Update app.py

app.py

@@ -23,7 +23,7 @@ GITHUB_PRIVATE_KEY = os.getenv("GITHUB_PRIVATE_KEY")
 # In-memory storage for our repository data
 repo_data_store = {}
 
-# --- NEW: Code Processing and Vectorization Class ---
+# --- Code Processing and Vectorization Class ---
 
 class CodeProcessor:
     def __init__(self, model_name='all-MiniLM-L6-v2'):
@@ -32,24 +32,30 @@ class CodeProcessor:
         print("Model initialized.")
 
     def parse_python_file(self, file_path):
-        """Parses a Python file to extract functions and their docstrings using AST."""
-        with open(file_path, 'r', encoding='utf-8') as source:
-            try:
-                tree = ast.parse(source.read())
-            except SyntaxError:
-                return [] # Ignore files with syntax errors
+        """Parses a Python file to extract functions and their source code using AST."""
+        # --- THE FIX IS HERE ---
+        # First, read the entire file content into a string variable.
+        with open(file_path, 'r', encoding='utf-8') as file:
+            source_code = file.read()
+        
+        try:
+            # Parse the source code string.
+            tree = ast.parse(source_code)
+        except SyntaxError:
+            print(f"SyntaxError in {file_path}, skipping.")
+            return []
         
         functions = []
         for node in ast.walk(tree):
             if isinstance(node, ast.FunctionDef):
-                # We extract the function's name and its full source code
-                # The source code provides more context than just the docstring
-                function_source = ast.get_source_segment(tree, node)
-                functions.append({
-                    "name": node.name,
-                    "source": function_source,
-                    "file_path": file_path
-                })
+                # Now, pass the correct `source_code` string to get the segment.
+                function_source = ast.get_source_segment(source_code, node)
+                if function_source: # Ensure we got a valid segment
+                    functions.append({
+                        "name": node.name,
+                        "source": function_source,
+                        "file_path": file_path.replace(tempfile.gettempdir(), "") # Store relative path
+                    })
         return functions
 
     def process_repo(self, repo_path):
@@ -59,44 +65,44 @@ class CodeProcessor:
         for root, _, files in os.walk(repo_path):
             for file in files:
                 if file.endswith('.py'):
-                    file_path = os.path.join(root, file)
-                    functions = self.parse_python_file(file_path)
-                    all_functions.extend(functions)
+                    # Create the full path relative to the repo root for parsing
+                    full_path = os.path.join(root, file)
+                    # Create a "clean" path for storage in metadata
+                    clean_path = os.path.relpath(full_path, repo_path)
+                    
+                    parsed_funcs = self.parse_python_file(full_path)
+                    for func in parsed_funcs:
+                        func['file_path'] = clean_path # Overwrite with the clean path
+                    all_functions.extend(parsed_funcs)
         
         if not all_functions:
             print("No Python functions found in the repository.")
             return None, None
 
         print(f"Found {len(all_functions)} functions. Generating embeddings...")
-        # We create embeddings for the source code of each function
         function_sources = [f["source"] for f in all_functions]
         embeddings = self.model.encode(function_sources, show_progress_bar=False)
         
-        # Create a FAISS index
         embedding_dim = embeddings.shape[1]
         index = faiss.IndexFlatL2(embedding_dim)
         index.add(np.array(embeddings, dtype=np.float32))
         
         print(f"FAISS index created successfully with {index.ntotal} vectors.")
         
-        # We return the index and the metadata (the list of function details)
         return index, all_functions
 
 # --- Initialize the processor globally ---
-# This ensures the model is loaded only once when the app starts.
 code_processor = CodeProcessor()
 
 
-# --- GitHub App Authentication & Repo Management (Largely Unchanged) ---
-
+# --- GitHub App Authentication & Repo Management (Unchanged) ---
+# ... (The rest of the file from the previous correct version remains the same) ...
 def create_jwt(app_id, private_key):
-    # (This function is unchanged)
     now = int(time.time())
     payload = {"iat": now, "exp": now + (10 * 60), "iss": app_id}
     return jwt.encode(payload, private_key, algorithm="RS256")
 
 def get_installation_access_token(installation_id, app_id, private_key):
-    # (This function is unchanged)
     app_jwt = create_jwt(app_id, private_key)
     headers = {"Authorization": f"Bearer {app_jwt}", "Accept": "application/vnd.github.v3+json"}
     url = f"https://api.github.com/app/installations/{installation_id}/access_tokens"
@@ -113,11 +119,9 @@ def process_repository(repo_url, token, repo_full_name):
         clone_url = repo_url.replace("https://", f"https://x-access-token:{token}@")
         Repo.clone_from(clone_url, temp_dir)
         
-        # --- NEW: Use the CodeProcessor ---
         faiss_index, metadata = code_processor.process_repo(temp_dir)
         
         if faiss_index:
-            # Store the live FAISS index and metadata in our in-memory store
             repo_data_store[repo_full_name] = {
                 "status": "processed",
                 "faiss_index": faiss_index,
@@ -130,16 +134,13 @@ def process_repository(repo_url, token, repo_full_name):
     except Exception as e:
         print(f"Failed to process repository {repo_full_name}: {e}")
     finally:
-        # Clean up the cloned repository from disk
         print(f"Cleaning up temporary directory: {temp_dir}")
         shutil.rmtree(temp_dir)
 
-
 # --- FastAPI App (Webhook handler is unchanged) ---
 app = FastAPI()
 
 async def verify_signature(request: Request):
-    # (This function is unchanged)
     if not GITHUB_WEBHOOK_SECRET: raise HTTPException(status_code=500, detail="Webhook secret not configured.")
     signature_header = request.headers.get("X-Hub-Signature-256")
     if not signature_header: raise HTTPException(status_code=400, detail="X-Hub-Signature-256 header is missing.")
@@ -154,7 +155,6 @@ def read_root():
 
 @app.post("/api/github/webhook")
 async def github_webhook(request: Request):
-    # (This webhook handler is unchanged from the last step)
     await verify_signature(request)
     event_type = request.headers.get("X-GitHub-Event")
     payload = await request.json()