nlm-memory 0.20.0 → 0.21.0

package/dist/core/hook/select.js

@@ -5,9 +5,16 @@
  * surfaced in this conversation, then the per-fire cap bounded by the
  * remaining per-conversation budget. Hits are assumed pre-ranked best-first.
  */
+function medianScore(hits) {
+    if (hits.length === 0)
+        return 0;
+    const sorted = hits.map((h) => h.matchScore).sort((a, b) => a - b);
+    return sorted[Math.floor(sorted.length / 2)] ?? 0;
+}
 export function selectHits(params) {
-    const { hits, surfaced, scoreThreshold, perFireCap, perConversationCap } = params;
-    const eligible = hits.filter((h) => h.matchScore >= scoreThreshold && !surfaced.has(h.id));
+    const { hits, surfaced, scoreThreshold, perFireCap, perConversationCap, relativeFloor = 0 } = params;
+    const relCut = relativeFloor > 0 ? relativeFloor * medianScore(hits) : 0;
+    const eligible = hits.filter((h) => h.matchScore >= scoreThreshold && h.matchScore >= relCut && !surfaced.has(h.id));
     const budget = Math.max(0, perConversationCap - surfaced.size);
     const limit = Math.min(perFireCap, budget);
     return eligible.slice(0, limit);

package/dist/core/hook/select.js.map

@@ -1 +1 @@
-{"version":3,"file":"select.js","sourceRoot":"","sources":["../../../src/core/hook/select.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAkBH,MAAM,UAAU,UAAU,CAAC,MAAoB;IAC7C,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,cAAc,EAAE,UAAU,EAAE,kBAAkB,EAAE,GAAG,MAAM,CAAC;IAClF,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAC1B,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,IAAI,cAAc,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAC7D,CAAC;IACF,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,kBAAkB,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC/D,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;IAC3C,OAAO,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AAClC,CAAC"}
+{"version":3,"file":"select.js","sourceRoot":"","sources":["../../../src/core/hook/select.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AA0BH,SAAS,WAAW,CAAC,IAAmC;IACtD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAChC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACnE,OAAO,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;AACpD,CAAC;AAED,MAAM,UAAU,UAAU,CAAC,MAAoB;IAC7C,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,cAAc,EAAE,UAAU,EAAE,kBAAkB,EAAE,aAAa,GAAG,CAAC,EAAE,GAAG,MAAM,CAAC;IACrG,MAAM,MAAM,GAAG,aAAa,GAAG,CAAC,CAAC,CAAC,CAAC,aAAa,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACzE,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAC1B,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,IAAI,cAAc,IAAI,CAAC,CAAC,UAAU,IAAI,MAAM,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CACvF,CAAC;IACF,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,kBAAkB,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IAC/D,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;IAC3C,OAAO,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AAClC,CAAC"}

package/dist/core/hook/transcript.js

@@ -17,7 +17,7 @@
  * Fail-quiet: a malformed file yields nulls/empty rather than throwing —
  * the Stop hook must never break on transcript I/O.
  */
-import { existsSync, readFileSync } from "node:fs";
+import { closeSync, existsSync, fstatSync, openSync, readSync } from "node:fs";
 const EMPTY_TURN = { text: "", toolUses: [] };
 function parseTurn(parsed) {
     if (parsed.type !== "assistant" || !parsed.message)
@@ -42,11 +42,25 @@ function parseTurn(parsed) {
         return null;
     return { text: textParts.join("\n"), toolUses };
 }
+const MAX_TRANSCRIPT_BYTES = 256 * 1024;
 function readLines(transcriptPath) {
     if (!transcriptPath || !existsSync(transcriptPath))
         return null;
     try {
-        return readFileSync(transcriptPath, "utf8").split("\n");
+        const fd = openSync(transcriptPath, "r");
+        try {
+            const size = fstatSync(fd).size;
+            const start = size > MAX_TRANSCRIPT_BYTES ? size - MAX_TRANSCRIPT_BYTES : 0;
+            const len = size - start;
+            if (len <= 0)
+                return [];
+            const buf = Buffer.alloc(len);
+            readSync(fd, buf, 0, len, start);
+            return buf.toString("utf8").split("\n");
+        }
+        finally {
+            closeSync(fd);
+        }
     }
     catch {
         return null;

package/dist/core/hook/transcript.js.map

@@ -1 +1 @@
-{"version":3,"file":"transcript.js","sourceRoot":"","sources":["../../../src/core/hook/transcript.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AA0BnD,MAAM,UAAU,GAAkB,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AAE7D,SAAS,SAAS,CAAC,MAAsB;IACvC,IAAI,MAAM,CAAC,IAAI,KAAK,WAAW,IAAI,CAAC,MAAM,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAChE,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC;IACvC,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;QAChC,OAAO,OAAO,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;IAC1D,CAAC;IACD,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IACzC,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,MAAM,QAAQ,GAAmB,EAAE,CAAC;IACpC,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,OAAO,KAAK,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YAC5D,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU,IAAI,OAAO,KAAK,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YACvE,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACjE,OAAO,EAAE,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,CAAC;AAClD,CAAC;AAED,SAAS,SAAS,CAAC,cAAsB;IACvC,IAAI,CAAC,cAAc,IAAI,CAAC,UAAU,CAAC,cAAc,CAAC;QAAE,OAAO,IAAI,CAAC;IAChE,IAAI,CAAC;QACH,OAAO,YAAY,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,cAAsB;IAEtB,MAAM,KAAK,GAAG,SAAS,CAAC,cAAc,CAAC,CAAC;IACxC,IAAI,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IACtB,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;QACxB,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAsB,CAAC;QAC3B,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAmB,CAAC;QAC9C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC7B,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,cAAsB;IAC1D,MAAM,KAAK,GAAG,SAAS,CAAC,cAAc,CAAC,CAAC;IACxC,IAAI,CAAC,KAAK;QAAE,OAAO,UAAU,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC9B,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAsB,CAAC;QAC3B,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAmB,CAAC;QAC9C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,OAAO,IAAI,CAAC;IACxB,CAAC;IACD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,yDAAyD;AACzD,MAAM,UAAU,qBAAqB,CAAC,cAAsB;IAC1D,MAAM,IAAI,GAAG,qBAAqB,CAAC,cAAc,CAAC,CAAC;IACnD,OAAO,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC;AAC3B,CAAC"}
+{"version":3,"file":"transcript.js","sourceRoot":"","sources":["../../../src/core/hook/transcript.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,EAAE,SAAS,EAAE,UAAU,EAAE,SAAS,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AA0B/E,MAAM,UAAU,GAAkB,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC;AAE7D,SAAS,SAAS,CAAC,MAAsB;IACvC,IAAI,MAAM,CAAC,IAAI,KAAK,WAAW,IAAI,CAAC,MAAM,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAChE,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC;IACvC,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;QAChC,OAAO,OAAO,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;IAC1D,CAAC;IACD,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IACzC,MAAM,SAAS,GAAa,EAAE,CAAC;IAC/B,MAAM,QAAQ,GAAmB,EAAE,CAAC;IACpC,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;QAC5B,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM,IAAI,OAAO,KAAK,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YAC5D,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;aAAM,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU,IAAI,OAAO,KAAK,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;YACvE,QAAQ,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;QAC1D,CAAC;IACH,CAAC;IACD,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACjE,OAAO,EAAE,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,CAAC;AAClD,CAAC;AAED,MAAM,oBAAoB,GAAG,GAAG,GAAG,IAAI,CAAC;AAExC,SAAS,SAAS,CAAC,cAAsB;IACvC,IAAI,CAAC,cAAc,IAAI,CAAC,UAAU,CAAC,cAAc,CAAC;QAAE,OAAO,IAAI,CAAC;IAChE,IAAI,CAAC;QACH,MAAM,EAAE,GAAG,QAAQ,CAAC,cAAc,EAAE,GAAG,CAAC,CAAC;QACzC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,SAAS,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,GAAG,oBAAoB,CAAC,CAAC,CAAC,IAAI,GAAG,oBAAoB,CAAC,CAAC,CAAC,CAAC,CAAC;YAC5E,MAAM,GAAG,GAAG,IAAI,GAAG,KAAK,CAAC;YACzB,IAAI,GAAG,IAAI,CAAC;gBAAE,OAAO,EAAE,CAAC;YACxB,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;YAC9B,QAAQ,CAAC,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC;YACjC,OAAO,GAAG,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC1C,CAAC;gBAAS,CAAC;YACT,SAAS,CAAC,EAAE,CAAC,CAAC;QAChB,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,cAAsB;IAEtB,MAAM,KAAK,GAAG,SAAS,CAAC,cAAc,CAAC,CAAC;IACxC,IAAI,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IACtB,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACxB,MAAM,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;QACxB,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAsB,CAAC;QAC3B,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAmB,CAAC;QAC9C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC7B,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,cAAsB;IAC1D,MAAM,KAAK,GAAG,SAAS,CAAC,cAAc,CAAC,CAAC;IACxC,IAAI,CAAC,KAAK;QAAE,OAAO,UAAU,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC;QAC9B,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,IAAI,MAAsB,CAAC;QAC3B,IAAI,CAAC;YACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAmB,CAAC;QAC9C,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,OAAO,IAAI,CAAC;IACxB,CAAC;IACD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,yDAAyD;AACzD,MAAM,UAAU,qBAAqB,CAAC,cAAsB;IAC1D,MAAM,IAAI,GAAG,qBAAqB,CAAC,cAAc,CAAC,CAAC;IACnD,OAAO,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC;AAC3B,CAAC"}

