app_whatsapp.py

import uvicorn
import asyncio
import sqlite3
from fastapi import FastAPI, Form
from dotenv import load_dotenv
from src.channels.whatsapp import WhatsAppChannel
from src.agents.personal_assistant import PersonalAssistant
from src.utils import get_current_date_time

# Load .env variables from the environment file
load_dotenv()

# Initiate FastAPI app
app = FastAPI()

# Initialize sqlite3 DB for saving agent memory
conn = sqlite3.connect("db/checkpoints.sqlite", check_same_thread=False)

# Initiate personal assistant instance
personal_assistant = PersonalAssistant()

# Configuration for the Langgraph agent, specifying thread ID
config = {"configurable": {"thread_id": "1"}}

async def process_message_async(to_whatsapp_number, incoming_message):
    """
    Processes the incoming message asynchronously:
    1. Formats the message with the current date and time.
    2. Invokes the personal assistant to get a response.
    3. Sends the response to the provided WhatsApp number.
    """
    # Format the message with current date/time
    message = (
        f"Message: {incoming_message}\n"
        f"Current Date/time: {get_current_date_time()}"
    )
    
    # Invoke the personal assistant to generate a response
    answer = personal_assistant.invoke(message, config=config)

    # Send the response via Twilio WhatsApp
    whatsapp = WhatsAppChannel()
    await asyncio.to_thread(
        whatsapp.send_message,
        to_number=to_whatsapp_number,
        body=answer
    )

@app.post("/whatsapp/webhook")
async def whatsapp_webhook(Body: str = Form(...), From: str = Form(...)):
    """
    Webhook endpoint that handles incoming messages from WhatsApp.
    Receives the message and triggers an asynchronous task to process it.
    """
    incoming_message = Body
    from_number = From
    print(f"Message received from {from_number}: {incoming_message}")

    # Create an asynchronous task to process the message without blocking
    asyncio.create_task(process_message_async(from_number, incoming_message))

    # Respond with a status indicating that the message was received
    return "Message received", 200

if __name__ == "__main__":
    # Start the FastAPI application on the specified host and port
    uvicorn.run(app, host="0.0.0.0", port=5000)