frontend.py

import streamlit as st
from backend import interpret_command_with_gpt, process_interpreted_command, writeAssistantResponse

# Initialize session state for messages, chatbot visibility, and context
if "messages" not in st.session_state:
    st.session_state.messages = []
if "show_chatbot" not in st.session_state:
    st.session_state.show_chatbot = False
if "context" not in st.session_state:
    st.session_state.context = ""  # Initialize context to store conversation history

# Title of the app
st.title("My First Chatbot")

# Button to start chatting with the bot
if st.button("Start Chatting"):
    st.session_state.show_chatbot = True
    # Initial message from the assistant when chat starts
    response = "Hi there, How can I assist you?"
    st.session_state.messages.append({"role": "assistant", "content": {
        "response": response
    } })
    st.session_state.context += f"Bot: {response}\n"  # Add response to the context

# Display the chatbot interface once the chat starts
if st.session_state.show_chatbot:
    # Display each message from the session state
    for message in st.session_state.messages:
        print(message)
        with st.chat_message(message["role"]):
            if "code" in message:
                st.code(message["code"], language=message["language"])
            else:
                st.markdown(message["content"]["response"])

    # User prompt input
    if prompt := st.chat_input("What is up?"):
        # Display the user input in the chat window
        with st.chat_message("user"):
            st.markdown(f"{prompt}")
        # Save user input to session state
        st.session_state.messages.append({"role": "user", "content": {
            "response": prompt
        }})

        # Append user input to context
        st.session_state.context += f"User: {prompt}\n"

        # Interpret the command with GPT, passing the full conversation context
        full_prompt = f"Previous chat context: {st.session_state.context}\n\nCurrent message: {prompt}"
        interpreted_command = interpret_command_with_gpt(full_prompt)

        # Process the interpreted command (generates response)
        response = process_interpreted_command(interpreted_command)

        # Append assistant's response to context
        st.session_state.context += f"Bot: {response}\n"

        # Display the assistant's response with typing animation
        if "code" in response:  # Simple heuristic to detect code-related output
            print(response)
            with st.chat_message("assistant"):
                a = response["language"]
                st.code(response["code"], language=a)  # Display response as a Python code block

                st.session_state.messages.append(
                    {
                        "role": "assistant",
                        "code": response["code"],
                        "language": response["language"]
                    }
                )
        else:
            # Display the assistant's response with typing animation
            writeAssistantResponse(response["response"])