# app.py from fastapi import FastAPI, Request, status from fastapi.responses import JSONResponse from fastapi.middleware.cors import CORSMiddleware # If you need CORS from contextlib import asynccontextmanager import uvicorn import logging from slowapi import Limiter, _rate_limit_exceeded_handler from slowapi.util import get_remote_address from slowapi.errors import RateLimitExceeded from slowapi.middleware import SlowAPIMiddleware import sys # Import sys module import config from services import generation from routers import ideas, images, videos # --- Logging Setup --- logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s') logger = logging.getLogger(__name__) # --- Rate Limiting Setup --- limiter = Limiter(key_func=get_remote_address, default_limits=[config.RATE_LIMIT]) # --- Lifespan Management (Model Loading/Unloading) --- @asynccontextmanager async def lifespan(app: FastAPI): # Startup: Assign essential state FIRST app.state.limiter = limiter logger.info("Rate limiter assigned to app state.") models_loaded_successfully = False try: logger.info("Application startup: Loading models...") generation.load_models() # This might raise an exception models_loaded_successfully = True # Set flag ONLY if load_models completes logger.info("Models loaded successfully.") except Exception as e: logger.error(f"FATAL: Model loading failed during startup: {e}", exc_info=True) # Option 1: Exit the application (cleaner for Docker environments) # logger.critical("Exiting application due to model loading failure.") # sys.exit(1) # Option 2: Raise the exception again to make uvicorn aware of failure # This might depend on how uvicorn handles lifespan exceptions # raise # Re-raise the exception # >>> Only yield if models loaded <<< if models_loaded_successfully: yield # Application is now ready to serve requests else: # If models didn't load, we don't yield, preventing Uvicorn # from reporting "Application startup complete." # You might need to manually stop the process if Option 1 above isn't used. logger.error("Application startup failed due to model loading errors. Server will not serve requests effectively.") # Keep the process running but indicate failure. Or use sys.exit(1) above. # If you don't yield or exit, Uvicorn might hang or exit depending on version. # Testing needed here - sys.exit(1) is often simplest in Docker. # For now, just logging the failure and not yielding. # --- Shutdown Logic --- # This part might not be reached if sys.exit was called logger.info("Application shutdown sequence starting.") if "generation" in globals() and hasattr(generation, 'model_cache'): generation.model_cache.clear() # Add any other cleanup here logger.info("Resources cleaned up.") # --- FastAPI App Initialization --- app = FastAPI( title="AI Content Generation API", description="API for generating content ideas, images, and videos using Hugging Face models.", version="1.0.0", lifespan=lifespan # Use the lifespan context manager ) # --- Middleware --- # Rate Limiting Middleware - Now it can safely access app.state.limiter app.add_middleware(SlowAPIMiddleware) # CORS Middleware (Uncomment and configure if needed for browser-based clients) # origins = [ # "http://localhost", # "http://localhost:8080", # "https://your-frontend-domain.com", # Add your frontend domain # ] # app.add_middleware( # CORSMiddleware, # allow_origins=origins, # allow_credentials=True, # allow_methods=["*"], # allow_headers=["*"], # ) # --- Exception Handlers --- app.add_exception_handler(RateLimitExceeded, _rate_limit_exceeded_handler) @app.exception_handler(Exception) async def general_exception_handler(request: Request, exc: Exception): logger.error(f"Unhandled exception: {exc}", exc_info=True) return JSONResponse( status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, content={"detail": "An internal server error occurred."}, ) # --- API Routers --- app.include_router(ideas.router) app.include_router(images.router) app.include_router(videos.router) # --- Root Endpoint --- @app.get("/", tags=["Status"]) async def read_root(): """Root endpoint providing basic API information.""" return { "message": "Welcome to the AI Content Generation API!", "docs": "/docs", "models": { "text": config.TEXT_MODEL_NAME, "image": config.IMAGE_MODEL_NAME, "video": config.VIDEO_MODEL_NAME }, "status": "OK" } # --- Main Execution (for local testing) --- if __name__ == "__main__": # When running locally: uvicorn app:app --reload # The following is mostly for structuring; direct execution isn't typical for deployment print("To run locally, use: uvicorn app:app --reload --host 0.0.0.0 --port 7860") # uvicorn.run(app, host="0.0.0.0", port=7860) # Port 7860 is common for HF Spaces