package/dist/core/ingest/backfill-derivables.d.ts

@@ -0,0 +1,92 @@
+/**
+ * backfill-derivables — one-time retroactive stamp of agent_persona +
+ * parent_session_id (#352 phase 2) onto sessions ingested before the
+ * scheduler started stamping them at classify time (see
+ * src/core/scheduler/scheduler.ts, commit 34968a8).
+ *
+ * Mirrors the live stamp site exactly so backfilled rows match
+ * freshly-ingested ones: claude-code rows go through deriveSubagentMeta
+ * (reversing the `<parent>/agent-<id>` runtimeSessionId + `[subagent <slug>]`
+ * label encoding); every other runtime has no subagent concept, so persona
+ * is just the runtime name and parent is null.
+ *
+ * Idempotent: selection is WHERE agent_persona IS NULL, so a re-run only
+ * touches rows a prior run (or live ingest) hasn't stamped yet.
+ *
+ * Known data quirk (Task 2 review): some subagent transcripts encode a
+ * literal `unknown` parent (`unknown/agent-<id>`), which derives
+ * parentSessionId = "unknown" — a non-joining placeholder, not a real
+ * session id. The backfill still writes it, for consistency with live
+ * stamping, but counts it separately (`unknownParent`) so operators can see
+ * how much of the corpus has an unresolvable parent link.
+ *
+ * Persona is frequently unrecoverable for pre-existing subagent rows: the
+ * live stamp site derives persona from the transient pre-classification
+ * chunk.label (which carries the `[subagent <slug>]` prefix), but the
+ * `label` column this backfill reads is the classifier's generated title —
+ * that prefix is already gone by the time a row lands in `sessions`. In
+ * practice this means most historical subagent rows get parent_session_id
+ * populated but agent_persona stays NULL. Only genuinely new ingests (which
+ * derive from the live chunk before classification overwrites it) get both.
+ * Such rows stay selectable (agent_persona IS NULL) forever, so the run
+ * compares derived values against stored ones and counts a row `updated`
+ * only when the write would change something; a re-run over a fully
+ * backfilled corpus reports updated=0 with the unrecoverable rows under
+ * `skippedNoop`.
+ *
+ * Task 5 addition: `--with-transcript-scan` optionally also re-scans each
+ * candidate's transcript file for primary_model/total_tokens/skill via
+ * scanTranscriptDerivables. Off by default — scanning streams every
+ * candidate's transcript off disk and is far slower than the persona/parent
+ * derivation above, which only touches already-loaded columns. When the
+ * flag is set, candidacy widens beyond WHERE agent_persona IS NULL with an
+ * OR arm keyed on the scan's own columns (claude-code-jsonl rows with a
+ * transcript_path and all three scan columns still NULL) — otherwise a row
+ * whose persona was stamped by an earlier flagless run would never be
+ * scannable, leaving the flag shipped but effectively unwired for exactly
+ * the rows Task 3 already processed. Rows selected only via that arm get a
+ * scan-only write (persona/parent untouched); either way the scan columns
+ * COALESCE-preserve, and a re-run reports updated=0: a stamped scan column
+ * excludes the row from the OR arm on the next pass, and rows that stay
+ * selectable via the persona arm (persona-unrecoverable subagent rows) skip
+ * the re-scan once any scan column is stamped and count as no-op when the
+ * scan cannot change the row (all-null result, or derivables already
+ * stored).
+ */
+import type { Database } from "better-sqlite3";
+export interface BackfillOptions {
+    readonly dryRun?: boolean;
+    /**
+     * Also scan each candidate's transcript file for
+     * primary_model/total_tokens/skill (#352 phase 2, Task 5). Off by
+     * default: slower than the persona/parent derivation (streams every
+     * candidate's transcript off disk).
+     */
+    readonly withTranscriptScan?: boolean;
+}
+export interface BackfillReport {
+    /** All rows in the sessions table, stamped or not. */
+    readonly total: number;
+    /** Rows whose write would change something (or did, when not --dry-run). */
+    readonly updated: number;
+    /**
+     * Rows not selected at all: agent_persona already stamped and (when
+     * withTranscriptScan) not eligible for the scan arm either.
+     */
+    readonly skippedAlreadyStamped: number;
+    /**
+     * Selected rows whose derived values match what's already stored — persona
+     * unrecoverable (NULL derives NULL), parent either already stamped by a
+     * prior run or also underivable, and (when withTranscriptScan) the
+     * transcript scan derived nothing either. Attempted, nothing to write.
+     */
+    readonly skippedNoop: number;
+    /** Of the selected rows, how many derived a real parent link (subagent-shaped runtimeSessionId). */
+    readonly subagentCandidates: number;
+    /** Of the selected rows, how many derived a literal "unknown" parent placeholder. */
+    readonly unknownParent: number;
+    /** Rows whose transcript was scanned and derived at least one non-null
+     *  field. Always 0 when withTranscriptScan is not set. */
+    readonly transcriptScanned: number;
+}
+export declare function backfillDerivables(db: Database, opts?: BackfillOptions): Promise<BackfillReport>;

