app.py

import chainlit as cl
from legal_agent.components.full_workflow import run_user_query


async def user_query_func(user_question):
    response = run_user_query(user_question)
    # Replace this with your actual logic for processing the user query
    # It could involve interacting with an LLM, searching web documents, etc.
    # For illustration purposes, let's just return a simple response
    return response


@cl.on_chat_start
def start():
    # user_query

    print("Chat started!")


@cl.on_message
async def main(message: cl.Message):
    # user_query
    user_question = message.content
    # response = user_query(user_question)
    # response = await user_query_func("What happended to the birds")
    response = await user_query_func(user_question)
    print(user_question, "see")
    # user_query = cl.make_async(user_query)

    # await user_query("What happended to the birds")
    # print(user_question, "see22222222")

    # Use LangchainCallbackHandler to capture the final answer
    # callback_handler = LangchainCallbackHandler(stream_final_answer=True)
    # response = await cl.make_async(user_query)(user_question)
    # response = await cl.make_async(user_query)(user_question)

    # await message.reply(response)
    await cl.Message(content=response["response"]).send()