bulltrackers-module 1.0.738 → 1.0.740

package/functions/computation-system-v2/framework/data/DataFetcher.js

@@ -6,10 +6,18 @@
  * * V2.3 FIX: "Insufficient History" bug.
  * - fetchBatched now orders by Entity ID to keep historical rows together.
  * - Implemented "Entity-Atomic Batching" to prevent splitting a user's history across batches.
+ * * V2.4 FIX: Runaway Query Cost Prevention [Fix #3].
+ * * V2.5 UPDATE: Super-Entity Monitoring [Safety Valve for Fix #6].
+ * - Warns if a single entity exceeds reasonable batch limits (Memory Risk).
  */
 
 const { BigQuery } = require('@google-cloud/bigquery');
 
+// FIX #3: Hard limit to prevent cost spirals
+const MAX_LOOKBACK_DAYS = 30;
+// FIX #6 (Alternative): Warn if an entity is massive (e.g. > 5x batch size)
+const BATCH_GROWTH_WARNING_THRESHOLD = 5; 
+
 class DataFetcher {
     constructor(config, queryBuilder, logger = null) {
         this.projectId = config.projectId;
@@ -151,6 +159,12 @@ class DataFetcher {
     
     async fetch(options) {
         const { table, targetDate, lookback = 0, filter = {}, fields = null, entities = null } = options;
+        
+        // FIX #3: Prevent Runaway Costs
+        if (lookback > MAX_LOOKBACK_DAYS) {
+            throw new Error(`[DataFetcher] COST GUARD: Lookback of ${lookback} days exceeds limit of ${MAX_LOOKBACK_DAYS}. Table: ${table}`);
+        }
+
         const tableConfig = this.tables[table] || {};
         const { dateField, entityField, dataField } = tableConfig;
         
@@ -168,6 +182,12 @@ class DataFetcher {
 
     async *fetchBatched(options, batchSize = 1000) {
         const { table, targetDate, lookback = 0, filter = {}, fields = null, entities = null } = options;
+
+        // FIX #3: Prevent Runaway Costs
+        if (lookback > MAX_LOOKBACK_DAYS) {
+            throw new Error(`[DataFetcher] COST GUARD: Lookback of ${lookback} days exceeds limit of ${MAX_LOOKBACK_DAYS}. Table: ${table}`);
+        }
+
         const tableConfig = this.tables[table] || {};
         const { dateField, entityField, dataField } = tableConfig;
 
@@ -181,6 +201,7 @@ class DataFetcher {
         
         let batch = [];
         let currentEntity = null;
+        let batchHasWarned = false; // Flag to prevent log spam for a single massive batch
 
         for await (const row of rowStream) {
             // FIX #2: Entity-Atomic Batching
@@ -188,12 +209,23 @@ class DataFetcher {
             if (entityField) {
                 const rowEntity = String(row[entityField]);
                 
-                // If batch is full AND we have moved to a new entity, yield the batch
-                // This ensures the current entity (which might have many rows) stays together
+                // Check if we should yield
+                // Condition: Batch is full AND we are on a NEW entity
                 if (batch.length >= batchSize && rowEntity !== currentEntity && currentEntity !== null) {
                     yield this._transform(batch, { lookback, dateField, entityField, dataField });
                     batch = [];
+                    batchHasWarned = false;
                 }
+                
+                // SAFETY VALVE (Fix #6 Alternative):
+                // If batch grows huge (Super Entity) and we CANNOT split (same entity), warn the admin.
+                if (batch.length > batchSize * BATCH_GROWTH_WARNING_THRESHOLD && !batchHasWarned) {
+                    this._log('WARN', `SUPER ENTITY DETECTED: Entity '${currentEntity}' in table '${table}' has >${batch.length} rows. ` + 
+                              `This exceeds batch size ${batchSize} by ${BATCH_GROWTH_WARNING_THRESHOLD}x. ` +
+                              `Risk of OOM or Timeouts. Consider filtering this entity.`);
+                    batchHasWarned = true;
+                }
+                
                 currentEntity = rowEntity;
             } else {
                 // Fallback for non-entity tables (strict count)

