@jigyasudham/veto 2.9.0 → 3.1.0

package/dist/transcripts/embed.js

@@ -0,0 +1,303 @@
+// Pure-JS Model2Vec inference — the semantic half of transcript recall.
+//
+// The model is frozen DATA, not an API: a quantized matrix plus a vocabulary.
+// So this is a reimplementation against a pinned contract, not a translation
+// of upstream code. The contract lives in the model package's `model.json`
+// and every assumption below is asserted at load time, because the failure
+// mode we care about is not a crash — it is plausible-looking wrong vectors.
+//
+// Correctness is pinned by tests/transcripts/embed.test.ts, which replays the
+// golden vectors shipped alongside the weights (fixed text -> expected
+// output, generated by the reference Python implementation in
+// scripts/repack-model.py). If this file drifts, those fail.
+//
+// Zero dependencies, and nothing here runs at import: the model is resolved
+// on first use and never at server boot (condition A2), so a missing or
+// corrupt copy degrades search to BM25 instead of taking Veto down.
+import { readFileSync } from 'node:fs';
+import { createRequire } from 'node:module';
+import { dirname, join } from 'node:path';
+const _require = createRequire(import.meta.url);
+/** Dim of the shipped table. Cross-checked against the header on load. */
+const EXPECTED_FORMAT = 'veto-model/1';
+// model2vec normalizes with this epsilon; matching it keeps golden vectors
+// bit-comparable for zero-norm inputs instead of producing NaN.
+const NORM_EPSILON = 1e-32;
+let _model;
+/** Package root holding the payload. Env override exists for tests and for
+ *  local work before the model package is published. */
+function modelDir() {
+    if (process.env.VETO_MODEL_DIR)
+        return join(process.env.VETO_MODEL_DIR, 'model');
+    // The package's exports map points at the header; the payload sits beside it.
+    return dirname(_require.resolve('@jigyasudham/veto-model/model.json'));
+}
+// ---------------------------------------------------------------------------
+// BertNormalizer
+//
+// Order is upstream's and matters: clean -> CJK spacing -> accents -> case.
+// ---------------------------------------------------------------------------
+// Upstream treats tab/newline/CR as text, not control chars, and strips the
+// rest of the C categories along with NUL and the replacement char.
+const CONTROL = /\p{C}/u;
+const PUNCT = /[!-/:-@[-`{-~]|\p{P}/u;
+function isControl(ch) {
+    if (ch === '\t' || ch === '\n' || ch === '\r')
+        return false;
+    return CONTROL.test(ch);
+}
+function isWhitespace(ch) {
+    return ch === ' ' || ch === '\t' || ch === '\n' || ch === '\r' || /\s/u.test(ch);
+}
+// The CJK blocks BERT spaces out so every ideograph becomes its own word.
+// Hiragana/katakana are deliberately absent upstream — they tokenize as runs.
+function isChinese(cp) {
+    return ((cp >= 0x4e00 && cp <= 0x9fff) ||
+        (cp >= 0x3400 && cp <= 0x4dbf) ||
+        (cp >= 0x20000 && cp <= 0x2a6df) ||
+        (cp >= 0x2a700 && cp <= 0x2b73f) ||
+        (cp >= 0x2b740 && cp <= 0x2b81f) ||
+        (cp >= 0x2b820 && cp <= 0x2ceaf) ||
+        (cp >= 0xf900 && cp <= 0xfaff) ||
+        (cp >= 0x2f800 && cp <= 0x2fa1f));
+}
+function normalize(text, m) {
+    let out = text;
+    if (m.cleanText) {
+        let buf = '';
+        for (const ch of out) {
+            // NUL and U+FFFD as escapes, matching upstream's filter. Literals
+            // here embed raw control bytes and make tools treat this file as binary.
+            if (ch === '\u0000' || ch === '\uFFFD' || isControl(ch))
+                continue;
+            buf += isWhitespace(ch) ? ' ' : ch;
+        }
+        out = buf;
+    }
+    if (m.handleChineseChars) {
+        let buf = '';
+        for (const ch of out) {
+            const cp = ch.codePointAt(0);
+            buf += isChinese(cp) ? ` ${ch} ` : ch;
+        }
+        out = buf;
+    }
+    // `strip_accents: null` in tokenizer.json does NOT mean "off" — upstream
+    // resolves it to the value of `lowercase`. Reading the null literally is
+    // the silent-divergence bug this whole file is most likely to have; it only
+    // shows on accented text, so the golden set carries "café naïve résumé".
+    if (m.stripAccents)
+        out = out.normalize('NFD').replace(/\p{Mn}/gu, '');
+    if (m.lowercase)
+        out = out.toLowerCase();
+    return out;
+}
+// ---------------------------------------------------------------------------
+// BertPreTokenizer — split on whitespace, then isolate punctuation as its own
+// word. "Isolated" means the punctuation is KEPT as a token, not discarded.
+// ---------------------------------------------------------------------------
+function preTokenize(text) {
+    const words = [];
+    for (const chunk of text.split(/\s+/u)) {
+        if (!chunk)
+            continue;
+        let cur = '';
+        for (const ch of chunk) {
+            if (PUNCT.test(ch)) {
+                if (cur) {
+                    words.push(cur);
+                    cur = '';
+                }
+                words.push(ch);
+            }
+            else {
+                cur += ch;
+            }
+        }
+        if (cur)
+            words.push(cur);
+    }
+    return words;
+}
+// ---------------------------------------------------------------------------
+// WordPiece — greedy longest-match-first, continuation pieces carry "##".
+// Operates on code points, not UTF-16 units, so astral characters (emoji)
+// are matched and length-capped as single characters like upstream does.
+// ---------------------------------------------------------------------------
+function wordPiece(word, m, out) {
+    const chars = Array.from(word);
+    if (chars.length > m.maxWordChars) {
+        if (m.unkId !== undefined)
+            out.push(m.unkId);
+        return;
+    }
+    const pieces = [];
+    let start = 0;
+    while (start < chars.length) {
+        let end = chars.length;
+        let id;
+        while (start < end) {
+            const sub = (start > 0 ? m.prefix : '') + chars.slice(start, end).join('');
+            id = m.vocab.get(sub);
+            if (id !== undefined)
+                break;
+            end -= 1;
+        }
+        // No prefix of the remainder is in the vocabulary: upstream discards every
+        // piece found so far and emits a single UNK for the whole word.
+        if (id === undefined) {
+            if (m.unkId !== undefined)
+                out.push(m.unkId);
+            return;
+        }
+        pieces.push(id);
+        start = end;
+    }
