app.py

import streamlit as st
from streamlit_chat import message
from langchain.callbacks.base import BaseCallbackHandler
from langchain.chains import ConversationChain
from utils import DataScienceConsultant


st.set_page_config(page_title='🤖 Data Generator Assistant', layout='centered', page_icon='🤖')
st.title("🤖 Chat with AI")

# initial message
INIT_MESSAGE = {"role": "assistant",
                "content": "Hello! I am a Data Science Consultant. I will help you create the right data for your product. "}


def init_conversationchain() -> ConversationChain:
    chat_executor = DataScienceConsultant()

    # Store LLM generated responses
    if "messages" not in st.session_state:
        st.session_state.messages = [INIT_MESSAGE]

    return chat_executor

def generate_response(conversation: ConversationChain, input_text: str) -> str:
    try:
        response = conversation.predict(input_text)
    except ValueError as e:
        print("################",str(e))
        response = str(e)
        if not response.startswith("Could not parse LLM output: `"):
            response = "There were some error in answering this question. "
        else:
            response = response.removeprefix("Could not parse LLM output: `").removesuffix("`")
    return response


# Re-initialize the chat
def new_chat() -> None:
    st.session_state["messages"] = [INIT_MESSAGE]
    st.session_state["langchain_messages"] = []
    conv_chain = init_conversationchain()


# Add a button to start a new chat
st.sidebar.button("New Chat", on_click=new_chat, type='primary')


# Initialize the conversation chain
conversation = init_conversationchain()

# Display chat messages
for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])

# Get user input
user_input = st.chat_input(placeholder="Your message ....", key="input")

# display user input
if user_input:
    st.session_state.messages.append({"role": "user", "content": user_input})
    user_message = st.chat_message("user")
    user_message.write(user_input)

# Generate response
if st.session_state.messages[-1]["role"] != "assistant":
    response = generate_response(conversation, user_input)
    st.session_state.messages.append({"role": "assistant", "content": response})
    assistant_message = st.chat_message("assistant")
    assistant_message.write(response)