@soulcraft/brainy 3.37.6 → 3.37.8

package/dist/storage/adapters/gcsStorage.js

@@ -146,6 +146,12 @@ export class GcsStorage extends BaseStorage {
             });
             // Initialize counts from storage
             await this.initializeCounts();
+            // CRITICAL FIX (v3.37.7): Clear any stale cache entries from previous runs
+            // This prevents cache poisoning from causing silent failures on container restart
+            prodLog.info('🧹 Clearing cache from previous run to prevent cache poisoning');
+            this.nounCacheManager.clear();
+            this.verbCacheManager.clear();
+            prodLog.info('✅ Cache cleared - starting fresh');
             this.isInitialized = true;
         }
         catch (error) {
@@ -336,8 +342,14 @@ export class GcsStorage extends BaseStorage {
                 contentType: 'application/json',
                 resumable: false // For small objects, non-resumable is faster
             });
-            // Update cache
-            this.nounCacheManager.set(node.id, node);
+            // CRITICAL FIX (v3.37.8): Only cache nodes with non-empty vectors
+            // This prevents cache pollution from HNSW's lazy-loading nodes (vector: [])
+            if (node.vector && Array.isArray(node.vector) && node.vector.length > 0) {
+                this.nounCacheManager.set(node.id, node);
+            }
+            else {
+                prodLog.warn(`[saveNode] Not caching node ${node.id.substring(0, 8)}... with empty vector (HNSW lazy mode)`);
+            }
             // Increment noun count
             const metadata = await this.getNounMetadata(node.id);
             if (metadata && metadata.type) {
@@ -380,11 +392,53 @@ export class GcsStorage extends BaseStorage {
      */
     async getNode(id) {
         await this.ensureInitialized();
-        // Check cache first
-        const cached = this.nounCacheManager.get(id);
-        if (cached) {
-            this.logger.trace(`Cache hit for noun ${id}`);
-            return cached;
+        // Check cache first WITH LOGGING
+        const cached = await this.nounCacheManager.get(id);
+        // DIAGNOSTIC LOGGING: Reveal cache poisoning
+        prodLog.info(`[getNode] 🔍 Cache check for ${id.substring(0, 8)}...:`, {
+            hasCached: cached !== undefined,
+            isNull: cached === null,
+            isObject: cached !== null && typeof cached === 'object',
+            type: typeof cached
+        });
+        // CRITICAL FIX (v3.37.8): Validate cached object before returning
+        if (cached !== undefined && cached !== null) {
+            // Log cached object structure to diagnose incomplete objects
+            prodLog.info(`[getNode] Cached object structure:`, {
+                hasId: !!cached.id,
+                idMatches: cached.id === id,
+                hasVector: !!cached.vector,
+                vectorLength: cached.vector?.length,
+                hasConnections: !!cached.connections,
+                connectionsType: typeof cached.connections,
+                hasLevel: cached.level !== undefined,
+                level: cached.level,
+                objectKeys: Object.keys(cached || {})
+            });
+            // Validate cached object has required fields (including non-empty vector!)
+            if (!cached.id || !cached.vector || !Array.isArray(cached.vector) || cached.vector.length === 0) {
+                prodLog.error(`[getNode] ❌ INVALID cached object for ${id.substring(0, 8)}...:`, {
+                    reason: !cached.id ? 'missing id' :
+                        !cached.vector ? 'missing vector' :
+                            !Array.isArray(cached.vector) ? 'vector not array' :
+                                cached.vector.length === 0 ? 'vector is empty array' :
+                                    'unknown'
+                });
+                prodLog.error(`[getNode] Removing invalid object from cache and loading from GCS`);
+                this.nounCacheManager.delete(id);
+                // Fall through to load from GCS
+            }
+            else {
+                prodLog.info(`[getNode] ✅ Valid cached object - returning`);
+                this.logger.trace(`Cache hit for noun ${id}`);
+                return cached;
+            }
+        }
+        else if (cached === null) {
+            prodLog.warn(`[getNode] ⚠️ Cache contains NULL for ${id.substring(0, 8)}... - ignoring and loading from GCS`);
+        }
+        else {
+            prodLog.info(`[getNode] ❌ Cache MISS - loading from GCS for ${id.substring(0, 8)}...`);
         }
         // Apply backpressure
         const requestId = await this.applyBackpressure();
@@ -420,8 +474,14 @@ export class GcsStorage extends BaseStorage {
                 level: data.level || 0
                 // NO metadata field - retrieved separately for scalability
             };
-            // Update cache
-            this.nounCacheManager.set(id, node);
+            // CRITICAL FIX: Only cache valid nodes with non-empty vectors (never cache null or empty)
+            if (node && node.id && node.vector && Array.isArray(node.vector) && node.vector.length > 0) {
+                this.nounCacheManager.set(id, node);
+                prodLog.info(`[getNode] 💾 Cached node ${id.substring(0, 8)}... successfully`);
+            }
+            else {
+                prodLog.warn(`[getNode] ⚠️ NOT caching invalid node for ${id.substring(0, 8)}... (missing id/vector or empty vector)`);
+            }
             this.logger.trace(`Successfully retrieved node ${id}`);
             this.releaseBackpressure(true, requestId);
             return node;
@@ -440,7 +500,8 @@ export class GcsStorage extends BaseStorage {
             prodLog.error(`[getNode]   Error object:`, JSON.stringify(error, null, 2));
             // Check if this is a "not found" error
             if (error.code === 404) {
-                prodLog.warn(`[getNode] Identified as 404 error - returning null`);
+                prodLog.warn(`[getNode] Identified as 404 error - returning null WITHOUT caching`);
+                // CRITICAL FIX: Do NOT cache null values
                 return null;
             }
             // Handle throttling

