@jigyasudham/veto 2.9.0 → 3.1.0

package/dist/transcripts/search.js

@@ -0,0 +1,332 @@
+// Portable BM25 search over indexed events (VERSION-3 item 6, Step 10 — redone).
+//
+// Plain-SQLite inverted index (search_docs / search_terms / search_postings),
+// scored in JS. Core SQL only — no FTS5, no math functions, nothing behind a
+// compile-time flag; that is the lesson of PR #28 (`no such module: fts5`).
+// Query terms are bound parameters resolved against the term dictionary, so
+// there is no query language and no injection surface at all.
+//
+// Contract change vs the FTS5 version: score is positive, HIGHER is better
+// (FTS5's bm25() was negative-lower-is-better). Snippets are built in JS from
+// events.text for the returned rows only.
+import { getTranscriptsDb } from './store.js';
+import { normalizeProjectDir } from '../memory/local.js';
+import { tokenize } from './tokenize.js';
+import { STRIDE_CHARS, WINDOW_CHARS } from './chunk.js';
+import { searchVectors } from './vectors.js';
+import { getDocCache, scopeMask } from './cache.js';
+const K1 = 1.2;
+const B = 0.75;
+// Terms present in more than half the corpus carry ~no signal but force
+// materializing huge posting lists (there is no LIMIT that is safe before
+// ranking). Dropped at the dictionary, before any postings are read — unless
+// every query term is that common, in which case they are all kept.
+const DF_CAP_RATIO = 0.5;
+const SNIPPET_TOKENS = 12;
+/** BM25, scored in JS because log() in SQL is a compile-time lottery. */
+function bm25(tf, df, n, len, avgdl) {
+    const idf = Math.log(1 + (n - df + 0.5) / (df + 0.5));
+    return idf * (tf * (K1 + 1)) / (tf + K1 * (1 - B + B * (avgdl > 0 ? len / avgdl : 1)));
+}
+/**
+ * Build a highlighted excerpt in JS — the moral equivalent of FTS5's
+ * snippet(…, '[', ']', '…', 12): the SNIPPET_TOKENS-token window containing
+ * the most matched terms, matches wrapped in [ ], ellipses when clipped.
+ */
+export function buildSnippet(text, terms) {
+    const re = /[\p{L}\p{N}_\-./@+]+/gu;
+    const tokens = [];
+    for (const m of text.matchAll(re)) {
+        const raw = m[0];
+        const matched = tokenize(raw).some(t => terms.has(t));
+        tokens.push({ start: m.index, end: m.index + raw.length, matched });
+    }
+    if (tokens.length === 0)
+        return text.slice(0, 120);
+    // Best SNIPPET_TOKENS-token window = most matches, earliest wins ties.
+    let best = 0;
+    let bestCount = -1;
+    for (let i = 0; i < tokens.length; i += 1) {
+        const windowEnd = Math.min(i + SNIPPET_TOKENS, tokens.length);
+        let count = 0;
+        for (let j = i; j < windowEnd; j += 1)
+            if (tokens[j].matched)
+                count += 1;
+        if (count > bestCount) {
+            bestCount = count;
+            best = i;
+        }
+        if (windowEnd === tokens.length)
+            break;
+    }
+    const windowEnd = Math.min(best + SNIPPET_TOKENS, tokens.length);
+    let out = '';
+    let pos = tokens[best].start;
+    for (let j = best; j < windowEnd; j += 1) {
+        const t = tokens[j];
+        out += text.slice(pos, t.start);
+        out += t.matched ? `[${text.slice(t.start, t.end)}]` : text.slice(t.start, t.end);
+        pos = t.end;
+    }
+    const prefix = tokens[best].start > 0 ? '…' : '';
+    const suffix = windowEnd < tokens.length || tokens[windowEnd - 1].end < text.length ? '…' : '';
+    return (prefix + out.replace(/\s+/g, ' ').trim() + suffix);
+}
+export function searchEvents(query, opts = {}) {
+    const db = getTranscriptsDb();
+    const terms = [...new Set(tokenize(query))];
+    if (terms.length === 0)
+        return [];
+    const docs = getDocCache();
+    const stats = { n: docs.count, avgdl: docs.avgdl };
+    if (stats.n === 0)
+        return [];
+    // Resolve terms against the dictionary in one round trip.
+    const marks = terms.map(() => '?').join(', ');
+    const found = db.prepare(`SELECT id, term FROM search_terms WHERE term IN (${marks})`).all(...terms);
+    if (found.length === 0)
+        return [];
+    // df per term — an index-only range count on the postings PK.
+    const dfStmt = db.prepare(`SELECT COUNT(*) AS n FROM search_postings WHERE term_id = ?`);
+    const withDf = found.map(t => ({ ...t, df: dfStmt.get(t.id).n }));
+    let kept = withDf.filter(t => t.df / stats.n <= DF_CAP_RATIO);
+    if (kept.length === 0)
+        kept = withDf; // every term is that common — keep them all
+    // Postings only — three narrow columns, no join. Joining search_docs per
+    // POSTING to fetch columns needed for the final handful of rows cost 7.1 ms
+    // per common term at 50 sessions; this costs 2.6 ms. Scope and length come
+    // from the cached metadata instead, and full rows are read for the top-K.
+    const rows = db.prepare(`SELECT term_id, doc_id, tf FROM search_postings WHERE term_id IN (${kept.map(() => '?').join(', ')})`).all(...kept.map(t => t.id));
+    if (rows.length === 0)
+        return [];
+    const mask = scopeMask(docs, {
+        projectDir: opts.projectDir ? normalizeProjectDir(opts.projectDir) : undefined,
+        sourceSessionId: opts.sourceSessionId,
+    });
+    const dfById = new Map(kept.map(t => [t.id, t.df]));
+    const acc = new Map();
+    for (const r of rows) {
+        if (mask && mask[r.doc_id] !== 1)
+            continue;
+        const df = dfById.get(r.term_id) ?? 1;
+        const inc = bm25(r.tf, df, stats.n, docs.len[r.doc_id] ?? 0, stats.avgdl);
+        acc.set(r.doc_id, (acc.get(r.doc_id) ?? 0) + inc);
+    }
+    if (acc.size === 0)
+        return [];
+    const limit = Math.min(Math.max(opts.limit ?? 10, 1), 50);
+    const top = [...acc.entries()].sort((a, b) => b[1] - a[1]).slice(0, limit);
+    // Metadata and text for the returned rows only.
+    const docIds = top.map(([id]) => id);
+    const marks2 = docIds.map(() => '?').join(', ');
+    const metaRows = db.prepare(`SELECT d.id, d.event_id, d.archive_id, d.source_session_id, d.seq, d.kind, e.text
+       FROM search_docs d JOIN events e ON e.id = d.event_id
+      WHERE d.id IN (${marks2})`).all(...docIds);
+    const metaById = new Map(metaRows.map(r => [r.id, r]));
+    const termSet = new Set(kept.map(t => t.term));
+    const out = [];
+    for (const [docId, score] of top) {
+        const m = metaById.get(docId);
+        if (!m)
+            continue;
+        out.push({
+            eventId: m.event_id,
+            archiveId: m.archive_id,
+            sourceSessionId: m.source_session_id,
+            seq: m.seq,
+            kind: m.kind,
+            snippet: buildSnippet(m.text ?? '', termSet),
+            score,
+        });
+    }
+    return out;
+}
+// --- Hybrid retrieval -------------------------------------------------------
+// How deep each ranker is consulted before fusing. Deeper than the returned
+// limit so a result ranked mid-list by one ranker can still be lifted by the
+// other — the entire mechanism by which a paraphrase beats keyword overlap.
+const CANDIDATE_DEPTH = 50;
+/**
+ * Reciprocal-rank fusion constant.
+ *
+ * NOT 60. That value comes from the RRF paper's TREC runs over ~1000-document
+ * lists; at our candidate depth it is actively harmful. With k=60 and d=50 the
+ * whole ranking compresses into a 1.8x spread (rank 1 = 0.0164, rank 50 =
+ * 0.0091), which destroys rank information and lets agreement dominate
+ * quality: two mid-list agreements (0.011 + 0.011) outscore one top-ranked
+ * single-ranker find (0.0164).
+ *
+ * That is fatal for the case this layer exists to serve. A strict paraphrase
+ * shares no vocabulary with its target, so the target is found by the semantic
+ * ranker ALONE and earns exactly one contribution — while any keyword-adjacent
+ * event earns a comparable one for free. The semantic discovery loses to noise.
+ *
+ * Requirement: a rank-1 hit from a single ranker must be able to outrank a pair