+    for (const p of pieces)
+        out.push(p);
+}
+/** Token ids for `text`, with no special tokens added (the shipped contract). */
+export function tokenizeForEmbedding(text, model = load()) {
+    const ids = [];
+    for (const word of preTokenize(normalize(text, model)))
+        wordPiece(word, model, ids);
+    // Upstream drops UNK entirely rather than pooling it (model2vec model.py:166,
+    // "necessary for word-level models"), so an out-of-vocab word contributes
+    // nothing instead of dragging every vector toward a shared UNK row. Filtering
+    // here rather than skipping at emit time matches upstream for the odd case of
+    // a literal "[UNK]" in the input, which WordPiece would otherwise keep.
+    return model.unkId === undefined ? ids : ids.filter((id) => id !== model.unkId);
+}
+// ---------------------------------------------------------------------------
+// Load
+// ---------------------------------------------------------------------------
+function load() {
+    if (_model)
+        return _model;
+    const dir = modelDir();
+    const header = JSON.parse(readFileSync(join(dir, 'model.json'), 'utf8'));
+    if (header.format !== EXPECTED_FORMAT) {
+        throw new Error(`veto-model: unsupported payload format ${header.format} (expected ${EXPECTED_FORMAT})`);
+    }
+    // model2vec multiplies token rows by an optional per-token weight vector
+    // before pooling. potion-base-8M ships none, and this implementation pools
+    // unweighted — so a model that DOES ship them must not load silently.
+    if (header.inference.per_token_weights != null) {
+        throw new Error('veto-model: payload declares per-token weights, which this implementation does not apply');
+    }
+    if (header.inference.add_special_tokens) {
+        throw new Error('veto-model: payload expects special tokens, which this implementation does not add');
+    }
+    if (header.inference.pooling !== 'mean' || !header.inference.normalize) {
+        throw new Error(`veto-model: unsupported inference contract (${header.inference.pooling}, normalize=${header.inference.normalize})`);
+    }
+    const { vocab_size: vocabSize, dim } = header;
+    const tableBuf = readFileSync(join(dir, 'embeddings.int8.bin'));
+    const scaleBuf = readFileSync(join(dir, 'scales.f32.bin'));
+    if (tableBuf.length !== vocabSize * dim) {
+        throw new Error(`veto-model: table is ${tableBuf.length} bytes, expected ${vocabSize * dim}`);
+    }
+    if (scaleBuf.length !== vocabSize * 4) {
+        throw new Error(`veto-model: scales are ${scaleBuf.length} bytes, expected ${vocabSize * 4}`);
+    }
+    const tok = JSON.parse(readFileSync(join(dir, 'tokenizer.json'), 'utf8'));
+    const norm = tok.normalizer ?? {};
+    const tm = tok.model ?? {};
+    if (tm.type !== 'WordPiece' || norm.type !== 'BertNormalizer' || tok.pre_tokenizer?.type !== 'BertPreTokenizer') {
+        throw new Error(`veto-model: unsupported tokenizer (${norm.type} / ${tok.pre_tokenizer?.type} / ${tm.type})`);
+    }
+    const vocab = new Map(Object.entries(tm.vocab ?? {}));
+    if (vocab.size !== vocabSize) {
+        throw new Error(`veto-model: vocab has ${vocab.size} entries, table has ${vocabSize} rows`);
+    }
+    const lowercase = norm.lowercase !== false;
+    const model = {
+        header,
+        dim,
+        table: new Int8Array(tableBuf.buffer, tableBuf.byteOffset, tableBuf.length),
+        // Little-endian float32; Node is LE on every supported platform, and a
+        // byteOffset-aligned view would be brittle, so copy through a DataView.
+        scales: readScales(scaleBuf, vocabSize),
+        vocab,
+        unkId: tm.unk_token ? vocab.get(tm.unk_token) : undefined,
+        prefix: tm.continuing_subword_prefix ?? '##',
+        maxWordChars: tm.max_input_chars_per_word ?? 100,
+        lowercase,
+        // Resolve null the way upstream does — see normalize().
+        stripAccents: norm.strip_accents == null ? lowercase : norm.strip_accents,
+        cleanText: norm.clean_text !== false,
+        handleChineseChars: norm.handle_chinese_chars !== false,
+    };
+    if (model.stripAccents !== header.tokenizer.strip_accents_effective) {
+        throw new Error('veto-model: resolved strip_accents disagrees with the payload header');
+    }
+    _model = model;
+    return model;
+}
+function readScales(buf, n) {
+    const view = new DataView(buf.buffer, buf.byteOffset, buf.length);
+    const out = new Float32Array(n);
+    for (let i = 0; i < n; i++)
+        out[i] = view.getFloat32(i * 4, true);
+    return out;
+}
+// ---------------------------------------------------------------------------
+// Embed
+// ---------------------------------------------------------------------------
+/**
+ * Embed `text` into a unit-length vector.
+ *
+ * Gather each token's row, dequantize it by that row's scale, mean-pool, then
+ * L2-normalize. Text with no tokens (empty or punctuation-free-of-vocab)
+ * yields the zero vector rather than NaN — upstream's behaviour, and the
+ * reason the epsilon is in the denominator.
+ */
+export function embed(text, model = load()) {
+    const ids = tokenizeForEmbedding(text, model);
+    const { dim, table, scales } = model;
+    const out = new Float32Array(dim);
+    if (ids.length === 0)
+        return out;
+    for (const id of ids) {
+        const base = id * dim;
+        const scale = scales[id];
+        for (let d = 0; d < dim; d++)
+            out[d] += table[base + d] * scale;
+    }
+    let norm = 0;
+    for (let d = 0; d < dim; d++) {
+        out[d] /= ids.length;
+        norm += out[d] * out[d];
+    }
+    norm = Math.sqrt(norm) + NORM_EPSILON;
+    for (let d = 0; d < dim; d++)
+        out[d] /= norm;
+    return out;
+}
+/** Cosine similarity of two unit vectors — a dot product, given the inputs. */
+export function cosine(a, b) {
+    let s = 0;
+    for (let i = 0; i < a.length; i++)
+        s += a[i] * b[i];
+    return s;
+}
+/** True when the model package resolves and its payload passes validation. */
+export function embeddingsAvailable() {
+    try {
+        load();
+        return true;
+    }
+    catch {
+        return false;
+    }
+}
+/** Upstream model id + revision of the loaded payload, for recall provenance. */
+export function modelProvenance() {
+    const { upstream } = load().header;
+    return { model_id: upstream.model_id, revision: upstream.revision };
+}
+/** Drop the cached table. Tests only — production loads once per process. */
+export function resetModelCache() {
+    _model = undefined;
+}
+//# sourceMappingURL=embed.js.map

