@lh8ppl/claude-memory-kit 0.2.4 → 0.3.1

package/src/repair.mjs

@@ -159,12 +159,25 @@ function repairLocks({ projectRoot, userDir, staleLockMs, now, ts }) {
  * @param {Function} [opts.reindexer]  test-injected reindex function; defaults to import('./index-rebuild.mjs').reindexFull
  */
 async function repairIndex({ projectRoot, userDir, reindexer }) {
+  // Production reindexFull requires a `db` (it calls db.exec) — repairIndex
+  // must open + own + close it, exactly like runReindex does. The earlier
+  // code called reindexFull({projectRoot,userDir}) with NO db, so
+  // `cmk repair --index`/`--all` threw "undefined (reading 'exec')" since
+  // Task 49 (cut-gate v0.3.1 finding — every test mocked the reindexer, so
+  // the real call-shape was never exercised). An injected reindexer (tests)
+  // takes whatever args it wants; we only open a db for the real one.
   let reindexFn = reindexer;
+  let db = null;
   if (!reindexFn) {
-    const mod = await import('./index-rebuild.mjs');
-    reindexFn = mod.reindexFull;
+    const [{ reindexFull }, { openIndexDb }] = await Promise.all([
+      import('./index-rebuild.mjs'),
+      import('./index-db.mjs'),
+    ]);
+    reindexFn = reindexFull;
+    db = openIndexDb({ projectRoot });
   }
   if (typeof reindexFn !== 'function') {
+    if (db) db.close();
     return {
       kind: 'index',
       changed: false,
@@ -172,7 +185,9 @@ async function repairIndex({ projectRoot, userDir, reindexer }) {
     };
   }
   try {
-    const r = await reindexFn({ projectRoot, userDir });
+    const r = db
+      ? await reindexFn({ projectRoot, userDir, db })
+      : await reindexFn({ projectRoot, userDir });
     return {
       kind: 'index',
       changed: true,
@@ -184,6 +199,8 @@ async function repairIndex({ projectRoot, userDir, reindexer }) {
       changed: false,
       error: `reindex failed: ${err?.message ?? err}`,
     };
+  } finally {
+    if (db) db.close();
   }
 }
 

package/src/result-shapes.mjs

@@ -104,7 +104,7 @@ export const ERROR_CATEGORIES = Object.freeze({
   POISON_GUARD: 'poison_guard',
 
   // `cmk search` requested --mode=semantic or --mode=hybrid but the
-  // Layer-5b semantic backend is not yet shipped (Task 30, design
+  // Layer-5b semantic backend's optional embedder is not installed (Task 30/65, design
   // §9.3). Pairs with `process.exitCode = 2` in subcommands.mjs per
   // tasks.md 30.2's explicit "exit 2 when unavailable" contract.
   // NO silent fallback to keyword — the user asked for semantic,

package/src/scratchpad.mjs

@@ -180,7 +180,9 @@ function findSectionRange(lines, sectionTitle) {
 }
 
 function insertIntoSection(text, sectionTitle, bullet) {
-  const lines = text.split('\n');
+  // Task 139 (D-126): CRLF-tolerant read; the join below re-emits \n,
+  // so a CRLF-converted scratchpad self-heals on the next write.
+  const lines = text.split(/\r?\n/);
   const range = findSectionRange(lines, sectionTitle);
   if (!range) return null;
   // Insert before the next `## ` heading; skip trailing blank lines so the
@@ -208,7 +210,7 @@ function insertIntoSection(text, sectionTitle, bullet) {
 export function ensureSectionExists(scratchpadPath, sectionTitle) {
   if (!existsSync(scratchpadPath)) return { created: false, error: 'no-file' };
   const text = readFileSync(scratchpadPath, 'utf8');
-  if (findSectionRange(text.split('\n'), sectionTitle)) return { created: false };
+  if (findSectionRange(text.split(/\r?\n/), sectionTitle)) return { created: false }; // Task 139: CRLF-tolerant
   const body = text.trimEnd(); // drop trailing whitespace/blank lines (no `\s+$` regex — trips ReDoS heuristics)
   // No leading blank lines for an empty/whitespace-only file (the scaffolded
   // scratchpads are never empty, but keep the output clean if one ever is).
@@ -220,7 +222,7 @@ export function ensureSectionExists(scratchpadPath, sectionTitle) {
 const EVICTED_ID_RE = /^- \(([PUL]-[A-Za-z0-9]+)\)/;
 
