app.py

import os
import threading
import asyncio
import discord
from dotenv import load_dotenv
from llama_cpp import Llama  # Library for GGUF models

# Load environment variables (set these via Hugging Face Secrets)
load_dotenv()
DISCORD_TOKEN = os.getenv("DISCORD_TOKEN")

if not DISCORD_TOKEN:
    raise ValueError("Discord bot token is missing. Set DISCORD_TOKEN in the environment variables.")

# Set the local path to your quantized model file.
# Ensure that this file (e.g. DeepScaleR-1.5B-Preview-Q8_0.gguf) is uploaded to your repository.
MODEL_PATH = "./DeepScaleR-1.5B-Preview-Q8_0.gguf"

# Initialize the model with appropriate settings.
# Adjust n_threads and other parameters as needed.
llm = Llama(model_path=MODEL_PATH, n_threads=4)

def generate_response(prompt):
    # Generate text using llama-cpp's Llama instance.
    # Adjust parameters (max_tokens, temperature, top_p) for speed/quality tradeoffs.
    output = llm(prompt=prompt, max_tokens=200, temperature=0.7, top_p=0.9, echo=False)
    response = output["text"]
    # Optionally enforce your bot identity:
    response = response.replace("DeepScaleR", "Shiv Yantra AI")
    return response

# ----------------------------
# Discord Bot Setup
# ----------------------------
intents = discord.Intents.default()
intents.message_content = True  # Enable reading message content
client = discord.Client(intents=intents)

@client.event
async def on_ready():
    print(f"Logged in as {client.user}")

@client.event
async def on_message(message):
    # Ignore messages from the bot itself.
    if message.author == client.user:
        return

    user_input = message.content.strip()
    if user_input:
        try:
            # Run the generate_response function in a separate thread so as not to block Discord's event loop.
            ai_response = await asyncio.to_thread(generate_response, user_input)
        except Exception as e:
            print(f"Error during generation: {e}")
            ai_response = "Error processing your request."
        await message.channel.send(ai_response)

def run_discord_bot():
    client.run(DISCORD_TOKEN)

# ----------------------------
# Start the Discord Bot
# ----------------------------
if __name__ == "__main__":
    # Start the Discord bot in a separate daemon thread.
    threading.Thread(target=run_discord_bot, daemon=True).start()
    
    # Keep the main thread alive.
    while True:
        pass