package/dist/transcripts/embed.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"embed.js","sourceRoot":"","sources":["../../src/transcripts/embed.ts"],"names":[],"mappings":"AAAA,wEAAwE;AACxE,EAAE;AACF,8EAA8E;AAC9E,6EAA6E;AAC7E,2EAA2E;AAC3E,2EAA2E;AAC3E,6EAA6E;AAC7E,EAAE;AACF,8EAA8E;AAC9E,uEAAuE;AACvE,8DAA8D;AAC9D,6DAA6D;AAC7D,EAAE;AACF,4EAA4E;AAC5E,wEAAwE;AACxE,oEAAoE;AAEpE,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAE1C,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAEhD,0EAA0E;AAC1E,MAAM,eAAe,GAAG,cAAc,CAAC;AACvC,2EAA2E;AAC3E,gEAAgE;AAChE,MAAM,YAAY,GAAG,KAAK,CAAC;AAiC3B,IAAI,MAAyB,CAAC;AAE9B;wDACwD;AACxD,SAAS,QAAQ;IACf,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc;QAAE,OAAO,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,EAAE,OAAO,CAAC,CAAC;IACjF,8EAA8E;IAC9E,OAAO,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,oCAAoC,CAAC,CAAC,CAAC;AACzE,CAAC;AAED,8EAA8E;AAC9E,iBAAiB;AACjB,EAAE;AACF,4EAA4E;AAC5E,8EAA8E;AAE9E,4EAA4E;AAC5E,oEAAoE;AACpE,MAAM,OAAO,GAAG,QAAQ,CAAC;AACzB,MAAM,KAAK,GAAG,uBAAuB,CAAC;AAEtC,SAAS,SAAS,CAAC,EAAU;IAC3B,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IAC5D,OAAO,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC1B,CAAC;AAED,SAAS,YAAY,CAAC,EAAU;IAC9B,OAAO,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,IAAI,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACnF,CAAC;AAED,0EAA0E;AAC1E,8EAA8E;AAC9E,SAAS,SAAS,CAAC,EAAU;IAC3B,OAAO,CACL,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC;QAC9B,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC;QAC9B,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC;QAChC,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC;QAChC,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC;QAChC,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC;QAChC,CAAC,EAAE,IAAI,MAAM,IAAI,EAAE,IAAI,MAAM,CAAC;QAC9B,CAAC,EAAE,IAAI,OAAO,IAAI,EAAE,IAAI,OAAO,CAAC,CACjC,CAAC;AACJ,CAAC;AAED,SAAS,SAAS,CAAC,IAAY,EAAE,CAAQ;IACvC,IAAI,GAAG,GAAG,IAAI,CAAC;IAEf,IAAI,CAAC,CAAC,SAAS,EAAE,CAAC;QAChB,IAAI,GAAG,GAAG,EAAE,CAAC;QACb,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;YACrB,kEAAkE;YAClE,yEAAyE;YACzE,IAAI,EAAE,KAAK,QAAQ,IAAI,EAAE,KAAK,QAAQ,IAAI,SAAS,CAAC,EAAE,CAAC;gBAAE,SAAS;YAClE,GAAG,IAAI,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;QACrC,CAAC;QACD,GAAG,GAAG,GAAG,CAAC;IACZ,CAAC;IAED,IAAI,CAAC,CAAC,kBAAkB,EAAE,CAAC;QACzB,IAAI,GAAG,GAAG,EAAE,CAAC;QACb,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;YACrB,MAAM,EAAE,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,CAAE,CAAC;YAC9B,GAAG,IAAI,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;QACxC,CAAC;QACD,GAAG,GAAG,GAAG,CAAC;IACZ,CAAC;IAED,yEAAyE;IACzE,yEAAyE;IACzE,4EAA4E;IAC5E,yEAAyE;IACzE,IAAI,CAAC,CAAC,YAAY;QAAE,GAAG,GAAG,GAAG,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IAEvE,IAAI,CAAC,CAAC,SAAS;QAAE,GAAG,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC;IAEzC,OAAO,GAAG,CAAC;AACb,CAAC;AAED,8EAA8E;AAC9E,8EAA8E;AAC9E,4EAA4E;AAC5E,8EAA8E;AAE9E,SAAS,WAAW,CAAC,IAAY;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC;QACvC,IAAI,CAAC,KAAK;YAAE,SAAS;QACrB,IAAI,GAAG,GAAG,EAAE,CAAC;QACb,KAAK,MAAM,EAAE,IAAI,KAAK,EAAE,CAAC;YACvB,IAAI,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC;gBACnB,IAAI,GAAG,EAAE,CAAC;oBAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBAAC,GAAG,GAAG,EAAE,CAAC;gBAAC,CAAC;gBACvC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;iBAAM,CAAC;gBACN,GAAG,IAAI,EAAE,CAAC;YACZ,CAAC;QACH,CAAC;QACD,IAAI,GAAG;YAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC3B,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,8EAA8E;AAC9E,0EAA0E;AAC1E,0EAA0E;AAC1E,yEAAyE;AACzE,8EAA8E;AAE9E,SAAS,SAAS,CAAC,IAAY,EAAE,CAAQ,EAAE,GAAa;IACtD,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/B,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,YAAY,EAAE,CAAC;QAClC,IAAI,CAAC,CAAC,KAAK,KAAK,SAAS;YAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;QAC7C,OAAO;IACT,CAAC;IAED,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QAC5B,IAAI,GAAG,GAAG,KAAK,CAAC,MAAM,CAAC;QACvB,IAAI,EAAsB,CAAC;QAC3B,OAAO,KAAK,GAAG,GAAG,EAAE,CAAC;YACnB,MAAM,GAAG,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAC3E,EAAE,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACtB,IAAI,EAAE,KAAK,SAAS;gBAAE,MAAM;YAC5B,GAAG,IAAI,CAAC,CAAC;QACX,CAAC;QACD,2EAA2E;QAC3E,gEAAgE;QAChE,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;YACrB,IAAI,CAAC,CAAC,KAAK,KAAK,SAAS;gBAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;YAC7C,OAAO;QACT,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAChB,KAAK,GAAG,GAAG,CAAC;IACd,CAAC;IACD,KAAK,MAAM,CAAC,IAAI,MAAM;QAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AACtC,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,oBAAoB,CAAC,IAAY,EAAE,KAAK,GAAG,IAAI,EAAE;IAC/D,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;QAAE,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC;IAEpF,8EAA8E;IAC9E,0EAA0E;IAC1E,8EAA8E;IAC9E,8EAA8E;IAC9E,wEAAwE;IACxE,OAAO,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,KAAK,KAAK,CAAC,KAAK,CAAC,CAAC;AAClF,CAAC;AAED,8EAA8E;AAC9E,OAAO;AACP,8EAA8E;AAE9E,SAAS,IAAI;IACX,IAAI,MAAM;QAAE,OAAO,MAAM,CAAC;IAE1B,MAAM,GAAG,GAAG,QAAQ,EAAE,CAAC;IACvB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,YAAY,CAAC,EAAE,MAAM,CAAC,CAAW,CAAC;IAEnF,IAAI,MAAM,CAAC,MAAM,KAAK,eAAe,EAAE,CAAC;QACtC,MAAM,IAAI,KAAK,CAAC,0CAA0C,MAAM,CAAC,MAAM,cAAc,eAAe,GAAG,CAAC,CAAC;IAC3G,CAAC;IACD,yEAAyE;IACzE,2EAA2E;IAC3E,sEAAsE;IACtE,IAAI,MAAM,CAAC,SAAS,CAAC,iBAAiB,IAAI,IAAI,EAAE,CAAC;QAC/C,MAAM,IAAI,KAAK,CAAC,0FAA0F,CAAC,CAAC;IAC9G,CAAC;IACD,IAAI,MAAM,CAAC,SAAS,CAAC,kBAAkB,EAAE,CAAC;QACxC,MAAM,IAAI,KAAK,CAAC,oFAAoF,CAAC,CAAC;IACxG,CAAC;IACD,IAAI,MAAM,CAAC,SAAS,CAAC,OAAO,KAAK,MAAM,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,SAAS,EAAE,CAAC;QACvE,MAAM,IAAI,KAAK,CAAC,+CAA+C,MAAM,CAAC,SAAS,CAAC,OAAO,eAAe,MAAM,CAAC,SAAS,CAAC,SAAS,GAAG,CAAC,CAAC;IACvI,CAAC;IAED,MAAM,EAAE,UAAU,EAAE,SAAS,EAAE,GAAG,EAAE,GAAG,MAAM,CAAC;IAC9C,MAAM,QAAQ,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,qBAAqB,CAAC,CAAC,CAAC;IAChE,MAAM,QAAQ,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,gBAAgB,CAAC,CAAC,CAAC;IAC3D,IAAI,QAAQ,CAAC,MAAM,KAAK,SAAS,GAAG,GAAG,EAAE,CAAC;QACxC,MAAM,IAAI,KAAK,CAAC,wBAAwB,QAAQ,CAAC,MAAM,oBAAoB,SAAS,GAAG,GAAG,EAAE,CAAC,CAAC;IAChG,CAAC;IACD,IAAI,QAAQ,CAAC,MAAM,KAAK,SAAS,GAAG,CAAC,EAAE,CAAC;QACtC,MAAM,IAAI,KAAK,CAAC,0BAA0B,QAAQ,CAAC,MAAM,oBAAoB,SAAS,GAAG,CAAC,EAAE,CAAC,CAAC;IAChG,CAAC;IAED,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,gBAAgB,CAAC,EAAE,MAAM,CAAC,CAIvE,CAAC;IACF,MAAM,IAAI,GAAG,GAAG,CAAC,UAAU,IAAI,EAAE,CAAC;IAClC,MAAM,EAAE,GAAG,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC;IAC3B,IAAI,EAAE,CAAC,IAAI,KAAK,WAAW,IAAI,IAAI,CAAC,IAAI,KAAK,gBAAgB,IAAI,GAAG,CAAC,aAAa,EAAE,IAAI,KAAK,kBAAkB,EAAE,CAAC;QAChH,MAAM,IAAI,KAAK,CAAC,sCAAsC,IAAI,CAAC,IAAI,MAAM,GAAG,CAAC,aAAa,EAAE,IAAI,MAAM,EAAE,CAAC,IAAI,GAAG,CAAC,CAAC;IAChH,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,GAAG,CAAiB,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,CAAC;IACtE,IAAI,KAAK,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,MAAM,IAAI,KAAK,CAAC,yBAAyB,KAAK,CAAC,IAAI,uBAAuB,SAAS,OAAO,CAAC,CAAC;IAC9F,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,KAAK,KAAK,CAAC;IAC3C,MAAM,KAAK,GAAU;QACnB,MAAM;QACN,GAAG;QACH,KAAK,EAAE,IAAI,SAAS,CAAC,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,UAAU,EAAE,QAAQ,CAAC,MAAM,CAAC;QAC3E,uEAAuE;QACvE,wEAAwE;QACxE,MAAM,EAAE,UAAU,CAAC,QAAQ,EAAE,SAAS,CAAC;QACvC,KAAK;QACL,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS;QACzD,MAAM,EAAE,EAAE,CAAC,yBAAyB,IAAI,IAAI;QAC5C,YAAY,EAAE,EAAE,CAAC,wBAAwB,IAAI,GAAG;QAChD,SAAS;QACT,wDAAwD;QACxD,YAAY,EAAE,IAAI,CAAC,aAAa,IAAI,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa;QACzE,SAAS,EAAE,IAAI,CAAC,UAAU,KAAK,KAAK;QACpC,kBAAkB,EAAE,IAAI,CAAC,oBAAoB,KAAK,KAAK;KACxD,CAAC;IAEF,IAAI,KAAK,CAAC,YAAY,KAAK,MAAM,CAAC,SAAS,CAAC,uBAAuB,EAAE,CAAC;QACpE,MAAM,IAAI,KAAK,CAAC,sEAAsE,CAAC,CAAC;IAC1F,CAAC;IAED,MAAM,GAAG,KAAK,CAAC;IACf,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,UAAU,CAAC,GAAW,EAAE,CAAS;IACxC,MAAM,IAAI,GAAG,IAAI,QAAQ,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,UAAU,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAClE,MAAM,GAAG,GAAG,IAAI,YAAY,CAAC,CAAC,CAAC,CAAC;IAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;QAAE,GAAG,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC;IAClE,OAAO,GAAG,CAAC;AACb,CAAC;AAED,8EAA8E;AAC9E,QAAQ;AACR,8EAA8E;AAE9E;;;;;;;GAOG;AACH,MAAM,UAAU,KAAK,CAAC,IAAY,EAAE,KAAK,GAAG,IAAI,EAAE;IAChD,MAAM,GAAG,GAAG,oBAAoB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC9C,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,KAAK,CAAC;IACrC,MAAM,GAAG,GAAG,IAAI,YAAY,CAAC,GAAG,CAAC,CAAC;IAClC,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,GAAG,CAAC;IAEjC,KAAK,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;QACrB,MAAM,IAAI,GAAG,EAAE,GAAG,GAAG,CAAC;QACtB,MAAM,KAAK,GAAG,MAAM,CAAC,EAAE,CAAE,CAAC;QAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE;YAAE,GAAG,CAAC,CAAC,CAAE,IAAI,KAAK,CAAC,IAAI,GAAG,CAAC,CAAE,GAAG,KAAK,CAAC;IACpE,CAAC;IAED,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC;QAC7B,GAAG,CAAC,CAAC,CAAE,IAAI,GAAG,CAAC,MAAM,CAAC;QACtB,IAAI,IAAI,GAAG,CAAC,CAAC,CAAE,GAAG,GAAG,CAAC,CAAC,CAAE,CAAC;IAC5B,CAAC;IACD,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC;IACtC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,EAAE,CAAC,EAAE;QAAE,GAAG,CAAC,CAAC,CAAE,IAAI,IAAI,CAAC;IAE9C,OAAO,GAAG,CAAC;AACb,CAAC;AAED,+EAA+E;AAC/E,MAAM,UAAU,MAAM,CAAC,CAAe,EAAE,CAAe;IACrD,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAE,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC;IACtD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,mBAAmB;IACjC,IAAI,CAAC;QAAC,IAAI,EAAE,CAAC;QAAC,OAAO,IAAI,CAAC;IAAC,CAAC;IAAC,MAAM,CAAC;QAAC,OAAO,KAAK,CAAC;IAAC,CAAC;AACtD,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,eAAe;IAC7B,MAAM,EAAE,QAAQ,EAAE,GAAG,IAAI,EAAE,CAAC,MAAM,CAAC;IACnC,OAAO,EAAE,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE,CAAC;AACtE,CAAC;AAED,6EAA6E;AAC7E,MAAM,UAAU,eAAe;IAC7B,MAAM,GAAG,SAAS,CAAC;AACrB,CAAC"}

