Allow calls to batch() with 0 length arrays

[nfcampos]

This can happen if eg the input to batch is a list generated dynamically, where a 0-length list might be a valid use case

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Ignored Deployment

Name	Status	Preview	Comments	Updated (UTC)
langchain	⬜️ Ignored (Inspect)	Visit Preview		Sep 15, 2023 2:26pm

[eyurtsev]

I think additional code is needed in default batch implementation since it assumes that the list of configs cannot be empty

        """
        configs = get_config_list(config, len(inputs))

        def invoke(input: Input, config: RunnableConfig) -> Union[Output, Exception]:
            if return_exceptions:
                try:
                    return self.invoke(input, config, **kwargs)
                except Exception as e:
                    return e
            else:
                return self.invoke(input, config, **kwargs)

        # If there's only one input, don't bother with the executor
        if len(inputs) == 1:
            return cast(List[Output], [invoke(inputs[0], configs[0])])

        with get_executor_for_config(configs[0]) as executor:
            return cast(List[Output], list(executor.map(invoke, inputs, configs)))

[nfcampos]

Good point! updated

[eyurtsev]

from langchain.chat_models.openai import ChatOpenAI
llm = ChatOpenAI()
llm.batch([])

.batch([]) won't produce any traces on langsmith