package/dist/core/ingest/backfill-derivables.js

@@ -0,0 +1,175 @@
+/**
+ * backfill-derivables — one-time retroactive stamp of agent_persona +
+ * parent_session_id (#352 phase 2) onto sessions ingested before the
+ * scheduler started stamping them at classify time (see
+ * src/core/scheduler/scheduler.ts, commit 34968a8).
+ *
+ * Mirrors the live stamp site exactly so backfilled rows match
+ * freshly-ingested ones: claude-code rows go through deriveSubagentMeta
+ * (reversing the `<parent>/agent-<id>` runtimeSessionId + `[subagent <slug>]`
+ * label encoding); every other runtime has no subagent concept, so persona
+ * is just the runtime name and parent is null.
+ *
+ * Idempotent: selection is WHERE agent_persona IS NULL, so a re-run only
+ * touches rows a prior run (or live ingest) hasn't stamped yet.
+ *
+ * Known data quirk (Task 2 review): some subagent transcripts encode a
+ * literal `unknown` parent (`unknown/agent-<id>`), which derives
+ * parentSessionId = "unknown" — a non-joining placeholder, not a real
+ * session id. The backfill still writes it, for consistency with live
+ * stamping, but counts it separately (`unknownParent`) so operators can see
+ * how much of the corpus has an unresolvable parent link.
+ *
+ * Persona is frequently unrecoverable for pre-existing subagent rows: the
+ * live stamp site derives persona from the transient pre-classification
+ * chunk.label (which carries the `[subagent <slug>]` prefix), but the
+ * `label` column this backfill reads is the classifier's generated title —
+ * that prefix is already gone by the time a row lands in `sessions`. In
+ * practice this means most historical subagent rows get parent_session_id
+ * populated but agent_persona stays NULL. Only genuinely new ingests (which
+ * derive from the live chunk before classification overwrites it) get both.
+ * Such rows stay selectable (agent_persona IS NULL) forever, so the run
+ * compares derived values against stored ones and counts a row `updated`
+ * only when the write would change something; a re-run over a fully
+ * backfilled corpus reports updated=0 with the unrecoverable rows under
+ * `skippedNoop`.
+ *
+ * Task 5 addition: `--with-transcript-scan` optionally also re-scans each
+ * candidate's transcript file for primary_model/total_tokens/skill via
+ * scanTranscriptDerivables. Off by default — scanning streams every
+ * candidate's transcript off disk and is far slower than the persona/parent
+ * derivation above, which only touches already-loaded columns. When the
+ * flag is set, candidacy widens beyond WHERE agent_persona IS NULL with an
+ * OR arm keyed on the scan's own columns (claude-code-jsonl rows with a
+ * transcript_path and all three scan columns still NULL) — otherwise a row
+ * whose persona was stamped by an earlier flagless run would never be
+ * scannable, leaving the flag shipped but effectively unwired for exactly
+ * the rows Task 3 already processed. Rows selected only via that arm get a
+ * scan-only write (persona/parent untouched); either way the scan columns
+ * COALESCE-preserve, and a re-run reports updated=0: a stamped scan column
+ * excludes the row from the OR arm on the next pass, and rows that stay
+ * selectable via the persona arm (persona-unrecoverable subagent rows) skip
+ * the re-scan once any scan column is stamped and count as no-op when the
+ * scan cannot change the row (all-null result, or derivables already
+ * stored).
+ */
+import { deriveSubagentMeta } from "../adapters/claude-code.js";
+import { scanTranscriptDerivables } from "./transcript-derivables.js";
+const BATCH_SIZE = 500;
+const NULL_TRANSCRIPT_DERIVABLES = {
+    primaryModel: null,
+    totalTokens: null,
+    skill: null,
+};
+// Widened OR arm for --with-transcript-scan: rows already persona-stamped
+// (e.g. by an earlier flagless Task 3 run) but never transcript-scanned.
+// Scoped to claude-code-jsonl since scanTranscriptDerivables derives nothing
+// for any other kind; requires ALL three scan columns NULL so any stamped
+// scan column excludes the row on the next pass (re-run idempotency).
+const SCAN_CANDIDATE_ARM = " OR (transcript_path IS NOT NULL AND transcript_kind = 'claude-code-jsonl'" +
+    " AND primary_model IS NULL AND total_tokens IS NULL AND skill IS NULL)";
+function selectCandidates(db, withTranscriptScan) {
+    return db
+        .prepare("SELECT id, runtime, runtime_session_id, label, agent_persona, parent_session_id, transcript_path, transcript_kind, primary_model, total_tokens, skill" +
+        " FROM sessions WHERE agent_persona IS NULL" +
+        (withTranscriptScan ? SCAN_CANDIDATE_ARM : ""))
+        .all();
+}
+// The sessions.runtime column stores the adapter's versioned runtime string
+// (e.g. "claude-code/1.0", "hermes/1.0" — see each adapter's runtimeVersion
+// field), but the live stamp site in scheduler.ts branches on the bare
+// adapter name ("claude-code", "hermes", ...). Strip the version suffix so
+// the backfill branches identically to a fresh ingest.
+function bareRuntimeName(runtime) {
+    const slash = runtime.indexOf("/");
+    return slash === -1 ? runtime : runtime.slice(0, slash);
+}
+function deriveMeta(row) {
+    const runtimeName = bareRuntimeName(row.runtime);
+    if (runtimeName === "claude-code") {
+        return deriveSubagentMeta(row.runtime_session_id ?? "", row.label);
+    }
+    return { persona: runtimeName, parentSessionId: null };
+}
+export async function backfillDerivables(db, opts = {}) {
+    const total = db.prepare("SELECT COUNT(*) AS n FROM sessions").get().n;
+    const candidates = selectCandidates(db, Boolean(opts.withTranscriptScan));
+    const skippedAlreadyStamped = total - candidates.length;
+    let subagentCandidates = 0;
+    let unknownParent = 0;
+    let skippedNoop = 0;
+    let transcriptScanned = 0;
+    const toWrite = [];
+    for (const row of candidates) {
+        const meta = row.agent_persona === null ? deriveMeta(row) : null;
+        if (meta) {
+            if (meta.parentSessionId !== null)
+                subagentCandidates++;
+            if (meta.parentSessionId === "unknown")
+                unknownParent++;
+        }
+        // Any non-null stored scan column means a completed prior scan already
+        // stamped every field this transcript can derive (all three co-derive
+        // from the same pass, at ingest and here) - re-scanning would only
+        // re-derive values COALESCE discards. Skipping is what keeps
+        // persona-unrecoverable rows (agent_persona NULL forever, re-selected
+        // every run via the persona arm) from re-streaming their transcript and
+        // churning as "updated" on every pass.
+        const scanAlreadyRan = row.primary_model !== null || row.total_tokens !== null || row.skill !== null;
+        let transcript = NULL_TRANSCRIPT_DERIVABLES;
+        if (opts.withTranscriptScan && row.transcript_path && !scanAlreadyRan) {
+            transcript = await scanTranscriptDerivables(row.transcript_path, row.transcript_kind ?? "");
+            if (transcript.primaryModel !== null || transcript.totalTokens !== null || transcript.skill !== null) {
+                transcriptScanned++;
+            }
+        }
+        // Persona-null rows: the meta write is a no-op exactly when the derived
+        // persona is also NULL and the derived parent matches what a prior run
+        // already stamped (or is equally NULL). Scan-only rows have nothing to
+        // write for meta by definition.
+        const metaIsNoop = meta === null || (meta.persona === null && meta.parentSessionId === row.parent_session_id);
+        // The scan write is COALESCE-guarded, so it changes the row only where a
+        // stored column is NULL and the scan derived a value for it. An all-null
+        // scan (missing/unreadable transcript) or a scan whose derivables are
+        // already stored is a no-op - counting those as "updated" was the run-N
+        // churn bug (updated=N repeating forever with byte-identical results).
+        const transcriptWrites = (row.primary_model === null && transcript.primaryModel !== null) ||
+            (row.total_tokens === null && transcript.totalTokens !== null) ||
+            (row.skill === null && transcript.skill !== null);
+        if (metaIsNoop && !transcriptWrites) {
+            skippedNoop++;
+            continue;
+        }
+        toWrite.push({ id: row.id, meta, transcript });
+    }
+    if (opts.dryRun) {
+        return { total, updated: toWrite.length, skippedAlreadyStamped, skippedNoop, subagentCandidates, unknownParent, transcriptScanned };
+    }
+    const metaAndScanStmt = db.prepare(`UPDATE sessions SET
+       agent_persona = ?, parent_session_id = ?,
+       primary_model = COALESCE(primary_model, ?),
+       total_tokens = COALESCE(total_tokens, ?),
+       skill = COALESCE(skill, ?)
+     WHERE id = ? AND agent_persona IS NULL`);
+    const scanOnlyStmt = db.prepare(`UPDATE sessions SET
+       primary_model = COALESCE(primary_model, ?),
+       total_tokens = COALESCE(total_tokens, ?),
+       skill = COALESCE(skill, ?)
+     WHERE id = ?`);
+    for (let i = 0; i < toWrite.length; i += BATCH_SIZE) {
+        const batch = toWrite.slice(i, i + BATCH_SIZE);
+        const runBatch = db.transaction((rows) => {
+            for (const row of rows) {
+                if (row.meta) {
+                    metaAndScanStmt.run(row.meta.persona, row.meta.parentSessionId, row.transcript.primaryModel, row.transcript.totalTokens, row.transcript.skill, row.id);
+                }
+                else {
+                    scanOnlyStmt.run(row.transcript.primaryModel, row.transcript.totalTokens, row.transcript.skill, row.id);
+                }
+            }
+        });
+        runBatch(batch);
+    }
+    return { total, updated: toWrite.length, skippedAlreadyStamped, skippedNoop, subagentCandidates, unknownParent, transcriptScanned };
+}
+//# sourceMappingURL=backfill-derivables.js.map