+ * of mid-list agreements. 1/(k+1) > 2/(k+d/2) solves to k < d/2 - 2, i.e.
+ * k < 23 at d=50. 10 sits comfortably inside that bound and keeps the top of
+ * the list sharply separated (rank 1 = 0.091, rank 10 = 0.050, rank 50 = 0.017).
+ */
+const RRF_K = 10;
+/**
+ * BM25 contributes only candidates it is actually confident about: anything
+ * below this fraction of its own top score is tail noise that would otherwise
+ * occupy fusion slots and crowd out semantic finds.
+ *
+ * Applied to BM25 only, deliberately. The two rankers have different score
+ * geometries: BM25 decays fast and spans orders of magnitude, so relative
+ * score is meaningful. Cosines sit in a narrow band — measured on the real
+ * corpus, off-topic queries reach 0.40 while on-topic start at 0.45 — so a
+ * relative floor there would be noise itself. Depth is the semantic side's
+ * only honest control.
+ *
+ * Only the tail is affected, so a strong keyword hit can never be dropped and
+ * lexical survival is preserved by construction.
+ */
+const BM25_RELATIVE_FLOOR = 0.25;
+/**
+ * One ranker's contribution for a 0-based rank.
+ *
+ * Exported so the property the fix turns on can be asserted directly rather
+ * than through retrieval, where it is not reliably constructible: with a
+ * bag-of-words embedding model, lexical overlap implies semantic similarity,
+ * so a document that matches a query's words but not its meaning barely
+ * exists. The arithmetic, however, is exact.
+ */
+export function rrfWeight(rank0) {
+    return 1 / (RRF_K + rank0 + 1);
+}
+/**
+ * Fuse ranked id lists by reciprocal rank. Pure, for testing and reuse.
+ *
+ * The property that matters: a top-ranked find from ONE ranker must be able
+ * to outrank a document both rankers place mid-list. That is the case the
+ * semantic layer exists to serve — a strict paraphrase is found by the
+ * semantic ranker alone — and it is exactly what k=60 at depth 50 broke.
+ */
+export function fuseByRank(lists) {
+    const acc = new Map();
+    for (const list of lists) {
+        list.forEach((id, i) => acc.set(id, (acc.get(id) ?? 0) + rrfWeight(i)));
+    }
+    return [...acc.entries()]
+        .map(([id, score]) => ({ id, score }))
+        .sort((a, b) => b.score - a.score);
+}
+/**
+ * BM25 and semantic recall, fused by reciprocal rank.
+ *
+ * RRF over raw score blending is deliberate: BM25 scores are unbounded and
+ * corpus-dependent while cosines sit in [-1, 1], so any weighted sum would be
+ * a hidden tuning knob that drifts with corpus size. Ranks have neither
+ * problem, and lexical survival is measurable (a keyword hit must not be
+ * pushed out of the fused list by semantic noise).
+ *
+ * Degrades silently to BM25 alone when no vectors are stored or the model is
+ * not installed (A2) — the caller cannot tell the difference except by score.
+ */
+export function searchEventsHybrid(query, opts = {}) {
+    const limit = Math.min(Math.max(opts.limit ?? 10, 1), 50);
+    const lexicalAll = searchEvents(query, { ...opts, limit: CANDIDATE_DEPTH });
+    const lexFloor = (lexicalAll[0]?.score ?? 0) * BM25_RELATIVE_FLOOR;
+    const lexical = lexicalAll.filter(h => h.score >= lexFloor);
+    // Cosine ranking has no notion of "no match" — it always returns its nearest
+    // neighbours, however far away they are. A similarity floor cannot fix that
+    // here: measured on the real corpus, off-topic queries reach 0.40 while true
+    // strict paraphrases go as low as 0.21, so the distributions overlap and any
+    // threshold that suppresses noise also cuts the recall this layer exists for.
+    //
+    // What DOES separate them is whether the query has any footing in the corpus
+    // at all. A paraphrase shares no words with its target but still speaks the
+    // corpus's language; `zzz-nonexistent-term` shares nothing with anything. So
+    // the semantic layer is allowed to generalize, never to extrapolate from a
+    // query the corpus has no word of — and "nothing matched" stays expressible.
+    // Unfiltered on the fallback paths: the relative floor exists to stop tail
+    // noise occupying FUSION slots. With no fusion happening it would just be
+    // hiding results the lexical ranker legitimately found.
+    if (!queryHasCorpusTerms(query))
+        return lexicalAll.slice(0, limit);
+    let semantic = [];
+    try {
+        semantic = searchVectors(query, {
+            projectDir: opts.projectDir ? normalizeProjectDir(opts.projectDir) : undefined,
+            sourceSessionId: opts.sourceSessionId,
+            limit: CANDIDATE_DEPTH,
+        });
+    }
+    catch {
+        semantic = []; // a broken model must not take the lexical path down with it
+    }
+    if (semantic.length === 0)
+        return lexicalAll.slice(0, limit);
+    const db = getTranscriptsDb();
+    // Resolve the semantic hits' doc ids to event metadata in one round trip.
+    const docIds = semantic.map(s => s.docId);
+    const docRows = db.prepare(`SELECT id, event_id, archive_id, source_session_id, seq, kind
+       FROM search_docs WHERE id IN (${docIds.map(() => '?').join(', ')})`).all(...docIds);
+    const docById = new Map(docRows.map(r => [r.id, r]));
+    const fused = new Map();
+    lexical.forEach((hit, i) => {
+        fused.set(hit.eventId, { hit, rrf: rrfWeight(i) });
+    });
+    semantic.forEach((s, i) => {
+        const doc = docById.get(s.docId);
+        if (!doc)
+            return;
+        const prev = fused.get(doc.event_id);
+        const inc = rrfWeight(i);
+        if (prev) {
+            prev.rrf += inc;
+            prev.chunkIndex = s.chunkIndex;
+        }
+        else
+            fused.set(doc.event_id, { doc, chunkIndex: s.chunkIndex, rrf: inc });
+    });
+    const ranked = [...fused.entries()]
+        .sort((a, b) => b[1].rrf - a[1].rrf)
+        .slice(0, limit);
+    // Text is needed only for entries that arrived semantically — lexical hits
+    // already carry a snippet built from their matched terms.
+    const needText = ranked.filter(([, e]) => !e.hit).map(([eventId]) => eventId);
+    const textById = new Map();
+    if (needText.length > 0) {
+        const rows = db.prepare(`SELECT id, text FROM events WHERE id IN (${needText.map(() => '?').join(', ')})`).all(...needText);
+        for (const r of rows)
+            textById.set(r.id, r.text ?? '');
+    }
+    const terms = new Set(tokenize(query));
+    return ranked.map(([eventId, e]) => {
+        if (e.hit)
+            return { ...e.hit, score: e.rrf };
+        const doc = e.doc;
+        const text = textById.get(eventId) ?? '';
+        return {
+            eventId,
+            archiveId: doc.archive_id,
+            sourceSessionId: doc.source_session_id,
+            seq: doc.seq,
+            kind: doc.kind,
+            // Aim the excerpt at the window that actually matched, so a long event
+            // does not open on an unrelated first paragraph (C6's opportunity).
+            snippet: buildSnippet(windowAround(text, e.chunkIndex ?? 0), terms),
+            score: e.rrf,
+        };
+    });
+}
+/** True when at least one query token appears in the corpus dictionary. */
+function queryHasCorpusTerms(query) {
+    const terms = [...new Set(tokenize(query))];
+    if (terms.length === 0)
+        return false;
+    const db = getTranscriptsDb();
+    const hit = db.prepare(`SELECT 1 AS present FROM search_terms WHERE term IN (${terms.map(() => '?').join(', ')}) LIMIT 1`).get(...terms);
+    return hit !== undefined;
+}
+/** The slice of `text` the winning chunk covered, widened by one stride each
+ *  way so the snippet has context to work with. */
+function windowAround(text, chunkIndex) {
+    if (text.length <= WINDOW_CHARS)
+        return text;
+    const start = Math.max(0, chunkIndex * STRIDE_CHARS - STRIDE_CHARS);
+    return text.slice(start, start + WINDOW_CHARS + STRIDE_CHARS * 2);
+}
+//# sourceMappingURL=search.js.map

