app.py

import gradio as gr
from random import randint
from all_models import models  # Import the list of available models
from externalmod import gr_Interface_load
import asyncio
import os
from threading import RLock
from flask import Flask, request, jsonify, send_file
from flask_cors import CORS
import tempfile

lock = RLock()
HF_TOKEN = os.environ.get("HF_TOKEN")

app = Flask(__name__)
CORS(app)  # Enable CORS for all routes

# Function to load models
def load_fn(models):
    global models_load
    models_load = {}
    
    for model in models:
        if model not in models_load.keys():
            try:
                m = gr_Interface_load(f'models/{model}', hf_token=HF_TOKEN)
            except Exception as error:
                print(error)
                m = gr.Interface(lambda: None, ['text'], ['image'])
            models_load.update({model: m})

load_fn(models)

num_models = 6  # Number of models to load initially
MAX_SEED = 3999999999
default_models = models[:num_models]  # Load the first few models for inference
inference_timeout = 600

# Asynchronous function to perform inference
async def infer(model_str, prompt, seed=1, timeout=inference_timeout):
    kwargs = {"seed": seed}
    task = asyncio.create_task(asyncio.to_thread(models_load[model_str].fn, prompt=prompt, **kwargs, token=HF_TOKEN))
    await asyncio.sleep(0)
    try:
        result = await asyncio.wait_for(task, timeout=timeout)
    except (Exception, asyncio.TimeoutError) as e:
        print(e)
        print(f"Task timed out: {model_str}")
        if not task.done(): 
            task.cancel()
        result = None
    if task.done() and result is not None:
        with lock:
            temp_image = tempfile.NamedTemporaryFile(suffix=".png", delete=False)
            result.save(temp_image.name)  # Save result as a temporary file
        return temp_image.name  # Return the path to the saved image
    return None

# Flask route for the API endpoint
@app.route('/generate_api', methods=['POST'])
def generate_api():
    data = request.get_json()

    # Extract required fields from the request
    model_str = data.get('model_str', default_models[0])  # Default to first model if not provided
    prompt = data.get('prompt', '')
    seed = data.get('seed', 1)

    if not prompt:
        return jsonify({"error": "Prompt is required"}), 400
    
    try:
        # Call the async inference function
        result_path = asyncio.run(infer(model_str, prompt, seed))
        if result_path:
            return send_file(result_path, mimetype='image/png')  # Send back the generated image file
        else:
            return jsonify({"error": "Failed to generate image"}), 500
    except Exception as e:
        return jsonify({"error": str(e)}), 500

if __name__ == '__main__':
    app.run(debug=True)