bulltrackers-module 1.0.57 → 1.0.59

package/functions/user-activity-sampler/helpers/sampler_helpers.js

@@ -1,6 +1,7 @@
 /**
  * @fileoverview Main pipe: pipe.maintenance.runUserActivitySampler
  * Samples user activity levels based on the Rankings API.
+ * REVISED: Removed exclusion list logic for simplification.
  */
 const { FieldValue } = require('@google-cloud/firestore');
 
@@ -11,12 +12,13 @@ const { FieldValue } = require('@google-cloud/firestore');
  * @returns {Promise<object>} Summary of the sampling process.
  */
 exports.runUserActivitySampler = async (config, dependencies) => {
+    // NOTE: firestoreUtils is still in dependencies but getExclusionIds is no longer called from here.
     const { db, logger, headerManager, proxyManager, firestoreUtils } = dependencies;
     const today = new Date().toISOString().slice(0, 10);
-    logger.log('INFO', '[UserActivitySampler] Starting user activity sampling...');
+    logger.log('INFO', '[UserActivitySampler] Starting user activity sampling (simplified)...');
 
-    // Validate configuration
-    if (!config.rankingsApiUrl || !config.targetPublicUsersPerBlock || !config.apiBatchSize || !config.outputCollectionName || !config.allHighValueBlocks || !config.exclusionConfig) {
+    // Validate configuration (Removed exclusionConfig validation)
+    if (!config.rankingsApiUrl || !config.targetPublicUsersPerBlock || !config.apiBatchSize || !config.outputCollectionName || !config.allHighValueBlocks ) {
         logger.log('ERROR', '[UserActivitySampler] Missing required configuration.');
         throw new Error('Missing required configuration for User Activity Sampler.');
     }
@@ -24,14 +26,28 @@ exports.runUserActivitySampler = async (config, dependencies) => {
     try {
         await firestoreUtils.resetProxyLocks(dependencies, config); // Use shared config for proxy path
 
-        const exclusionIds = await firestoreUtils.getExclusionIds(dependencies, config.exclusionConfig, 'sampler'); // Pass 'sampler' or similar unique type
-        logger.log('INFO', `[UserActivitySampler] Fetched ${exclusionIds.size} exclusion IDs.`);
+        // --- REMOVED Exclusion ID fetching ---
+        // const exclusionIds = await firestoreUtils.getExclusionIds(dependencies, config.exclusionConfig, 'sampler');
+        // logger.log('INFO', `[UserActivitySampler] Fetched ${exclusionIds.size} exclusion IDs.`);
+        const processedInThisRun = new Set(); // Keep track of IDs processed in this run to avoid duplicates within batches
 
         const blockResults = [];
         let totalPublicSampled = 0;
         let totalPrivateEstimate = 0;
 
+        // Ensure allHighValueBlocks is defined and is an array
+        if (!config.allHighValueBlocks || !Array.isArray(config.allHighValueBlocks)) {
+             logger.log('ERROR', '[UserActivitySampler] Configuration error: allHighValueBlocks is missing or not an array.');
+             throw new Error('Configuration error: allHighValueBlocks must be an array.');
+        }
+
+
         for (const block of config.allHighValueBlocks) {
+             // Ensure block and block.startId exist
+             if (!block || typeof block.startId === 'undefined') {
+                logger.log('WARN', '[UserActivitySampler] Skipping invalid block configuration:', block);
+                continue;
+            }
             const blockId = block.startId;
             logger.log('INFO', `[UserActivitySampler] Processing block ${blockId}...`);
 
@@ -45,13 +61,14 @@ exports.runUserActivitySampler = async (config, dependencies) => {
             while (N_public_sampled_block < config.targetPublicUsersPerBlock && attempts < MAX_ATTEMPTS) {
                 attempts++;
                 const cidsToSample = [];
-                // Generate CIDs, ensuring they are not excluded
+                // Generate CIDs, ensuring they are not already processed in this run
                 while (cidsToSample.length < config.apiBatchSize) {
                     const randomId = String(Math.floor(Math.random() * 1000000) + blockId);
-                    if (!exclusionIds.has(randomId)) {
+                    // --- REMOVED exclusionIds check, added check for processedInThisRun ---
+                    if (!processedInThisRun.has(randomId)) {
                         cidsToSample.push(parseInt(randomId, 10));
-                        // Add temporarily to exclusion set for this run to avoid duplicates within batches
-                        exclusionIds.add(randomId);
+                        // Add temporarily to processed set for this run to avoid duplicates within batches
+                        processedInThisRun.add(randomId);
                     }
                 }
                 N_sampled_total_block += cidsToSample.length;
@@ -70,13 +87,11 @@ exports.runUserActivitySampler = async (config, dependencies) => {
 