package/dist/transcripts/expand.d.ts

@@ -0,0 +1,20 @@
+import type { ArchiveRow } from './schema.js';
+export declare function _resetExpandCache(): void;
+export type ExpandResult = {
+    ok: boolean;
+    reason?: string;
+    text?: string;
+    secretsRedacted?: number;
+    source?: string;
+    sourceSessionId?: string;
+    provenance?: string;
+    fromSeq?: number;
+    toSeq?: number;
+};
+/** Serve a byte range from an archive, masked. Bounds are validated first. */
+export declare function expandBytes(archive: ArchiveRow, offset: number, length: number): ExpandResult;
+/** Expand a single event to its exact source line (masked), with a turn/timestamp citation. */
+export declare function expandEvent(eventId: string): ExpandResult;
+/** Expand a seq range to the exact contiguous L0 lines it covers (masked). */
+export declare function expandRange(archiveId: string, fromSeq: number, toSeq: number): ExpandResult;
+//# sourceMappingURL=expand.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"expand.d.ts","sourceRoot":"","sources":["../../src/transcripts/expand.ts"],"names":[],"mappings":"AAcA,OAAO,KAAK,EAAE,UAAU,EAAY,MAAM,aAAa,CAAC;AAYxD,wBAAgB,iBAAiB,IAAI,IAAI,CAAmB;AAE5D,MAAM,MAAM,YAAY,GAAG;IACzB,EAAE,EAAE,OAAO,CAAC;IACZ,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB,CAAC;AAkBF,8EAA8E;AAC9E,wBAAgB,WAAW,CAAC,OAAO,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,YAAY,CAa7F;AAED,+FAA+F;AAC/F,wBAAgB,WAAW,CAAC,OAAO,EAAE,MAAM,GAAG,YAAY,CASzD;AAED,8EAA8E;AAC9E,wBAAgB,WAAW,CAAC,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,YAAY,CAa3F"}

package/dist/transcripts/expand.js

@@ -0,0 +1,97 @@
+// Mask-on-expansion: serve exact L0 bytes, always masked (VERSION-3 item 6, Step 7).
+//
+// L0 is stored RAW (Rule 0), but nothing leaves this module unmasked. Every byte
+// range served out of an archive passes through mask() first, so even though a
+// secret may sit in the raw archive on disk, it can never transit into an AI
+// context through Veto. Offsets are validated against the archive length before
+// any slice (S6) so a corrupt index can never read out of bounds. This is the
+// read-side counterpart to Step 6's mask-on-write, and it is tested here BEFORE
+// the recall tool (Step 11) is allowed to trust it.
+import { readFileSync, existsSync, statSync } from 'node:fs';
+import { gunzipSync } from 'node:zlib';
+import { getTranscriptsDb } from './store.js';
+import { mask } from './mask.js';
+// Small 1-entry cache so a recall that expands several ranges gunzips once.
+let _cache = null;
+function loadArchive(path) {
+    const { mtimeMs } = statSync(path);
+    if (_cache && _cache.path === path && _cache.mtimeMs === mtimeMs)
+        return _cache.buf;
+    const buf = gunzipSync(readFileSync(path));
+    _cache = { path, mtimeMs, buf };
+    return buf;
+}
+export function _resetExpandCache() { _cache = null; }
+function getArchiveById(id) {
+    const db = getTranscriptsDb();
+    return db.prepare(`SELECT * FROM archives WHERE id = ?`).get(id) ?? null;
+}
+function cite(archive, loc) {
+    const parts = [archive.source, `session ${archive.source_session_id}`];
+    if (loc.seq !== undefined) {
+        parts.push(loc.toSeq !== undefined && loc.toSeq !== loc.seq ? `turns ${loc.seq}-${loc.toSeq}` : `turn ${loc.seq}`);
+    }
+    if (loc.ts)
+        parts.push(loc.ts);
+    if (loc.offset !== undefined && loc.length !== undefined)
+        parts.push(`bytes ${loc.offset}-${loc.offset + loc.length}`);
+    return `[${parts.join(' · ')}]`;
+}
+/** Serve a byte range from an archive, masked. Bounds are validated first. */
+export function expandBytes(archive, offset, length) {
+    if (!existsSync(archive.archive_path))
+        return { ok: false, reason: 'archive_missing' };
+    let buf;
+    try {
+        buf = loadArchive(archive.archive_path);
+    }
+    catch {
+        return { ok: false, reason: 'unreadable' };
+    }
+    if (!Number.isInteger(offset) || !Number.isInteger(length) || offset < 0 || length < 0 || offset + length > buf.length) {
+        return { ok: false, reason: 'out_of_bounds' };
+    }
+    const m = mask(buf.toString('utf8', offset, offset + length));
+    return {
+        ok: true, text: m.text, secretsRedacted: m.count,
+        source: archive.source, sourceSessionId: archive.source_session_id,
+        provenance: cite(archive, { offset, length }),
+    };
+}
+/** Expand a single event to its exact source line (masked), with a turn/timestamp citation. */
+export function expandEvent(eventId) {
+    const db = getTranscriptsDb();
+    const ev = db.prepare(`SELECT * FROM events WHERE id = ?`).get(eventId) ?? null;
+    if (!ev)
+        return { ok: false, reason: 'event_not_found' };
+    const archive = getArchiveById(ev.archive_id);
+    if (!archive)
+        return { ok: false, reason: 'archive_not_found' };
+    const r = expandBytes(archive, ev.raw_offset, ev.raw_length);
+    if (r.ok) {
+        r.provenance = cite(archive, { seq: ev.seq, ts: ev.ts_utc });
+        r.fromSeq = ev.seq;
+        r.toSeq = ev.seq;
+    }
+    return r;
+}
+/** Expand a seq range to the exact contiguous L0 lines it covers (masked). */
+export function expandRange(archiveId, fromSeq, toSeq) {
+    const db = getTranscriptsDb();
+    const archive = getArchiveById(archiveId);
+    if (!archive)
+        return { ok: false, reason: 'archive_not_found' };
+    const rows = db.prepare(`SELECT raw_offset, raw_length FROM events WHERE archive_id = ? AND seq BETWEEN ? AND ? ORDER BY seq`).all(archiveId, Math.min(fromSeq, toSeq), Math.max(fromSeq, toSeq));
+    if (rows.length === 0)
+        return { ok: false, reason: 'no_events' };
+    const start = Math.min(...rows.map(r => r.raw_offset));
+    const end = Math.max(...rows.map(r => r.raw_offset + r.raw_length));
+    const r = expandBytes(archive, start, end - start);
+    if (r.ok) {
+        r.provenance = cite(archive, { seq: fromSeq, toSeq });
+        r.fromSeq = fromSeq;
+        r.toSeq = toSeq;
+    }
+    return r;
+}
+//# sourceMappingURL=expand.js.map