package/functions/computation-system-v2/framework/data/SchemaRegistry.js

@@ -1,12 +1,12 @@
 /**
  * @fileoverview Schema Registry - Dynamic schema discovery with caching
- * 
- * Core innovation of v2: No hardcoded schemas. Instead, we:
+ * * Core innovation of v2: No hardcoded schemas. Instead, we:
  * 1. Fetch schemas from BigQuery INFORMATION_SCHEMA on first access
  * 2. Cache them with configurable TTL
  * 3. Validate all queries against cached schemas BEFORE sending to BigQuery
- * 
- * This prevents runtime query failures and eliminates schema maintenance burden.
+ * * This prevents runtime query failures and eliminates schema maintenance burden.
+ * * UPDATE: Implemented Request Coalescing (Fix #5) to prevent "Thundering Herd"
+ * on startup or cache expiry.
  */
 
 const { BigQuery } = require('@google-cloud/bigquery');
@@ -45,18 +45,21 @@ class SchemaRegistry {
         
         this.client = new BigQuery({ projectId: this.projectId });
         this.cache = new Map();
+        this.pendingFetches = new Map(); // FIX: Track in-flight requests
         
         // Track schema fetch stats for monitoring
         this.stats = {
             hits: 0,
             misses: 0,
             refreshes: 0,
-            errors: 0
+            errors: 0,
+            coalesced: 0 // New metric
         };
     }
     
     /**
      * Get schema for a table, fetching from BigQuery if not cached.
+     * Implements Request Coalescing to handle concurrent access.
      * @param {string} tableName - Table name (without dataset prefix)
      * @returns {Promise<TableSchema>}
      */
@@ -67,6 +70,13 @@ class SchemaRegistry {
             this.stats.hits++;
             return cached;
         }
+
+        // FIX: Check for pending fetch (Request Coalescing)
+        if (this.pendingFetches.has(tableName)) {
+            this.stats.coalesced++;
+            // this._log('DEBUG', `Coalescing request for ${tableName}`);
+            return this.pendingFetches.get(tableName);
+        }
         
         if (cached) {
             this.stats.refreshes++;
@@ -76,7 +86,16 @@ class SchemaRegistry {
             this._log('DEBUG', `Schema cache miss for ${tableName}, fetching...`);
         }
         
-        return await this._fetchAndCacheSchema(tableName);
+        // Create the promise and store it
+        const fetchPromise = this._fetchAndCacheSchema(tableName);
+        this.pendingFetches.set(tableName, fetchPromise);
+
+        try {
+            return await fetchPromise;
+        } finally {
+            // Always clean up pending map, success or failure
+            this.pendingFetches.delete(tableName);
+        }
     }
     
     /**
@@ -158,6 +177,8 @@ class SchemaRegistry {
     async warmCache(tableNames) {
         const results = { success: [], failed: [] };
         
+        // With request coalescing, we can just map and wait.
+        // Simultaneous calls for the same table will automatically merge.
         await Promise.all(tableNames.map(async (tableName) => {
             try {
                 await this.getSchema(tableName);
@@ -178,9 +199,11 @@ class SchemaRegistry {
     clearCache(tableName = null) {
         if (tableName) {
             this.cache.delete(tableName);
+            this.pendingFetches.delete(tableName); // Also clear pending if forced
             this._log('DEBUG', `Cleared schema cache for ${tableName}`);
         } else {
             this.cache.clear();
+            this.pendingFetches.clear();
             this._log('DEBUG', 'Cleared entire schema cache');
         }
     }
@@ -193,6 +216,7 @@ class SchemaRegistry {
         return {
             ...this.stats,
             cachedTables: this.cache.size,
+            pendingRequests: this.pendingFetches.size,
             cacheContents: Array.from(this.cache.keys())
         };
     }
@@ -284,4 +308,4 @@ class SchemaRegistry {
     }
 }
 
-module.exports = { SchemaRegistry };
+module.exports = { SchemaRegistry };

package/functions/computation-system-v2/framework/execution/Orchestrator.js

@@ -9,6 +9,8 @@
  * * * UPDATE: Includes Global vs Batch Data Split to fix "Identity Crisis".
  * * * UPDATE: Implemented FORCE logic to bypass "up-to-date" checks for testing.
  * * * UPDATE: Aggregates performance reporting to prevent log spam.
+ * * * FIX: Resolved N+1 Dependency Fetching (Strict Mode in Streaming).
+ * * * FIX: Added missing 'skipped' property to return types for type safety.
  */
 
