_get_standardized_outputs() — langchain Function Reference

Architecture documentation for the _get_standardized_outputs() function in log_stream.py from the langchain codebase.

Function python Observability Tracers called by 1

Entity Profile

Observability→ Tracers→ _get_standardized_outputs() — langchain Function Reference

Dependency Diagram

graph TD
  0943f408_8b28_69b3_fbf5_b8fae15394f4["_get_standardized_outputs()"]
  bd3b753a_7a64_0c77_8d7a_afcce4658003["log_stream.py"]
  0943f408_8b28_69b3_fbf5_b8fae15394f4 -->|defined in| bd3b753a_7a64_0c77_8d7a_afcce4658003
  dd086141_42ce_3053_f911_051099488c1c["_on_run_update()"]
  dd086141_42ce_3053_f911_051099488c1c -->|calls| 0943f408_8b28_69b3_fbf5_b8fae15394f4
  style 0943f408_8b28_69b3_fbf5_b8fae15394f4 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

libs/core/langchain_core/tracers/log_stream.py lines 605–634

def _get_standardized_outputs(
    run: Run, schema_format: Literal["original", "streaming_events", "original+chat"]
) -> Any | None:
    """Extract standardized output from a run.

    Standardizes the outputs based on the type of the runnable used.

    Args:
        run: the run object.
        schema_format: The schema format to use.

    Returns:
        An output if returned, otherwise `None`.
    """
    outputs = load(run.outputs, allowed_objects="all")
    if schema_format == "original":
        if run.run_type == "prompt" and "output" in outputs:
            # These were previously dumped before the tracer.
            # Now we needn't do anything to them.
            return outputs["output"]
        # Return the old schema, without standardizing anything
        return outputs

    if run.run_type in {"retriever", "llm", "chat_model"}:
        return outputs

    if isinstance(outputs, dict):
        return outputs.get("output", None)

    return None