app.py

import os
from flask import Flask, render_template, request, jsonify
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

app = Flask(__name__)

# Define cache directory
cache_dir = "/app/cache"
os.environ["HF_HOME"] = cache_dir

# Load Myanmarsar-GPT (1.42B params) from Hugging Face
MODEL_NAME = "simbolo-ai/Myanmarsar-GPT"
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, cache_dir=cache_dir)
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, cache_dir=cache_dir)

# Function to generate chatbot responses
def generate_response(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    with torch.no_grad():
        output = model.generate(**inputs, max_length=200)
    return tokenizer.decode(output[0], skip_special_tokens=True)

# Serve the HTML page
@app.route("/")
def home():
    return render_template("index.html")

# API route for chatbot responses
@app.route("/chat", methods=["POST"])
def chat():
    try:
        if not request.is_json:
            print("Error: Request is not JSON")
            return jsonify({"error": "Request must be JSON"}), 415

        data = request.get_json()
        user_message = data.get("message", "")

        if not user_message:
            print("Error: No message received")
            return jsonify({"error": "No message provided"}), 400

        print(f"Received message: {user_message}")

        bot_reply = generate_response(user_message)
        print(f"AI response: {bot_reply}")

        return jsonify({"reply": bot_reply})

    except Exception as e:
        print(f"Error processing request: {e}")
        return jsonify({"error": str(e)}), 500

if __name__ == "__main__":
    port = int(os.environ.get("PORT", 7860))  # Default to 7860, but use any assigned port
    app.run(host="0.0.0.0", port=port)