 const crypto = require('crypto');
@@ -316,21 +318,41 @@ class Orchestrator {
 
             const { data: batchLocalData, entityIds } = batch;
             const combinedData = { ...batchLocalData, ...globalData };
+            
+            // STRICT FIX: Prefetch dependencies for the batch.
             const batchDeps = await this._prefetchBatchDependencies(entry, dateStr, depResults, entityIds);
+            
             const { rules } = this.ruleInjector.createContext(); 
             const batchResults = {};
             
             await Promise.all(entityIds.map(entityId => limit(async () => {
                 const instance = new entry.class();
                 const entityData = this._filterDataForEntity(combinedData, entityId, driverEntityField);
+                
                 const context = {
                     computation: entry, date: dateStr, entityId, data: entityData,
+                    
+                    // STRICT FIX: No fallback to _lazyLoadDependency.
                     getDependency: (depName, targetId) => {
-                        if (batchDeps[depName] && batchDeps[depName].has(targetId || entityId)) {
-                            return batchDeps[depName].get(targetId || entityId);
+                        const id = targetId || entityId;
+                        
+                        // 1. Look in Batch-Prefetched Dependencies (Priority)
+                        if (batchDeps[depName] && batchDeps[depName].has(id)) {
+                            return batchDeps[depName].get(id);
+                        }
+                        
+                        // 2. Look in Global/Preloaded Dependencies
+                        if (depResults[depName]) {
+                             if (depResults[depName][id] !== undefined) return depResults[depName][id];
                         }
-                        return this._lazyLoadDependency(dateStr, depName, targetId || entityId, depResults);
+                        
+                        // 3. STRICT MODE: Throw Error
+                        throw new Error(
+                            `[Strict Dependency] Dependency '${depName}' (ID: ${id}) not found in batch context. ` +
+                            `Ensure '${depName}' is listed in ${entry.name}.getConfig().dependencies.`
+                        );
                     },
+                    
                     previousResult, rules, references: this.referenceDataCache,
                     config: this.config, dataFetcher: this.dataFetcher
                 };
@@ -357,22 +379,18 @@ class Orchestrator {
             if (cp) await checkpointer.complete(dateStr, entry.name, cp.id);
         }
 
-        return { count: totalCount, hash: rollingHash.digest('hex').substring(0, 16) };
+        // FIX: Return valid object shape including skipped: false
+        return { count: totalCount, hash: rollingHash.digest('hex').substring(0, 16), skipped: false };
     }
 
     /**
      * Determine if a computation should use remote workers
-     * 
-     * @param {Object} entry - Manifest entry
+     * * @param {Object} entry - Manifest entry
      * @param {Object} options - Execution options
      * @param {boolean} [options.useWorkerPool] - Runtime override (true/false/undefined)
      * @param {boolean} [options.forceLocal] - Force local execution
      */
     _shouldUseRemoteWorkers(entry, options) {
-        // Runtime override takes precedence (for admin testing)
-        // useWorkerPool: true  -> force use worker pool
-        // useWorkerPool: false -> force local execution
-        // useWorkerPool: undefined -> use config
         if (options.useWorkerPool === true) {
             if (!this.remoteRunner) {
                 this._log('WARN', 'useWorkerPool=true but remoteRunner not initialized');
@@ -384,30 +402,22 @@ class Orchestrator {
             return false;
         }
         
-        // No remote runner configured
         if (!this.remoteRunner) return false;
-        
-        // Force local execution via options
         if (options.forceLocal) return false;
         
         const poolConfig = this.config.workerPool || {};
         
-        // Exclusion list
         if (poolConfig.excludeComputations?.includes(entry.name) || 
             poolConfig.excludeComputations?.includes(entry.originalName)) {
             return false;
         }
         
-        // Force list (override threshold)
         if (poolConfig.forceOffloadComputations?.includes(entry.name) || 
             poolConfig.forceOffloadComputations?.includes(entry.originalName)) {
             return true;
         }
         
-        // Only per-entity computations can be offloaded
         if (entry.type !== 'per-entity') return false;
-        
-        // Default: use remote if worker pool is enabled
         return true;
     }
 
@@ -492,7 +502,6 @@ class Orchestrator {
                 this._log('WARN', `[Remote] Batch ${batchIndex}: ${errors.length} entities failed`);
                 totalErrors += errors.length;
                 
-                // Log first few errors for debugging
                 errors.slice(0, 3).forEach(e => {
                     this._log('DEBUG', `  - ${e.entityId}: ${e.error}`);
                 });
@@ -524,7 +533,8 @@ class Orchestrator {
             this._log('WARN', `[Remote] Completed with ${totalErrors} total errors out of ${totalCount + totalErrors} entities`);
         }
 
