Add unit tests for dashboard repository and enhance BSON handling

internal/dashboard/statistics_repository.go

@@ -20,6 +20,12 @@ const noticesCollectionName = "notices"
 // keeping counts aligned with the tender panel (MinIO is the source of truth for scraped docs).
 const scrapedScopeCacheTTL = 5 * time.Minute
 
+const (
+	scrapedScopeResolveTimeout = 2 * time.Minute
+	statisticsQueryTimeout     = 2 * time.Minute
+	maxScrapedFolderInClause   = 5000
+)
+
 // scrapedTendersScope is the resolved Mongo filter for tenders with scraped documents.
 // zero is set when MinIO reports no procedure folders; callers must return zero without querying.
 type scrapedTendersScope struct {
@@ -28,12 +34,15 @@ type scrapedTendersScope struct {
 }
 
 func (r *repository) Statistics(ctx context.Context, days int, startUnix int64) (*StatisticsReportResponse, error) {
+	queryCtx, cancel := context.WithTimeout(context.WithoutCancel(ctx), statisticsQueryTimeout)
+	defer cancel()
+
 	now := time.Now().UTC()
 	endDay := time.Date(now.Year(), now.Month(), now.Day(), 0, 0, 0, 0, time.UTC)
 	startDay := endDay.AddDate(0, 0, -(days - 1))
 
 	// Resolve scraped-document scope from MinIO (cached); falls back to Mongo flags on error.
-	scrapedScope, err := r.cachedScrapedTendersScope(ctx)
+	scrapedScope, err := r.cachedScrapedTendersScope(queryCtx)
 	if err != nil {
 		return nil, err
 	}
@@ -46,7 +55,7 @@ func (r *repository) Statistics(ctx context.Context, days int, startUnix int64)
 		totalTranslated   int64
 	)
 
-	g, gctx := errgroup.WithContext(ctx)
+	g, gctx := errgroup.WithContext(queryCtx)
 
 	g.Go(func() error {
 		// Use created_at (first ingest time). processing_metadata.scraped_at is reset on
@@ -221,7 +230,10 @@ func (r *repository) cachedScrapedTendersScope(ctx context.Context) (scrapedTend
 		}
 		r.scrapedScopeMu.Unlock()
 
-		scope, resolveErr := r.resolveScrapedTendersScope(ctx)
+		resolveCtx, cancel := context.WithTimeout(context.Background(), scrapedScopeResolveTimeout)
+		defer cancel()
+
+		scope, resolveErr := r.resolveScrapedTendersScope(resolveCtx)
 		if resolveErr != nil {
 			r.logger.Warn("MinIO scraped scope unavailable, using Mongo fallback for dashboard statistics", map[string]interface{}{
 				"error": resolveErr.Error(),
@@ -250,19 +262,18 @@ func (r *repository) resolveScrapedTendersScope(ctx context.Context) (scrapedTen
 		return scrapedTendersScope{}, err
 	}
 	if len(folderIDs) > 0 {
+		if len(folderIDs) > maxScrapedFolderInClause {
+			r.logger.Warn("Too many MinIO procedure folders for dashboard statistics, using Mongo fallback", map[string]interface{}{
+				"folder_count": len(folderIDs),
+				"max_in_clause": maxScrapedFolderInClause,
+			})
+			return mongoScrapedTendersScope(), nil
+		}
 		return scrapedTendersScope{
 			match: bson.M{"contract_folder_id": bson.M{"$in": folderIDs}},
 		}, nil
 	}
-	if r.procedureLister != nil {
-		return scrapedTendersScope{zero: true}, nil
-	}
-	return scrapedTendersScope{
-		match: bson.M{
-			"processing_metadata.documents_scraped": true,
-			"scraped_documents":                   bson.M{"$exists": true, "$ne": bson.A{}},
-		},
-	}, nil
+	return mongoScrapedTendersScope(), nil
 }
 
 func (r *repository) scrapedDocumentsFolderIDs(ctx context.Context) ([]string, error) {