package/dist/transcripts/search.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"search.js","sourceRoot":"","sources":["../../src/transcripts/search.ts"],"names":[],"mappings":"AAAA,iFAAiF;AACjF,EAAE;AACF,8EAA8E;AAC9E,6EAA6E;AAC7E,4EAA4E;AAC5E,4EAA4E;AAC5E,8DAA8D;AAC9D,EAAE;AACF,2EAA2E;AAC3E,8EAA8E;AAC9E,0CAA0C;AAE1C,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAC9C,OAAO,EAAE,mBAAmB,EAAE,MAAM,oBAAoB,CAAC;AACzD,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AACxD,OAAO,EAAE,aAAa,EAAE,MAAM,cAAc,CAAC;AAC7C,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,YAAY,CAAC;AAmBpD,MAAM,EAAE,GAAG,GAAG,CAAC;AACf,MAAM,CAAC,GAAG,IAAI,CAAC;AACf,wEAAwE;AACxE,0EAA0E;AAC1E,6EAA6E;AAC7E,oEAAoE;AACpE,MAAM,YAAY,GAAG,GAAG,CAAC;AACzB,MAAM,cAAc,GAAG,EAAE,CAAC;AAE1B,yEAAyE;AACzE,SAAS,IAAI,CAAC,EAAU,EAAE,EAAU,EAAE,CAAS,EAAE,GAAW,EAAE,KAAa;IACzE,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC;IACtD,OAAO,GAAG,GAAG,CAAC,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACzF,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,YAAY,CAAC,IAAY,EAAE,KAAkB;IAC3D,MAAM,EAAE,GAAG,wBAAwB,CAAC;IACpC,MAAM,MAAM,GAAuD,EAAE,CAAC;IACtE,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC;QAClC,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACjB,MAAM,OAAO,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACtD,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC;IACtE,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAEnD,uEAAuE;IACvE,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,IAAI,SAAS,GAAG,CAAC,CAAC,CAAC;IACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC1C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,cAAc,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC;QAC9D,IAAI,KAAK,GAAG,CAAC,CAAC;QACd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,IAAI,CAAC;YAAE,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC,OAAO;gBAAE,KAAK,IAAI,CAAC,CAAC;QACzE,IAAI,KAAK,GAAG,SAAS,EAAE,CAAC;YAAC,SAAS,GAAG,KAAK,CAAC;YAAC,IAAI,GAAG,CAAC,CAAC;QAAC,CAAC;QACvD,IAAI,SAAS,KAAK,MAAM,CAAC,MAAM;YAAE,MAAM;IACzC,CAAC;IACD,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,cAAc,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC;IAEjE,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,IAAI,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC;IAC7B,KAAK,IAAI,CAAC,GAAG,IAAI,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACzC,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QACpB,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC;QAChC,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC;QAClF,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC;IACd,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;IACjD,MAAM,MAAM,GAAG,SAAS,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;IAC/F,OAAO,CAAC,MAAM,GAAG,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,GAAG,MAAM,CAAC,CAAC;AAC7D,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,KAAa,EAAE,OAAsB,EAAE;IAClE,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAC5C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAElC,MAAM,IAAI,GAAG,WAAW,EAAE,CAAC;IAC3B,MAAM,KAAK,GAAG,EAAE,CAAC,EAAE,IAAI,CAAC,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,CAAC;IACnD,IAAI,KAAK,CAAC,CAAC,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE7B,0DAA0D;IAC1D,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC9C,MAAM,KAAK,GAAG,EAAE,CAAC,OAAO,CACtB,oDAAoD,KAAK,GAAG,CAC7D,CAAC,GAAG,CAAC,GAAG,KAAK,CAAmC,CAAC;IAClD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAElC,8DAA8D;IAC9D,MAAM,MAAM,GAAG,EAAE,CAAC,OAAO,CAAC,6DAA6D,CAAC,CAAC;IACzF,MAAM,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,EAAE,EAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAmB,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACrF,IAAI,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,CAAC,IAAI,YAAY,CAAC,CAAC;IAC9D,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,IAAI,GAAG,MAAM,CAAC,CAAC,4CAA4C;IAElF,yEAAyE;IACzE,4EAA4E;IAC5E,2EAA2E;IAC3E,0EAA0E;IAC1E,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACrB,qEAAqE,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CACvG,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAsD,CAAC;IACnF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,EAAE;QAC3B,UAAU,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS;QAC9E,eAAe,EAAE,IAAI,CAAC,eAAe;KACtC,CAAC,CAAC;IAEH,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACpD,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,KAAK,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;QACrB,IAAI,IAAI,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;YAAE,SAAS;QAC3C,MAAM,EAAE,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACtC,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QAC1E,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IACD,IAAI,GAAG,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAE9B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC1D,MAAM,GAAG,GAAG,CAAC,GAAG,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAE3E,gDAAgD;IAChD,MAAM,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC;IACrC,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAChD,MAAM,QAAQ,GAAG,EAAE,CAAC,OAAO,CACzB;;uBAEmB,MAAM,GAAG,CAC7B,CAAC,GAAG,CAAC,GAAG,MAAM,CAA8H,CAAC;IAC9I,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACvD,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAE/C,MAAM,GAAG,GAAgB,EAAE,CAAC;IAC5B,KAAK,MAAM,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,GAAG,EAAE,CAAC;QACjC,MAAM,CAAC,GAAG,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QAC9B,IAAI,CAAC,CAAC;YAAE,SAAS;QACjB,GAAG,CAAC,IAAI,CAAC;YACP,OAAO,EAAE,CAAC,CAAC,QAAQ;YACnB,SAAS,EAAE,CAAC,CAAC,UAAU;YACvB,eAAe,EAAE,CAAC,CAAC,iBAAiB;YACpC,GAAG,EAAE,CAAC,CAAC,GAAG;YACV,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,OAAO,EAAE,YAAY,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,EAAE,OAAO,CAAC;YAC5C,KAAK;SACN,CAAC,CAAC;IACL,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,+EAA+E;AAE/E,4EAA4E;AAC5E,6EAA6E;AAC7E,4EAA4E;AAC5E,MAAM,eAAe,GAAG,EAAE,CAAC;AAE3B;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,KAAK,GAAG,EAAE,CAAC;AAEjB;;;;;;;;;;;;;;GAcG;AACH,MAAM,mBAAmB,GAAG,IAAI,CAAC;AAEjC;;;;;;;;GAQG;AACH,MAAM,UAAU,SAAS,CAAC,KAAa;IACrC,OAAO,CAAC,GAAG,CAAC,KAAK,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC;AACjC,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,UAAU,CAAC,KAAiB;IAC1C,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC,OAAO,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1E,CAAC;IACD,OAAO,CAAC,GAAG,GAAG,CAAC,OAAO,EAAE,CAAC;SACtB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,CAAC;SACrC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;AACvC,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,kBAAkB,CAAC,KAAa,EAAE,OAAsB,EAAE;IACxE,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC1D,MAAM,UAAU,GAAG,YAAY,CAAC,KAAK,EAAE,EAAE,GAAG,IAAI,EAAE,KAAK,EAAE,eAAe,EAAE,CAAC,CAAC;IAC5E,MAAM,QAAQ,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,CAAC,GAAG,mBAAmB,CAAC;IACnE,MAAM,OAAO,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,IAAI,QAAQ,CAAC,CAAC;IAE5D,6EAA6E;IAC7E,4EAA4E;IAC5E,6EAA6E;IAC7E,6EAA6E;IAC7E,8EAA8E;IAC9E,EAAE;IACF,6EAA6E;IAC7E,4EAA4E;IAC5E,6EAA6E;IAC7E,2EAA2E;IAC3E,6EAA6E;IAC7E,2EAA2E;IAC3E,0EAA0E;IAC1E,wDAAwD;IACxD,IAAI,CAAC,mBAAmB,CAAC,KAAK,CAAC;QAAE,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAEnE,IAAI,QAAQ,GAAqC,EAAE,CAAC;IACpD,IAAI,CAAC;QACH,QAAQ,GAAG,aAAa,CAAC,KAAK,EAAE;YAC9B,UAAU,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS;YAC9E,eAAe,EAAE,IAAI,CAAC,eAAe;YACrC,KAAK,EAAE,eAAe;SACvB,CAAC,CAAC;IACL,CAAC;IAAC,MAAM,CAAC;QACP,QAAQ,GAAG,EAAE,CAAC,CAAC,6DAA6D;IAC9E,CAAC;IACD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAE7D,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAE9B,0EAA0E;IAC1E,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;IAC1C,MAAM,OAAO,GAAG,EAAE,CAAC,OAAO,CACxB;uCACmC,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CACvE,CAAC,GAAG,CAAC,GAAG,MAAM,CAAoG,CAAC;IACpH,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAGrD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAiB,CAAC;IAEvC,OAAO,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE;QACzB,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,OAAO,EAAE,EAAE,GAAG,EAAE,GAAG,EAAE,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IACrD,CAAC,CAAC,CAAC;IACH,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACxB,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;QACjC,IAAI,CAAC,GAAG;YAAE,OAAO;QACjB,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACrC,MAAM,GAAG,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC;QACzB,IAAI,IAAI,EAAE,CAAC;YAAC,IAAI,CAAC,GAAG,IAAI,GAAG,CAAC;YAAC,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CAAC;QAAC,CAAC;;YACzD,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,EAAE,EAAE,GAAG,EAAE,UAAU,EAAE,CAAC,CAAC,UAAU,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC;IAC5E,CAAC,CAAC,CAAC;IAEH,MAAM,MAAM,GAAG,CAAC,GAAG,KAAK,CAAC,OAAO,EAAE,CAAC;SAChC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;SACnC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAEnB,2EAA2E;IAC3E,0DAA0D;IAC1D,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC;IAC9E,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC3C,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACrB,4CAA4C,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAClF,CAAC,GAAG,CAAC,GAAG,QAAQ,CAA0C,CAAC;QAC5D,KAAK,MAAM,CAAC,IAAI,IAAI;YAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,CAAC;IACzD,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;IAEvC,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,EAAE,EAAE;QACjC,IAAI,CAAC,CAAC,GAAG;YAAE,OAAO,EAAE,GAAG,CAAC,CAAC,GAAG,EAAE,KAAK,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC;QAC7C,MAAM,GAAG,GAAG,CAAC,CAAC,GAAI,CAAC;QACnB,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACzC,OAAO;YACL,OAAO;YACP,SAAS,EAAE,GAAG,CAAC,UAAU;YACzB,eAAe,EAAE,GAAG,CAAC,iBAAiB;YACtC,GAAG,EAAE,GAAG,CAAC,GAAG;YACZ,IAAI,EAAE,GAAG,CAAC,IAAI;YACd,uEAAuE;YACvE,oEAAoE;YACpE,OAAO,EAAE,YAAY,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,CAAC,UAAU,IAAI,CAAC,CAAC,EAAE,KAAK,CAAC;YACnE,KAAK,EAAE,CAAC,CAAC,GAAG;SACb,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAED,2EAA2E;AAC3E,SAAS,mBAAmB,CAAC,KAAa;IACxC,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAC5C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IACrC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,GAAG,GAAG,EAAE,CAAC,OAAO,CACpB,wDAAwD,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,WAAW,CACnG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAoC,CAAC;IACnD,OAAO,GAAG,KAAK,SAAS,CAAC;AAC3B,CAAC;AAED;mDACmD;AACnD,SAAS,YAAY,CAAC,IAAY,EAAE,UAAkB;IACpD,IAAI,IAAI,CAAC,MAAM,IAAI,YAAY;QAAE,OAAO,IAAI,CAAC;IAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,GAAG,YAAY,GAAG,YAAY,CAAC,CAAC;IACpE,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,YAAY,GAAG,YAAY,GAAG,CAAC,CAAC,CAAC;AACpE,CAAC"}

