feat: add model_name

Author: harvey_xiang

src/memos/llms/openai.py

@@ -28,12 +28,14 @@ def __init__(self, config: OpenAILLMConfig):
         )
         logger.info("OpenAI LLM instance initialized")
 
-    @timed_with_status(log_prefix="OpenAI LLM", log_args=["model_name_or_path"])
+    @timed_with_status(
+        log_prefix="OpenAI LLM",
+        log_extra_args=lambda self, messages, **kwargs: {
+            "model_name_or_path": kwargs.get("model_name_or_path", self.config.model_name_or_path)
+        },
+    )
     def generate(self, messages: MessageList, **kwargs) -> str:
         """Generate a response from OpenAI LLM, optionally overriding generation params."""
-        logger.info(
-            f"LLM Model: {self.config.model_name_or_path} {kwargs.get('model_name_or_path')}"
-        )
         response = self.client.chat.completions.create(
             model=kwargs.get("model_name_or_path", self.config.model_name_or_path),
             messages=messages,
@@ -58,7 +60,12 @@ def generate(self, messages: MessageList, **kwargs) -> str:
             return reasoning_content + response_content
         return response_content
 
-    @timed_with_status(log_prefix="OpenAI LLM", log_args=["model_name_or_path"])
+    @timed_with_status(
+        log_prefix="OpenAI LLM",
+        log_extra_args=lambda self, messages, **kwargs: {
+            "model_name_or_path": self.config.model_name_or_path
+        },
+    )
     def generate_stream(self, messages: MessageList, **kwargs) -> Generator[str, None, None]:
         """Stream response from OpenAI LLM with optional reasoning support."""
         if kwargs.get("tools"):

src/memos/utils.py

@@ -19,8 +19,10 @@ def timed_with_status(
     Parameters:
     - log: enable timing logs (default True)
     - log_prefix: prefix; falls back to function name
-    - log_args: names to include in logs (str or list/tuple of str).
-    - log_extra_args: extra arguments to include in logs (dict).
+    - log_args: names to include in logs (str or list/tuple of str), values are taken from kwargs by name.
+    - log_extra_args:
+        - can be a dict: fixed contextual fields that are always attached to logs;
+        - or a callable: like `fn(*args, **kwargs) -> dict`, used to dynamically generate contextual fields at runtime.
     """
 
     if isinstance(log_args, str):
@@ -51,12 +53,24 @@ def wrapper(*args, **kwargs):
                 elapsed_ms = (time.perf_counter() - start) * 1000.0
 
                 ctx_parts = []
+                # 1) Collect parameters from kwargs by name
                 for key in effective_log_args:
                     val = kwargs.get(key)
                     ctx_parts.append(f"{key}={val}")
 
-                if log_extra_args:
-                    ctx_parts.extend(f"{key}={val}" for key, val in log_extra_args.items())
+                # 2) Support log_extra_args as dict or callable, so we can dynamically
+                #    extract values from self or other runtime context
+                extra_items = {}
+                try:
+                    if callable(log_extra_args):
+                        extra_items = log_extra_args(*args, **kwargs) or {}
+                    elif isinstance(log_extra_args, dict):
+                        extra_items = log_extra_args
+                except Exception as e:
+                    logger.warning(f"[TIMER_WITH_STATUS] log_extra_args callback error: {e!r}")
+
+                if extra_items:
+                    ctx_parts.extend(f"{key}={val}" for key, val in extra_items.items())
 
                 ctx_str = f" [{', '.join(ctx_parts)}]" if ctx_parts else ""