_process_next_step_output() — langchain Function Reference

Architecture documentation for the _process_next_step_output() function in agent_iterator.py from the langchain codebase.

Function python AgentOrchestration ToolInterface calls 1 called by 1

Entity Profile

AgentOrchestration→ ToolInterface→ _process_next_step_output() — langchain Function Reference

Dependency Diagram

graph TD
  d5f8bce5_089f_ce88_6683_139f191b2eb6["_process_next_step_output()"]
  490b6bce_9d4c_017c_89ae_b13ea74223ea["AgentExecutorIterator"]
  d5f8bce5_089f_ce88_6683_139f191b2eb6 -->|defined in| 490b6bce_9d4c_017c_89ae_b13ea74223ea
  a9a06b32_02d8_cc0e_a2b9_4c37ea88936c["__iter__()"]
  a9a06b32_02d8_cc0e_a2b9_4c37ea88936c -->|calls| d5f8bce5_089f_ce88_6683_139f191b2eb6
  4d87dde3_a32d_10d0_6e36_d6572758987b["_return()"]
  d5f8bce5_089f_ce88_6683_139f191b2eb6 -->|calls| 4d87dde3_a32d_10d0_6e36_d6572758987b
  style d5f8bce5_089f_ce88_6683_139f191b2eb6 fill:#6366f1,stroke:#818cf8,color:#fff

Relationship Graph

Source Code

libs/langchain/langchain_classic/agents/agent_iterator.py lines 318–345

    def _process_next_step_output(
        self,
        next_step_output: AgentFinish | list[tuple[AgentAction, str]],
        run_manager: CallbackManagerForChainRun,
    ) -> AddableDict:
        """Process the output of the next step.

        Process the output of the next step,
        handling AgentFinish and tool return cases.
        """
        logger.debug("Processing output of Agent loop step")
        if isinstance(next_step_output, AgentFinish):
            logger.debug(
                "Hit AgentFinish: _return -> on_chain_end -> run final output logic",
            )
            return self._return(next_step_output, run_manager=run_manager)

        self.intermediate_steps.extend(next_step_output)
        logger.debug("Updated intermediate_steps with step output")

        # Check for tool return
        if len(next_step_output) == 1:
            next_step_action = next_step_output[0]
            tool_return = self.agent_executor._get_tool_return(next_step_action)  # noqa: SLF001
            if tool_return is not None:
                return self._return(tool_return, run_manager=run_manager)

        return AddableDict(intermediate_step=next_step_output)