bulltrackers-module 1.0.652 → 1.0.653

package/functions/computation-system/helpers/computation_dispatcher.js

@@ -5,14 +5,16 @@
  * UPDATED: Enforces Strict One-Shot Policy (Standard -> HighMem -> Dead Letter).
  * UPDATED: Generates Google Cloud Trace Context (traceId/spanId) for end-to-end monitoring.
  * UPDATED: Added Schedule Awareness (Daily, Weekly, Monthly) to filter tasks by date.
+ * UPDATED: Force Run now validates Root Data Availability before dispatching.
  */
 
 const { getExpectedDateStrings, getEarliestDataDates, normalizeName, DEFINITIVE_EARLIEST_DATES } = require('../utils/utils.js');
 const { groupByPass, analyzeDateExecution } = require('../WorkflowOrchestrator.js');
 const { PubSubUtils } = require('../../core/utils/pubsub_utils');
 const { fetchComputationStatus } = require('../persistence/StatusRepository');
-const { checkRootDataAvailability } = require('../data/AvailabilityChecker');
-const { runFinalSweepCheck } = require('../tools/FinalSweepReporter'); // [NEW]
+// [UPDATED] Imported getAvailabilityWindow and checkRootDependencies
+const { checkRootDataAvailability, getAvailabilityWindow, checkRootDependencies } = require('../data/AvailabilityChecker');
+const { runFinalSweepCheck } = require('../tools/FinalSweepReporter'); 
 const crypto = require('crypto');
 
 const OOM_THRESHOLD_MB = 1500; // Unused
@@ -233,7 +235,12 @@ async function handleForceRun(config, dependencies, computationManifest, reqBody
     let targetDates = [];
     if (dateInput) {
         // Single Date Mode
-        targetDates = [dateInput];
+        // We still perform the check, but for a single item array
+        if (Array.isArray(dateInput)) {
+            targetDates = dateInput;
+        } else {
+            targetDates = [dateInput];
+        }
     } else {
         // All Dates Mode (Backfill)
         logger.log('INFO', `[ForceRun] No date provided. Calculating date range for ${computationName}...`);
@@ -242,11 +249,54 @@ async function handleForceRun(config, dependencies, computationManifest, reqBody
         targetDates = getExpectedDateStrings(earliestDates.absoluteEarliest, new Date());
     }
     
-    logger.log('WARN', `[ForceRun] 🚨 MANUALLY Triggering ${computationName} for ${targetDates.length} dates. Pass: ${manifestItem.pass}`);
+    // [NEW] 3. Filter Impossible Dates (Availability Check)
+    logger.log('INFO', `[ForceRun] 🔍 Validating data availability for ${targetDates.length} candidate dates...`);
+    
+    // Sort dates to get efficient min/max for range query
+    targetDates.sort();
+    const startDate = targetDates[0];
+    const endDate = targetDates[targetDates.length - 1];
+
+    const availabilityMap = await getAvailabilityWindow(dependencies, startDate, endDate);
+    const validDates = [];
+    const skippedStats = {};
+
+    for (const date of targetDates) {
+        const status = availabilityMap.get(date);
+        
+        // If no index exists, we pass an empty object (AvailabilityChecker will fail all checks)
+        // This effectively filters out dates where we have NO knowledge of data
+        const effectiveStatus = status || {}; 
+        
+        const check = checkRootDependencies(manifestItem, effectiveStatus);
+        
+        if (check.canRun) {
+            validDates.push(date);
+        } else {
+            const reason = check.missing.length > 0 ? `Missing: ${check.missing.join(', ')}` : 'Dependencies not met';
+            skippedStats[reason] = (skippedStats[reason] || 0) + 1;
+        }
+    }
+
+    if (validDates.length === 0) {
+        logger.log('WARN', `[ForceRun] 🛑 ABORTING: No valid dates found for ${computationName} out of ${targetDates.length} requested.`);
+        return {
+            status: 'ABORTED',
+            computation: computationName,
+            reason: 'NO_DATA_AVAILABLE',
+            skippedDetails: skippedStats
+        };
+    }
+
+    if (validDates.length < targetDates.length) {
+        logger.log('INFO', `[ForceRun] ⚠️ Filtered impossible dates: ${targetDates.length} requested -> ${validDates.length} valid.`, { skippedStats });
+    }
+
+    logger.log('WARN', `[ForceRun] 🚨 MANUALLY Triggering ${computationName} for ${validDates.length} VALID dates. Pass: ${manifestItem.pass}`);
 
-    // 3. Construct Tasks
+    // 4. Construct Tasks
     const dispatchId = crypto.randomUUID();
-    const tasks = targetDates.map(date => {
+    const tasks = validDates.map(date => {
         const traceId = crypto.randomBytes(16).toString('hex');
         const spanId = crypto.randomBytes(8).toString('hex');
         return {
@@ -262,7 +312,7 @@ async function handleForceRun(config, dependencies, computationManifest, reqBody
         };
     });
 
-    // 4. Batch Publish (Chunked to stay under Pub/Sub limits)
+    // 5. Batch Publish (Chunked to stay under Pub/Sub limits)
     const CHUNK_SIZE = 250; // Safe batch size
     const topic = (reqBody.resources === 'high-mem') 
         ? (config.computationTopicHighMem || 'computation-tasks-highmem')
@@ -287,8 +337,10 @@ async function handleForceRun(config, dependencies, computationManifest, reqBody
     return { 
         status: 'FORCED',
         computation: computationName,
-        mode: dateInput ? 'SINGLE_DATE' : 'ALL_DATES',
+        mode: dateInput ? 'SINGLE/ARRAY_DATE' : 'ALL_DATES',
+        datesRequested: targetDates.length,
         datesTriggered: dispatchedCount,
+        skippedImpossible: targetDates.length - dispatchedCount,
         targetTopic: topic
     };
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bulltrackers-module",
-  "version": "1.0.652",
+  "version": "1.0.653",
   "description": "Helper Functions for Bulltrackers.",
   "main": "index.js",
   "files": [