compat_urllib_parse_unquote,
compat_urllib_parse_unquote_plus,
compat_urllib_parse_urlparse,
- compat_urllib_request,
compat_urlparse,
compat_str,
)
from ..utils import (
clean_html,
encode_dict,
+ error_to_compat_str,
ExtractorError,
float_or_none,
get_element_by_attribute,
get_element_by_id,
int_or_none,
+ mimetype2ext,
orderedSet,
parse_duration,
+ remove_quotes,
remove_start,
+ sanitized_Request,
smuggle_url,
str_to_int,
unescapeHTML,
login_data = compat_urllib_parse.urlencode(encode_dict(login_form_strs)).encode('ascii')
- req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
+ req = sanitized_Request(self._LOGIN_URL, login_data)
login_results = self._download_webpage(
req, None,
note='Logging in', errnote='unable to log in', fatal=False)
tfa_data = compat_urllib_parse.urlencode(encode_dict(tfa_form_strs)).encode('ascii')
- tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data)
+ tfa_req = sanitized_Request(self._TWOFACTOR_URL, tfa_data)
tfa_results = self._download_webpage(
tfa_req, None,
note='Submitting TFA code', errnote='unable to submit tfa', fatal=False)
|(?: # or the v= param in all its forms
(?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
(?:\?|\#!?) # the params delimiter ? or # or #!
- (?:.*?&)?? # any other preceding param (like /?s=tuff&v=xxxx)
+ (?:.*?[&;])?? # any other preceding param (like /?s=tuff&v=xxxx or ?s=tuff&v=V36LpHqtcDY)
v=
)
))
$"""
_NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
_formats = {
- '5': {'ext': 'flv', 'width': 400, 'height': 240},
- '6': {'ext': 'flv', 'width': 450, 'height': 270},
- '13': {'ext': '3gp'},
- '17': {'ext': '3gp', 'width': 176, 'height': 144},
- '18': {'ext': 'mp4', 'width': 640, 'height': 360},
- '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
- '34': {'ext': 'flv', 'width': 640, 'height': 360},
- '35': {'ext': 'flv', 'width': 854, 'height': 480},
- '36': {'ext': '3gp', 'width': 320, 'height': 240},
- '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
- '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
- '43': {'ext': 'webm', 'width': 640, 'height': 360},
- '44': {'ext': 'webm', 'width': 854, 'height': 480},
- '45': {'ext': 'webm', 'width': 1280, 'height': 720},
- '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
- '59': {'ext': 'mp4', 'width': 854, 'height': 480},
- '78': {'ext': 'mp4', 'width': 854, 'height': 480},
-
-
- # 3d videos
- '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
- '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
- '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
- '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
- '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
- '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
- '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
+ '5': {'ext': 'flv', 'width': 400, 'height': 240, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'},
+ '6': {'ext': 'flv', 'width': 450, 'height': 270, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'},
+ '13': {'ext': '3gp', 'acodec': 'aac', 'vcodec': 'mp4v'},
+ '17': {'ext': '3gp', 'width': 176, 'height': 144, 'acodec': 'aac', 'abr': 24, 'vcodec': 'mp4v'},
+ '18': {'ext': 'mp4', 'width': 640, 'height': 360, 'acodec': 'aac', 'abr': 96, 'vcodec': 'h264'},
+ '22': {'ext': 'mp4', 'width': 1280, 'height': 720, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
+ '34': {'ext': 'flv', 'width': 640, 'height': 360, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
+ '35': {'ext': 'flv', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
+ '36': {'ext': '3gp', 'width': 320, 'height': 240, 'acodec': 'aac', 'abr': 32, 'vcodec': 'mp4v'},
+ '37': {'ext': 'mp4', 'width': 1920, 'height': 1080, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
+ '38': {'ext': 'mp4', 'width': 4096, 'height': 3072, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'},
+ '43': {'ext': 'webm', 'width': 640, 'height': 360, 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8'},
+ '44': {'ext': 'webm', 'width': 854, 'height': 480, 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8'},
+ '45': {'ext': 'webm', 'width': 1280, 'height': 720, 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8'},
+ '46': {'ext': 'webm', 'width': 1920, 'height': 1080, 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8'},
+ '59': {'ext': 'mp4', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
+ '78': {'ext': 'mp4', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'},
+
+
+ # 3D videos
+ '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -20},
+ '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -20},
+ '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264', 'preference': -20},
+ '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264', 'preference': -20},
+ '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8', 'preference': -20},