package/dist/core/ingest/backfill-derivables.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"backfill-derivables.js","sourceRoot":"","sources":["../../../src/core/ingest/backfill-derivables.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAsDG;AAGH,OAAO,EAAE,kBAAkB,EAAE,MAAM,+BAA+B,CAAC;AACnE,OAAO,EAAE,wBAAwB,EAA6B,MAAM,4BAA4B,CAAC;AAEjG,MAAM,UAAU,GAAG,GAAG,CAAC;AAEvB,MAAM,0BAA0B,GAAyB;IACvD,YAAY,EAAE,IAAI;IAClB,WAAW,EAAE,IAAI;IACjB,KAAK,EAAE,IAAI;CACZ,CAAC;AAqDF,0EAA0E;AAC1E,yEAAyE;AACzE,6EAA6E;AAC7E,0EAA0E;AAC1E,sEAAsE;AACtE,MAAM,kBAAkB,GACtB,4EAA4E;IAC5E,wEAAwE,CAAC;AAE3E,SAAS,gBAAgB,CAAC,EAAY,EAAE,kBAA2B;IACjE,OAAO,EAAE;SACN,OAAO,CACN,uJAAuJ;QACvJ,4CAA4C;QAC5C,CAAC,kBAAkB,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,CAAC,EAAE,CAAC,CAC/C;SACA,GAAG,EAAE,CAAC;AACX,CAAC;AAED,4EAA4E;AAC5E,4EAA4E;AAC5E,uEAAuE;AACvE,2EAA2E;AAC3E,uDAAuD;AACvD,SAAS,eAAe,CAAC,OAAe;IACtC,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IACnC,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AAC1D,CAAC;AAED,SAAS,UAAU,CAAC,GAAiB;IACnC,MAAM,WAAW,GAAG,eAAe,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IACjD,IAAI,WAAW,KAAK,aAAa,EAAE,CAAC;QAClC,OAAO,kBAAkB,CAAC,GAAG,CAAC,kBAAkB,IAAI,EAAE,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC;IACrE,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,WAAW,EAAE,eAAe,EAAE,IAAI,EAAE,CAAC;AACzD,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,EAAY,EAAE,OAAwB,EAAE;IAC/E,MAAM,KAAK,GAAI,EAAE,CAAC,OAAO,CAAC,oCAAoC,CAAC,CAAC,GAAG,EAAoB,CAAC,CAAC,CAAC;IAC1F,MAAM,UAAU,GAAG,gBAAgB,CAAC,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,CAAC;IAC1E,MAAM,qBAAqB,GAAG,KAAK,GAAG,UAAU,CAAC,MAAM,CAAC;IAExD,IAAI,kBAAkB,GAAG,CAAC,CAAC;IAC3B,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,IAAI,WAAW,GAAG,CAAC,CAAC;IACpB,IAAI,iBAAiB,GAAG,CAAC,CAAC;IAC1B,MAAM,OAAO,GAMP,EAAE,CAAC;IAET,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,GAAG,CAAC,aAAa,KAAK,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;QACjE,IAAI,IAAI,EAAE,CAAC;YACT,IAAI,IAAI,CAAC,eAAe,KAAK,IAAI;gBAAE,kBAAkB,EAAE,CAAC;YACxD,IAAI,IAAI,CAAC,eAAe,KAAK,SAAS;gBAAE,aAAa,EAAE,CAAC;QAC1D,CAAC;QAED,uEAAuE;QACvE,sEAAsE;QACtE,mEAAmE;QACnE,6DAA6D;QAC7D,sEAAsE;QACtE,wEAAwE;QACxE,uCAAuC;QACvC,MAAM,cAAc,GAClB,GAAG,CAAC,aAAa,KAAK,IAAI,IAAI,GAAG,CAAC,YAAY,KAAK,IAAI,IAAI,GAAG,CAAC,KAAK,KAAK,IAAI,CAAC;QAChF,IAAI,UAAU,GAAG,0BAA0B,CAAC;QAC5C,IAAI,IAAI,CAAC,kBAAkB,IAAI,GAAG,CAAC,eAAe,IAAI,CAAC,cAAc,EAAE,CAAC;YACtE,UAAU,GAAG,MAAM,wBAAwB,CAAC,GAAG,CAAC,eAAe,EAAE,GAAG,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC;YAC5F,IAAI,UAAU,CAAC,YAAY,KAAK,IAAI,IAAI,UAAU,CAAC,WAAW,KAAK,IAAI,IAAI,UAAU,CAAC,KAAK,KAAK,IAAI,EAAE,CAAC;gBACrG,iBAAiB,EAAE,CAAC;YACtB,CAAC;QACH,CAAC;QAED,wEAAwE;QACxE,uEAAuE;QACvE,uEAAuE;QACvE,gCAAgC;QAChC,MAAM,UAAU,GAAG,IAAI,KAAK,IAAI,IAAI,CAAC,IAAI,CAAC,OAAO,KAAK,IAAI,IAAI,IAAI,CAAC,eAAe,KAAK,GAAG,CAAC,iBAAiB,CAAC,CAAC;QAC9G,yEAAyE;QACzE,yEAAyE;QACzE,sEAAsE;QACtE,wEAAwE;QACxE,uEAAuE;QACvE,MAAM,gBAAgB,GACpB,CAAC,GAAG,CAAC,aAAa,KAAK,IAAI,IAAI,UAAU,CAAC,YAAY,KAAK,IAAI,CAAC;YAChE,CAAC,GAAG,CAAC,YAAY,KAAK,IAAI,IAAI,UAAU,CAAC,WAAW,KAAK,IAAI,CAAC;YAC9D,CAAC,GAAG,CAAC,KAAK,KAAK,IAAI,IAAI,UAAU,CAAC,KAAK,KAAK,IAAI,CAAC,CAAC;QACpD,IAAI,UAAU,IAAI,CAAC,gBAAgB,EAAE,CAAC;YACpC,WAAW,EAAE,CAAC;YACd,SAAS;QACX,CAAC;QACD,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,GAAG,CAAC,EAAE,EAAE,IAAI,EAAE,UAAU,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;QAChB,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,CAAC,MAAM,EAAE,qBAAqB,EAAE,WAAW,EAAE,kBAAkB,EAAE,aAAa,EAAE,iBAAiB,EAAE,CAAC;IACtI,CAAC;IAED,MAAM,eAAe,GAAG,EAAE,CAAC,OAAO,CAChC;;;;;4CAKwC,CACzC,CAAC;IACF,MAAM,YAAY,GAAG,EAAE,CAAC,OAAO,CAC7B;;;;kBAIc,CACf,CAAC;IAEF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,IAAI,UAAU,EAAE,CAAC;QACpD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,CAAC;QAC/C,MAAM,QAAQ,GAAG,EAAE,CAAC,WAAW,CAAC,CAAC,IAAkB,EAAE,EAAE;YACrD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;gBACvB,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;oBACb,eAAe,CAAC,GAAG,CACjB,GAAG,CAAC,IAAI,CAAC,OAAO,EAChB,GAAG,CAAC,IAAI,CAAC,eAAe,EACxB,GAAG,CAAC,UAAU,CAAC,YAAY,EAC3B,GAAG,CAAC,UAAU,CAAC,WAAW,EAC1B,GAAG,CAAC,UAAU,CAAC,KAAK,EACpB,GAAG,CAAC,EAAE,CACP,CAAC;gBACJ,CAAC;qBAAM,CAAC;oBACN,YAAY,CAAC,GAAG,CACd,GAAG,CAAC,UAAU,CAAC,YAAY,EAC3B,GAAG,CAAC,UAAU,CAAC,WAAW,EAC1B,GAAG,CAAC,UAAU,CAAC,KAAK,EACpB,GAAG,CAAC,EAAE,CACP,CAAC;gBACJ,CAAC;YACH,CAAC;QACH,CAAC,CAAC,CAAC;QACH,QAAQ,CAAC,KAAK,CAAC,CAAC;IAClB,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,CAAC,MAAM,EAAE,qBAAqB,EAAE,WAAW,EAAE,kBAAkB,EAAE,aAAa,EAAE,iBAAiB,EAAE,CAAC;AACtI,CAAC"}

