fix(langchain): handle optional serialized values gracefully

[hassiebp]

Important

Handle optional serialized values gracefully in langchain.py and extract_model.py by updating functions to accept None and adding error handling.

• Behavior:
  • Update get_langchain_run_name() in langchain.py to handle None for serialized argument, using try-except blocks for key access.
  • Modify _extract_model_name() in extract_model.py to accept None for serialized, adjusting logic to prevent errors.
• Functions:
  • Change serialized parameter type from Dict[str, Any] to Optional[Dict[str, Any]] in multiple functions in langchain.py and extract_model.py, including on_chain_start(), on_llm_start(), and _extract_model_by_path_for_id().
  • Add checks for None in _extract_model_from_repr_by_pattern() and _extract_model_by_path() in extract_model.py.
• Misc:
  • Update docstrings in langchain.py to reflect changes in parameter types and behavior.

^{This description was created by for b1a3ce6. It will automatically update as commits are pushed.}

[greptile-apps]

Disclaimer: Experimental PR review

PR Summary

This PR improves error handling and type safety for optional serialized values in the Langchain integration of langfuse-python.

• Fixed unsafe dictionary access in extract_model.py by adding proper null checks for serialized parameter
• Added proper type hints for optional parameters in langchain.py callback handler methods
• Refactored get_langchain_run_name to use try/except blocks instead of dict.get() for safer error handling
• Improved handling of potentially missing keys in AzureOpenAI model extraction logic
• Added documentation clarifying optional parameter behavior in callback methods

_{2 file(s) reviewed, no comment(s)
Edit PR Review Bot Settings | Greptile}