test_thinking() — langchain Function Reference

Architecture documentation for the test_thinking() function in test_chat_models.py from the langchain codebase.

Function python LangChainCore LanguageModelBase

Entity Profile

LangChainCore→ LanguageModelBase→ test_thinking() — langchain Function Reference

Dependency Diagram

graph TD
  bde73317_7d7c_4953_a071_7c85c3be53e7["test_thinking()"]
  f27640dd_3870_5548_d153_f9504ae1021f["test_chat_models.py"]
  bde73317_7d7c_4953_a071_7c85c3be53e7 -->|defined in| f27640dd_3870_5548_d153_f9504ae1021f
  style bde73317_7d7c_4953_a071_7c85c3be53e7 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

libs/partners/anthropic/tests/integration_tests/test_chat_models.py lines 1006–1043

def test_thinking() -> None:
    llm = ChatAnthropic(
        model="claude-sonnet-4-5-20250929",  # type: ignore[call-arg]
        max_tokens=5_000,  # type: ignore[call-arg]
        thinking={"type": "enabled", "budget_tokens": 2_000},
    )

    input_message = {"role": "user", "content": "Hello"}
    response = llm.invoke([input_message])
    assert any("thinking" in block for block in response.content)
    for block in response.content:
        assert isinstance(block, dict)
        if block["type"] == "thinking":
            assert set(block.keys()) == {"type", "thinking", "signature"}
            assert block["thinking"]
            assert isinstance(block["thinking"], str)
            assert block["signature"]
            assert isinstance(block["signature"], str)

    # Test streaming
    full: BaseMessageChunk | None = None
    for chunk in llm.stream([input_message]):
        full = cast("BaseMessageChunk", chunk) if full is None else full + chunk
    assert isinstance(full, AIMessageChunk)
    assert isinstance(full.content, list)
    assert any("thinking" in block for block in full.content)
    for block in full.content:
        assert isinstance(block, dict)
        if block["type"] == "thinking":
            assert set(block.keys()) == {"type", "thinking", "signature", "index"}
            assert block["thinking"]
            assert isinstance(block["thinking"], str)
            assert block["signature"]
            assert isinstance(block["signature"], str)

    # Test pass back in
    next_message = {"role": "user", "content": "How are you?"}
    _ = llm.invoke([input_message, full, next_message])