Issue with Chat with histories

[yw791772]

I'm experimenting on the message history pattern backed with dynamoDB.

On the client side, I'm following the code here, I have it as

prompt = "blah"
async with aconnect_sse(
      async_client,
      "POST",
      "stream",
      json={
          "input": prompt
          "config": {"session_id": "foo"}
      },
  ) as event_source:
     ...

on the server side I have:

with_message_history = RunnableWithMessageHistory(
        ChatOpenAI(),
        get_session_history,
        input_messages_key="input",
        history_messages_key="history",
    )

    add_routes(
        app,
        with_message_history,
        path="/openai/history",
    )

I'm getting following error:

ERROR:    Exception in ASGI application
Traceback (most recent call last):
  File "/home/vscode/.local/share/virtualenvs/langserve-3zV0jjiH/lib/python3.10/site-packages/pydantic/v1/main.py", line 716, in validate
    value_as_dict = dict(value)
ValueError: dictionary update sequence element #0 has length 1; 2 is required

The above exception was the direct cause of the following exception:

It's been working fine for a normal chat completion end point for me for a while now. I imagine it's because runnableWithHistory's input_message_key is set, that expect a different input format?

What else I tried:

• I look at the examples and the client.py code, change the input to {human_input: prompt}, no pandandic errors, and I add a configurables on the server side to hard code a session_id to the configurable dict. And then I hit this error:

ValueError: Invalid input type <class 'langchain_core.runnables.utils.AddableDict'>. Must be a PromptValue, str, or list of BaseMessages.

• I notices that the client code is doing a _lc_serializer.dumpd(input), so I try that and see if that will help, still a no

I'm kind lost on what the correct format I should be using, wondering if anyone could help me

[yw791772]

I'm dumb, I have a layer of code doing the template, and didn't realize that I need to chain the template with llm