package/dist/transcripts/store.d.ts

@@ -0,0 +1,21 @@
+import type { DatabaseSync } from 'node:sqlite';
+import { TRANSCRIPTS_SCHEMA_VERSION } from './schema.js';
+/** True when node:sqlite can be loaded — capture is silently unavailable otherwise. */
+export declare function transcriptsAvailable(): boolean;
+/** Root dir for all transcript data (archives + sidecar DB). Env override for tests. */
+export declare function transcriptsRoot(): string;
+/** Path to the sidecar index DB. Env override for tests (mirrors VETO_TEST_DB). */
+export declare function transcriptsDbPath(): string;
+/** Dir holding the L0 .gz archive files (populated in Step 5). */
+export declare function archiveDir(): string;
+export declare function resetTranscriptsDb(): void;
+export declare function getTranscriptsDb(): DatabaseSync;
+/**
+ * Apply every migration newer than the DB's current user_version, each in its
+ * own transaction. Idempotent: re-running is a no-op once caught up. Returns the
+ * resulting schema version.
+ */
+export declare function runMigrations(db: DatabaseSync): number;
+export declare function schemaVersion(db?: DatabaseSync): number;
+export { TRANSCRIPTS_SCHEMA_VERSION };
+//# sourceMappingURL=store.d.ts.map

package/dist/transcripts/store.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"store.d.ts","sourceRoot":"","sources":["../../src/transcripts/store.ts"],"names":[],"mappings":"AASA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAIhD,OAAO,EAAc,0BAA0B,EAAE,MAAM,aAAa,CAAC;AAYrE,uFAAuF;AACvF,wBAAgB,oBAAoB,IAAI,OAAO,CAE9C;AAED,wFAAwF;AACxF,wBAAgB,eAAe,IAAI,MAAM,CAExC;AAED,mFAAmF;AACnF,wBAAgB,iBAAiB,IAAI,MAAM,CAE1C;AAED,kEAAkE;AAClE,wBAAgB,UAAU,IAAI,MAAM,CAEnC;AAKD,wBAAgB,kBAAkB,IAAI,IAAI,CAMzC;AAED,wBAAgB,gBAAgB,IAAI,YAAY,CAc/C;AAED;;;;GAIG;AACH,wBAAgB,aAAa,CAAC,EAAE,EAAE,YAAY,GAAG,MAAM,CAiBtD;AAOD,wBAAgB,aAAa,CAAC,EAAE,GAAE,YAAiC,GAAG,MAAM,CAE3E;AAED,OAAO,EAAE,0BAA0B,EAAE,CAAC"}