package/dist/transcripts/expand.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"expand.js","sourceRoot":"","sources":["../../src/transcripts/expand.ts"],"names":[],"mappings":"AAAA,qFAAqF;AACrF,EAAE;AACF,iFAAiF;AACjF,+EAA+E;AAC/E,6EAA6E;AAC7E,gFAAgF;AAChF,8EAA8E;AAC9E,gFAAgF;AAChF,oDAAoD;AAEpD,OAAO,EAAE,YAAY,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAC7D,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAC9C,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAGjC,4EAA4E;AAC5E,IAAI,MAAM,GAA0D,IAAI,CAAC;AACzE,SAAS,WAAW,CAAC,IAAY;IAC/B,MAAM,EAAE,OAAO,EAAE,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IACnC,IAAI,MAAM,IAAI,MAAM,CAAC,IAAI,KAAK,IAAI,IAAI,MAAM,CAAC,OAAO,KAAK,OAAO;QAAE,OAAO,MAAM,CAAC,GAAG,CAAC;IACpF,MAAM,GAAG,GAAG,UAAU,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC;IAC3C,MAAM,GAAG,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,EAAE,CAAC;IAChC,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,UAAU,iBAAiB,KAAW,MAAM,GAAG,IAAI,CAAC,CAAC,CAAC;AAc5D,SAAS,cAAc,CAAC,EAAU;IAChC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,OAAQ,EAAE,CAAC,OAAO,CAAC,qCAAqC,CAAC,CAAC,GAAG,CAAC,EAAE,CAA4B,IAAI,IAAI,CAAC;AACvG,CAAC;AAGD,SAAS,IAAI,CAAC,OAAmB,EAAE,GAAQ;IACzC,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,WAAW,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;IACvE,IAAI,GAAG,CAAC,GAAG,KAAK,SAAS,EAAE,CAAC;QAC1B,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,KAAK,SAAS,IAAI,GAAG,CAAC,KAAK,KAAK,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC;IACrH,CAAC;IACD,IAAI,GAAG,CAAC,EAAE;QAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAC/B,IAAI,GAAG,CAAC,MAAM,KAAK,SAAS,IAAI,GAAG,CAAC,MAAM,KAAK,SAAS;QAAE,KAAK,CAAC,IAAI,CAAC,SAAS,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC;IACvH,OAAO,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;AAClC,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,WAAW,CAAC,OAAmB,EAAE,MAAc,EAAE,MAAc;IAC7E,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,YAAY,CAAC;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,iBAAiB,EAAE,CAAC;IACvF,IAAI,GAAW,CAAC;IAChB,IAAI,CAAC;QAAC,GAAG,GAAG,WAAW,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC;IAAC,CAAC;IAAC,MAAM,CAAC;QAAC,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,YAAY,EAAE,CAAC;IAAC,CAAC;IACtG,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,MAAM,GAAG,CAAC,IAAI,MAAM,GAAG,CAAC,IAAI,MAAM,GAAG,MAAM,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACvH,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,eAAe,EAAE,CAAC;IAChD,CAAC;IACD,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC;IAC9D,OAAO;QACL,EAAE,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,eAAe,EAAE,CAAC,CAAC,KAAK;QAChD,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,eAAe,EAAE,OAAO,CAAC,iBAAiB;QAClE,UAAU,EAAE,IAAI,CAAC,OAAO,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;KAC9C,CAAC;AACJ,CAAC;AAED,+FAA+F;AAC/F,MAAM,UAAU,WAAW,CAAC,OAAe;IACzC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,EAAE,GAAI,EAAE,CAAC,OAAO,CAAC,mCAAmC,CAAC,CAAC,GAAG,CAAC,OAAO,CAA0B,IAAI,IAAI,CAAC;IAC1G,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,iBAAiB,EAAE,CAAC;IACzD,MAAM,OAAO,GAAG,cAAc,CAAC,EAAE,CAAC,UAAU,CAAC,CAAC;IAC9C,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,mBAAmB,EAAE,CAAC;IAChE,MAAM,CAAC,GAAG,WAAW,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC;IAC7D,IAAI,CAAC,CAAC,EAAE,EAAE,CAAC;QAAC,CAAC,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,EAAE,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,EAAE,EAAE,EAAE,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC;QAAC,CAAC,CAAC,OAAO,GAAG,EAAE,CAAC,GAAG,CAAC;QAAC,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC,GAAG,CAAC;IAAC,CAAC;IACjH,OAAO,CAAC,CAAC;AACX,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,WAAW,CAAC,SAAiB,EAAE,OAAe,EAAE,KAAa;IAC3E,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,OAAO,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC;IAC1C,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,mBAAmB,EAAE,CAAC;IAChE,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACrB,qGAAqG,CACtG,CAAC,GAAG,CAAC,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,KAAK,CAAC,CAAiD,CAAC;IACrH,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IACjE,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC;IACvD,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,UAAU,GAAG,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC;IACpE,MAAM,CAAC,GAAG,WAAW,CAAC,OAAO,EAAE,KAAK,EAAE,GAAG,GAAG,KAAK,CAAC,CAAC;IACnD,IAAI,CAAC,CAAC,EAAE,EAAE,CAAC;QAAC,CAAC,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,EAAE,EAAE,GAAG,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC,CAAC;QAAC,CAAC,CAAC,OAAO,GAAG,OAAO,CAAC;QAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC;IAAC,CAAC;IAC1G,OAAO,CAAC,CAAC;AACX,CAAC"}

package/dist/transcripts/ingest.d.ts

@@ -0,0 +1,17 @@
+import { type EventRow } from './schema.js';
+export declare const PARSER_VERSION = 2;
+export type IngestStatus = 'indexed' | 'already_indexed' | 'not_found' | 'archive_missing' | 'session_mismatch' | 'error';
+export type IngestResult = {
+    status: IngestStatus;
+    events?: number;
+    secretsRedacted?: number;
+    reason?: string;
+};
+/** (Re)derive events for one archive. Best-effort: never throws. */
+export declare function ingestArchive(archiveId: string): IngestResult;
+/** Ensure an archive is indexed at the current parser version (lazy trigger for recall). */
+export declare function ensureIndexed(archiveId: string): IngestResult;
+/** Ingest by source session id (looks up its archive). */
+export declare function ingestSession(sourceSessionId: string, source?: string): IngestResult;
+export declare function getEvents(archiveId: string): EventRow[];
+//# sourceMappingURL=ingest.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"ingest.d.ts","sourceRoot":"","sources":["../../src/transcripts/ingest.ts"],"names":[],"mappings":"AAgBA,OAAO,EAAqC,KAAK,QAAQ,EAAE,MAAM,aAAa,CAAC;AAO/E,eAAO,MAAM,cAAc,IAAI,CAAC;AAEhC,MAAM,MAAM,YAAY,GACpB,SAAS,GAAG,iBAAiB,GAAG,WAAW,GAAG,iBAAiB,GAAG,kBAAkB,GAAG,OAAO,CAAC;AAEnG,MAAM,MAAM,YAAY,GAAG;IACzB,MAAM,EAAE,YAAY,CAAC;IACrB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB,CAAC;AAWF,oEAAoE;AACpE,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,GAAG,YAAY,CA0F7D;AAED,4FAA4F;AAC5F,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,GAAG,YAAY,CAE7D;AAED,0DAA0D;AAC1D,wBAAgB,aAAa,CAAC,eAAe,EAAE,MAAM,EAAE,MAAM,SAAW,GAAG,YAAY,CAOtF;AAED,wBAAgB,SAAS,CAAC,SAAS,EAAE,MAAM,GAAG,QAAQ,EAAE,CAGvD"}

