from langchain_core.messages.tool import default_tool_parser, default_tool_chunk_parser

Case 1: function value is None

raw = [{"function": None, "id": "call_1"}] default_tool_parser(raw) # TypeError

Case 2: function dict missing keys

raw = [{"function": {}, "id": "call_1"}] default_tool_parser(raw) # KeyError: name

Case 3: One malformed call drops ALL valid calls

raw = [ {"function": None, "id": "bad"}, {"function": {"name": "good_tool", "arguments": "{"a": 1}"}, "id": "good"}, ] default_tool_parser(raw) # Crashes on first item, good call never parsed

TL;DR

Modify default_tool_parser and default_tool_chunk_parser to validate the function field and handle missing or malformed data using the .get() method.

Guidance

• Validate the function field in default_tool_parser and default_tool_chunk_parser to ensure it is a dictionary with the expected keys before attempting to access its values.
• Use the .get() method to safely retrieve values from the function dictionary, avoiding KeyError exceptions.
• Modify the parsing loop to continue processing valid tool calls even if a single malformed call is encountered, utilizing the invalid_tool_calls return list as intended.
• Review the documentation for default_tool_parser to reflect its actual behavior and consider updating it to match the intended best-effort parsing approach.

Example

function_name = raw_tool_call["function"].get("name")
if function_name is None:
    # Handle missing or malformed function name
    invalid_tool_calls.append(raw_tool_call)
    continue

Notes

The provided solution focuses on modifying the default_tool_parser and default_tool_chunk_parser functions to improve their robustness against malformed input data. However, a comprehensive solution might require additional changes, such as updating the documentation and potentially adjusting the error handling in the caller function (ai.py) to better handle exceptions.

Recommendation

Apply workaround: Modify the default_tool_parser and default_tool_chunk_parser functions as described to improve their handling of malformed input data and to align with the intended best-effort parsing behavior.