main.py

from flask import Flask,request,make_response
import os
import logging
from dotenv import load_dotenv
from heyoo import WhatsApp
import assemblyai as aai
import openai
from utility import parse_multiple_transactions



# load env data
load_dotenv()





# messenger object
messenger = WhatsApp(
  os.environ["whatsapp_token"],
  phone_number_id=os.environ["phone_number_id"] )

# aai.settings.api_key = os.environ["aai_key"]
# transcriber = aai.Transcriber()





app = Flask(__name__)

VERIFY_TOKEN = "30cca545-3838-48b2-80a7-9e43b1ae8ce4"


client = openai.OpenAI(
    api_key=os.environ.get("sambanova_api_key"),
    base_url="https://api.sambanova.ai/v1",
)

def generateResponse(prompt):
  #----- Call API to classify and extract relevant transaction information
    # These templates help provide a unified response format for use as context clues when
    # parsing the AI generated response into a structured data format
    relevant_info_template = """
          Intent: The CRUD operation
          Transaction Type: The type of transaction
          Details: as a sublist of the key details like name of item, amount, description, among other details you are able to extract. 
      """
    sample_response_template = """
      The information provided indicates that you want to **create/record** a new transaction.

      **Extracted Information**:

      **Intent**: Create

      Transaction 1: 

      **Transaction Type**: Purchase
      **Details**: 
              - Item: Car
              - Purpose: Business
              - Amount: 1000
              - Tax: 200
              - Note: A new car for business

      Transaction 2:

      **Transaction Type**: Expense  
      **Details**:  
            - Item: Office Chair  
            - Amount: 300 USD  
            - Category: Furniture  
      """
    response = client.chat.completions.create(
        model='Meta-Llama-3.1-70B-Instruct',
        messages=[
              {"role": "system", "content": f"You are a helpful assistant that classifies transactions written in natural language into CRUD operations (Create, Read, Update, and Delete) and extracts relevant information. Format the relevant information extracted from the transaction text in this format: {relevant_info_template}. You can use markdown syntax to present a nicely formated and readable response to the user, but make sure the user does not see the markdown keyword. Keywords and field names must be in bold face. A sample response could look like this: {sample_response_template}. Delineate multiple transactions with the label 'Transaction 1' before the start of the relevant information for each transaction. There should be only one intent even in the case of multiple transactions."},
              {"role": "user", "content": prompt}
        ]
    )
    #----- Process response
    try:
      response = response.choices[0].message.content
    except Exception as e:
        print(f'An error occurred: {str(e)}')
        response = None

    return response




def respond(query_str:str):
    response = "hello, I don't have a brain yet"
    return response


@app.route("/", methods=["GET", "POST"])
def hook():
  if request.method == "GET":
    if request.args.get("hub.verify_token") == VERIFY_TOKEN:
        logging.info("Verified webhook")
        response = make_response(request.args.get("hub.challenge"), 200)
        response.mimetype = "text/plain"
        return response
    logging.error("Webhook Verification failed")
    return "Invalid verification token"

  # get message update..
  data = request.get_json()
  changed_field = messenger.changed_field(data)

  if changed_field == "messages":
    new_message = messenger.get_mobile(data)
    if new_message:
      mobile = messenger.get_mobile(data)
      message_type = messenger.get_message_type(data)

      if message_type == "text":
        message = messenger.get_message(data)
        # Handle greetings
        if message.lower() in ("hi", "hello", "help", "how are you"):
          response = "Hi there! My name is BuzyHelper. How can I help you today?"
          messenger.send_message(message=f"{response}",recipient_id=mobile)
        else:
          response = str(generateResponse(message))
          # Parse AI generated response into a structured format
          parsed_trans_data = parse_multiple_transactions(response)
          print("Response:", response)
          logging.info(f"\nAnswer: {response}\n")
        # Handle cases where response is not a valid image path
          messenger.send_message(message=f"{response}\n\n {parsed_trans_data}", recipient_id=mobile)
            
      # elif message_type == "audio":
      #   audio = messenger.get_audio(data)
      #   audio_id, mime_type = audio["id"], audio["mime_type"]
      #   audio_url = messenger.query_media_url(audio_id)
      #   audio_filename = messenger.download_media(audio_url, mime_type)
      #   transcript =transcriber.transcribe(audio_filename)
      #   print(audio_filename)
      #   print(transcript.text)
      #   res = transcript.text
      #   logging.info(f"\nAudio: {audio}\n")
      #   response = str(generateResponse(res))
      #   if isinstance(response, str):
      #     messenger.send_message(message=f"{response}", recipient_id=mobile)
      #   elif isinstance(response, str) and os.path.isfile(response):
      #     messenger.send_image(image_path=response, recipient_id=mobile)

      else:
        messenger.send_message(message="Please send me text or audio messages",recipient_id=mobile)

  return "ok"

if __name__ == '__main__':
  app.run(debug=True,host="0.0.0.0", port=7860)