-        return { count: totalCount, hash: rollingHash.digest('hex').substring(0, 16) };
+        // FIX: Return valid object shape including skipped: false
+        return { count: totalCount, hash: rollingHash.digest('hex').substring(0, 16), skipped: false };
     }
 
     async _executeGlobal(entry, dateStr, depResults, previousResult, options, forceEntities) {
@@ -569,7 +579,8 @@ class Orchestrator {
             await this.storageManager.finalizeResults(dateStr, entry);
         }
 
-        return { count: Object.keys(results || {}).length, hash: finalHash };
+        // FIX: Return valid object shape including skipped: false
+        return { count: Object.keys(results || {}).length, hash: finalHash, skipped: false };
     }
 
     _printExecutionSummary(summary) {
@@ -668,6 +679,10 @@ class Orchestrator {
     async _lazyLoadDependency(dateStr, depName, entityId, preloaded) {
         if (preloaded[depName] && !entityId) return preloaded[depName];
         if (preloaded[depName] && entityId) return preloaded[depName][entityId];
+        
+        // WARN: This is the slow path that we removed from Streaming
+        this._log('WARN', `LAZY LOAD: Fetching single entity '${entityId}' for '${depName}'. This is slow.`);
+        
         if (entityId) return this.stateRepository.getEntityResult(dateStr, depName, entityId);
         return this.stateRepository.getResult(dateStr, depName);
     }

package/functions/computation-system-v2/framework/execution/RemoteTaskRunner.js

@@ -1,20 +1,18 @@
 /**
  * @fileoverview Remote Task Runner (Serverless Worker Pool Client)
- * 
- * RESPONSIBILITIES:
+ * * RESPONSIBILITIES:
  * 1. Package entity data and context into GCS files
  * 2. Invoke remote worker functions in parallel
  * 3. Collect results and errors
  * 4. Handle retries for transient failures
- * 
- * DATA FLOW:
+ * * DATA FLOW:
  * Orchestrator calls runBatch() -> Upload to GCS -> Invoke Workers -> Collect Results
- * 
- * DESIGN PRINCIPLES:
+ * * DESIGN PRINCIPLES:
  * - Workers are stateless - all context is passed via GCS
  * - High parallelism - hundreds of concurrent invocations
  * - Fault isolation - one entity failure doesn't affect others
  * - Cost efficient - workers scale to zero between runs
+ * - RESILIENCE: Implements Circuit Breaker to prevent Retry Cost Spirals [Fix #2]
  */
 
 const { Storage } = require('@google-cloud/storage');
@@ -36,6 +34,13 @@ class RemoteTaskRunner {
         this.timeout = poolConfig.timeout || 60000; // 60s default
         this.retries = poolConfig.retries || 2;
         
+        // Circuit Breaker Config [Fix #2]
+        this.cbConfig = {
+            minInvocations: 20,       // Minimum calls before checking rate
+            failureThreshold: 0.30,   // Trip if failure rate > 30%
+            ...poolConfig.circuitBreaker
+        };
+        
         // Local mode for testing
         this.localMode = poolConfig.localMode || process.env.WORKER_LOCAL_MODE === 'true';
         
@@ -53,8 +58,7 @@ class RemoteTaskRunner {
 
     /**
      * Execute a batch of entities remotely (or locally for testing)
-     * 
-     * @param {Object} entry - Manifest entry for the computation
+     * * @param {Object} entry - Manifest entry for the computation
      * @param {string} dateStr - Target date (YYYY-MM-DD)
      * @param {Object} baseContext - Shared context (references, config)
      * @param {string[]} entityIds - Entity IDs to process
@@ -126,11 +130,21 @@ class RemoteTaskRunner {
         const errors = [];
         const uploadedPaths = [];
 
