YouTube Stream Downloader

youtube-dl.py

"""YouTube Stream Downloader
2021-08-16 v1.0 by Cees Timmerman
2021-09-18 v1.1 Better error handling. Revert to XML captions as conversion to SRT is broken in pytube 11.0.1 for https://www.youtube.com/watch?v=AOZw1tgD8dA
"""
import logging, os, re

import pytube
from pytube.cli import on_progress


# Least gnarly solution for printing the source of any error or exception; traceback module only handled the latter.
logging.basicConfig(format='%(levelname)s:%(message)s', level=os.environ.get("YT_DL_LOGLEVEL", "INFO"))
log = logging.getLogger()


def fix_filename(name, max_length=255):
    """
    Replace invalid characters on Linux/Windows/MacOS with underscores.
    List from https://stackoverflow.com/a/31976060/819417
    Trailing spaces & periods are ignored on Windows.
    >>> fix_filename("  COM1  ")
    '_ COM1 _'
    >>> fix_filename("COM10")
    'COM10'
    >>> fix_filename("COM1,")
    'COM1,'
    >>> fix_filename("COM1.txt")
    '_.txt'
    >>> all('_' == fix_filename(chr(i)) for i in list(range(32)))
    True
    """
    return re.sub(r'[/\\:|<>"?*\0-\x1f]|^(AUX|COM[1-9]|CON|LPT[1-9]|NUL|PRN)(?![^.])|^\s|[\s.]$', "_", name[:max_length], flags=re.IGNORECASE)
    # Includes . return re.sub(r'[/\\:|<>"?*\0-\31]|^(AUX|COM[1-9]|CON|LPT[1-9]|NUL|PRN)(\.|$)|^\s|[\s.]$', "_", name[:max_length], flags=re.IGNORECASE)


def download_youtube_stream(url, folder='output-folder'):
    youtube = pytube.YouTube(url, on_progress_callback=on_progress)
    print(youtube.title)
    """
    ['__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_age_restricted', '_author', '_embed_html', '_fmt_streams', '_initial_data', '_js', '_js_url', '_metadata', '_player_config_args', '_publish_date', '_title', '_vid_info', '_watch_html', 'age_restricted', 'allow_oauth_cache', 'author', 'bypass_age_gate', 'caption_tracks', 'captions', 'channel_id', 'channel_url', 'check_availability', 'description', 'embed_html', 'embed_url', 'fmt_streams', 'initial_data', 'js', 'js_url', 'keywords', 'length', 'metadata', 'publish_date', 'rating', 'register_on_complete_callback', 'register_on_progress_callback', 'stream_monostate', 'streaming_data', 'streams', 'thumbnail_url', 'title', 'use_oauth', 'vid_info', 'video_id', 'views', 'watch_html', 'watch_url']
    """
    streams = youtube.streams
    """
    ['__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_filesize', '_monostate', 'abr', 'audio_codec', 'bitrate', 'codecs', 'default_filename', 'download', 'exists_at_path', 'expiration', 'filesize', 'filesize_approx', 'get_file_path', 'includes_audio_track', 'includes_video_track', 'is_3d', 'is_adaptive', 'is_dash', 'is_hdr', 'is_live', 'is_otf', 'is_progressive', 'itag', 'mime_type', 'on_complete', 'on_progress', 'parse_codecs', 'resolution', 'stream_to_buffer', 'subtype', 'title', 'type', 'url', 'video_codec']
    """
    for i, s in enumerate(streams):
        if s.includes_audio_track and s.includes_video_track:
            print(f"{i + 1:>2}: {s.resolution:>5} {s.bitrate / 1000:>4,.0f}kbps {s.abr:>7} {s.mime_type:<10} {str(s.codecs):<28} {s.filesize / 1000_000:>7,.2f} MB {s.filesize / 1024 / 1024:>7,.2f} MiB")

    choice = int(input("Which stream? ")) - 1
    stream = streams[choice]
    print("Downloading...")
    stream.download(folder, skip_existing=True)
    if 'en' in youtube.captions:
        caption = youtube.captions['en']
        print(caption)
        # xml->srt broken in pytube 11.0.1 on https://www.youtube.com/watch?v=AOZw1tgD8dA
        caption.download(title=fix_filename(youtube.title), output_path=folder, srt=False)
    print(f"\nDone. {folder}\\{s.default_filename}")


if __name__ == "__main__":
    try:
        # Python 2 is dangerous (use raw_input) but unsupported.
        download_youtube_stream(input("YouTube URL: "))
    except Exception as ex:
        log.exception(ex)
    
    input()  # Keep console open until user presses enter.