pi-mega-compact 0.8.26 → 0.9.1

package/src/store/vectorIndex.ts

@@ -40,6 +40,8 @@ export interface VectorIndexHit {
   score: number;
 }
 
+import { withOpenTimeout } from "./pgOpenGuard.js";
+
 let db: PGliteInstance | undefined;
 let initPromise: Promise<PGliteInstance | undefined> | undefined;
 let disabled = false;
@@ -131,12 +133,19 @@ async function openPgLite(
     if (!mod) return undefined;
     const dir = indexDir();
     mkdirSync(dir, { recursive: true });
-    const pg = await new mod.PGlite({
-      dataDir: dir,
-      extensions: { vector: mod.vector },
-    });
-    await pg.exec("CREATE EXTENSION IF NOT EXISTS vector;");
-    await pg.exec(`
+    // Bounded open: PGlite is single-writer over a shared dataDir, so a second
+    // pi process on the same dir can block here forever. Without the ceiling the
+    // never-settling promise gets cached in initPromise and every later caller
+    // awaits it — which is how a stalled index wedged a whole pi turn.
+    let openTimedOut = false;
+    const pg = await withOpenTimeout(
+      (async () => {
+        const inst = await new mod.PGlite({
+          dataDir: dir,
+          extensions: { vector: mod.vector },
+        });
+        await inst.exec("CREATE EXTENSION IF NOT EXISTS vector;");
+        await inst.exec(`
       CREATE TABLE IF NOT EXISTS vector_index (
         repo_id       TEXT NOT NULL,
         session_id    TEXT NOT NULL,
@@ -145,10 +154,27 @@ async function openPgLite(
         PRIMARY KEY (repo_id, session_id, checkpoint_id)
       );
     `);
-    // HNSW index over cosine distance for fast NN. Created idempotently.
-    await pg.exec(
-      "CREATE INDEX IF NOT EXISTS vector_index_hnsw ON vector_index USING hnsw (embedding vector_cosine_ops);",
+        // HNSW index over cosine distance for fast NN. Created idempotently.
+        await inst.exec(
+          "CREATE INDEX IF NOT EXISTS vector_index_hnsw ON vector_index USING hnsw (embedding vector_cosine_ops);",
+        );
+        return inst;
+      })(),
+      (reason) => {
+        openTimedOut = true;
+        logWarn(`init ${reason}`);
+      },
     );
+    if (!pg) {
+      if (openTimedOut) {
+        // Don't leave the dead open cached, and don't retry on the next call —
+        // a contended dataDir would just burn another full timeout per caller.
+        // Same terminal state as any other init failure: fall back to the scan.
+        initPromise = undefined;
+        disabled = true;
+      }
+      return undefined;
+    }
     db = pg;
     return pg;
   } catch (err) {

package/src/vector-search.ts

@@ -14,6 +14,7 @@
 import { cosineSimilarity } from "./embedder.js";
 import { normalizeSessionId } from "./store.js";
 import { mmrRerank, type MmrItem } from "./dedup/mmr.js";
+import { stateDirForRepo } from "./store/repoKey.js";
 import { topK } from "./dedup/topk.js";
 import {
 	listCheckpoints,
@@ -320,13 +321,17 @@ export async function vectorSearchAsync(
 		// Index empty/unavailable → synchronous per-session fallback (this repo).
 		return vectorSearch(store, sid, query, k);
 	}
-	// Hydrate each index hit from the authoritative node:sqlite store. repoId is
-	// that repo's stateDir, so cross-repo hits resolve against their own store.
-	// Tag cross-repo hits with their source repoId so the recall block can label
-	// them ("from repo <name>"); same-repo hits stay unlabeled.
+	// Hydrate each index hit from the authoritative node:sqlite store. The
+	// index keys on repo_id (S25: git root via repoKey; legacy rows keyed by
+	// stateDir) — resolve repo_id → stateDir via stateDirForRepo, and skip
+	// unresolvable/foreign hits (degrade, never crash). Cross-repo hits carry
+	// their source repoId so the recall block can label them; same-repo stays
+	// unlabeled.
 	const hydrated: SearchHit[] = [];
 	for (const h of indexHits) {
-		const cp = getCheckpoint(h.sessionId, h.checkpointId, h.repoId);
+		const hitStateDir = stateDirForRepo(h.repoId);
+		if (!hitStateDir) continue;
+		const cp = getCheckpoint(h.sessionId, h.checkpointId, hitStateDir);
 		if (cp && cp.dedupStatus !== "removed") {
 			const crossRepo =
 				opts.crossRepo && selfRepo && h.repoId && h.repoId !== selfRepo;

package/src/vectorStore.ts

@@ -17,6 +17,7 @@ import {
 	type DedupTier,
 } from "./config/dedup.js";
 import { logDecision } from "./monitoring.js";
+import { repoKey } from "./store/repoKey.js";
 import type { StoredCheckpoint } from "./store.js";
 import { getStateDir, normalizeSessionId, compressSmart } from "./store.js";
 import { computeContentDigest } from "./dedup/digest.js";
@@ -149,7 +150,9 @@ export class VectorStore {
 	) {
 		this.embedder = opts.embedder ?? defaultEmbedder();
 		this.stateDir = opts.stateDir ?? getStateDir();
-		this.repoId = opts.repoId ?? this.stateDir;
+		// S25: single repo-scope key shared with the memory index (git-root
+		// scoped; falls back to stateDir outside git).
+		this.repoId = opts.repoId ?? repoKey(this.stateDir);
 		// Sprint 14: all tier flags/thresholds flow from the single config source
 		// (DedupConfig). The legacy opts.dedupSim / opts.l2Enabled remain accepted
 		// for backward-compat callers but flags are authoritative via `cfg`.