package/dist/transcripts/store.js

@@ -0,0 +1,105 @@
+// Sidecar transcripts.db connection + additive migration runner
+// (VERSION-3 item 6, Step 2).
+//
+// Opened lazily (node:sqlite is a Node 22.5+ built-in; loaded via createRequire
+// so bundlers skip it and older runtimes degrade gracefully). WAL + busy_timeout
+// let the HUD keep reading veto.db while capture writes here without lock
+// contention. This module NEVER opens veto.db — the two DBs stay isolated.
+import { createRequire } from 'node:module';
+import { join, dirname } from 'node:path';
+import { mkdirSync } from 'node:fs';
+import { effectiveTranscriptsDir } from './config.js';
+import { MIGRATIONS, TRANSCRIPTS_SCHEMA_VERSION } from './schema.js';
+const _require = createRequire(import.meta.url);
+let _DbSync = null;
+function requireDbSync() {
+    if (!_DbSync) {
+        _DbSync = _require('node:sqlite').DatabaseSync;
+    }
+    return _DbSync;
+}
+/** True when node:sqlite can be loaded — capture is silently unavailable otherwise. */
+export function transcriptsAvailable() {
+    try {
+        requireDbSync();
+        return true;
+    }
+    catch {
+        return false;
+    }
+}
+/** Root dir for all transcript data (archives + sidecar DB). Env override for tests. */
+export function transcriptsRoot() {
+    return process.env.VETO_TRANSCRIPTS_DIR ?? effectiveTranscriptsDir();
+}
+/** Path to the sidecar index DB. Env override for tests (mirrors VETO_TEST_DB). */
+export function transcriptsDbPath() {
+    return process.env.VETO_TRANSCRIPTS_DB ?? join(transcriptsRoot(), 'index.db');
+}
+/** Dir holding the L0 .gz archive files (populated in Step 5). */
+export function archiveDir() {
+    return join(transcriptsRoot(), 'archives');
+}
+let _db = null;
+let _openedPath = null;
+export function resetTranscriptsDb() {
+    if (_db) {
+        try {
+            _db.close();
+        }
+        catch { /* ignore */ }
+        _db = null;
+        _openedPath = null;
+    }
+}
+export function getTranscriptsDb() {
+    const path = transcriptsDbPath();
+    if (_db && _openedPath === path)
+        return _db;
+    if (_db)
+        resetTranscriptsDb();
+    const DbSync = requireDbSync();
+    if (path !== ':memory:')
+        mkdirSync(dirname(path), { recursive: true });
+    const db = new DbSync(path);
+    db.exec('PRAGMA journal_mode = WAL');
+    db.exec('PRAGMA busy_timeout = 5000');
+    db.exec('PRAGMA foreign_keys = ON');
+    runMigrations(db);
+    _db = db;
+    _openedPath = path;
+    return db;
+}
+/**
+ * Apply every migration newer than the DB's current user_version, each in its
+ * own transaction. Idempotent: re-running is a no-op once caught up. Returns the
+ * resulting schema version.
+ */
+export function runMigrations(db) {
+    let applied = readUserVersion(db);
+    for (const m of MIGRATIONS) {
+        if (m.version > applied) {
+            db.exec('BEGIN');
+            try {
+                db.exec(m.up);
+                db.exec(`PRAGMA user_version = ${m.version}`);
+                db.exec('COMMIT');
+                applied = m.version;
+            }
+            catch (e) {
+                db.exec('ROLLBACK');
+                throw e;
+            }
+        }
+    }
+    return applied;
+}
+function readUserVersion(db) {
+    const row = db.prepare('PRAGMA user_version').get();
+    return row?.user_version ?? 0;
+}
+export function schemaVersion(db = getTranscriptsDb()) {
+    return readUserVersion(db);
+}
+export { TRANSCRIPTS_SCHEMA_VERSION };
+//# sourceMappingURL=store.js.map

package/dist/transcripts/store.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"store.js","sourceRoot":"","sources":["../../src/transcripts/store.ts"],"names":[],"mappings":"AAAA,gEAAgE;AAChE,8BAA8B;AAC9B,EAAE;AACF,gFAAgF;AAChF,iFAAiF;AACjF,0EAA0E;AAC1E,2EAA2E;AAE3E,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,SAAS,EAAE,MAAM,SAAS,CAAC;AACpC,OAAO,EAAE,uBAAuB,EAAE,MAAM,aAAa,CAAC;AACtD,OAAO,EAAE,UAAU,EAAE,0BAA0B,EAAE,MAAM,aAAa,CAAC;AAErE,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAChD,IAAI,OAAO,GAAqD,IAAI,CAAC;AAErE,SAAS,aAAa;IACpB,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO,GAAI,QAAQ,CAAC,aAAa,CAAkC,CAAC,YAAY,CAAC;IACnF,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,uFAAuF;AACvF,MAAM,UAAU,oBAAoB;IAClC,IAAI,CAAC;QAAC,aAAa,EAAE,CAAC;QAAC,OAAO,IAAI,CAAC;IAAC,CAAC;IAAC,MAAM,CAAC;QAAC,OAAO,KAAK,CAAC;IAAC,CAAC;AAC/D,CAAC;AAED,wFAAwF;AACxF,MAAM,UAAU,eAAe;IAC7B,OAAO,OAAO,CAAC,GAAG,CAAC,oBAAoB,IAAI,uBAAuB,EAAE,CAAC;AACvE,CAAC;AAED,mFAAmF;AACnF,MAAM,UAAU,iBAAiB;IAC/B,OAAO,OAAO,CAAC,GAAG,CAAC,mBAAmB,IAAI,IAAI,CAAC,eAAe,EAAE,EAAE,UAAU,CAAC,CAAC;AAChF,CAAC;AAED,kEAAkE;AAClE,MAAM,UAAU,UAAU;IACxB,OAAO,IAAI,CAAC,eAAe,EAAE,EAAE,UAAU,CAAC,CAAC;AAC7C,CAAC;AAED,IAAI,GAAG,GAAwB,IAAI,CAAC;AACpC,IAAI,WAAW,GAAkB,IAAI,CAAC;AAEtC,MAAM,UAAU,kBAAkB;IAChC,IAAI,GAAG,EAAE,CAAC;QACR,IAAI,CAAC;YAAC,GAAG,CAAC,KAAK,EAAE,CAAC;QAAC,CAAC;QAAC,MAAM,CAAC,CAAC,YAAY,CAAC,CAAC;QAC3C,GAAG,GAAG,IAAI,CAAC;QACX,WAAW,GAAG,IAAI,CAAC;IACrB,CAAC;AACH,CAAC;AAED,MAAM,UAAU,gBAAgB;IAC9B,MAAM,IAAI,GAAG,iBAAiB,EAAE,CAAC;IACjC,IAAI,GAAG,IAAI,WAAW,KAAK,IAAI;QAAE,OAAO,GAAG,CAAC;IAC5C,IAAI,GAAG;QAAE,kBAAkB,EAAE,CAAC;IAC9B,MAAM,MAAM,GAAG,aAAa,EAAE,CAAC;IAC/B,IAAI,IAAI,KAAK,UAAU;QAAE,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACvE,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC;IAC5B,EAAE,CAAC,IAAI,CAAC,2BAA2B,CAAC,CAAC;IACrC,EAAE,CAAC,IAAI,CAAC,4BAA4B,CAAC,CAAC;IACtC,EAAE,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC;IACpC,aAAa,CAAC,EAAE,CAAC,CAAC;IAClB,GAAG,GAAG,EAAE,CAAC;IACT,WAAW,GAAG,IAAI,CAAC;IACnB,OAAO,EAAE,CAAC;AACZ,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,EAAgB;IAC5C,IAAI,OAAO,GAAG,eAAe,CAAC,EAAE,CAAC,CAAC;IAClC,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;QAC3B,IAAI,CAAC,CAAC,OAAO,GAAG,OAAO,EAAE,CAAC;YACxB,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACjB,IAAI,CAAC;gBACH,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;gBACd,EAAE,CAAC,IAAI,CAAC,yBAAyB,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;gBAC9C,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBAClB,OAAO,GAAG,CAAC,CAAC,OAAO,CAAC;YACtB,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACX,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;gBACpB,MAAM,CAAC,CAAC;YACV,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,eAAe,CAAC,EAAgB;IACvC,MAAM,GAAG,GAAG,EAAE,CAAC,OAAO,CAAC,qBAAqB,CAAC,CAAC,GAAG,EAA2C,CAAC;IAC7F,OAAO,GAAG,EAAE,YAAY,IAAI,CAAC,CAAC;AAChC,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,KAAmB,gBAAgB,EAAE;IACjE,OAAO,eAAe,CAAC,EAAE,CAAC,CAAC;AAC7B,CAAC;AAED,OAAO,EAAE,0BAA0B,EAAE,CAAC"}

