app.py

import chainlit as cl
from dotenv import load_dotenv
from langchain.schema import StrOutputParser
from langchain.schema.runnable.config import RunnableConfig

from domain.models.part import Part
from domain.server import Server
from domain.workflows.part_workflow import PartWorkflow


@cl.on_chat_start
async def on_chat_start():    
    # runnable = agent | StrOutputParser()
    server = build_server()
    cl.user_session.set("server", server)

@cl.on_message
async def on_message(message: cl.Message):
    server = cl.user_session.get("server")
    msg = cl.Message(content="")
    async for chunk in server.get_response_from(message.content):
        await msg.stream_token(chunk)

    await msg.send()

@cl.set_starters
async def set_starters():
    return [
        cl.Starter(
            label="Start a session",
            message="Let's start a session.",
            ),
        cl.Starter(
            label="What is IFS?",
            message="What is IFS?",
            ),
        cl.Starter(
            label="Is this therapy?",
            message="Does this application count as therapy?",
            )
        ]

def build_server():
    workflows = [] # [ PartWorkflow(Part()) ]
    return Server(workflows)