fix logger

app.py

@@ -420,6 +420,7 @@ def _format_body_for_log(content_type: str, body: bytes) -> str:
 async def log_received_request(request: Request, call_next):
     """Log incoming requests and responses with basic metadata."""
     request_id = uuid.uuid4().hex[:12]
+    request.state.request_id = request_id
     start = time.perf_counter()
 
     body_text = ""
@@ -450,9 +451,9 @@ async def log_received_request(request: Request, call_next):
         elapsed_ms = (time.perf_counter() - start) * 1000
         logger.info(f"⬅️  {request_id} {response.status_code} {elapsed_ms:.1f}ms")
         return response
-    except Exception as exc:
+    except Exception:
         elapsed_ms = (time.perf_counter() - start) * 1000
-        logger.error(f"⬅️  {request_id} 500 {elapsed_ms:.1f}ms error={exc}")
+        logger.exception(f"⬅️  {request_id} 500 {elapsed_ms:.1f}ms unhandled error")
         raise
 
 
@@ -873,7 +874,7 @@ async def switch_model(request: ModelSwitchRequest):
     summary="Chat Completions",
     description="High-performance OpenAI-compatible chat completions with connection pooling."
 )
-async def chat_completions(request: ChatCompletionRequest):
+async def chat_completions(request: ChatCompletionRequest, raw_request: Request):
     """
     OpenAI-compatible chat completions with performance optimizations.
 
@@ -882,9 +883,13 @@ async def chat_completions(request: ChatCompletionRequest):
     - HTTP connection pooling
     - Request metrics tracking
     """
+    request_id = getattr(raw_request.state, "request_id", "-")
     try:
         request_start = time.time()
 
+        if not http_session or http_session.closed:
+            raise HTTPException(status_code=500, detail="HTTP session not initialized")
+
         # Get current model from cache
         cached_model = model_cache.get(current_model)
         if not cached_model:
@@ -910,7 +915,11 @@ async def chat_completions(request: ChatCompletionRequest):
 
         return result
     except aiohttp.ClientError as e:
+        logger.exception(f"request_id={request_id} llama-server error")
         raise HTTPException(status_code=500, detail=f"llama-server error: {str(e)}")
+    except Exception:
+        logger.exception(f"request_id={request_id} chat_completions error")
+        raise
 
 
 async def search_web_async(query: str, max_results: int = 5) -> list[dict]:
@@ -975,7 +984,7 @@ def format_search_context(query: str, search_results: list[dict]) -> str:
     summary="Web-Augmented Chat Completions",
     description="Chat completions with real-time web search and result caching."
 )
-async def web_chat_completions(request: WebChatRequest):
+async def web_chat_completions(request: WebChatRequest, raw_request: Request):
     """
     Chat completions with web search augmentation.
 
@@ -984,6 +993,7 @@ async def web_chat_completions(request: WebChatRequest):
     - LRU cache for search results (1 hour TTL)
     - Parallel execution where possible
     """
+    request_id = getattr(raw_request.state, "request_id", "-")
     try:
         # Get the last user message as search query
         user_messages = [msg for msg in request.messages if msg.get("role") == "user"]
@@ -1015,6 +1025,9 @@ Always cite sources when using information from the search results."""
 
         augmented_messages.insert(-1, system_prompt)
 
+        if not http_session or http_session.closed:
+            raise HTTPException(status_code=500, detail="HTTP session not initialized")
+
         # Get current model from cache
         cached_model = model_cache.get(current_model)
         if not cached_model:
@@ -1043,8 +1056,12 @@ Always cite sources when using information from the search results."""
         return result
 
     except aiohttp.ClientError as e:
+        logger.exception(f"request_id={request_id} llama-server error")
         raise HTTPException(status_code=500, detail=f"llama-server error: {str(e)}")
+    except HTTPException:
+        raise
     except Exception as e:
+        logger.exception(f"request_id={request_id} web_chat_completions error")
         raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
 
 

logger.py

@@ -80,12 +80,10 @@ class Logger:
         Returns:
             Configured logger instance
         """
-        # Create or get logger
-        logger = logging.getLogger(name)
-
-        # Only configure handlers once for the root logger
-        if not cls._initialized and name == "AGI":
-            logger.setLevel(level)
+        # Always return a named logger, but configure root handlers once
+        if not cls._initialized:
+            root_logger = logging.getLogger()
+            root_logger.setLevel(level)
 
             # Console handler with colors
             console_handler = logging.StreamHandler(sys.stdout)
@@ -95,7 +93,7 @@ class Logger:
                 datefmt='%Y-%m-%d %H:%M:%S'
             )
             console_handler.setFormatter(console_formatter)
-            logger.addHandler(console_handler)
+            root_logger.addHandler(console_handler)
 
             # File handler with rotation (if enabled)
             if log_file:
@@ -113,13 +111,11 @@ class Logger:
                     datefmt='%Y-%m-%d %H:%M:%S'
                 )
                 file_handler.setFormatter(file_formatter)
-                logger.addHandler(file_handler)
+                root_logger.addHandler(file_handler)
 
-            # Prevent propagation to avoid duplicate logs
-            logger.propagate = False
             cls._initialized = True
 
-        return logger
+        return logging.getLogger(name)
 
 
 # Convenience function for easy import
@@ -161,4 +157,4 @@ if __name__ == "__main__":
     api_logger.info("API logger initialized")
 
     client_logger = get_logger("client")
-    client_logger.info("Client logger initialized")
+    client_logger.info("Client logger initialized")