demo.py

import gradio as gr
from gradio_calendar import Calendar
import pandas as pd
import psycopg2
from psycopg2 import sql
from openai import OpenAI
from pprint import pprint
from user import User
from dotenv import load_dotenv
import os
from datetime import timedelta
import json
import traceback
import shutil

# Load environment variables from .env file
load_dotenv()

# Function to fetch all users (name and created_at) for display
def fetch_users():
    db_params = {
        'dbname': 'ourcoach',
        'user': 'ourcoach',
        'password': 'hvcTL3kN3pOG5KteT17T',
        'host': 'staging-ourcoach.cx8se8o0iaiy.ap-southeast-1.rds.amazonaws.com',
        'port': '5432'
    }
    try:
        with psycopg2.connect(**db_params) as conn:
            with conn.cursor() as cursor:
                query = sql.SQL("SELECT id, name, created_at FROM {table}").format(
                    table=sql.Identifier('public', 'users')
                )
                cursor.execute(query)
                rows = cursor.fetchall()
                colnames = [desc[0] for desc in cursor.description]
                df = pd.DataFrame(rows, columns=colnames)
                return df[['id', 'name', 'created_at']]
    except psycopg2.Error as e:
        print(f"An error occurred: {e}")
        return pd.DataFrame()

def get_user_info(user_id):
    db_params = {
        'dbname': 'ourcoach',
        'user': 'ourcoach',
        'password': 'hvcTL3kN3pOG5KteT17T',
        'host': 'staging-ourcoach.cx8se8o0iaiy.ap-southeast-1.rds.amazonaws.com',
        'port': '5432'
    }
    try:
        with psycopg2.connect(**db_params) as conn:
            with conn.cursor() as cursor:
                query = sql.SQL("SELECT * FROM {table} WHERE id = %s").format(
                    table=sql.Identifier('public', 'users')
                )
                cursor.execute(query, (user_id,))
                row = cursor.fetchone()
                if row:
                    colnames = [desc[0] for desc in cursor.description]
                    user_data = dict(zip(colnames, row))
                    ### MODIFY THE FORMAT OF USER DATA
                    user_data_clean = json.loads(user_data['onboarding'])
                    doLiving = "\n".join([f"- {item['question']} : {item['answer']}" for item in user_data_clean['doLiving']])
                    whoImportant = "\n".join([f"- {item['question']} : {item['answer']}" for item in user_data_clean['whoImportant']])
                    user_data_formatted = f"""
                        ### USER PROFILE ###
                        
                        Name: {user_data_clean['firstName']}
                        {user_data_clean['firstName']}'s Legend Persona: {user_data_clean['legendPersona']}
                        Pronouns: {user_data_clean['pronouns']}
                        Birthday: {user_data_clean['birthDate']}
                        {user_data_clean['firstName']}'s MBTI: {user_data_clean['mbti']}
                        {user_data_clean['firstName']}'s Love Language: {user_data_clean['loveLanguage']}
                        Has {user_data_clean['firstName']} tried coaching before: {user_data_clean['triedCoaching']}
                        Belief in Astrology: {user_data_clean['astrology']}
                        What matters the most in {user_data_clean['firstName']}'s life: {user_data_clean['mattersMost'][0]}
                        Other things that also matters (ordered from the most to least important):
                         1. {user_data_clean['mattersMost'][1]}
                         2. {user_data_clean['mattersMost'][2]}
                         3. {user_data_clean['mattersMost'][3]}
                         4. {user_data_clean['mattersMost'][4]} (Matters the least)
                        What does {user_data_clean['firstName']} do for a living:
                        {doLiving}
                        {user_data_clean['firstName']}'s current situation: {user_data_clean['mySituation']}
                        {user_data_clean['firstName']}'s most important person:
                        {whoImportant}
                    """
                    
                    return user_data_formatted
                else:
                    return None
    except psycopg2.Error as e:
        print(f"An error occurred: {e}")
        return None
    
# Load the initial user data
user_df = fetch_users()

with gr.Blocks() as app:
    gr.Markdown("## Demo 2")