package/dist/transcripts/toc.d.ts

@@ -0,0 +1,18 @@
+export type Segment = {
+    index: number;
+    fromSeq: number;
+    toSeq: number;
+    firstTs: string | null;
+    lastTs: string | null;
+    title: string;
+    userMessages: number;
+    assistantMessages: number;
+    toolCalls: number;
+    files: string[];
+    tools: string[];
+    hasError: boolean;
+};
+export declare function buildTOC(archiveId: string): Segment[];
+/** Compact one-line-per-segment rendering for the recall tool's first response. */
+export declare function renderTOC(segments: Segment[]): string;
+//# sourceMappingURL=toc.d.ts.map

package/dist/transcripts/toc.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"toc.d.ts","sourceRoot":"","sources":["../../src/transcripts/toc.ts"],"names":[],"mappings":"AAcA,MAAM,MAAM,OAAO,GAAG;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,KAAK,EAAE,MAAM,CAAC;IACd,YAAY,EAAE,MAAM,CAAC;IACrB,iBAAiB,EAAE,MAAM,CAAC;IAC1B,SAAS,EAAE,MAAM,CAAC;IAClB,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,QAAQ,EAAE,OAAO,CAAC;CACnB,CAAC;AASF,wBAAgB,QAAQ,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,EAAE,CAkDrD;AAED,mFAAmF;AACnF,wBAAgB,SAAS,CAAC,QAAQ,EAAE,OAAO,EAAE,GAAG,MAAM,CAQrD"}

package/dist/transcripts/toc.js

@@ -0,0 +1,96 @@
+// Metadata table-of-contents / phase segmentation (VERSION-3 item 6, Step 9).
+//
+// B1 of the vectorless-retrieval pipeline: split a session into phases so the
+// recall tool can hand the host AI a compact map to navigate BEFORE any text
+// search. Boundaries are deterministic — a new segment begins at each real user
+// message (a user prompt + the assistant's response + tool activity until the
+// next prompt). Events before the first user prompt form a preamble segment.
+// Computed on demand from the events table (no extra storage).
+import { getTranscriptsDb } from './store.js';
+import { EDIT_TOOLS, FILE_RE, ERROR_RE } from './pyramid.js';
+const TITLE_MAX = 80;
+function preview(text) {
+    const t = text.replace(/\s+/g, ' ').trim();
+    return t.length > TITLE_MAX ? t.slice(0, TITLE_MAX - 1) + '…' : t;
+}
+export function buildTOC(archiveId) {
+    const db = getTranscriptsDb();
+    const rows = db.prepare(`SELECT seq, kind, tool_name, text, ts_utc FROM events WHERE archive_id = ? ORDER BY seq`).all(archiveId);
+    if (rows.length === 0)
+        return [];
+    const segments = [];
+    let cur = null;
+    const open = (first, title) => {
+        cur = {
+            index: segments.length, fromSeq: first.seq, toSeq: first.seq,
+            firstTs: first.ts_utc, lastTs: first.ts_utc, title,
+            userMessages: 0, assistantMessages: 0, toolCalls: 0, files: [], tools: [], hasError: false,
+            _files: new Set(), _tools: new Set(),
+        };
+        segments.push(cur);
+    };
+    const close = () => {
+        if (cur) {
+            cur.files = [...cur._files];
+            cur.tools = [...cur._tools];
+        }
+    };
+    for (const r of rows) {
+        if (r.kind === 'user_message') {
+            close();
+            open(r, preview(r.text ?? '') || `segment ${segments.length + 1}`);
+        }
+        else if (!cur) {
+            // Preamble: events before the first user message.
+            open(r, '(session start)');
+        }
+        const c = cur;
+        c.toSeq = r.seq;
+        if (r.ts_utc) {
+            if (!c.firstTs || r.ts_utc < c.firstTs)
+                c.firstTs = r.ts_utc;
+            if (!c.lastTs || r.ts_utc > c.lastTs)
+                c.lastTs = r.ts_utc;
+        }
+        switch (r.kind) {
+            case 'user_message':
+                c.userMessages++;
+                break;
+            case 'assistant_message':
+                c.assistantMessages++;
+                break;
+            case 'tool_call': {
+                c.toolCalls++;
+                const name = r.tool_name ?? 'tool';
+                c._tools.add(name);
+                if (r.text && EDIT_TOOLS.has(name)) {
+                    const m = r.text.match(FILE_RE);
+                    if (m)
+                        c._files.add(m[1]);
+                }
+                break;
+            }
+            case 'tool_result':
+                if (r.text && ERROR_RE.test(r.text))
+                    c.hasError = true;
+                break;
+            default: break;
+        }
+    }
+    close();
+    return segments;
+}
+/** Compact one-line-per-segment rendering for the recall tool's first response. */
+export function renderTOC(segments) {
+    return segments.map(s => {
+        const bits = [`#${s.index} [${s.fromSeq}-${s.toSeq}]`];
+        if (s.hasError)
+            bits.push('⚠err');
+        if (s.files.length)
+            bits.push(`files:${s.files.slice(0, 4).join(',')}`);
+        if (s.tools.length)
+            bits.push(`tools:${s.tools.slice(0, 4).join(',')}`);
+        return `${bits.join(' ')}  ${s.title}`;
+    }).join('\n');
+}
+//# sourceMappingURL=toc.js.map

