hls_download.py

import subprocess
import os

def download_clips(stream_url, out_dir, start_time, end_time, resize=True):
    output_file = os.path.join(out_dir, f"train_{len(os.listdir(out_dir))}.mp4")
    tmp_file = os.path.join(out_dir, f"train_{len(os.listdir(out_dir))}_tmp.mp4")
    yt_dlp_cmd = [
        'yt-dlp',
        '--download-sections', f'*{start_time}-{end_time}',
        stream_url,
        '-o', tmp_file
    ]

    print(' '.join(yt_dlp_cmd))
    
    try:
        subprocess.run(yt_dlp_cmd, check=True, capture_output=True, text=True)
    except subprocess.CalledProcessError as e:
        print(f"Error occurred: {e}")
        print(f"yt-dlp output: {e.output}")
        return None
    print(f"Downloaded {output_file}")
    if resize:  # resize and convert to 30 fps
        ffmpeg_cmd = [
                'ffmpeg',
                '-i', tmp_file,
                '-c:v', 'libx264',
                '-preset', 'veryfast',
                '-crf', '23',
                '-r', '30',
                '-maxrate', '2M',
                '-bufsize', '4M',
                '-vf', f"scale=-2:300",
                '-c:a', 'aac',
                '-b:a', '128k',
                '-y',
                output_file
            ]
        print(' '.join(ffmpeg_cmd))
        try:
            subprocess.run(ffmpeg_cmd, check=True, capture_output=True, text=True)
        except subprocess.CalledProcessError as e:
            print(f"Error occurred: {e}")
            print(f"ffmpeg output: {e.output}")
            return None
    else:
        os.rename(tmp_file, output_file)
    print(f"Converted {output_file}")

    return output_file