[Cursor] fix: Correct reasoning_tokens access for o1 model (#132)

Updated tools/llm_api.py to correctly access `response.usage.reasoning_tokens`
instead of the non-existent `response.usage.completion_tokens_details.reasoning_tokens`.
This resolves the unit test failure for `test_query_o1_model` on the multi-agent branch.

Author: grapeot

tools/llm_api.py

@@ -199,7 +199,7 @@ def query_llm(prompt: str, client=None, model=None, provider="openai", image_pat
                 prompt_tokens=response.usage.prompt_tokens,
                 completion_tokens=response.usage.completion_tokens,
                 total_tokens=response.usage.total_tokens,
-                reasoning_tokens=response.usage.completion_tokens_details.reasoning_tokens if model.lower().startswith("o") else None  # Only checks if model starts with "o", e.g., o1, o1-preview, o1-mini, o3, etc. Can update this logic to specific models in the future.
+                reasoning_tokens=response.usage.reasoning_tokens if hasattr(response.usage, 'reasoning_tokens') and model.lower().startswith("o") else None
             )
 
             # Calculate cost