+ '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8', 'preference': -20},
+ '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8', 'preference': -20},
# Apple HTTP Live Streaming
- '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
- '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
- '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
- '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
- '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
- '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
- '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
+ '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 48, 'vcodec': 'h264', 'preference': -10},
+ '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -10},
+ '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -10},
+ '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 256, 'vcodec': 'h264', 'preference': -10},
+ '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 256, 'vcodec': 'h264', 'preference': -10},
+ '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 48, 'vcodec': 'h264', 'preference': -10},
+ '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 24, 'vcodec': 'h264', 'preference': -10},
# DASH mp4 video
- '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559)
- '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
- '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
- '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'},
+ '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559)
+ '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
+ '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'fps': 60, 'preference': -40},
+ '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'fps': 60, 'preference': -40},
+ '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40},
# Dash mp4 audio
'139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'},
'170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40},
'218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40},
'219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40},
- '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'container': 'webm', 'vcodec': 'vp9'},
- '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
- '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
- '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
- '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
- '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'vp9'},
- '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
+ '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp9', 'preference': -40},
+ '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ # itag 272 videos are either 3840x2160 (e.g. RtoitU2A-3E) or 7680x4320 (sLprVF6d7Ug)
+ '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40},
+ '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40},
+ '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40},
+ '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40},
+ '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40},
# Dash webm audio
- '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50},
- '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
+ '171': {'ext': 'webm', 'acodec': 'vorbis', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50},
+ '172': {'ext': 'webm', 'acodec': 'vorbis', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
# Dash webm audio with opus inside
'249': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 50, 'preference': -50},
'ext': 'mp4',
'upload_date': '20120506',
'title': 'Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]',
+ 'alt_title': 'I Love It (feat. Charli XCX)',
'description': 'md5:782e8651347686cba06e58f71ab51773',
'tags': ['Icona Pop i love it', 'sweden', 'pop music', 'big beat records', 'big beat', 'charli',
'xcx', 'charli xcx', 'girls', 'hbo', 'i love it', "i don't care", 'icona', 'pop',
'iconic ep', 'iconic', 'love', 'it'],
'uploader': 'Icona Pop',
'uploader_id': 'IconaPop',
+ 'creator': 'Icona Pop',
}
},
{
'ext': 'mp4',
'upload_date': '20130703',
'title': 'Justin Timberlake - Tunnel Vision (Explicit)',
+ 'alt_title': 'Tunnel Vision',
'description': 'md5:64249768eec3bc4276236606ea996373',
'uploader': 'justintimberlakeVEVO',
'uploader_id': 'justintimberlakeVEVO',
+ 'creator': 'Justin Timberlake',
'age_limit': 18,
}
},
'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012',
'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7',
'uploader': 'SET India',
- 'uploader_id': 'setindia'
+ 'uploader_id': 'setindia',
+ 'age_limit': 18,
}
},
{
'id': 'nfWlot6h_JM',
'ext': 'm4a',
'title': 'Taylor Swift - Shake It Off',
+ 'alt_title': 'Shake It Off',
'description': 'md5:95f66187cd7c8b2c13eb78e1223b63c3',
'uploader': 'TaylorSwiftVEVO',
'uploader_id': 'TaylorSwiftVEVO',
'upload_date': '20140818',