package/dist/core/ingest/ingest-session.d.ts

@@ -28,6 +28,11 @@ export interface IngestDeps {
     readonly embedder: LLMClient;
     readonly store: SqliteSessionStore | PgSessionStore;
     readonly factStore?: SqliteFactStore | PgFactStore;
+    /** Provider and model name of the active classifier, for provenance stamping. */
+    readonly classifierDescriptor?: {
+        readonly provider: string;
+        readonly model: string;
+    };
     /** Optional logger — defaults to console.error. */
     readonly log?: (msg: string) => void;
 }

package/dist/core/ingest/ingest-session.js

@@ -8,6 +8,7 @@
  */
 import { createHash } from "node:crypto";
 import { extractFacts } from "../facts/extract-facts.js";
+import { scanTranscriptDerivables } from "../ingest/transcript-derivables.js";
 import { PgSessionStore } from "../storage/pg-session-store.js";
 const BODY_CAP = 200_000;
 const CONFIDENCE_FLOOR = 0.3;
@@ -44,6 +45,14 @@ export async function ingestSession(input, deps) {
             confidence: classification.confidence,
         };
     }
+    // Task #352 phase 2 (Task 5): scan the transcript file (if any) for
+    // primary_model/total_tokens/skill. Webhook pushes carry transcriptKind
+    // "webhook", not "claude-code-jsonl", so this always resolves all-null
+    // today -- wired for parity with the scheduler stamp site and any future
+    // webhook kind that does carry a scannable transcript.
+    const transcriptDerivables = input.transcriptPath
+        ? await scanTranscriptDerivables(input.transcriptPath, "webhook")
+        : { primaryModel: null, totalTokens: null, skill: null };
     const record = {
         id,
         runtime: input.runtime,
@@ -62,6 +71,20 @@ export async function ingestSession(input, deps) {
         entities: classification.entities,
         decisions: classification.decisions,
         openQuestions: classification.open,
+        // Push ingest carries no project directory; the scope backfill command
+        // derives it later from the transcript when evidence exists.
+        scope: null,
+        // Webhook pushes have no adapter chunk to derive subagent lineage from;
+        // stamp the runtime name as persona, matching every non-claude-code
+        // adapter at the scheduler stamp site (#352 phase 2).
+        agentPersona: input.runtime,
+        parentSessionId: null,
+        primaryModel: transcriptDerivables.primaryModel,
+        totalTokens: transcriptDerivables.totalTokens,
+        skill: transcriptDerivables.skill,
+        ...(deps.classifierDescriptor !== undefined
+            ? { classifier: { ...deps.classifierDescriptor, confidence: classification.confidence } }
+            : {}),
     };
     const facts = deps.factStore
         ? extractFacts(classification, id, startedAt)