                     if (!response || typeof response.json !== 'function') {
                          logger.log('WARN', `[UserActivitySampler] Invalid response structure from proxy for block ${blockId}, batch attempt ${attempts}. Skipping batch.`);
-                         // Consider retrying or specific error handling
                          continue; // Skip to next attempt
                     }
 
 
                     if (!response.ok) {
-                        // Log API errors but continue sampling other batches/blocks
                         const errorText = await response.text();
                         logger.log('WARN', `[UserActivitySampler] API error ${response.status} for block ${blockId}, batch attempt ${attempts}. Skipping batch. Error: ${errorText}`);
                         wasSuccess = false;
@@ -102,13 +117,11 @@ exports.runUserActivitySampler = async (config, dependencies) => {
                 } catch (fetchError) {
                     logger.log('ERROR', `[UserActivitySampler] Fetch failed for block ${blockId}, batch attempt ${attempts}. Skipping batch.`, { errorMessage: fetchError.message });
                     wasSuccess = false; // Mark as failure for header performance
-                    // Continue to next attempt
                 } finally {
                     if (selectedHeader) {
                         headerManager.updatePerformance(selectedHeader.id, wasSuccess);
                     }
                 }
-                 // Small delay between batches to avoid overwhelming API/proxies
                  await new Promise(resolve => setTimeout(resolve, config.delayBetweenBatchesMs || 200));
 
             } // End while loop for block sampling
@@ -125,7 +138,6 @@ exports.runUserActivitySampler = async (config, dependencies) => {
             const now = new Date();
             const oneDayAgo = new Date(now.getTime() - (24 * 60 * 60 * 1000));
             const oneWeekAgo = new Date(now.getTime() - (7 * 24 * 60 * 60 * 1000));
-            const oneMonthAgo = new Date(now.getTime() - (30 * 24 * 60 * 60 * 1000)); // Approx
              const threeMonthsAgo = new Date(now.getTime() - (90 * 24 * 60 * 60 * 1000)); // Approx
 
 
@@ -143,7 +155,7 @@ exports.runUserActivitySampler = async (config, dependencies) => {
 
                     if (lastActivityDate >= oneDayAgo) counts.A1++;
                     else if (lastActivityDate >= oneWeekAgo) counts.A2++;
-                    else if (lastActivityDate >= threeMonthsAgo) counts.A3++; // Changed from oneMonthAgo based on description
+                    else if (lastActivityDate >= threeMonthsAgo) counts.A3++;
                     else counts.A4++;
                 } catch(e) {
                      logger.log('WARN', `[UserActivitySampler] Error parsing LastActivity date '${user.LastActivity}' for user ${user.CID}. Counting as A4.`);
@@ -157,8 +169,7 @@ exports.runUserActivitySampler = async (config, dependencies) => {
                 fractions[category] = N_public_sampled_block > 0 ? (counts[category] / N_public_sampled_block) : 0;
             }
 
-            // TODO: Fetch or define N_block (total users for this block)
-            const N_block = config.knownTotalUsersPerBlock?.[blockId] || 1000000; // Example fallback
+            const N_block = 1000000; // Total users assumed per block
 
             const estimatedCounts = {};
             for (const category in fractions) {
@@ -170,7 +181,7 @@ exports.runUserActivitySampler = async (config, dependencies) => {
                 sampledDate: today,
                 f_private: f_private_block,
                 publicSampleSize: N_public_sampled_block,
-                totalUsersInBlock: N_block, // Add this if available
+                totalUsersInBlock: N_block,
                 activityCounts_Sample: counts,
                 activityFractions_Sample: fractions,
                 estimatedCounts_TotalBlock: estimatedCounts,
@@ -200,8 +211,7 @@ exports.runUserActivitySampler = async (config, dependencies) => {
 
     } catch (error) {
         logger.log('ERROR', '[UserActivitySampler] Fatal error during sampling process.', { errorMessage: error.message, errorStack: error.stack });
-        // Attempt to flush performance even on error
         try { await headerManager.flushPerformanceUpdates(); } catch (e) { /* ignore flush error */ }
-        throw error; // Re-throw to indicate failure
+        throw error;
     }
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bulltrackers-module",
-  "version": "1.0.57",
+  "version": "1.0.59",
   "description": "Helper Functions for Bulltrackers.",
   "main": "index.js",
   "files": [