""" Download videos listed in video_urls.json using yt-dlp. Supports resuming (skips already downloaded videos). Uses parallel workers for speed. """ import json import subprocess import sys import os import time import argparse from pathlib import Path from concurrent.futures import ThreadPoolExecutor, as_completed SCRIPT_DIR = Path(__file__).parent URL_FILE = SCRIPT_DIR / "video_urls.json" VIDEO_DIR = SCRIPT_DIR / "videos" LOG_FILE = SCRIPT_DIR / "download_log.jsonl" DENO_PATH = Path.home() / ".deno" / "bin" def download_one(video_id: str, url: str, output_dir: Path, cookies_file: str | None = None, use_oauth2: bool = False, sleep_interval: float = 0) -> dict: """Download a single video. Returns status dict.""" video_dir = output_dir / video_id video_dir.mkdir(parents=True, exist_ok=True) existing = list(video_dir.glob("*.mp4")) + list(video_dir.glob("*.mkv")) + list(video_dir.glob("*.webm")) if existing: return {"video_id": video_id, "status": "skipped", "msg": "already exists"} cmd = [ "yt-dlp", "--no-playlist", "--remote-components", "ejs:github", "-f", "bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best", "--merge-output-format", "mp4", "-o", str(video_dir / "%(id)s.%(ext)s"), "--retries", "3", "--socket-timeout", "30", "--no-overwrites", "--quiet", "--no-warnings", ] if cookies_file: cmd.extend(["--cookies", cookies_file]) if use_oauth2: cmd.extend(["--username", "oauth2", "--password", "", "--extractor-args", "youtube:player_skip=initial_data"]) if sleep_interval > 0: cmd.extend(["--sleep-interval", str(int(sleep_interval)), "--max-sleep-interval", str(int(sleep_interval * 2))]) cmd.append(url) env = os.environ.copy() if DENO_PATH.exists(): env["PATH"] = str(DENO_PATH) + ":" + env.get("PATH", "") try: result = subprocess.run(cmd, capture_output=True, text=True, timeout=600, env=env) if result.returncode == 0: return {"video_id": video_id, "status": "ok"} else: return {"video_id": video_id, "status": "failed", "msg": result.stderr.strip()[-200:]} except subprocess.TimeoutExpired: return {"video_id": video_id, "status": "timeout"} except Exception as e: return {"video_id": video_id, "status": "error", "msg": str(e)} def load_done_ids(log_file: Path, include_failed: bool = True) -> set: """Load IDs to skip. If include_failed=True, also skip previously failed IDs.""" done = set() if log_file.exists(): with open(log_file) as f: for line in f: try: entry = json.loads(line) status = entry.get("status", "") if status in ("ok", "skipped"): done.add(entry["video_id"]) elif include_failed and status in ("failed", "timeout", "error"): done.add(entry["video_id"]) except json.JSONDecodeError: pass return done def main(): parser = argparse.ArgumentParser(description="Download NitroGen videos") parser.add_argument("--workers", type=int, default=4, help="parallel download workers (default: 4)") parser.add_argument("--limit", type=int, default=0, help="max videos to download (0 = all)") parser.add_argument("--source", type=str, default="", help="filter by source (e.g. 'youtube', 'twitch')") parser.add_argument("--cookies", type=str, default="", help="path to cookies.txt (Netscape format) for YouTube") parser.add_argument("--reset-failed", action="store_true", help="retry previously failed downloads") parser.add_argument("--sleep", type=float, default=0, help="sleep seconds between requests per worker (anti-throttle)") parser.add_argument("--oauth2", action="store_true", help="use YouTube OAuth2 authentication instead of cookies") parser.add_argument("--retry-ids", type=str, default="", help="JSON file with list of video IDs to retry") args = parser.parse_args() if not URL_FILE.exists(): print(f"ERROR: {URL_FILE} not found. Run extract_video_urls.py first.") sys.exit(1) with open(URL_FILE) as f: all_videos = json.load(f) if args.source: all_videos = {k: v for k, v in all_videos.items() if v.get("source") == args.source} cookies_file = args.cookies if args.cookies else None if cookies_file and not Path(cookies_file).exists(): print(f"ERROR: cookies file not found: {cookies_file}") sys.exit(1) VIDEO_DIR.mkdir(parents=True, exist_ok=True) if args.retry_ids: with open(args.retry_ids) as f: retry_set = set(json.load(f)) todo = {k: v for k, v in all_videos.items() if k in retry_set} else: done_ids = load_done_ids(LOG_FILE, include_failed=not args.reset_failed) todo = {k: v for k, v in all_videos.items() if k not in done_ids} if args.limit > 0: todo = dict(list(todo.items())[:args.limit]) print(f"Total videos: {len(all_videos)}") print(f"Already done: {len(all_videos) - len(todo)}") print(f"To download: {len(todo)}") print(f"Workers: {args.workers}") print(f"Cookies: {cookies_file or 'None'}") print(f"Output dir: {VIDEO_DIR}") print() if not todo: print("Nothing to download.") return ok_count = 0 fail_count = 0 rate_limit_count = 0 with open(LOG_FILE, "a") as log_f: if args.workers == 1 and args.sleep > 0: for i, (vid, info) in enumerate(todo.items(), 1): result = download_one(vid, info["url"], VIDEO_DIR, cookies_file, args.oauth2, args.sleep) log_f.write(json.dumps(result) + "\n") log_f.flush() status_icon = "OK" if result["status"] in ("ok", "skipped") else "FAIL" if result["status"] in ("ok", "skipped"): ok_count += 1 else: fail_count += 1 if "rate limit" in result.get("msg", "").lower() or "add a delay" in result.get("msg", "").lower(): rate_limit_count += 1 if rate_limit_count >= 3: wait = min(rate_limit_count * 30, 300) print(f" Rate limited {rate_limit_count} times, backing off {wait}s...") time.sleep(wait) print(f"[{i:5d}/{len(todo)}] {status_icon} {result['video_id']} - {result['status']}" + (f" ({result.get('msg', '')[:80]})" if result.get("msg") else "")) if result["status"] in ("ok",): rate_limit_count = 0 time.sleep(args.sleep) elif result["status"] == "skipped": pass else: time.sleep(args.sleep) else: with ThreadPoolExecutor(max_workers=args.workers) as executor: futures = {} for vid, info in todo.items(): fut = executor.submit(download_one, vid, info["url"], VIDEO_DIR, cookies_file, args.oauth2, args.sleep) futures[fut] = vid for i, fut in enumerate(as_completed(futures), 1): result = fut.result() log_f.write(json.dumps(result) + "\n") log_f.flush() status_icon = "OK" if result["status"] in ("ok", "skipped") else "FAIL" if result["status"] in ("ok", "skipped"): ok_count += 1 else: fail_count += 1 print(f"[{i:5d}/{len(todo)}] {status_icon} {result['video_id']} - {result['status']}" + (f" ({result.get('msg', '')[:80]})" if result.get("msg") else "")) print(f"\nDone! OK: {ok_count}, Failed: {fail_count}") if __name__ == "__main__": main()