+ 'creator': 'Taylor Swift',
},
'params': {
'youtube_include_dash_manifest': True,
'ext': 'mp4',
'upload_date': '20100430',
'uploader_id': 'deadmau5',
+ 'creator': 'deadmau5',
'description': 'md5:12c56784b8032162bb936a5f76d55360',
'uploader': 'deadmau5',
'title': 'Deadmau5 - Some Chords (HD)',
+ 'alt_title': 'Some Chords',
},
'expected_warnings': [
'DASH manifest missing',
'info_dict': {
'id': 'lqQg6PlCWgI',
'ext': 'mp4',
- 'upload_date': '20120724',
+ 'upload_date': '20150827',
'uploader_id': 'olympic',
'description': 'HO09 - Women - GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
'uploader': 'Olympics',
},
'params': {
'skip_download': 'requires avconv',
- }
+ },
+ 'skip': 'This live event has ended.',
},
# Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097)
{
},
{
# Title with JS-like syntax "};" (see https://github.com/rg3/youtube-dl/issues/7468)
+ # Also tests cut-off URL expansion in video description (see
+ # https://github.com/rg3/youtube-dl/issues/1892,
+ # https://github.com/rg3/youtube-dl/issues/8164)
'url': 'https://www.youtube.com/watch?v=lsguqyKfVQg',
'info_dict': {
'id': 'lsguqyKfVQg',
'ext': 'mp4',
'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21',
+ 'alt_title': 'Dark Walk',
'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a',
'upload_date': '20151119',
'uploader_id': 'IronSoulElf',
'uploader': 'IronSoulElf',
+ 'creator': 'Todd Haberman, Daniel Law Heath & Aaron Kaplan',
},
'params': {
'skip_download': True,
'url': 'https://www.youtube.com/watch?v=Ms7iBXnlUO8',
'only_matching': True,
},
+ {
+ # Video with yt:stretch=17:0
+ 'url': 'https://www.youtube.com/watch?v=Q39EVAstoRM',
+ 'info_dict': {
+ 'id': 'Q39EVAstoRM',
+ 'ext': 'mp4',
+ 'title': 'Clash Of Clans#14 Dicas De Ataque Para CV 4',
+ 'description': 'md5:ee18a25c350637c8faff806845bddee9',
+ 'upload_date': '20151107',
+ 'uploader_id': 'UCCr7TALkRbo3EtFzETQF1LA',
+ 'uploader': 'CH GAMER DROID',
+ },
+ 'params': {
+ 'skip_download': True,
+ },
+ },
+ {
+ 'url': 'https://www.youtube.com/watch?feature=player_embedded&v=V36LpHqtcDY',
+ 'only_matching': True,
+ }
]
def __init__(self, *args, **kwargs):
'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
video_id, note=False)
except ExtractorError as err:
- self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err))
+ self._downloader.report_warning('unable to download video subtitles: %s' % error_to_compat_str(err))
return {}
sub_lang_list = {}
def _get_ytplayer_config(self, video_id, webpage):
patterns = (
+ # User data may contain arbitrary character sequences that may affect
+ # JSON extraction with regex, e.g. when '};' is contained the second
+ # regex won't capture the whole JSON. Yet working around by trying more
+ # concrete regex first keeping in mind proper quoted string handling
+ # to be implemented in future that will replace this workaround (see
+ # https://github.com/rg3/youtube-dl/issues/7468,
+ # https://github.com/rg3/youtube-dl/pull/7599)
r';ytplayer\.config\s*=\s*({.+?});ytplayer',
r';ytplayer\.config\s*=\s*({.+?});',
)
try:
args = player_config['args']
caption_url = args['ttsurl']
+ if not caption_url:
+ self._downloader.report_warning(err_msg)
+ return {}
timestamp = args['timestamp']
# We get the available subtitles
list_params = compat_urllib_parse.urlencode({
full_info.update(f)
codecs = r.attrib.get('codecs')
if codecs:
- if full_info.get('acodec') == 'none' and 'vcodec' not in full_info:
+ if full_info.get('acodec') == 'none':
full_info['vcodec'] = codecs
- elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info:
+ elif full_info.get('vcodec') == 'none':
full_info['acodec'] = codecs
formats.append(full_info)
else:
video_description = re.sub(r'''(?x)
<a\s+
(?:[a-zA-Z-]+="[^"]+"\s+)*?
- title="([^"]+)"\s+
+ (?:title|href)="([^"]+)"\s+
(?:[a-zA-Z-]+="[^"]+"\s+)*?
- class="yt-uix-redirect-link"\s*>
- [^<]+
+ class="(?:yt-uix-redirect-link|yt-uix-sessionlink[^"]*)"[^>]*>
+ [^<]+\.{3}\s*
</a>
''', r'\1', video_description)
video_description = clean_html(video_description)
upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
upload_date = unified_strdate(upload_date)
+ m_music = re.search(
+ r'<h4[^>]+class="title"[^>]*>\s*Music\s*</h4>\s*<ul[^>]*>\s*<li>(?P<title>.+?) by (?P<creator>.+?)(?:\(.+?\))?</li',
+ video_webpage)
+ if m_music:
+ video_alt_title = remove_quotes(unescapeHTML(m_music.group('title')))
+ video_creator = clean_html(m_music.group('creator'))
+ else:
+ video_alt_title = video_creator = None
+
m_cat_container = self._search_regex(
r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
video_webpage, 'categories', default=None)
if 'ratebypass' not in url:
url += '&ratebypass=yes'
+ dct = {
+ 'format_id': format_id,
+ 'url': url,
+ 'player_url': player_url,
+ }
+ if format_id in self._formats:
+ dct.update(self._formats[format_id])