 function consolidate(text, { nowDate }) {
-  const lines = text.split('\n');
+  const lines = text.split(/\r?\n/); // Task 139: CRLF-tolerant
   const removeIdx = new Set();
   const evicted = [];
   const staleCutoff = new Date(nowDate.getTime() - STALE_AFTER_DAYS * 24 * 60 * 60 * 1000);

package/src/search.mjs

@@ -11,7 +11,7 @@
 //             ~100ms for 10k bullets. Always available — the keyword
 //             backend ships in v0.1.0 with no extra install.
 //
-//   semantic  the Layer-5b semantic backend (not yet shipped — the embedded
+//   semantic  the Layer-5b semantic backend (Task 65: sqlite-vec + local ONNX embedder; the embedded
 //             vector backend is a future release; the DI seam below is the
 //             drop-in point). Until then this mode errors with
 //             ERROR_CATEGORIES.SEMANTIC_UNAVAILABLE when the caller
@@ -54,6 +54,15 @@ export const SEARCH_MODES = Object.freeze({
 export const DEFAULT_LIMIT = 20;
 const MAX_LIMIT = 1000;
 
+// Task 104.2 (D-117) — search scopes. 'facts' = the curated observation
+// index (L1, the default). 'transcripts' = the SEPARATE raw-transcript
+// chunk index (the L3 last-resort tier) — reached ONLY when explicitly
+// asked, so raw history never pollutes curated results.
+export const SEARCH_SCOPES = Object.freeze({
+  FACTS: 'facts',
+  TRANSCRIPTS: 'transcripts',
+});
+
 const TRUST_ORDINAL = Object.freeze({
   low: 1,
   medium: 2,
@@ -107,7 +116,24 @@ function validateInput(opts) {
       errors.push(`limit: must be a positive integer ≤ ${MAX_LIMIT}`);
     }
   }
-  return { errors, mode };
+  const scope = opts.scope ?? SEARCH_SCOPES.FACTS;
+  if (scope !== SEARCH_SCOPES.FACTS && scope !== SEARCH_SCOPES.TRANSCRIPTS) {
+    errors.push(`scope: must be one of facts/transcripts (got ${JSON.stringify(scope)})`);
+  }
+  if (scope === SEARCH_SCOPES.TRANSCRIPTS) {
+    // Chunks carry no tier/trust/created_at — rejecting these is more honest
+    // than silently ignoring them (the explicit-vs-configured asymmetry rule).
+    for (const [key, label] of [
+      ['tier', 'tier'],
+      ['minTrust', 'minTrust'],
+      ['since', 'since'],
+    ]) {
+      if (opts[key] !== undefined) {
+        errors.push(`${label}: not supported under the transcripts scope (raw chunks carry no ${label})`);
+      }
+    }
+  }
+  return { errors, mode, scope };
 }
 
 // --- Keyword (FTS5 BM25) backend --------------------------------------
@@ -211,6 +237,60 @@ function runKeywordSearch(db, opts) {
   }));
 }
 