package/dist/core/ingest/ingest-session.js.map

@@ -1 +1 @@
-{"version":3,"file":"ingest-session.js","sourceRoot":"","sources":["../../../src/core/ingest/ingest-session.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,8BAA8B,CAAC;AAG5D,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AAKnE,MAAM,QAAQ,GAAG,OAAO,CAAC;AACzB,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAgC7B,MAAM,UAAU,eAAe,CAAC,OAAe,EAAE,SAAiB,EAAE,IAAY;IAC9E,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC;SAC9B,MAAM,CAAC,OAAO,CAAC;SACf,MAAM,CAAC,GAAG,CAAC;SACX,MAAM,CAAC,SAAS,CAAC;SACjB,MAAM,CAAC,GAAG,CAAC;SACX,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;SAC5B,MAAM,CAAC,KAAK,CAAC;SACb,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAChB,OAAO,WAAW,IAAI,EAAE,CAAC;AAC3B,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CAAC,KAAkB,EAAE,IAAgB;IACtE,MAAM,SAAS,GAAG,KAAK,CAAC,SAAS,IAAI,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC9D,MAAM,EAAE,GAAG,KAAK,CAAC,EAAE,IAAI,eAAe,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;IAC7E,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1D,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAEtB,IAAI,cAAc,CAAC;IACnB,IAAI,CAAC;QACH,cAAc,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC9D,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,MAAM,KAAK,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACzD,GAAG,CAAC,kCAAkC,EAAE,KAAK,KAAK,EAAE,CAAC,CAAC;QACtD,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,mBAAmB,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,CAAC;IAChF,CAAC;IAED,IAAI,cAAc,CAAC,UAAU,GAAG,gBAAgB,EAAE,CAAC;QACjD,OAAO;YACL,EAAE;YACF,MAAM,EAAE,gBAAgB;YACxB,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;YAC1B,UAAU,EAAE,cAAc,CAAC,UAAU;SACtC,CAAC;IACJ,CAAC;IAED,MAAM,MAAM,GAAiB;QAC3B,EAAE;QACF,OAAO,EAAE,KAAK,CAAC,OAAO;QACtB,gBAAgB,EAAE,KAAK,CAAC,gBAAgB,IAAI,IAAI;QAChD,SAAS;QACT,OAAO,EAAE,KAAK,CAAC,OAAO,IAAI,IAAI;QAC9B,WAAW,EAAE,IAAI;QACjB,KAAK,EAAE,cAAc,CAAC,KAAK;QAC3B,OAAO,EAAE,cAAc,CAAC,OAAO;QAC/B,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;QACnC,MAAM,EAAE,QAAQ;QAChB,cAAc,EAAE,SAAS;QACzB,cAAc,EAAE,KAAK,CAAC,cAAc,IAAI,IAAI;QAC5C,gBAAgB,EAAE,IAAI;QACtB,gBAAgB,EAAE,IAAI;QACtB,QAAQ,EAAE,cAAc,CAAC,QAAQ;QACjC,SAAS,EAAE,cAAc,CAAC,SAAS;QACnC,aAAa,EAAE,cAAc,CAAC,IAAI;KACnC,CAAC;IAEF,MAAM,KAAK,GAAwB,IAAI,CAAC,SAAS;QAC/C,CAAC,CAAC,YAAY,CAAC,cAAc,EAAE,EAAE,EAAE,SAAS,CAAC;QAC7C,CAAC,CAAC,EAAE,CAAC;IAEP,0EAA0E;IAC1E,wEAAwE;IACxE,IAAI,IAAI,CAAC,KAAK,YAAY,cAAc,EAAE,CAAC;QACzC,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EACxD,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,SAAwB,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACjF,CAAC;SAAM,CAAC;QACN,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EACxD,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,SAA4B,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACrF,CAAC;IACD,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,UAAU,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,UAAU,EAAE,cAAc,CAAC,UAAU,EAAE,CAAC;AACvG,CAAC"}
+{"version":3,"file":"ingest-session.js","sourceRoot":"","sources":["../../../src/core/ingest/ingest-session.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,8BAA8B,CAAC;AAC5D,OAAO,EAAE,wBAAwB,EAAE,MAAM,uCAAuC,CAAC;AAGjF,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AAKnE,MAAM,QAAQ,GAAG,OAAO,CAAC;AACzB,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAkC7B,MAAM,UAAU,eAAe,CAAC,OAAe,EAAE,SAAiB,EAAE,IAAY;IAC9E,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC;SAC9B,MAAM,CAAC,OAAO,CAAC;SACf,MAAM,CAAC,GAAG,CAAC;SACX,MAAM,CAAC,SAAS,CAAC;SACjB,MAAM,CAAC,GAAG,CAAC;SACX,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;SAC5B,MAAM,CAAC,KAAK,CAAC;SACb,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAChB,OAAO,WAAW,IAAI,EAAE,CAAC;AAC3B,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CAAC,KAAkB,EAAE,IAAgB;IACtE,MAAM,SAAS,GAAG,KAAK,CAAC,SAAS,IAAI,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC9D,MAAM,EAAE,GAAG,KAAK,CAAC,EAAE,IAAI,eAAe,CAAC,KAAK,CAAC,OAAO,EAAE,SAAS,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;IAC7E,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1D,MAAM,EAAE,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAEtB,IAAI,cAAc,CAAC;IACnB,IAAI,CAAC;QACH,cAAc,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC9D,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,MAAM,KAAK,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACzD,GAAG,CAAC,kCAAkC,EAAE,KAAK,KAAK,EAAE,CAAC,CAAC;QACtD,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,mBAAmB,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,KAAK,EAAE,CAAC;IAChF,CAAC;IAED,IAAI,cAAc,CAAC,UAAU,GAAG,gBAAgB,EAAE,CAAC;QACjD,OAAO;YACL,EAAE;YACF,MAAM,EAAE,gBAAgB;YACxB,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE;YAC1B,UAAU,EAAE,cAAc,CAAC,UAAU;SACtC,CAAC;IACJ,CAAC;IAED,oEAAoE;IACpE,wEAAwE;IACxE,uEAAuE;IACvE,yEAAyE;IACzE,uDAAuD;IACvD,MAAM,oBAAoB,GAAG,KAAK,CAAC,cAAc;QAC/C,CAAC,CAAC,MAAM,wBAAwB,CAAC,KAAK,CAAC,cAAc,EAAE,SAAS,CAAC;QACjE,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;IAE3D,MAAM,MAAM,GAAiB;QAC3B,EAAE;QACF,OAAO,EAAE,KAAK,CAAC,OAAO;QACtB,gBAAgB,EAAE,KAAK,CAAC,gBAAgB,IAAI,IAAI;QAChD,SAAS;QACT,OAAO,EAAE,KAAK,CAAC,OAAO,IAAI,IAAI;QAC9B,WAAW,EAAE,IAAI;QACjB,KAAK,EAAE,cAAc,CAAC,KAAK;QAC3B,OAAO,EAAE,cAAc,CAAC,OAAO;QAC/B,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;QACnC,MAAM,EAAE,QAAQ;QAChB,cAAc,EAAE,SAAS;QACzB,cAAc,EAAE,KAAK,CAAC,cAAc,IAAI,IAAI;QAC5C,gBAAgB,EAAE,IAAI;QACtB,gBAAgB,EAAE,IAAI;QACtB,QAAQ,EAAE,cAAc,CAAC,QAAQ;QACjC,SAAS,EAAE,cAAc,CAAC,SAAS;QACnC,aAAa,EAAE,cAAc,CAAC,IAAI;QAClC,uEAAuE;QACvE,6DAA6D;QAC7D,KAAK,EAAE,IAAI;QACX,wEAAwE;QACxE,oEAAoE;QACpE,sDAAsD;QACtD,YAAY,EAAE,KAAK,CAAC,OAAO;QAC3B,eAAe,EAAE,IAAI;QACrB,YAAY,EAAE,oBAAoB,CAAC,YAAY;QAC/C,WAAW,EAAE,oBAAoB,CAAC,WAAW;QAC7C,KAAK,EAAE,oBAAoB,CAAC,KAAK;QACjC,GAAG,CAAC,IAAI,CAAC,oBAAoB,KAAK,SAAS;YACzC,CAAC,CAAC,EAAE,UAAU,EAAE,EAAE,GAAG,IAAI,CAAC,oBAAoB,EAAE,UAAU,EAAE,cAAc,CAAC,UAAU,EAAE,EAAE;YACzF,CAAC,CAAC,EAAE,CAAC;KACR,CAAC;IAEF,MAAM,KAAK,GAAwB,IAAI,CAAC,SAAS;QAC/C,CAAC,CAAC,YAAY,CAAC,cAAc,EAAE,EAAE,EAAE,SAAS,CAAC;QAC7C,CAAC,CAAC,EAAE,CAAC;IAEP,0EAA0E;IAC1E,wEAAwE;IACxE,IAAI,IAAI,CAAC,KAAK,YAAY,cAAc,EAAE,CAAC;QACzC,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EACxD,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,SAAwB,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACjF,CAAC;SAAM,CAAC;QACN,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EACxD,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,SAA4B,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IACrF,CAAC;IACD,OAAO,EAAE,EAAE,EAAE,MAAM,EAAE,UAAU,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,UAAU,EAAE,cAAc,CAAC,UAAU,EAAE,CAAC;AACvG,CAAC"}

package/dist/core/ingest/reclassify-oversized.d.ts

@@ -16,6 +16,11 @@ export interface ReclassifyDeps {
     readonly embedder: LLMClient;
     readonly classifier: LLMClient;
     readonly adapters: ReadonlyArray<TranscriptAdapter>;
+    /** Provider and model name of the active classifier, for provenance stamping. */
+    readonly classifierDescriptor?: {
+        readonly provider: string;
+        readonly model: string;
+    };
     readonly log?: (msg: string) => void;
 }
 export interface ReclassifyOptions {

package/dist/core/ingest/reclassify-oversized.js

@@ -1,5 +1,8 @@
 import { classifyAdaptive } from "../classifier/hierarchical-classify.js";
 import { extractFacts } from "../facts/extract-facts.js";
+import { deriveScope } from "../scope/derive-scope.js";
+import { loadAliasMap } from "../scope/alias-map.js";
+import { scopeStampEnabled } from "../scope/stamp-flag.js";
 const BODY_CAP = 200_000;
 const CONFIDENCE_FLOOR = 0.3;
 export function selectOversizedFailures(db, limit) {
@@ -78,6 +81,12 @@ export async function reclassifyOversized(deps, opts = {}) {
             entities: classification.entities,
             decisions: classification.decisions,
             openQuestions: classification.open,
+            scope: scopeStampEnabled() && chunk.projectDir
+                ? deriveScope(chunk.projectDir, loadAliasMap())
+                : null,
+            ...(deps.classifierDescriptor !== undefined
+                ? { classifier: { ...deps.classifierDescriptor, confidence: classification.confidence } }
+                : {}),
         };
         const extracted = extractFacts(classification, chunk.id, chunk.startedAt);
         try {

