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
// ============================================================
// XERJ Console — "which data classes does this engine hold?" probe
//
// Sibling to brains-probe.js. The AI / logs / vector / memory dashboards each
// fall back to FULL MOCK data when their backing index has zero rows (see
// data/backends/xerj.js: `if (!total) return null` → query.js → mock.js). So a
// fresh or brain-only engine that advertised those dashboards would be showing
// fabricated numbers behind a real-looking UI. This probe answers, per data
// class, "does the engine actually hold this?" so app.js can hide the nav entry
// until it does — exactly the pattern secondBrain already uses via brains-probe.
//
// One cheap `_cat/indices` call, classified client-side by the index-name
// literals the dashboards query (data/backends/xerj.js). Fail-CLOSED: absent on
// any transport failure or non-OK status, so we never CLAIM data we can't prove.
//
// Cached per baseUrl with a short TTL (same as brains-probe) so the boot probe
// and periodic re-probes don't hammer `_cat`, and a backend switch re-probes.
// ============================================================
const TTL_MS = 30_000;
/** baseUrl → { at: epoch-ms, value: features } */
const cache = new Map();
/** The data-class feature keys, all false. Also the shape app.js merges into
* state.liveFeatures, and the safe answer on any failure. */
export function emptyDataFeatures() {
return {
'chat-events': false,
'vector-ops': false,
'agent-memory': false,
'anomalies': false,
'logs-ingest-events': false,
'logs': false,
};
}
/** Classify a list of index names into data-class presence flags. The literals
* mirror the hardcoded `/<index>/_search` targets in data/backends/xerj.js. */
export function classifyIndices(names) {
const has = (n) => names.includes(n);
const anyPrefix = (p) => names.some((n) => n.startsWith(p));
return {
'chat-events': has('chat-events'), // ai-overview, rag-quality
'vector-ops': has('vector-ops'), // vector-index
'agent-memory': has('agent-memory'), // agent-memory
'anomalies': has('anomalies'), // anomaly-detect
'logs-ingest-events': has('logs-ingest-events'), // ingest-pipeline
'logs': anyPrefix('logs-'), // logs-overview (logs-*)
};
}
/**
* Return which data classes the engine at `baseUrl` holds. Never throws;
* returns all-false on transport failure, non-OK status, or empty baseUrl.
*/
export async function dataFeaturesPresent(baseUrl, signal) {
const base = (baseUrl || '').replace(/\/+$/, '');
if (!base) return emptyDataFeatures();
const hit = cache.get(base);
if (hit && Date.now() - hit.at < TTL_MS) return hit.value;
let value = emptyDataFeatures();
try {
// The data-class indices (chat-events, vector-ops, …) are non-dot, so the
// default `_cat/indices` lists them; no expand_wildcards needed.
const r = await fetch(`${base}/_cat/indices`, {
signal,
headers: { accept: 'text/plain, application/json' },
});
if (r.ok) {
value = classifyIndices(parseIndexNames(await r.text()));
}
} catch {
value = emptyDataFeatures(); // engine down / CORS / abort — no claim
}
cache.set(base, { at: Date.now(), value });
return value;
}
/**
* Parse a `_cat/indices` body into index names. This engine emits plain text
* lines (`health status NAME uuid …`); tolerate a JSON array in case a later
* build adds `?format=json`. Exported for the node self-test.
*/
export function parseIndexNames(text) {
const trimmed = (text || '').trim();
if (!trimmed) return [];
if (trimmed.startsWith('[')) {
try { return JSON.parse(trimmed).map((i) => i.index).filter(Boolean); } catch { return []; }
}
return trimmed.split('\n')
.map((l) => l.trim().split(/\s+/)[2])
.filter(Boolean);
}