test_usage_callback_async() — langchain Function Reference

Architecture documentation for the test_usage_callback_async() function in test_usage_callback.py from the langchain codebase.

Function python CoreAbstractions RunnableInterface

Entity Profile

CoreAbstractions→ RunnableInterface→ test_usage_callback_async() — langchain Function Reference

Dependency Diagram

graph TD
  5a5a2bd7_30bf_4b76_56ec_04f23e2ab781["test_usage_callback_async()"]
  b9fdf1fe_41aa_aa38_11e1_ae02e5bde19d["test_usage_callback.py"]
  5a5a2bd7_30bf_4b76_56ec_04f23e2ab781 -->|defined in| b9fdf1fe_41aa_aa38_11e1_ae02e5bde19d
  style 5a5a2bd7_30bf_4b76_56ec_04f23e2ab781 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

libs/core/tests/unit_tests/callbacks/test_usage_callback.py lines 100–122

async def test_usage_callback_async() -> None:
    llm = FakeChatModelWithResponseMetadata(
        messages=iter(messages), model_name="test_model"
    )

    # Test context manager
    with get_usage_metadata_callback() as cb:
        _ = await llm.ainvoke("Message 1")
        _ = await llm.ainvoke("Message 2")
        total_1_2 = add_usage(usage1, usage2)
        assert cb.usage_metadata == {"test_model": total_1_2}
        _ = await llm.ainvoke("Message 3")
        _ = await llm.ainvoke("Message 4")
        total_3_4 = add_usage(usage3, usage4)
        assert cb.usage_metadata == {"test_model": add_usage(total_1_2, total_3_4)}

    # Test via config
    llm = FakeChatModelWithResponseMetadata(
        messages=iter(messages[:2]), model_name="test_model"
    )
    callback = UsageMetadataCallbackHandler()
    _ = await llm.abatch(["Message 1", "Message 2"], config={"callbacks": [callback]})
    assert callback.usage_metadata == {"test_model": total_1_2}