[BUG]: LangChain expects at least one chunk from a streaming trace after timeout

[MartinGotelli]

Tracer Version(s)

3.14.2

Python Version(s)

3.11.4

Pip Version(s)

pip 24.3.1

Bug Report

Using ddtrace-run command with LangChain and ChatOpenAI, after a timeout, the ddtrace integration fails.

packages/ddtrace/contrib/internal/langchain/patch.py", line 421, in _on_span_finished
    joined_chunks = streamed_chunks[0]
                    ~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

Reproduction Code

ddtrace-run python my_script.py -> "Oops!"
python my_script.py -> "Success!"

from langchain_openai import ChatOpenAI
from openai import APITimeoutError


def run():
    chain = ChatOpenAI(
        api_key="my_key",
        timeout=0.0001,
    )
    result = chain.stream("Hello, my name is")
    try:
        next(result)
    except APITimeoutError as e:
        print(f"Success! Caught expected timeout error: {e}")
    except Exception as e:
        print(f"Oops! Caught unexpected error: {e}")


if __name__ == "__main__":
    run()

Error Logs

No response

Libraries in Use

langchain==0.3.24
langchain-aws==0.2.10
langchain-community==0.3.22
langchain-core==0.3.55
langchain-openai==0.3.10
langchain-postgres==0.0.13
langchain-text-splitters==0.3.8
langdetect==1.0.9
langgraph==0.2.61
langgraph-checkpoint==2.0.9
langgraph-sdk==0.1.51
langsmith==0.2.10
openai==1.68.2

Operating System

No response

[MartinGotelli]

Sorry, I had to test this before creating the bug, but using latest LangChain/LangSmith libraries, the result is different.

With these versions works:

langchain==0.3.27
langchain-aws==0.2.10
langchain-community==0.3.27
langchain-core==0.3.76
langchain-openai==0.3.33
langchain-postgres==0.0.13
langchain-text-splitters==0.3.11
langdetect==1.0.9
langgraph==0.2.61
langgraph-checkpoint==2.0.9
langgraph-sdk==0.1.51
langsmith==0.4.30
openai==1.109.0

[MartinGotelli]

Also...

This code works, but running ddtrace-run does not work.

from langchain_openai import ChatOpenAI
from openai import APITimeoutError


def run():
    chain = ChatOpenAI(
        api_key="my_key",
        timeout=0.0001,
    )
    result = chain.stream("Hello, my name is")
    try:
        next(result)
    except APITimeoutError as e:
        print(f"Success! Caught expected timeout error: {e}")
    except Exception as e:
        print(f"Oops! Caught unexpected error: {e}")


if __name__ == "__main__":
    from ddtrace import tracer

    with tracer.trace("ddtrace_langchain_stream"):
        run()

EDIT: I was missing the import ddtrace.auto. This code fails (with old and new langchain library versions):

if __name__ == "__main__":
    import ddtrace.auto
    from ddtrace import tracer

    with tracer.trace("ddtrace_langchain_stream"):
        run()