package/dist/transcripts/ingest.js

@@ -0,0 +1,120 @@
+// Ingest: derive normalized events from an L0 archive (VERSION-3 item 6, Step 6).
+//
+// Idempotent + lazy via the archive's parser_version + indexed_through_seq
+// watermark: a fresh or grown archive (Step 5 resets both) gets (re)indexed; an
+// unchanged, already-indexed archive is a no-op; a parser upgrade re-derives.
+// Session-id verification (council/devil): if the archive's session id does not
+// appear anywhere in the parsed file, the mapping pointed at the wrong transcript
+// and we SKIP rather than index a mismatched file.
+import { readFileSync, existsSync } from 'node:fs';
+import { gunzipSync } from 'node:zlib';
+import { randomUUID } from 'node:crypto';
+import { getTranscriptsDb } from './store.js';
+import { resetCaches } from './cache.js';
+import { parseClaudeTranscript } from './adapters/claude.js';
+import { tokenize } from './tokenize.js';
+import { SEARCHABLE_KINDS } from './schema.js';
+const SEARCHABLE = new Set(SEARCHABLE_KINDS);
+// Bump when the parser changes in a way that should re-derive existing archives.
+// v2: the portable search index replaced events_fts — archives ingested at v1
+// have events but no search_docs/postings rows, so they must re-derive.
+export const PARSER_VERSION = 2;
+function getArchiveById(archiveId) {
+    const db = getTranscriptsDb();
+    return db.prepare(`SELECT * FROM archives WHERE id = ?`).get(archiveId) ?? null;
+}
+function needsIndexing(row) {
+    return row.parser_version !== PARSER_VERSION;
+}
+/** (Re)derive events for one archive. Best-effort: never throws. */
+export function ingestArchive(archiveId) {
+    try {
+        const db = getTranscriptsDb();
+        const row = getArchiveById(archiveId);
+        if (!row)
+            return { status: 'not_found' };
+        if (!needsIndexing(row))
+            return { status: 'already_indexed', events: row.indexed_through_seq };
+        if (!existsSync(row.archive_path))
+            return { status: 'archive_missing' };
+        const buf = gunzipSync(readFileSync(row.archive_path));
+        const { events, sessionIds, secretsRedacted } = parseClaudeTranscript(buf);
+        if (sessionIds.size > 0 && !sessionIds.has(row.source_session_id)) {
+            return { status: 'session_mismatch', reason: 'archive session id not present in transcript' };
+        }
+        const ins = db.prepare(`INSERT INTO events (id, archive_id, source_session_id, seq, line_index, block_index, kind,
+         source_type, role, tool_name, text, secret_count, event_uuid, parent_uuid, is_sidechain,
+         ts_source, ts_utc, raw_offset, raw_length)
+       VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`);
+        const insDoc = db.prepare(`INSERT INTO search_docs (event_id, archive_id, source_session_id, project_dir, seq, kind, len)
+       VALUES (?, ?, ?, ?, ?, ?, ?) RETURNING id`);
+        const insTerm = db.prepare(`INSERT INTO search_terms (term) VALUES (?)
+       ON CONFLICT(term) DO UPDATE SET term = term RETURNING id`);
+        const insPosting = db.prepare(`INSERT INTO search_postings (term_id, doc_id, tf) VALUES (?, ?, ?)`);
+        // Term ids are cached per ingest run — a session reuses its vocabulary
+        // heavily, so most lookups never hit the upsert.
+        const termIds = new Map();
+        const termId = (term) => {
+            let id = termIds.get(term);
+            if (id === undefined) {
+                id = Number(insTerm.get(term).id);
+                termIds.set(term, id);
+            }
+            return id;
+        };
+        db.exec('BEGIN');
+        try {
+            db.prepare(`DELETE FROM events WHERE archive_id = ?`).run(archiveId);
+            db.prepare(`DELETE FROM search_postings WHERE doc_id IN (SELECT id FROM search_docs WHERE archive_id = ?)`).run(archiveId);
+            // Doc ids are reassigned by this re-derive, so stale vectors would not
+            // just orphan — they would silently re-attach to the wrong events.
+            db.prepare(`DELETE FROM search_vectors WHERE doc_id IN (SELECT id FROM search_docs WHERE archive_id = ?)`).run(archiveId);
+            db.prepare(`DELETE FROM search_docs WHERE archive_id = ?`).run(archiveId);
+            for (const e of events) {
+                const eventId = randomUUID();
+                ins.run(eventId, archiveId, row.source_session_id, e.seq, e.lineIndex, e.blockIndex, e.kind, e.sourceType, e.role, e.toolName, e.text, e.secretCount, e.eventUuid, e.parentUuid, e.isSidechain ? 1 : 0, e.tsSource, e.tsUtc, e.rawOffset, e.rawLength);
+                // Index the searchable, non-empty events (text is already masked).
+                if (SEARCHABLE.has(e.kind) && e.text && e.text.trim()) {
+                    const tokens = tokenize(e.text);
+                    const tf = new Map();
+                    for (const t of tokens)
+                        tf.set(t, (tf.get(t) ?? 0) + 1);
+                    const docId = Number(insDoc.get(eventId, archiveId, row.source_session_id, row.project_dir, e.seq, e.kind, tokens.length).id);
+                    for (const [term, count] of tf)
+                        insPosting.run(termId(term), docId, count);
+                }
+            }
+            // Vectors were just dropped with their docs; clearing the marker lets
+            // the lazy embed pass rebuild them without re-parsing the transcript.
+            db.prepare(`UPDATE archives SET indexed_through_seq = ?, parser_version = ?, chunker_version = 0, embed_model = NULL
+          WHERE id = ?`).run(events.length, PARSER_VERSION, archiveId);
+            db.exec('COMMIT');
+            resetCaches();
+        }
+        catch (e) {
+            db.exec('ROLLBACK');
+            throw e;
+        }
+        return { status: 'indexed', events: events.length, secretsRedacted };
+    }
+    catch (e) {
+        return { status: 'error', reason: e instanceof Error ? e.message : String(e) };
+    }
+}
+/** Ensure an archive is indexed at the current parser version (lazy trigger for recall). */
+export function ensureIndexed(archiveId) {
+    return ingestArchive(archiveId);
+}
+/** Ingest by source session id (looks up its archive). */
+export function ingestSession(sourceSessionId, source = 'claude') {
+    const db = getTranscriptsDb();
+    const row = db.prepare(`SELECT id FROM archives WHERE source = ? AND source_session_id = ?`).get(source, sourceSessionId) ?? null;
+    if (!row)
+        return { status: 'not_found' };
+    return ingestArchive(row.id);
+}
+export function getEvents(archiveId) {
+    const db = getTranscriptsDb();
+    return db.prepare(`SELECT * FROM events WHERE archive_id = ? ORDER BY seq`).all(archiveId);
+}
+//# sourceMappingURL=ingest.js.map

