pre>
+ <h3>Escaping Special Characters</h3>
+ <p>Special characters can be enclosed in quotes or escaped with a backslash
+ to be used in full-text search:</p><pre><code class="language-text-x-trilium-auto">"note.txt"
+\#hash
+#myLabel = 'Say "Hello World"'</code></pre>
+ <p>Three types of quotes are supported: single, double, and backtick.</p>
+ <h3>Type Coercion</h3>
+ <p>Label values are technically strings but can be coerced for numeric comparisons:</p><pre><code class="language-text-x-trilium-auto">note.dateCreated =* '2019-05'</code></pre>
+ <p>This finds notes created in May 2019. Numeric operators like <code spellcheck="false">#publicationYear &gt;= 1960</code> convert
+ string values to numbers for comparison.</p>
+ <h2>Auto-Trigger Search from URL</h2>
+ <p>You can open Trilium and automatically trigger a search by including the
+ search <a href="https://meyerweb.com/eric/tools/dencoder/">url encoded</a> string
+ in the URL:</p>
+ <p><code spellcheck="false">http://localhost:8080/#?searchString=abc</code>
+ </p>
+ <h2>Search Configuration</h2>
+ <h3>Parameters</h3>
+ <figure class="table">
+ <table>
+ <thead>
+ <tr>
+ <th>Parameter</th>
+ <th>Value</th>
+ <th>Description</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td><code spellcheck="false">MIN_FUZZY_TOKEN_LENGTH</code>
+ </td>
+ <td>3</td>
+ <td>Minimum characters for fuzzy matching</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">MAX_EDIT_DISTANCE</code>
+ </td>
+ <td>2</td>
+ <td>Ceiling on character changes, reached only by 6+ character terms; shorter
+ terms allow fewer. See the <em>Fuzzy tolerance (AUTO)</em> section above</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">RESULT_SUFFICIENCY_THRESHOLD</code>
+ </td>
+ <td>5</td>
+ <td>Minimum exact results before fuzzy fallback</td>
+ </tr>
+ <tr>
+ <td><code spellcheck="false">MAX_CONTENT_SIZE</code>
+ </td>
+ <td>10MB</td>
+ <td>Maximum note content size for search processing</td>
+ </tr>
+ </tbody>
+ </table>
+ </figure>
+ <h3>Limits</h3>
+ <ul>
+ <li>Searched note content is limited to 10MB per note to prevent performance
+ issues</li>
+ <li>Notes exceeding this limit will still be included in title and attribute
+ searches</li>
+ <li>Fuzzy matching requires tokens of at least 3 characters</li>
+ </ul>
\ No newline at end of file
diff --git a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Themes/Personalizing the font.html b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Themes/Personalizing the font.html
index b7e48caaa0..a402bcb77d 100644
--- a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Themes/Personalizing the font.html
+++ b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Basic Concepts and Features/Themes/Personalizing the font.html
@@ -17,28 +17,27 @@
them along with it.</p>
<p>The list of fonts contains the following items:</p>
<ul>
- <li class="ck-list-marker-bold"><strong>Generic fonts</strong>
+ <li><strong>Generic fonts</strong>
<ul>
<li><em>Theme defined</em> which uses the font that is embedded with the current
<a
href="#root/_help_Wy267RK4M69c">theme</a>. For example on the default modern theme, the font is <em>Inter</em>.
This does not require the font to be installed.</li>
<li><em>System default</em> which uses a combination of fonts that works best
- for the <a href="#root/_help_poXkQfguuA0U">desktop app</a>.
- For example on Windows it will use <em>Segoe UI</em>.</li>
+ for the <a href="#root/_help_poXkQfguuA0U">desktop app</a>. For example on Windows
+ it will use <em>Segoe UI</em>.</li>
<li>Generic font selections for fonts with serifs (e.g. <em>Times New Roman</em>-like),
no serifs and monospace.</li>
</ul>
</li>
<li><strong>Web fonts</strong> which appear on the&nbsp;<a class="reference-link"
- href="#root/_help_WOcw2SLH6tbX">Server Installation</a>,
- which is a predefined list of fonts and grouped into Sans-serif, serif,
- monospace and handwriting. The fonts that are not supported by your system
- (i.e. not installed) are not displayed.</li>
+ href="#root/_help_WOcw2SLH6tbX">Server Installation</a>, which is a predefined
+ list of fonts and grouped into Sans-serif, serif, monospace and handwriting.
+ The fonts that are not supported by your system (i.e. not installed) are
+ not displayed.</li>
<li><strong>System fonts</strong> which appear on the&nbsp;<a class="reference-link"
- href="#root/_help_poXkQfguuA0U">Desktop Installation</a>.
- Unlike the server fonts which are predefined, the system fonts are listed
- from your operating system.</li>
+ href="#root/_help_poXkQfguuA0U">Desktop Installation</a>. Unlike the server fonts
+ which are predefined, the system fonts are listed from your operating system.</li>
<li><strong>Custom fonts</strong>, allowing any font to be used provided it's
imported into Trilium. See the section below for more information.</li>
</ul>
diff --git a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Collections/Geo Map.html b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Collections/Geo Map.html
index 9d51597fe4..a0e634536f 100644
--- a/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Collections/Geo Map.html
+++ b/apps/server/src/assets/doc_notes/en/User Guide/User Guide/Collections/Geo Map.html
@@ -54,6 +54,13 @@
</ul>
<p>The position on the map and the zoom are saved inside the map note and
restored when visiting again the note.</p>
+<p>A map that has no position saved yet is framed around the markers it contains
+ when you open it, so they are all in view without having to go looking
+ for them. A map that contains no markers shows the whole world instead.</p>
+<p>In practice this applies to maps whose markers were added externally,
+ by a script or through <a class="reference-link" href="#root/_help_pgxEVkzLl1OP">ETAPI (REST API)</a>.
+ When you add markers yourself you have already moved the map to reach the
+ place you are marking, and that movement saves a position.</p>
<h2>Searching the map</h2>
<p>The search bar at the top-left of the map searches in two places. It always
searches the notes that are already on the map. It can also search for