agentic-qe 3.13.11 → 3.14.0

package/dist/learning/sqlite-persistence.js

@@ -16,7 +16,8 @@ import { v4 as uuidv4 } from 'uuid';
 import { safeJsonParse } from '../shared/safe-json.js';
 import { toErrorMessage } from '../shared/error-utils.js';
 import { getUnifiedMemory } from '../kernel/unified-memory.js';
-import { computeBatchEmbeddings, getEmbeddingDimension, isUsingEndpoint, } from './real-embeddings.js';
+import { computeBatchEmbeddings, getEmbeddingDimension, isUsingEndpoint, getActiveEmbeddingSpaceIdentity, } from './real-embeddings.js';
+import { EMBEDDING_SPACE_CANARY, inspectEmbeddingSpaceRows, } from './embedding-space.js';
 import { recordPatternUsage } from './pattern-usage-recorder.js';
 export const DEFAULT_SQLITE_CONFIG = {
     // LEGACY: Ignored when useUnified=true (the default). All data goes to memory.db
@@ -106,6 +107,8 @@ export class SQLitePatternStore {
             }
             // Run deduplication migration (one-time, idempotent)
             this.deduplicatePatterns();
+            // Legacy rows remain NULL/unverified; only new writes receive provenance.
+            this.ensureEmbeddingSpaceSchema();
             // Prepare statements
             this.prepareStatements();
             this.initialized = true;
@@ -143,11 +146,12 @@ export class SQLitePatternStore {
       );
 