package/dist/core/ingest/reclassify-oversized.js.map

@@ -1 +1 @@
-{"version":3,"file":"reclassify-oversized.js","sourceRoot":"","sources":["../../../src/core/ingest/reclassify-oversized.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,gBAAgB,EAAE,MAAM,2CAA2C,CAAC;AAC7E,OAAO,EAAE,YAAY,EAAE,MAAM,8BAA8B,CAAC;AAM5D,MAAM,QAAQ,GAAG,OAAO,CAAC;AACzB,MAAM,gBAAgB,GAAG,GAAG,CAAC;AA4B7B,MAAM,UAAU,uBAAuB,CACrC,EAAY,EACZ,KAAc;IAEd,MAAM,GAAG,GACP,sDAAsD;QACtD,yEAAyE;QACzE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,UAAU,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAC7D,OAAO,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,GAAG,EAA0D,CAAC;AACvF,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,IAAoB,EACpB,OAA0B,EAAE;IAE5B,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9D,MAAM,IAAI,GAAG,uBAAuB,CAAC,IAAI,CAAC,EAAE,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;IAE1D,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,oBAAoB,GAAG,CAAC,CAAC;IAC7B,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,IAAI,WAAW,GAAG,CAAC,CAAC;IACpB,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;QAC7C,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,+BAA+B,GAAG,CAAC,YAAY,EAAE,CAAC,CAAC;YACvD,SAAS;QACX,CAAC;QAED,IAAI,KAA0B,CAAC;QAC/B,IAAI,CAAC;YACH,KAAK,GAAG,MAAM,OAAO,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;QACtD,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,WAAW,EAAE,CAAC;YACd,GAAG,CAAC,6BAA6B,GAAG,CAAC,WAAW,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACnG,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,WAAW,EAAE,CAAC;YACd,SAAS;QACX,CAAC;QAED,IAAI,cAAc,CAAC;QACnB,IAAI,CAAC;YACH,cAAc,GAAG,MAAM,gBAAgB,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC;QACvE,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,gCAAgC,KAAK,CAAC,EAAE,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAC/F,SAAS;QACX,CAAC;QAED,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,oEAAoE;YACpE,gEAAgE;YAChE,SAAS;QACX,CAAC;QAED,IAAI,cAAc,CAAC,UAAU,GAAG,gBAAgB,EAAE,CAAC;YACjD,oBAAoB,EAAE,CAAC;YACvB,IAAI,CAAC,EAAE;iBACJ,OAAO,CAAC,uFAAuF,CAAC;iBAChG,GAAG,CAAC,GAAG,CAAC,YAAY,EAAE,GAAG,CAAC,WAAW,CAAC,CAAC;YAC1C,SAAS;QACX,CAAC;QAED,MAAM,MAAM,GAAiB;YAC3B,EAAE,EAAE,KAAK,CAAC,EAAE;YACZ,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,gBAAgB,EAAE,KAAK,CAAC,gBAAgB;YACxC,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,WAAW,EAAE,KAAK,CAAC,WAAW;YAC9B,KAAK,EAAE,cAAc,CAAC,KAAK;YAC3B,OAAO,EAAE,cAAc,CAAC,OAAO;YAC/B,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;YACnC,MAAM,EAAE,QAAQ;YAChB,cAAc,EAAE,OAAO,CAAC,cAAc;YACtC,cAAc,EAAE,GAAG,CAAC,WAAW;YAC/B,gBAAgB,EAAE,IAAI;YACtB,gBAAgB,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM;YACnC,QAAQ,EAAE,cAAc,CAAC,QAAQ;YACjC,SAAS,EAAE,cAAc,CAAC,SAAS;YACnC,aAAa,EAAE,cAAc,CAAC,IAAI;SACnC,CAAC;QAEF,MAAM,SAAS,GAAG,YAAY,CAAC,cAAc,EAAE,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAE1E,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EAAE;gBAC1D,SAAS,EAAE,IAAI,CAAC,SAAS;gBACzB,KAAK,EAAE,SAAS;aACjB,CAAC,CAAC;YACH,IAAI,CAAC,EAAE;iBACJ,OAAO,CACN,sFAAsF;gBACtF,4CAA4C,CAC7C;iBACA,GAAG,CAAC,KAAK,CAAC,EAAE,EAAE,GAAG,CAAC,YAAY,EAAE,GAAG,CAAC,WAAW,CAAC,CAAC;YACpD,QAAQ,EAAE,CAAC;YACX,QAAQ,IAAI,cAAc,CAAC,QAAQ,CAAC,MAAM,CAAC;YAC3C,SAAS,IAAI,cAAc,CAAC,SAAS,CAAC,MAAM,CAAC;YAC7C,KAAK,IAAI,SAAS,CAAC,MAAM,CAAC;QAC5B,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,8BAA8B,KAAK,CAAC,EAAE,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAC/F,CAAC;IACH,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,MAAM,EAAE,QAAQ,EAAE,oBAAoB,EAAE,MAAM,EAAE,WAAW,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;AACrH,CAAC"}
+{"version":3,"file":"reclassify-oversized.js","sourceRoot":"","sources":["../../../src/core/ingest/reclassify-oversized.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,gBAAgB,EAAE,MAAM,2CAA2C,CAAC;AAC7E,OAAO,EAAE,YAAY,EAAE,MAAM,8BAA8B,CAAC;AAC5D,OAAO,EAAE,WAAW,EAAE,MAAM,6BAA6B,CAAC;AAC1D,OAAO,EAAE,YAAY,EAAE,MAAM,0BAA0B,CAAC;AACxD,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAM9D,MAAM,QAAQ,GAAG,OAAO,CAAC;AACzB,MAAM,gBAAgB,GAAG,GAAG,CAAC;AA8B7B,MAAM,UAAU,uBAAuB,CACrC,EAAY,EACZ,KAAc;IAEd,MAAM,GAAG,GACP,sDAAsD;QACtD,yEAAyE;QACzE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,UAAU,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAC7D,OAAO,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,GAAG,EAA0D,CAAC;AACvF,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,IAAoB,EACpB,OAA0B,EAAE;IAE5B,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC,CAAS,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9D,MAAM,IAAI,GAAG,uBAAuB,CAAC,IAAI,CAAC,EAAE,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;IAE1D,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,oBAAoB,GAAG,CAAC,CAAC;IAC7B,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,IAAI,WAAW,GAAG,CAAC,CAAC;IACpB,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;QAC7C,IAAI,CAAC,OAAO,EAAE,CAAC;YACb,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,+BAA+B,GAAG,CAAC,YAAY,EAAE,CAAC,CAAC;YACvD,SAAS;QACX,CAAC;QAED,IAAI,KAA0B,CAAC;QAC/B,IAAI,CAAC;YACH,KAAK,GAAG,MAAM,OAAO,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;QACtD,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,WAAW,EAAE,CAAC;YACd,GAAG,CAAC,6BAA6B,GAAG,CAAC,WAAW,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACnG,SAAS;QACX,CAAC;QACD,IAAI,CAAC,KAAK,EAAE,CAAC;YACX,WAAW,EAAE,CAAC;YACd,SAAS;QACX,CAAC;QAED,IAAI,cAAc,CAAC;QACnB,IAAI,CAAC;YACH,cAAc,GAAG,MAAM,gBAAgB,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC;QACvE,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,gCAAgC,KAAK,CAAC,EAAE,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YAC/F,SAAS;QACX,CAAC;QAED,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,oEAAoE;YACpE,gEAAgE;YAChE,SAAS;QACX,CAAC;QAED,IAAI,cAAc,CAAC,UAAU,GAAG,gBAAgB,EAAE,CAAC;YACjD,oBAAoB,EAAE,CAAC;YACvB,IAAI,CAAC,EAAE;iBACJ,OAAO,CAAC,uFAAuF,CAAC;iBAChG,GAAG,CAAC,GAAG,CAAC,YAAY,EAAE,GAAG,CAAC,WAAW,CAAC,CAAC;YAC1C,SAAS;QACX,CAAC;QAED,MAAM,MAAM,GAAiB;YAC3B,EAAE,EAAE,KAAK,CAAC,EAAE;YACZ,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,gBAAgB,EAAE,KAAK,CAAC,gBAAgB;YACxC,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,OAAO,EAAE,KAAK,CAAC,OAAO;YACtB,WAAW,EAAE,KAAK,CAAC,WAAW;YAC9B,KAAK,EAAE,cAAc,CAAC,KAAK;YAC3B,OAAO,EAAE,cAAc,CAAC,OAAO;YAC/B,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC;YACnC,MAAM,EAAE,QAAQ;YAChB,cAAc,EAAE,OAAO,CAAC,cAAc;YACtC,cAAc,EAAE,GAAG,CAAC,WAAW;YAC/B,gBAAgB,EAAE,IAAI;YACtB,gBAAgB,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM;YACnC,QAAQ,EAAE,cAAc,CAAC,QAAQ;YACjC,SAAS,EAAE,cAAc,CAAC,SAAS;YACnC,aAAa,EAAE,cAAc,CAAC,IAAI;YAClC,KAAK,EAAE,iBAAiB,EAAE,IAAI,KAAK,CAAC,UAAU;gBAC5C,CAAC,CAAC,WAAW,CAAC,KAAK,CAAC,UAAU,EAAE,YAAY,EAAE,CAAC;gBAC/C,CAAC,CAAC,IAAI;YACR,GAAG,CAAC,IAAI,CAAC,oBAAoB,KAAK,SAAS;gBACzC,CAAC,CAAC,EAAE,UAAU,EAAE,EAAE,GAAG,IAAI,CAAC,oBAAoB,EAAE,UAAU,EAAE,cAAc,CAAC,UAAU,EAAE,EAAE;gBACzF,CAAC,CAAC,EAAE,CAAC;SACR,CAAC;QAEF,MAAM,SAAS,GAAG,YAAY,CAAC,cAAc,EAAE,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAE1E,IAAI,CAAC;YACH,MAAM,IAAI,CAAC,KAAK,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EAAE;gBAC1D,SAAS,EAAE,IAAI,CAAC,SAAS;gBACzB,KAAK,EAAE,SAAS;aACjB,CAAC,CAAC;YACH,IAAI,CAAC,EAAE;iBACJ,OAAO,CACN,sFAAsF;gBACtF,4CAA4C,CAC7C;iBACA,GAAG,CAAC,KAAK,CAAC,EAAE,EAAE,GAAG,CAAC,YAAY,EAAE,GAAG,CAAC,WAAW,CAAC,CAAC;YACpD,QAAQ,EAAE,CAAC;YACX,QAAQ,IAAI,cAAc,CAAC,QAAQ,CAAC,MAAM,CAAC;YAC3C,SAAS,IAAI,cAAc,CAAC,SAAS,CAAC,MAAM,CAAC;YAC7C,KAAK,IAAI,SAAS,CAAC,MAAM,CAAC;QAC5B,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,MAAM,EAAE,CAAC;YACT,GAAG,CAAC,8BAA8B,KAAK,CAAC,EAAE,KAAK,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAC/F,CAAC;IACH,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,MAAM,EAAE,QAAQ,EAAE,oBAAoB,EAAE,MAAM,EAAE,WAAW,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC;AACrH,CAAC"}

