from fastapi import FastAPI, HTTPException from fastapi.staticfiles import StaticFiles from pydantic import BaseModel, HttpUrl from typing import List import os import asyncio import uuid import aiohttp import re from urllib.parse import urlparse import shutil import aiofiles # Create FastAPI app app = FastAPI() # Create and mount staticfiles directory os.makedirs("staticfiles", exist_ok=True) app.mount("/static", StaticFiles(directory="staticfiles"), name="static") # Define input model for the request class SlideshowRequest(BaseModel): image_urls: List[HttpUrl] audio_url: HttpUrl duration: int zoom: bool = False def extract_google_drive_id(url): """Extract file ID from a Google Drive URL""" pattern = r'(?:/file/d/|id=|/open\?id=)([^/&]+)' match = re.search(pattern, str(url)) return match.group(1) if match else None async def download_file(url, local_path): """Download a file from URL to local path asynchronously""" try: # Handle Google Drive URLs if "drive.google.com" in str(url): file_id = extract_google_drive_id(url) if file_id: url = f"https://drive.google.com/uc?export=download&id={file_id}" async with aiohttp.ClientSession() as session: async with session.get(str(url)) as response: response.raise_for_status() async with aiofiles.open(local_path, 'wb') as f: while True: chunk = await response.content.read(8192) if not chunk: break await f.write(chunk) return True except Exception as e: print(f"Error downloading {url}: {str(e)}") return False async def create_slideshow(image_paths, audio_path, output_path, duration, zoom=False, zoom_ratio=0.04): """Generate slideshow from images and audio using ffmpeg asynchronously""" # Create temporary file list for ffmpeg concat (used in non-zoom mode) concat_file = "temp_concat.txt" if not zoom: # Original implementation without zoom effect async with aiofiles.open(concat_file, "w") as f: for img in image_paths: await f.write(f"file '{img}'\n") await f.write(f"duration {duration}\n") # Add the last image again without duration (required by ffmpeg) if image_paths: await f.write(f"file '{image_paths[-1]}'\n") # Run ffmpeg command to create slideshow with audio total_duration = len(image_paths) * duration cmd = [ "ffmpeg", "-f", "concat", "-safe", "0", "-i", concat_file, "-i", audio_path, "-c:v", "libx264", "-pix_fmt", "yuv420p", "-c:a", "aac", "-shortest", "-y", "-t", str(total_duration), output_path ] else: # Complex implementation with zoom effect # For each image, we remove the per-input duration and instead do a trimmed zoompan. filters = [] for i, img in enumerate(image_paths): filter_str = ( f"[{i}:v]scale=1920:1080:force_original_aspect_ratio=decrease," f"pad=1920:1080:(ow-iw)/2:(oh-ih)/2,setsar=1," f"zoompan=z='min(zoom+0.0015,1.5)':d={duration*25}:" f"x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)'," f"trim=duration={duration},setpts=PTS-STARTPTS[v{i}];" ) filters.append(filter_str) # Concatenate all zoompan outputs filter_complex = "".join(filters) for i in range(len(image_paths)): filter_complex += f"[v{i}]" filter_complex += f"concat=n={len(image_paths)}:v=1:a=0[outv]" cmd = ["ffmpeg", "-y"] # Remove duration here; let zoompan+trim control each clip's length. for img in image_paths: cmd.extend(["-loop", "1", "-i", img]) cmd.extend([ "-i", audio_path, "-filter_complex", filter_complex, "-map", "[outv]", "-map", f"{len(image_paths)}:a", "-c:v", "libx264", "-pix_fmt", "yuv420p", "-c:a", "aac", "-shortest", output_path ]) try: process = await asyncio.create_subprocess_exec( *cmd, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE ) stdout, stderr = await process.communicate() if os.path.exists(concat_file): os.remove(concat_file) if process.returncode != 0: print(f"FFmpeg error: {stderr.decode()}") return False return True except Exception as e: print(f"FFmpeg error: {str(e)}") if os.path.exists(concat_file): os.remove(concat_file) return False @app.post("/make_slideshow") async def make_slideshow(request: SlideshowRequest): """ Create a slideshow from images and audio with specified duration per image. Returns the URL of the generated video. """ # Create unique directory for this request request_id = str(uuid.uuid4()) request_dir = os.path.join("staticfiles", request_id) os.makedirs(request_dir, exist_ok=True) try: # Download images image_paths = [] download_tasks = [] for i, url in enumerate(request.image_urls): image_path = os.path.join(request_dir, f"image_{i:03d}.png") image_paths.append(image_path) download_tasks.append(download_file(url, image_path)) # Download audio audio_path = os.path.join(request_dir, "audio.mp3") download_tasks.append(download_file(request.audio_url, audio_path)) # Wait for all downloads to complete results = await asyncio.gather(*download_tasks) # Check if all downloads were successful if not all(results[:-1]): # All image downloads raise HTTPException(status_code=400, detail="Failed to download one or more images") if not results[-1]: # Audio download raise HTTPException(status_code=400, detail=f"Failed to download audio: {request.audio_url}") # Output video path output_path = os.path.join(request_dir, "slideshow.mp4") # Generate slideshow if not await create_slideshow(image_paths, audio_path, output_path, request.duration, request.zoom): raise HTTPException(status_code=500, detail="Failed to create slideshow") # Return URL to the video base_url = "https://saq1b-api.hf.space/static" video_url = f"{base_url}/{request_id}/slideshow.mp4" return {"url": video_url} except Exception as e: # Clean up on error if os.path.exists(request_dir): shutil.rmtree(request_dir) raise HTTPException(status_code=500, detail=f"Error: {str(e)}") if __name__ == "__main__": import uvicorn uvicorn.run(app, host="0.0.0.0", port=7860)