       -- Pattern embeddings table (BLOB storage for vectors)
-      CREATE TABLE IF NOT EXISTS qe_pattern_embeddings (
+      CREATE TABLE IF NOT EXISTS qe_pattern_embeddings (
         pattern_id TEXT PRIMARY KEY,
         embedding BLOB NOT NULL,
         dimension INTEGER NOT NULL,
-        model TEXT DEFAULT 'all-MiniLM-L6-v2',
+        model TEXT DEFAULT 'all-MiniLM-L6-v2',
+        space_id TEXT,
         created_at TEXT DEFAULT (datetime('now')),
         FOREIGN KEY (pattern_id) REFERENCES qe_patterns(id) ON DELETE CASCADE
       );
@@ -298,6 +302,15 @@ export class SQLitePatternStore {
     /**
      * Prepare commonly used statements
      */
+    ensureEmbeddingSpaceSchema() {
+        if (!this.db)
+            throw new Error('Database not initialized');
+        const columns = this.db.prepare("PRAGMA table_info('qe_pattern_embeddings')").all();
+        if (!columns.some((column) => column.name === 'space_id')) {
+            this.db.exec('ALTER TABLE qe_pattern_embeddings ADD COLUMN space_id TEXT');
+        }
+        this.db.exec('CREATE INDEX IF NOT EXISTS idx_qe_pattern_embeddings_space ON qe_pattern_embeddings(space_id)');
+    }
     prepareStatements() {
         if (!this.db)
             throw new Error('Database not initialized');
@@ -323,14 +336,14 @@ export class SQLitePatternStore {
         updated_at = datetime('now')
     `));
         this.prepared.set('insertEmbedding', this.db.prepare(`
-      INSERT OR REPLACE INTO qe_pattern_embeddings (pattern_id, embedding, dimension, model)
-      VALUES (?, ?, ?, ?)
+      INSERT OR REPLACE INTO qe_pattern_embeddings (pattern_id, embedding, dimension, model, space_id)
+      VALUES (?, ?, ?, ?, ?)
     `));
         this.prepared.set('getPattern', this.db.prepare(`
       SELECT * FROM qe_patterns WHERE id = ?
     `));
         this.prepared.set('getPatternWithEmbedding', this.db.prepare(`
-      SELECT p.*, e.embedding, e.dimension
+      SELECT p.*, e.embedding, e.dimension, e.space_id
       FROM qe_patterns p
       LEFT JOIN qe_pattern_embeddings e ON p.id = e.pattern_id
       WHERE p.id = ?
@@ -360,13 +373,13 @@ export class SQLitePatternStore {
       SELECT * FROM qe_patterns WHERE qe_domain = ? ORDER BY quality_score DESC LIMIT ?
     `));
         this.prepared.set('getAllEmbeddings', this.db.prepare(`
-      SELECT pattern_id, embedding, dimension FROM qe_pattern_embeddings
+      SELECT pattern_id, embedding, dimension, space_id FROM qe_pattern_embeddings
     `));
         // Recent-first variant used by the RVF backfill path. Order by
         // qe_patterns.last_used_at (then created_at) so the bounded backfill keeps
         // the most actively-used patterns when the cap is hit.
         this.prepared.set('getRecentEmbeddings', this.db.prepare(`
-      SELECT e.pattern_id, e.embedding, e.dimension
+      SELECT e.pattern_id, e.embedding, e.dimension, e.space_id
         FROM qe_pattern_embeddings e
         LEFT JOIN qe_patterns p ON p.id = e.pattern_id
        ORDER BY COALESCE(p.last_used_at, p.created_at, e.created_at) DESC
@@ -382,7 +395,7 @@ export class SQLitePatternStore {
     /**
      * Store a pattern
      */
-    storePattern(pattern, embedding) {
+    storePattern(pattern, embedding, embeddingSpaceId) {
         if (!this.db)
             throw new Error('Database not initialized');
         const id = pattern.id || uuidv4();
@@ -405,9 +418,13 @@ export class SQLitePatternStore {
                 .get(pattern.name, pattern.qeDomain ?? '', pattern.patternType ?? '');
             actualId = actualRow ? actualRow.id : id;
             if (embedding) {
+                const spaceId = embeddingSpaceId ?? getActiveEmbeddingSpaceIdentity()?.spaceId;
+                if (!spaceId) {
+                    throw new Error('VECTOR_SPACE_UNVERIFIED: refusing to persist an embedding without runtime provenance');
+                }
                 // Store embedding as BLOB (Float32Array)
                 const buffer = Buffer.from(new Float32Array(embedding).buffer);
-                insertEmbedding.run(actualId, buffer, embedding.length, 'all-MiniLM-L6-v2');
+                insertEmbedding.run(actualId, buffer, embedding.length, 'all-MiniLM-L6-v2', spaceId);
             }
         });
         transaction();
@@ -529,6 +546,7 @@ export class SQLitePatternStore {
         return rows.map(row => ({
             patternId: row.pattern_id,
             embedding: Array.from(new Float32Array(row.embedding.buffer, row.embedding.byteOffset, row.dimension)),
+            spaceId: row.space_id,
         }));
     }
     /**
@@ -549,8 +567,15 @@ export class SQLitePatternStore {
         return rows.map(row => ({
             patternId: row.pattern_id,
             embedding: Array.from(new Float32Array(row.embedding.buffer, row.embedding.byteOffset, row.dimension)),
+            spaceId: row.space_id,
         }));
     }
+    getEmbeddingSpaceDiagnostics(activeSpaceId = getActiveEmbeddingSpaceIdentity()?.spaceId ?? null) {
+        if (!this.db)
+            throw new Error('Database not initialized');
+        const rows = this.db.prepare('SELECT space_id AS spaceId FROM qe_pattern_embeddings').all();
+        return inspectEmbeddingSpaceRows(rows, activeSpaceId);
+    }
     /**
      * Record pattern usage.
      *
@@ -788,7 +813,7 @@ export class SQLitePatternStore {
     /**
      * Store embedding for an existing pattern
      */
-    storePatternEmbedding(patternId, embedding) {
+    storePatternEmbedding(patternId, embedding, embeddingSpaceId) {
         if (!this.db)
             throw new Error('Database not initialized');
         const insertEmbedding = this.prepared.get('insertEmbedding');
@@ -797,7 +822,11 @@ export class SQLitePatternStore {
         }
         // Convert embedding to Buffer for storage
         const buffer = Buffer.from(new Float32Array(embedding).buffer);
-        insertEmbedding.run(patternId, buffer, embedding.length, 'all-MiniLM-L6-v2');
+        const spaceId = embeddingSpaceId ?? getActiveEmbeddingSpaceIdentity()?.spaceId;
+        if (!spaceId) {
+            throw new Error('VECTOR_SPACE_UNVERIFIED: refusing to persist an embedding without runtime provenance');
+        }
+        insertEmbedding.run(patternId, buffer, embedding.length, 'all-MiniLM-L6-v2', spaceId);
     }
     /**
      * Record pattern reuse for metrics tracking (Phase 5.4)
@@ -935,14 +964,22 @@ export class SQLitePatternStore {
         if (!this.db)
             throw new Error('Database not initialized');
         const dimension = getEmbeddingDimension(); // 384
-        // Find patterns without embeddings (skip bench patterns)
-        const patternsWithout = this.db.prepare(`
-      SELECT p.id, p.name, p.description, p.pattern_type, p.qe_domain
-      FROM qe_patterns p
-      WHERE p.id NOT IN (SELECT pattern_id FROM qe_pattern_embeddings)
-        AND p.id NOT LIKE 'bench-%'
-      ORDER BY p.quality_score DESC
-    `).all();
+        // Initialize the executable identity before choosing repair candidates.
+        // The canary is cached by the embedding layer and is never persisted.
+        await computeBatchEmbeddings([EMBEDDING_SPACE_CANARY]);
+        const activeSpaceId = getActiveEmbeddingSpaceIdentity()?.spaceId;
+        if (!activeSpaceId) {
+            throw new Error('VECTOR_SPACE_UNVERIFIED: backfill embedder has no runtime provenance');
+        }
+        // Repair absent, legacy/unverified, and mismatched embeddings.
+        const patternsWithout = this.db.prepare(`
+      SELECT p.id, p.name, p.description, p.pattern_type, p.qe_domain
+      FROM qe_patterns p
+      LEFT JOIN qe_pattern_embeddings e ON e.pattern_id = p.id
+      WHERE (e.pattern_id IS NULL OR e.space_id IS NULL OR e.space_id <> ?)
+        AND p.id NOT LIKE 'bench-%'
+      ORDER BY p.quality_score DESC
+    `).all(activeSpaceId);
         const alreadyHad = this.db.prepare(`
       SELECT COUNT(*) as c FROM qe_pattern_embeddings
     `).get();
@@ -1012,7 +1049,7 @@ export class SQLitePatternStore {
                             continue;
                         }
                         const buffer = Buffer.from(new Float32Array(embedding).buffer);
-                        insertEmbedding.run(textsWithIds[j].id, buffer, dimension, embeddingTag);
+                        insertEmbedding.run(textsWithIds[j].id, buffer, dimension, embeddingTag, activeSpaceId);
                         processed++;
                     }
                     catch (e) {