api.py

from flask import Flask, request, jsonify
from app import generate_chat_completion
import time

app = Flask(__name__)

@app.route('/v1/chat/completions', methods=['POST'])
def chat_completions():
    data = request.json
    messages = data.get('messages', [])
    max_tokens = data.get('max_tokens', 560)
    temperature = data.get('temperature', 0.7)
    
    if not messages:
        return jsonify({"error": "messages array is required"}), 400
    
    try:
        result = generate_chat_completion(
            messages=messages,
            max_tokens=max_tokens,
            temperature=temperature
        )
        return jsonify(result)
    except Exception as e:
        return jsonify({"error": str(e)}), 500

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=8081)  # Note port 8081