package/dist/transcripts/toc.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"toc.js","sourceRoot":"","sources":["../../src/transcripts/toc.ts"],"names":[],"mappings":"AAAA,8EAA8E;AAC9E,EAAE;AACF,8EAA8E;AAC9E,6EAA6E;AAC7E,gFAAgF;AAChF,8EAA8E;AAC9E,6EAA6E;AAC7E,+DAA+D;AAE/D,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAC9C,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,cAAc,CAAC;AAE7D,MAAM,SAAS,GAAG,EAAE,CAAC;AAmBrB,SAAS,OAAO,CAAC,IAAY;IAC3B,MAAM,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;IAC3C,OAAO,CAAC,CAAC,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AACpE,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,SAAiB;IACxC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACrB,yFAAyF,CAC1F,CAAC,GAAG,CAAC,SAAS,CAAS,CAAC;IACzB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAEjC,MAAM,QAAQ,GAAc,EAAE,CAAC;IAC/B,IAAI,GAAG,GAAoE,IAAI,CAAC;IAEhF,MAAM,IAAI,GAAG,CAAC,KAAS,EAAE,KAAa,EAAE,EAAE;QACxC,GAAG,GAAG;YACJ,KAAK,EAAE,QAAQ,CAAC,MAAM,EAAE,OAAO,EAAE,KAAK,CAAC,GAAG,EAAE,KAAK,EAAE,KAAK,CAAC,GAAG;YAC5D,OAAO,EAAE,KAAK,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,KAAK;YAClD,YAAY,EAAE,CAAC,EAAE,iBAAiB,EAAE,CAAC,EAAE,SAAS,EAAE,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE,KAAK;YAC1F,MAAM,EAAE,IAAI,GAAG,EAAU,EAAE,MAAM,EAAE,IAAI,GAAG,EAAU;SACrD,CAAC;QACF,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACrB,CAAC,CAAC;IACF,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,IAAI,GAAG,EAAE,CAAC;YAAC,GAAG,CAAC,KAAK,GAAG,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC;YAAC,GAAG,CAAC,KAAK,GAAG,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC;QAAC,CAAC;IACxE,CAAC,CAAC;IAEF,KAAK,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;QACrB,IAAI,CAAC,CAAC,IAAI,KAAK,cAAc,EAAE,CAAC;YAC9B,KAAK,EAAE,CAAC;YACR,IAAI,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,IAAI,WAAW,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,CAAC;QACrE,CAAC;aAAM,IAAI,CAAC,GAAG,EAAE,CAAC;YAChB,kDAAkD;YAClD,IAAI,CAAC,CAAC,EAAE,iBAAiB,CAAC,CAAC;QAC7B,CAAC;QACD,MAAM,CAAC,GAAG,GAAI,CAAC;QACf,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC;QAChB,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC;YAAC,IAAI,CAAC,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,OAAO;gBAAE,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC,MAAM,CAAC;YAAC,IAAI,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM;gBAAE,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC;QAAC,CAAC;QAC1I,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAC;YACf,KAAK,cAAc;gBAAE,CAAC,CAAC,YAAY,EAAE,CAAC;gBAAC,MAAM;YAC7C,KAAK,mBAAmB;gBAAE,CAAC,CAAC,iBAAiB,EAAE,CAAC;gBAAC,MAAM;YACvD,KAAK,WAAW,CAAC,CAAC,CAAC;gBACjB,CAAC,CAAC,SAAS,EAAE,CAAC;gBACd,MAAM,IAAI,GAAG,CAAC,CAAC,SAAS,IAAI,MAAM,CAAC;gBACnC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;gBACnB,IAAI,CAAC,CAAC,IAAI,IAAI,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;oBAAC,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;oBAAC,IAAI,CAAC;wBAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBAAC,CAAC;gBACnG,MAAM;YACR,CAAC;YACD,KAAK,aAAa;gBAAE,IAAI,CAAC,CAAC,IAAI,IAAI,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;oBAAE,CAAC,CAAC,QAAQ,GAAG,IAAI,CAAC;gBAAC,MAAM;YAClF,OAAO,CAAC,CAAC,MAAM;QACjB,CAAC;IACH,CAAC;IACD,KAAK,EAAE,CAAC;IACR,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,mFAAmF;AACnF,MAAM,UAAU,SAAS,CAAC,QAAmB;IAC3C,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE;QACtB,MAAM,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;QACvD,IAAI,CAAC,CAAC,QAAQ;YAAE,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAClC,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM;YAAE,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACxE,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM;YAAE,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACxE,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,CAAC;IACzC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAChB,CAAC"}

package/dist/transcripts/tokenize.d.ts

@@ -0,0 +1,14 @@
+/** Sub-token expansion on by default; the ingest benchmark measures both ways. */
+export declare const SUBTOKENS_DEFAULT = true;
+/**
+ * Tokenize text for indexing or querying — always the same function.
+ *
+ * 1. Split on anything that is not a letter, digit, or `_ - . / @ +`.
+ * 2. Emit each token lowercased.
+ * 3. When it looks compound, also emit its lowercased sub-tokens:
+ *    `normalizeProjectDir` → normalizeprojectdir, normalize, project, dir;
+ *    `src/transcripts/search.ts` → the full path, src, transcripts, search, ts.
+ * 4. Drop blobs (see isBlob) and tokens with no letter or digit at all.
+ */
+export declare function tokenize(text: string, subtokens?: boolean): string[];
+//# sourceMappingURL=tokenize.d.ts.map

package/dist/transcripts/tokenize.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"tokenize.d.ts","sourceRoot":"","sources":["../../src/transcripts/tokenize.ts"],"names":[],"mappings":"AAgCA,kFAAkF;AAClF,eAAO,MAAM,iBAAiB,OAAO,CAAC;AAEtC;;;;;;;;;GASG;AACH,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,SAAS,GAAE,OAA2B,GAAG,MAAM,EAAE,CAsBvF"}