Update app.py

app.py

@@ -253,6 +253,15 @@ def semantic_search(query: str, top_k=1, category_filter=None, language='en'):
             if file_path not in best_scores or score > best_scores[file_path]:
                 best_scores[file_path] = score
         
+        # Debug logging for duplicate investigation
+        if query.lower() == "quality industrial zone":
+            logger.info(f"Debug - Query: {query}")
+            logger.info(f"Debug - Total matches from Pinecone: {len(res['matches'])}")
+            logger.info(f"Debug - Unique documents after deduplication: {len(best_scores)}")
+            logger.info(f"Debug - Document scores: {list(best_scores.items())[:5]}")
+            for file_path, score in list(best_scores.items())[:3]:
+                logger.info(f"Debug - Document: {file_path}, Score: {score}")
+        
         if not best_scores:
             return []
         
@@ -279,10 +288,14 @@ def semantic_search(query: str, top_k=1, category_filter=None, language='en'):
             max_docs = 5
         
         # Create a lookup dictionary for efficient metadata retrieval
+        # Store the match with the highest score for each file_path
         metadata_lookup = {}
         for match in res["matches"]:
             file_path_key = match["metadata"].get("file_path", "")
-            if file_path_key not in metadata_lookup:
+            score = match["score"]
+            
+            # Only store if this is the first match for this file_path or if it has a higher score
+            if file_path_key not in metadata_lookup or score > metadata_lookup[file_path_key]["score"]:
                 metadata_lookup[file_path_key] = match
         
         results = []
@@ -334,6 +347,12 @@ def semantic_search(query: str, top_k=1, category_filter=None, language='en'):
                     results.append(result)
                     doc_count += 1
         
+        # Debug logging for final results
+        if query.lower() == "quality industrial zone":
+            logger.info(f"Debug - Final results count: {len(results)}")
+            for i, result in enumerate(results):
+                logger.info(f"Debug - Result {i+1}: {result.get('title', 'No title')} - {result.get('file_path', 'No path')}")
+        
         return results
     except Exception as e:
         logger.error(f"Search error: {e}")