Debug about incremental indexing and also MCP

fastcode/loader.py

@@ -15,6 +15,7 @@
     is_supported_file,
     should_ignore_path,
     get_repo_name_from_url,
+    get_repo_name_from_path,
     normalize_path,
     ensure_dir,
 )
@@ -135,9 +136,12 @@ def load_from_path(self, path: str, target_dir: Optional[str] = None) -> str:
         if not os.path.isdir(path):
             raise ValueError(f"Path is not a directory: {path}")
 
-        source_path = os.path.abspath(path)
-        self.repo_name = os.path.basename(source_path)
         destination_root = os.path.abspath(target_dir) if target_dir else self.safe_repo_root
+        source_path = os.path.abspath(path)
+        self.repo_name = get_repo_name_from_path(
+            source_path,
+            workspace_root=destination_root,
+        )
         destination_path = os.path.join(destination_root, self.repo_name)
 
         # If source is already in workspace destination, use it directly.
@@ -381,4 +385,3 @@ def cleanup(self):
     def __del__(self):
         """Cleanup on deletion"""
         self.cleanup()
-

fastcode/query_processor.py

@@ -533,7 +533,7 @@ def _enhance_with_llm(self, query: str, intent: str,
             else:
                 return {}
 
-            print(f"LLM response of _enhance_with_llm: {response}")
+            self.logger.debug("LLM enhancement response received")
 
             # Parse LLM response
             enhancements = self._parse_llm_response(response, intent)
@@ -825,4 +825,3 @@ def _parse_rewritten_query(self, response: str) -> Optional[str]:
             rewritten = rewritten[1:-1]
 
         return rewritten if rewritten else None
-

fastcode/retriever.py

@@ -501,7 +501,7 @@ def _select_relevant_repositories(self, query: Union[str, List[str]], keywords:
         # 2. get top k
         for repo_name, scores in sorted_repos[:top_k]:
             selected_repos.append(repo_name)
-            print(
+            self.logger.info(
                 f"Selected repo: {repo_name} "
                 f"(semantic: {scores['semantic_score']:.3f}, "
                 f"bm25: {scores['bm25_score']:.3f}, "
@@ -510,7 +510,9 @@ def _select_relevant_repositories(self, query: Union[str, List[str]], keywords:
 
         # 3. no repo selected
         if not selected_repos:
-            print(f"No repositories met the minimum score threshold of {MIN_SCORE_THRESHOLD}")
+            self.logger.info(
+                f"No repositories met the minimum score threshold of {MIN_SCORE_THRESHOLD}"
+            )
 
         return selected_repos
 
@@ -1440,4 +1442,3 @@ def _apply_agency_mode(self, query: str, results: List[Dict[str, Any]],
             self.logger.error(traceback.format_exc())
             # Fallback to original results
             return results
-

fastcode/utils.py

@@ -310,6 +310,27 @@ def get_repo_name_from_url(url: str) -> str:
     return parts[-1] if parts else "unknown_repo"
 
 
+def get_repo_name_from_path(path: str, workspace_root: Optional[str] = None) -> str:
+    """Derive a stable local repository identifier from its absolute path."""
+    normalized = normalize_path(os.path.abspath(path or ""))
+    base_name = os.path.basename(normalized.rstrip("/")) or "local_repo"
+
+    if workspace_root:
+        normalized_root = normalize_path(os.path.abspath(workspace_root))
+        parent_dir = normalize_path(os.path.dirname(normalized.rstrip("/")))
+        suffix = base_name.rsplit("-", 1)
+        if (
+            parent_dir == normalized_root
+            and len(suffix) == 2
+            and len(suffix[1]) == 8
+            and all(ch in "0123456789abcdef" for ch in suffix[1].lower())
+        ):
+            return base_name
+
+    digest = hashlib.md5(normalized.encode("utf-8")).hexdigest()[:8]
+    return f"{base_name}-{digest}"
+
+
 def clean_docstring(docstring: str) -> str:
     """Clean and format docstring"""
     if not docstring:
@@ -335,4 +356,3 @@ def clean_docstring(docstring: str) -> str:
                  for line in lines]
 
     return "\n".join(lines).strip()
-

fastcode/vector_store.py

@@ -336,7 +336,9 @@ def search_repository_overviews(self, query_vector: np.ndarray, k: int = 5,
                 distance = float(np.linalg.norm(query_vector - embedding))
                 similarity = 1.0 / (1.0 + distance)
 
-            print(f"similarity: {similarity}, repo_name: {repo_name}")
+            self.logger.debug(
+                f"Repository overview similarity for {repo_name}: {similarity:.4f}"
+            )
 
             # Apply minimum score filter
             if min_score is not None and similarity < min_score:
@@ -752,4 +754,3 @@ def invalidate_scan_cache(self):
         """Invalidate the scan cache (call this when indexes change)"""
         self._index_scan_cache = None
         self.logger.debug("Invalidated index scan cache")
-