The agent engineering platform.
Python agentsaiai-agentsanthropicchatgptdeepagentsenterpriseframeworkgeminigenerative-ailangchainlanggraphllmmultiagentopen-sourceopenaipydanticpythonragtypescript
community[patch]: Fix `tool_calls` parsing when streaming from DeepInfra (#26813)
Tom2b83c7c3ab16
diff --git a/libs/community/langchain_community/chat_models/deepinfra.py b/libs/community/langchain_community/chat_models/deepinfra.pyindex 061bed9e8b..71389cc2c9 100644--- a/libs/community/langchain_community/chat_models/deepinfra.py+++ b/libs/community/langchain_community/chat_models/deepinfra.py@@ -144,13 +144,12 @@ def _convert_delta_to_message_chunk() -> BaseMessageChunk:role = _dict.get("role")content = _dict.get("content") or ""+ tool_calls = _dict.get("tool_calls") or []if role == "user" or default_class == HumanMessageChunk:return HumanMessageChunk(content=content)elif role == "assistant" or default_class == AIMessageChunk:- tool_calls = [- _parse_tool_calling(tool_call) for tool_call in _dict.get("tool_calls", [])- ]+ tool_calls = [_parse_tool_calling(tool_call) for tool_call in tool_calls]return AIMessageChunk(content=content, tool_calls=tool_calls)elif role == "system" or default_class == SystemMessageChunk:return SystemMessageChunk(content=content)