+        // Circuit Breaker Stats (scoped to this batch)
+        const batchStats = {
+            invocations: 0,
+            failures: 0,
+            tripped: false
+        };
+
         // Phase 1: Upload context packages to GCS
         this._log('INFO', 'Uploading context packages to GCS...');
         const uploadStart = Date.now();
         
         const uploadTasks = entityIds.map(entityId => uploadLimit(async () => {
+            // Check tripped status early to save uploads if massive failure occurring
+            if (batchStats.tripped) return;
+
             const contextPackage = this._buildContextPackage(
                 entry,
                 entityId,
@@ -158,13 +172,19 @@ class RemoteTaskRunner {
 
         const invokeTasks = uploadedPaths.map(({ entityId, path }) => 
             invokeLimit(async () => {
+                // FAIL FAST: If circuit tripped, do not invoke worker
+                if (batchStats.tripped) {
+                    errors.push({ entityId, error: 'Skipped: Circuit Breaker Tripped due to high failure rate' });
+                    return;
+                }
+
                 try {
                     const response = await this._invokeWorkerWithRetry({
                         computationName: entry.originalName || entry.name,
                         entityId,
                         date: dateStr,
                         dataUri: { bucket: this.bucketName, path }
-                    });
+                    }, 1, batchStats); // Pass stats object to retry logic
 
                     if (response.status === 'success' && response.result !== null) {
                         results[entityId] = response.result;
@@ -174,12 +194,18 @@ class RemoteTaskRunner {
                     // status === 'success' with result === null means skipped (filtered out)
                     
                 } catch (e) {
+                    // Circuit Breaker errors are thrown here
                     errors.push({ entityId, error: e.message });
                 }
             })
         );
 
         await Promise.all(invokeTasks);
+        
+        if (batchStats.tripped) {
+            this._log('ERROR', `Batch ABORTED by Circuit Breaker. Stats: ${batchStats.failures} failures / ${batchStats.invocations} invocations.`);
+        }
+        
         this._log('INFO', `Invocations complete in ${Date.now() - invokeStart}ms`);
 
         // Phase 3: Cleanup GCS (fire and forget)
@@ -237,12 +263,21 @@ class RemoteTaskRunner {
     }
 
     /**
-     * Invoke a worker with retry logic
+     * Invoke a worker with retry logic and Circuit Breaker
      */
-    async _invokeWorkerWithRetry(payload, attempt = 1) {
+    async _invokeWorkerWithRetry(payload, attempt = 1, stats = null) {
+        // Track Invocation (Cost)
+        if (stats) stats.invocations++;
+
         try {
             return await this._invokeWorker(payload);
         } catch (e) {
+            // Track Failure
+            if (stats) {
+                stats.failures++;
+                this._checkCircuitBreaker(stats);
+            }
+
             const isRetryable = this._isRetryableError(e);
             
             if (isRetryable && attempt < this.retries) {
@@ -250,14 +285,40 @@ class RemoteTaskRunner {
                 const delay = Math.min(1000 * Math.pow(2, attempt - 1), 10000);
                 await new Promise(r => setTimeout(r, delay));
                 
+                // Re-check circuit before retrying (another thread might have tripped it)
+                if (stats) this._checkCircuitBreaker(stats);
+
                 this._log('DEBUG', `Retrying ${payload.entityId} (attempt ${attempt + 1})`);
-                return this._invokeWorkerWithRetry(payload, attempt + 1);
+                return this._invokeWorkerWithRetry(payload, attempt + 1, stats);
             }
             
             throw e;
         }
     }
 
+    /**
+     * Check circuit breaker status and throw if tripped
+     */
+    _checkCircuitBreaker(stats) {
+        if (stats.tripped) {
+            throw new Error('Circuit Breaker: Batch aborted due to high failure rate');
+        }
+
+        // Only check after minimum invocations (warmup)
+        if (stats.invocations >= this.cbConfig.minInvocations) {
+            const failureRate = stats.failures / stats.invocations;
+            
+            if (failureRate > this.cbConfig.failureThreshold) {
+                stats.tripped = true;
+                const msg = `🚨 CIRCUIT BREAKER TRIPPED! Failure rate ${(failureRate * 100).toFixed(1)}% ` +
+                            `(${stats.failures}/${stats.invocations}) exceeds threshold of ${(this.cbConfig.failureThreshold * 100)}%`;
+                
+                this._log('ERROR', msg);
+                throw new Error(msg);
+            }
+        }
+    }
+
     /**
      * Invoke a single worker via HTTP
      */
@@ -324,4 +385,4 @@ class RemoteTaskRunner {
     }
 }
 
-module.exports = { RemoteTaskRunner };
+module.exports = { RemoteTaskRunner };

package/functions/computation-system-v2/handlers/adminTest.js

@@ -17,8 +17,7 @@
  *     -d '{"action": "run", "computation": "UserPortfolioSummary", "date": "2026-01-25"}'
  */
 
-const system = require('../index');
-
+const system = require('../core-api');
 /**
  * Admin test handler.
  */