    # State to hold user's chat content
    client = OpenAI(api_key=os.getenv('OPENAI_API_KEY'))
    user_message = gr.State("")
    user = gr.State()
    prev_date = gr.State(pd.Timestamp.now().date())

    # Dropdown for user selection
    user_dropdown = gr.Dropdown(
        choices=user_df['id'].apply(lambda x: f"{x} - {user_df.loc[user_df['id'] == x, 'name'].values[0]}").tolist(),
        label="Please Select User First",
        value=None,
    )

    assistant = gr.Dropdown(
        choices=[("Backup of Coach Steve","asst_SI8I6oLdqPAQTAiUL3tTO8E4"), ("Coach Steve","asst_C7oBmqDRi085X8V7DPdMxhRF")],
        label="Please Select an Assistant",
    )

    # Calendar component for selecting date/time
    date_picker = Calendar(type="string", label="Select Date and Time")

    # "Next Day" button
    change_day_button = gr.Button("Change Day")

    # Tabs for Chat, Mementos/Goals, and Message History
    with gr.Tabs():
        with gr.Tab("Chat"):
            # Chatbot component
            chatbot = gr.Chatbot(type='messages')
            chatbot_input = gr.Textbox(label="Your Message")
            send_button = gr.Button("Send")
            reset_button = gr.Button("Reset")
        with gr.Tab("Mementos/Goals"):
            # Display mementos/goals as a JSON list
            mementos_display = gr.JSON(label="Mementos / Goals")
        with gr.Tab("Raw Message History"):  # New tab for message history
            # Display history as a copyable text area
            message_history_display = gr.Textbox(label="Raw Message History", interactive=False, lines=20, show_copy_button=True, show_label=True, autofocus=True)

    # Action on button click
    def build_user(selected_user, assistant):
        print("Building user")
        user_id = selected_user.split(" - ")[0]  # Extract user ID
        user_info = get_user_info(user_id)
        return User(user_id, user_info, client, assistant), [{'role': 'assistant', 'content': f"Chatbot initialized for user {selected_user}."}]

    def change_assistant(selected_user, assistant, user):
        print("Building user")
        user_id = selected_user.split(" - ")[0]  # Extract user ID
        if user_id == user.user_id:
            user = user.change_assistant(assistant), [{'role': 'assistant', 'content': f"Assistant changed to:{assistant}."}]
            return user
        user_info = get_user_info(user_id)
        return User(user_id, user_info, client, assistant), [{'role': 'assistant', 'content': f"Assistant changed to:{assistant}."}]

    def reset(history, user):
        history = [{'role': 'assistant', 'content': "Chatbot reset."}]
        user.reset()
        # delete all mementos
        user_mementos_path = f"mementos/{user.user_id}"
        if os.path.exists(user_mementos_path):
            for memento_file in os.listdir(user_mementos_path):
                os.remove(f"{user_mementos_path}/{memento_file}")
        return history

    def update_message_history(user):
        # Compile message history into a plain text format for easy copy-pasting
        return "\n".join([f"{msg['role']}: {msg['content']}" for msg in user.get_messages(False)])

    def fetch_mementos(selected_user):
        # Load mementos for the selected user from a hypothetical JSON store
        user_id = selected_user.split(" - ")[0]
        mementos = []
        user_mementos_path = f"mementos/{user_id}"
        
        # Check if the directory exists
        if os.path.exists(user_mementos_path):
            for memento_file in os.listdir(user_mementos_path):
                with open(f"{user_mementos_path}/{memento_file}", 'r') as f:
                    mementos.append(json.load(f))
        else:
            mementos = [{"message": "No mementos found for this user."}]
            
        print("Fetched mementos:", mementos)  # Debug print to verify loading
        return mementos

    user_dropdown.change(
        fn=build_user,
        inputs=[user_dropdown, assistant],
        outputs=[user, chatbot]
    ).then(
        fn=fetch_mementos,
        inputs=user_dropdown,
        outputs=mementos_display  # Directly updating the mementos display component
    )
    # .then(
    #     fn=update_message_history,  # Update message history after user selection
    #     inputs=user,
    #     outputs=message_history_display
    # )