+
# Some itags are not included in DASH manifest thus corresponding formats will
# lack metadata (see https://github.com/rg3/youtube-dl/pull/5993).
# Trying to extract metadata from url_encoded_fmt_stream_map entry.
mobj = re.search(r'^(?P<width>\d+)[xX](?P<height>\d+)$', url_data.get('size', [''])[0])
width, height = (int(mobj.group('width')), int(mobj.group('height'))) if mobj else (None, None)
- dct = {
- 'format_id': format_id,
- 'url': url,
- 'player_url': player_url,
+
+ more_fields = {
'filesize': int_or_none(url_data.get('clen', [None])[0]),
'tbr': float_or_none(url_data.get('bitrate', [None])[0], 1000),
'width': width,
'fps': int_or_none(url_data.get('fps', [None])[0]),
'format_note': url_data.get('quality_label', [None])[0] or url_data.get('quality', [None])[0],
}
+ for key, value in more_fields.items():
+ if value:
+ dct[key] = value
type_ = url_data.get('type', [None])[0]
if type_:
type_split = type_.split(';')
kind_ext = type_split[0].split('/')
if len(kind_ext) == 2:
- kind, ext = kind_ext
- dct['ext'] = ext
+ kind, _ = kind_ext
+ dct['ext'] = mimetype2ext(type_split[0])
if kind in ('audio', 'video'):
codecs = None
for mobj in re.finditer(
if codecs:
codecs = codecs.split(',')
if len(codecs) == 2:
- acodec, vcodec = codecs[0], codecs[1]
+ acodec, vcodec = codecs[1], codecs[0]
else:
acodec, vcodec = (codecs[0], 'none') if kind == 'audio' else ('none', codecs[0])
dct.update({
'acodec': acodec,
'vcodec': vcodec,
})
- if format_id in self._formats:
- dct.update(self._formats[format_id])
formats.append(dct)
elif video_info.get('hlsvp'):
manifest_url = video_info['hlsvp'][0]
url_map = self._extract_from_m3u8(manifest_url, video_id)
formats = _map_to_format_list(url_map)
+ # Accept-Encoding header causes failures in live streams on Youtube and Youtube Gaming
+ for a_format in formats:
+ a_format.setdefault('http_headers', {})['Youtubedl-no-compression'] = 'True'
else:
+ unavailable_message = self._html_search_regex(
+ r'(?s)<h1[^>]+id="unavailable-message"[^>]*>(.+?)</h1>',
+ video_webpage, 'unavailable message', default=None)
+ if unavailable_message:
+ raise ExtractorError(unavailable_message, expected=True)
raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
# Look for the DASH manifest
r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">',
video_webpage)
if stretched_m:
- ratio = float(stretched_m.group('w')) / float(stretched_m.group('h'))
- for f in formats:
- if f.get('vcodec') != 'none':
- f['stretched_ratio'] = ratio
+ w = float(stretched_m.group('w'))
+ h = float(stretched_m.group('h'))
+ # yt:stretch may hold invalid ratio data (e.g. for Q39EVAstoRM ratio is 17:0).
+ # We will only process correct ratios.
+ if w > 0 and h > 0:
+ ratio = w / h
+ for f in formats:
+ if f.get('vcodec') != 'none':
+ f['stretched_ratio'] = ratio
self._sort_formats(formats)
'uploader': video_uploader,
'uploader_id': video_uploader_id,
'upload_date': upload_date,
+ 'creator': video_creator,
'title': video_title,
+ 'alt_title': video_alt_title,
'thumbnail': video_thumbnail,
'description': video_description,
'categories': video_categories,
youtube\.com/
(?:
(?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries)
- \? (?:.*?&)*? (?:p|a|list)=
+ \? (?:.*?[&;])*? (?:p|a|list)=
| p/
)
(
},
}]
+ @classmethod
+ def suitable(cls, url):
+ return False if YoutubePlaylistsIE.suitable(url) else super(YoutubeChannelIE, cls).suitable(url)
+
def _real_extract(self, url):
channel_id = self._match_id(url)
return super(YoutubeUserIE, cls).suitable(url)
-class YoutubeUserPlaylistsIE(YoutubePlaylistsBaseInfoExtractor):
- IE_DESC = 'YouTube.com user playlists'
- _VALID_URL = r'https?://(?:\w+\.)?youtube\.com/user/(?P<id>[^/]+)/playlists'
- IE_NAME = 'youtube:user:playlists'
+class YoutubePlaylistsIE(YoutubePlaylistsBaseInfoExtractor):
+ IE_DESC = 'YouTube.com user/channel playlists'
+ _VALID_URL = r'https?://(?:\w+\.)?youtube\.com/(?:user|channel)/(?P<id>[^/]+)/playlists'
+ IE_NAME = 'youtube:playlists'
_TESTS = [{
'url': 'http://www.youtube.com/user/ThirstForScience/playlists',
'id': 'igorkle1',
'title': 'Игорь Клейнер',
},
+ }, {
+ 'url': 'https://www.youtube.com/channel/UCiU1dHvZObB2iP6xkJ__Icw/playlists',
+ 'playlist_mincount': 17,
+ 'info_dict': {
+ 'id': 'UCiU1dHvZObB2iP6xkJ__Icw',
+ 'title': 'Chem Player',
+ },
}]
projects / youtube-dl / blobdiff