test_agent_iterator_failing_tool() — langchain Function Reference

Architecture documentation for the test_agent_iterator_failing_tool() function in test_agent_iterator.py from the langchain codebase.

Function python LangChainCore Runnables

Entity Profile

LangChainCore→ Runnables→ test_agent_iterator_failing_tool() — langchain Function Reference

Dependency Diagram

graph TD
  c45b4c9f_777c_d824_994d_4d769a269faa["test_agent_iterator_failing_tool()"]
  2423e003_42f4_7c6a_3fba_30db54c07a0c["test_agent_iterator.py"]
  c45b4c9f_777c_d824_994d_4d769a269faa -->|defined in| 2423e003_42f4_7c6a_3fba_30db54c07a0c
  style c45b4c9f_777c_d824_994d_4d769a269faa fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

libs/langchain/tests/unit_tests/agents/test_agent_iterator.py lines 355–386

def test_agent_iterator_failing_tool() -> None:
    """Test AgentExecutorIterator with a tool that raises an exception."""
    # Get agent for testing.
    bad_action_name = "FailingTool"
    responses = [
        f"I'm turning evil\nAction: {bad_action_name}\nAction Input: misalignment",
        "Oh well\nFinal Answer: curses foiled again",
    ]
    fake_llm = FakeListLLM(responses=responses)

    tools = [
        Tool(
            name="FailingTool",
            func=lambda _: 1 / 0,  # This tool will raise a ZeroDivisionError
            description="A tool that fails",
        ),
    ]

    agent = initialize_agent(
        tools,
        fake_llm,
        agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
        verbose=True,
    )

    agent_iter = agent.iter(inputs="when was langchain made")
    assert isinstance(agent_iter, AgentExecutorIterator)
    # initialize iterator
    iterator = iter(agent_iter)

    with pytest.raises(ZeroDivisionError):
        next(iterator)