package/dist/transcripts/ingest.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ingest.js","sourceRoot":"","sources":["../../src/transcripts/ingest.ts"],"names":[],"mappings":"AAAA,kFAAkF;AAClF,EAAE;AACF,2EAA2E;AAC3E,gFAAgF;AAChF,8EAA8E;AAC9E,gFAAgF;AAChF,kFAAkF;AAClF,mDAAmD;AAEnD,OAAO,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACnD,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAC9C,OAAO,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AACzC,OAAO,EAAE,qBAAqB,EAAE,MAAM,sBAAsB,CAAC;AAC7D,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAkC,MAAM,aAAa,CAAC;AAE/E,MAAM,UAAU,GAAG,IAAI,GAAG,CAAS,gBAAgB,CAAC,CAAC;AAErD,iFAAiF;AACjF,8EAA8E;AAC9E,wEAAwE;AACxE,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,CAAC;AAYhC,SAAS,cAAc,CAAC,SAAiB;IACvC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,OAAQ,EAAE,CAAC,OAAO,CAAC,qCAAqC,CAAC,CAAC,GAAG,CAAC,SAAS,CAA4B,IAAI,IAAI,CAAC;AAC9G,CAAC;AAED,SAAS,aAAa,CAAC,GAAe;IACpC,OAAO,GAAG,CAAC,cAAc,KAAK,cAAc,CAAC;AAC/C,CAAC;AAED,oEAAoE;AACpE,MAAM,UAAU,aAAa,CAAC,SAAiB;IAC7C,IAAI,CAAC;QACH,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;QAC9B,MAAM,GAAG,GAAG,cAAc,CAAC,SAAS,CAAC,CAAC;QACtC,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;QACzC,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC;YAAE,OAAO,EAAE,MAAM,EAAE,iBAAiB,EAAE,MAAM,EAAE,GAAG,CAAC,mBAAmB,EAAE,CAAC;QAC/F,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,YAAY,CAAC;YAAE,OAAO,EAAE,MAAM,EAAE,iBAAiB,EAAE,CAAC;QAExE,MAAM,GAAG,GAAG,UAAU,CAAC,YAAY,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC;QACvD,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,eAAe,EAAE,GAAG,qBAAqB,CAAC,GAAG,CAAC,CAAC;QAE3E,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,iBAAiB,CAAC,EAAE,CAAC;YAClE,OAAO,EAAE,MAAM,EAAE,kBAAkB,EAAE,MAAM,EAAE,8CAA8C,EAAE,CAAC;QAChG,CAAC;QAED,MAAM,GAAG,GAAG,EAAE,CAAC,OAAO,CACpB;;;wEAGkE,CACnE,CAAC;QACF,MAAM,MAAM,GAAG,EAAE,CAAC,OAAO,CACvB;iDAC2C,CAC5C,CAAC;QACF,MAAM,OAAO,GAAG,EAAE,CAAC,OAAO,CACxB;gEAC0D,CAC3D,CAAC;QACF,MAAM,UAAU,GAAG,EAAE,CAAC,OAAO,CAC3B,oEAAoE,CACrE,CAAC;QACF,uEAAuE;QACvE,iDAAiD;QACjD,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;QAC1C,MAAM,MAAM,GAAG,CAAC,IAAY,EAAU,EAAE;YACtC,IAAI,EAAE,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;YAC3B,IAAI,EAAE,KAAK,SAAS,EAAE,CAAC;gBACrB,EAAE,GAAG,MAAM,CAAE,OAAO,CAAC,GAAG,CAAC,IAAI,CAA6B,CAAC,EAAE,CAAC,CAAC;gBAC/D,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACxB,CAAC;YACD,OAAO,EAAE,CAAC;QACZ,CAAC,CAAC;QACF,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACjB,IAAI,CAAC;YACH,EAAE,CAAC,OAAO,CAAC,yCAAyC,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YACrE,EAAE,CAAC,OAAO,CACR,+FAA+F,CAChG,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YACjB,uEAAuE;YACvE,mEAAmE;YACnE,EAAE,CAAC,OAAO,CACR,8FAA8F,CAC/F,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YACjB,EAAE,CAAC,OAAO,CAAC,8CAA8C,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YAC1E,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;gBACvB,MAAM,OAAO,GAAG,UAAU,EAAE,CAAC;gBAC7B,GAAG,CAAC,GAAG,CACL,OAAO,EAAE,SAAS,EAAE,GAAG,CAAC,iBAAiB,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,IAAI,EACnF,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,UAAU,EAAE,CAAC,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EACzG,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,SAAS,CAC9C,CAAC;gBACF,mEAAmE;gBACnE,IAAI,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;oBACtD,MAAM,MAAM,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;oBAChC,MAAM,EAAE,GAAG,IAAI,GAAG,EAAkB,CAAC;oBACrC,KAAK,MAAM,CAAC,IAAI,MAAM;wBAAE,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;oBACxD,MAAM,KAAK,GAAG,MAAM,CAAE,MAAM,CAAC,GAAG,CAC9B,OAAO,EAAE,SAAS,EAAE,GAAG,CAAC,iBAAiB,EAAE,GAAG,CAAC,WAAW,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,EAAE,MAAM,CAAC,MAAM,CAC7D,CAAC,EAAE,CAAC,CAAC;oBAClC,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE;wBAAE,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;gBAC7E,CAAC;YACH,CAAC;YACD,sEAAsE;YACtE,sEAAsE;YACtE,EAAE,CAAC,OAAO,CACR;uBACe,CAChB,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,EAAE,cAAc,EAAE,SAAS,CAAC,CAAC;YAChD,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YAClB,WAAW,EAAE,CAAC;QAChB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YACpB,MAAM,CAAC,CAAC;QACV,CAAC;QAED,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,eAAe,EAAE,CAAC;IACvE,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;IACjF,CAAC;AACH,CAAC;AAED,4FAA4F;AAC5F,MAAM,UAAU,aAAa,CAAC,SAAiB;IAC7C,OAAO,aAAa,CAAC,SAAS,CAAC,CAAC;AAClC,CAAC;AAED,0DAA0D;AAC1D,MAAM,UAAU,aAAa,CAAC,eAAuB,EAAE,MAAM,GAAG,QAAQ;IACtE,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,MAAM,GAAG,GAAI,EAAE,CAAC,OAAO,CACrB,oEAAoE,CACrE,CAAC,GAAG,CAAC,MAAM,EAAE,eAAe,CAAgC,IAAI,IAAI,CAAC;IACtE,IAAI,CAAC,GAAG;QAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IACzC,OAAO,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,SAAS,CAAC,SAAiB;IACzC,MAAM,EAAE,GAAG,gBAAgB,EAAE,CAAC;IAC9B,OAAO,EAAE,CAAC,OAAO,CAAC,wDAAwD,CAAC,CAAC,GAAG,CAAC,SAAS,CAAe,CAAC;AAC3G,CAAC"}

package/dist/transcripts/manage.d.ts

@@ -0,0 +1,42 @@
+import { type Segment } from './toc.js';
+import { type SessionFacts } from './pyramid.js';
+export type ArchiveSummary = {
+    sourceSessionId: string;
+    source: string;
+    projectDir: string | null;
+    events: number;
+    sourceBytes: number;
+    archiveBytes: number;
+    capturedAt: string;
+    updatedAt: string;
+    indexed: boolean;
+};
+export declare function listArchives(opts?: {
+    projectDir?: string;
+    limit?: number;
+}): ArchiveSummary[];
+export type ArchiveDetail = {
+    summary: ArchiveSummary;
+    toc: Segment[];
+    facts: SessionFacts;
+} | null;
+export declare function showArchive(sourceSessionId: string, source?: string): ArchiveDetail;
+export type PurgeResult = {
+    archives: number;
+    events: number;
+    indexRows: number;
+    files: number;
+    mappings: number;
+};
+export declare function purgeSession(sourceSessionId: string, source?: string): PurgeResult;
+export declare function purgeProject(projectDir: string): PurgeResult;
+export declare function purgeAll(): PurgeResult;
+export type DiskUsage = {
+    archives: number;
+    archiveBytes: number;
+    dbBytes: number;
+    totalBytes: number;
+};
+export declare function transcriptsDiskUsage(): DiskUsage;
+export declare function fmtBytes(n: number): string;
+//# sourceMappingURL=manage.d.ts.map