    assistant.change(
        fn=change_assistant,
        inputs=[user_dropdown, assistant, user],
        outputs=[user, chatbot]
    ).then(
        fn=fetch_mementos,
        inputs=user_dropdown,
        outputs=mementos_display  # Directly updating the mementos display component
    ).then(
        fn=update_message_history,  # Update message history after user selection
        inputs=user,
        outputs=message_history_display
    )

    def user_input(user_message, history):
        user_content = user_message
        return "", history + [{"role": "user", "content": user_message}], user_content

    def chatbot_response(user, message):
        print(f"User message: {message}")
        user.send_message(message)
        return user.get_messages()

    chatbot_input.submit(
        fn=user_input,
        inputs=[chatbot_input, chatbot],
        outputs=[chatbot_input, chatbot, user_message],
        queue=False
    ).then(
        fn=chatbot_response,
        inputs=[user, user_message],
        outputs=chatbot
    ).then(
        fn=fetch_mementos,
        inputs=user_dropdown,
        outputs=mementos_display  # Directly updating the mementos display component
    ).then(
        fn=update_message_history,  # Update message history after a chat message
        inputs=user,
        outputs=message_history_display
    )

    send_button.click(
        fn=user_input,
        inputs=[chatbot_input, chatbot],
        outputs=[chatbot_input, chatbot, user_message],
        queue=False
    ).then(
        fn=chatbot_response,
        inputs=[user, user_message],
        outputs=chatbot
    ).then(
        fn=fetch_mementos,
        inputs=user_dropdown,
        outputs=mementos_display  # Directly updating the mementos display component
    ).then(
        fn=update_message_history,  # Update message history after a chat message
        inputs=user,
        outputs=message_history_display
    )

    reset_button.click(
        fn=reset,
        inputs=[chatbot, user],
        outputs=chatbot
    ).then(
        fn=update_message_history,  # Clear message history on reset
        inputs=user,
        outputs=message_history_display
    )

    def date_changed(selected_date, user):
        user.change_date(selected_date)
        history = user.get_messages()[:-1] + [{'role': 'assistant', 'content': f"Date changed to {selected_date}."}] + [user.get_messages()[-1]]
        return history
    

    def next_day(selected_date, prev_date, user):
        # Calculate the next date
        next_date = pd.to_datetime(selected_date).date()

        print("Selected date:", next_date)
        print("Previous date:", prev_date)
        if next_date == prev_date:
            next_date = pd.to_datetime(selected_date) + timedelta(days=1)
            next_date = next_date.date()

        # infer the follow up date for the user's mementos
        user.infer_memento_follow_ups()
        
        # Upload files from mementos/to_upload/{user_id} to vector store
        user_id = user.user_id
        to_upload_folder = f"mementos/to_upload/{user_id}"
        destination_folder = f"mementos/{user_id}"

        if os.path.exists(to_upload_folder):
            # Ensure the destination folder exists
            if not os.path.exists(destination_folder):
                os.makedirs(destination_folder)

            for filename in os.listdir(to_upload_folder):
                if filename.endswith('.json'):
                    file_path = os.path.join(to_upload_folder, filename)
                    try:
                        # Upload the file to the vector store
                        with open(file_path, "rb") as f:
                            memory_file = client.files.create(file=f, purpose="assistants")
                            vector_store_file = client.beta.vector_stores.files.create(
                                vector_store_id=user.conversations.user_personal_memory.id,
                                file_id=memory_file.id
                            )
                        # Move the file to mementos/{user_id}
                        shutil.move(file_path, os.path.join(destination_folder, filename))
                    except Exception as e:
                        print(f"Failed to upload and move file {filename}: {e}")
                        traceback.print_exc()

        # Call date_changed function
        return date_changed(next_date, user), next_date, next_date.strftime("%Y-%m-%d")

    change_day_button.click(
        fn=next_day,
        inputs=[date_picker, prev_date, user],
        outputs=[chatbot, prev_date, date_picker]
    ).then(
        fn=fetch_mementos,
        inputs=user_dropdown,
        outputs=mementos_display  # Directly updating the mementos display
    ).then(
        fn=update_message_history,  # Update message history on next day change
        inputs=user,
        outputs=message_history_display
    )

# Launch the app
app.launch()