package/dist/core/ingest/reprocess.d.ts

@@ -0,0 +1,96 @@
+/**
+ * reprocess: retroactive re-classification of prior sessions under a
+ * stronger configured lane.
+ *
+ * Selection: sessions with a non-empty body whose classifier_model is NULL,
+ * differs from the current lane model, or (when --min-confidence is given)
+ * whose stored confidence falls below the threshold even for the same model.
+ *
+ * Per session: classifyAdaptive on the stored body, then a full insertSession
+ * upsert (refreshes chunks + facts). A below-floor classification updates the
+ * session row (label, summary, markers, provenance) but preserves prior facts
+ * and their embeddings unchanged. It is counted in belowFloorOverwrites.
+ * Workstream binding (workstream_id) is NOT touched by the upsert.
+ *
+ * Resumable via a JSON state file (default ~/.nlm/reprocess.state). A lane
+ * change (different provider or model) invalidates the done-set.
+ *
+ * SQLite only. Pg variant deferred; file task against nlm#TODO in report.
+ */
+import type { Database } from "better-sqlite3";
+import type { EmbeddingConfigStore } from "../embedding/embedding-config.js";
+import type { SqliteFactStore } from "../storage/sqlite-fact-store.js";
+import type { SqliteSessionStore } from "../storage/sqlite-session-store.js";
+import type { LLMClient } from "../../ports/llm-client.js";
+export interface ReprocessDeps {
+    readonly db: Database;
+    readonly store: SqliteSessionStore;
+    readonly factStore: SqliteFactStore;
+    readonly embedder: LLMClient;
+    readonly classifier: LLMClient;
+    readonly classifierDescriptor: {
+        readonly provider: string;
+        readonly model: string;
+    };
+    readonly embeddingConfig?: EmbeddingConfigStore;
+    readonly embedderDescriptor?: {
+        readonly provider: string;
+        readonly model: string;
+    };
+    readonly log?: (msg: string) => void;
+}
+export interface ReprocessOptions {
+    readonly limit?: number;
+    readonly dryRun?: boolean;
+    readonly minConfidence?: number;
+    readonly onlyNull?: boolean;
+    readonly excludeModels?: string[];
+    readonly statePath?: string;
+    readonly classifyTimeoutMs?: number;
+    readonly verbose?: boolean;
+    readonly forceEmbed?: boolean;
+    readonly onProgress?: (i: number, n: number, sid: string, status: string) => void;
+}
+export interface CohortGroup {
+    readonly classifierProvider: string | null;
+    readonly classifierModel: string | null;
+    readonly confidenceBand: string;
+    readonly count: number;
+}
+export interface ReprocessReport {
+    readonly totalEligible: number;
+    readonly processed: number;
+    readonly succeeded: number;
+    readonly failed: number;
+    readonly skippedAlreadyDone: number;
+    readonly limitSkipped: number;
+    readonly belowFloorOverwrites: number;
+    readonly meanConfidenceOld: number | null;
+    readonly meanConfidenceNew: number | null;
+    readonly cohort?: ReadonlyArray<CohortGroup>;
+    readonly stateNote?: string;
+}
+interface ReprocessSessionRow {
+    readonly id: string;
+    readonly runtime: string;
+    readonly runtime_session_id: string | null;
+    readonly started_at: string;
+    readonly ended_at: string | null;
+    readonly duration_min: number | null;
+    readonly label: string;
+    readonly summary: string;
+    readonly status: string;
+    readonly transcript_kind: string | null;
+    readonly transcript_path: string | null;
+    readonly transcript_offset: number | null;
+    readonly transcript_length: number | null;
+    readonly classifier_provider: string | null;
+    readonly classifier_model: string | null;
+    readonly classifier_confidence: number | null;
+}
+export declare function selectReprocessCandidates(db: Database, model: string, minConfidence?: number, opts?: {
+    onlyNull?: boolean;
+    excludeModels?: string[];
+}): ReprocessSessionRow[];
+export declare function reprocess(deps: ReprocessDeps, opts?: ReprocessOptions): Promise<ReprocessReport>;
+export {};