package/dist/storage/adapters/s3CompatibleStorage.js

@@ -237,6 +237,16 @@ export class S3CompatibleStorage extends BaseStorage {
             await this.cleanupLegacyIndexFolder();
             // Initialize counts from storage
             await this.initializeCounts();
+            // CRITICAL FIX (v3.37.7): Clear any stale cache entries from previous runs
+            // This prevents cache poisoning from causing silent failures on container restart
+            const nodeCacheSize = this.nodeCache?.size || 0;
+            if (nodeCacheSize > 0) {
+                prodLog.info(`🧹 Clearing ${nodeCacheSize} cached node entries from previous run`);
+                this.nodeCache.clear();
+            }
+            else {
+                prodLog.info('🧹 Node cache is empty - starting fresh');
+            }
             this.isInitialized = true;
             this.logger.info(`Initialized ${this.serviceType} storage with bucket ${this.bucketName}`);
         }
@@ -808,6 +818,52 @@ export class S3CompatibleStorage extends BaseStorage {
      */
     async getNode(id) {
         await this.ensureInitialized();
+        // Check cache first WITH LOGGING
+        const cached = this.nodeCache.get(id);
+        // DIAGNOSTIC LOGGING: Reveal cache poisoning
+        prodLog.info(`[getNode] 🔍 Cache check for ${id.substring(0, 8)}...:`, {
+            hasCached: cached !== undefined,
+            isNull: cached === null,
+            isObject: cached !== null && typeof cached === 'object',
+            type: typeof cached
+        });
+        // CRITICAL FIX (v3.37.8): Validate cached object before returning
+        if (cached !== undefined && cached !== null) {
+            // Log cached object structure to diagnose incomplete objects
+            prodLog.info(`[getNode] Cached object structure:`, {
+                hasId: !!cached.id,
+                idMatches: cached.id === id,
+                hasVector: !!cached.vector,
+                vectorLength: cached.vector?.length,
+                hasConnections: !!cached.connections,
+                connectionsType: typeof cached.connections,
+                objectKeys: Object.keys(cached || {})
+            });
+            // Validate cached object has required fields (including non-empty vector!)
+            if (!cached.id || !cached.vector || !Array.isArray(cached.vector) || cached.vector.length === 0) {
+                prodLog.error(`[getNode] ❌ INVALID cached object for ${id.substring(0, 8)}...:`, {
+                    reason: !cached.id ? 'missing id' :
+                        !cached.vector ? 'missing vector' :
+                            !Array.isArray(cached.vector) ? 'vector not array' :
+                                cached.vector.length === 0 ? 'vector is empty array' :
+                                    'unknown'
+                });
+                prodLog.error(`[getNode] Removing invalid object from cache and loading from S3`);
+                this.nodeCache.delete(id);
+                // Fall through to load from S3
+            }
+            else {
+                prodLog.info(`[getNode] ✅ Valid cached object - returning`);
+                this.logger.trace(`Cache hit for node ${id}`);
+                return cached;
+            }
+        }
+        else if (cached === null) {
+            prodLog.warn(`[getNode] ⚠️ Cache contains NULL for ${id.substring(0, 8)}... - ignoring and loading from S3`);
+        }
+        else {
+            prodLog.info(`[getNode] ❌ Cache MISS - loading from S3 for ${id.substring(0, 8)}...`);
+        }
         try {
             // Import the GetObjectCommand only when needed
             const { GetObjectCommand } = await import('@aws-sdk/client-s3');
@@ -858,6 +914,14 @@ export class S3CompatibleStorage extends BaseStorage {
                 connections,
                 level: parsedNode.level || 0
             };
+            // CRITICAL FIX: Only cache valid nodes with non-empty vectors (never cache null or empty)
+            if (node && node.id && node.vector && Array.isArray(node.vector) && node.vector.length > 0) {
+                this.nodeCache.set(id, node);
+                prodLog.info(`[getNode] 💾 Cached node ${id.substring(0, 8)}... successfully`);
+            }
+            else {
+                prodLog.warn(`[getNode] ⚠️ NOT caching invalid node for ${id.substring(0, 8)}... (missing id/vector or empty vector)`);
+            }
             this.logger.trace(`Successfully retrieved node ${id}`);
             return node;
         }
@@ -876,7 +940,8 @@ export class S3CompatibleStorage extends BaseStorage {
             prodLog.error(`[getNode]   Error object:`, JSON.stringify(error, null, 2));
             // Check if this is a "not found" error (S3 uses "NoSuchKey")
             if (error?.name === 'NoSuchKey' || error?.Code === 'NoSuchKey' || error?.$metadata?.httpStatusCode === 404) {
-                prodLog.warn(`[getNode] Identified as 404/NoSuchKey error - returning null`);
+                prodLog.warn(`[getNode] Identified as 404/NoSuchKey error - returning null WITHOUT caching`);
+                // CRITICAL FIX: Do NOT cache null values
                 return null;
             }
             // Handle throttling

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@soulcraft/brainy",
-  "version": "3.37.6",
+  "version": "3.37.8",
   "description": "Universal Knowledge Protocol™ - World's first Triple Intelligence database unifying vector, graph, and document search in one API. 31 nouns × 40 verbs for infinite expressiveness.",
   "main": "dist/index.js",
   "module": "dist/index.js",