+// --- Transcript-scope keyword backend (Task 104.2, the L3 raw tier) ----
+
+const TRANSCRIPT_KEYWORD_SQL = `
+SELECT
+  t.source_file AS source_file,
+  t.source_line AS source_line,
+  t.heading AS heading,
+  transcript_chunks_fts.rank AS score,
+  snippet(transcript_chunks_fts, 0, '<b>', '</b>', '...', 16) AS snippet
+FROM transcript_chunks_fts
+JOIN transcript_chunks t ON t.rowid = transcript_chunks_fts.rowid
+WHERE transcript_chunks_fts MATCH @query
+ORDER BY transcript_chunks_fts.rank
+LIMIT @limit
+`;
+
+// Synthetic, readable id for a raw chunk (chunks are locations, not curated
+// facts — no [PUL]-XXXXXXXX identity). Also the RRF fusion key in hybrid
+// mode and the drill-back handle the memory-search skill surfaces.
+function transcriptHitId(row) {
+  return `T:${row.source_file}:${row.source_line}`;
+}
+
+function runTranscriptKeywordSearch(db, opts) {
+  let rows;
+  try {
+    rows = db
+      .prepare(TRANSCRIPT_KEYWORD_SQL)
+      .all({ query: opts.query, limit: opts.limit ?? DEFAULT_LIMIT });
+  } catch (err) {
+    if (err?.code === 'SQLITE_ERROR' || /fts5:|no such column:/i.test(err?.message ?? '')) {
+      throw new FTS5ParseError(err, opts.query);
+    }
+    throw err;
+  }
+  return rows.map((r) => ({
+    id: transcriptHitId(r),
+    // Raw turns contain newlines (dialogue + Tools blocks) — flatten so the
+    // one-line-per-hit output contract holds across scopes.
+    snippet: flattenSnippet(r.snippet),
+    source_file: r.source_file,
+    source_line: r.source_line,
+    heading: r.heading,
+    score: r.score,
+  }));
+}
+
+const TRANSCRIPT_SNIPPET_MAX = 240;
+
+function flattenSnippet(s) {
+  const flat = String(s ?? '').replace(/\s+/g, ' ').trim();
+  return flat.length > TRANSCRIPT_SNIPPET_MAX ? flat.slice(0, TRANSCRIPT_SNIPPET_MAX) + '…' : flat;
+}
+
 // --- Reciprocal-rank fusion (hybrid mode) -----------------------------
 
 /**
@@ -255,10 +335,15 @@ export function reciprocalRankFusion({
 // --- Public boundary --------------------------------------------------
 
 export function search(opts = {}) {
-  const { errors, mode } = validateInput(opts);
+  const { errors, mode, scope } = validateInput(opts);
   if (errors.length > 0) {
     return errorResult({ category: ERROR_CATEGORIES.SCHEMA, errors });
   }
+  // Scope dispatch (Task 104.2): the transcripts scope swaps the keyword
+  // backend; semantic/hybrid use the caller-prepared backend exactly like
+  // the facts scope (prepareSemanticBackend({scope}) embeds the right table).
+  const keywordBackend =
+    scope === SEARCH_SCOPES.TRANSCRIPTS ? runTranscriptKeywordSearch : runKeywordSearch;
 
   // Semantic + hybrid require an injected backend. Production v0.1.0
   // passes undefined → error with the not-yet-shipped hint. A future
@@ -268,7 +353,8 @@ export function search(opts = {}) {
       return errorResult({
         category: ERROR_CATEGORIES.SEMANTIC_UNAVAILABLE,
         errors: [
-          'the Layer-5b semantic backend is not yet shipped — semantic/hybrid search will land in a future release. ' +
+          'no semantic backend provided — semantic/hybrid need the embedded Layer-5b backend prepared by the caller ' +
+            '(the CLI/MCP do this automatically when the optional @huggingface/transformers embedder is installed). ' +
             'Use --mode=keyword for the always-available FTS5 search.',
         ],
       });
@@ -278,15 +364,16 @@ export function search(opts = {}) {
   let results;
   try {
     if (mode === SEARCH_MODES.KEYWORD) {
-      results = runKeywordSearch(opts.db, opts);
+      results = keywordBackend(opts.db, opts);
     } else if (mode === SEARCH_MODES.SEMANTIC) {
       // The semantic backend is an injected callable returning the same
-      // shape as runKeywordSearch (array of {id, snippet, source_file,
-      // source_line, tier, trust, score}).
+      // shape as the scope's keyword backend (facts: {id, snippet,
+      // source_file, source_line, tier, trust, score}; transcripts: the
+      // synthetic-T:-id shape without tier/trust).
       results = opts.semanticBackend(opts);
     } else {
       // hybrid: run both backends + fuse.
-      const keywordResults = runKeywordSearch(opts.db, opts);
+      const keywordResults = keywordBackend(opts.db, opts);
       const semanticResults = opts.semanticBackend(opts);
       const fused = reciprocalRankFusion({
         keywordResults,
@@ -308,5 +395,5 @@ export function search(opts = {}) {
     throw err;
   }
 
-  return { action: 'found', mode, results };
+  return { action: 'found', mode, scope, results };
 }