aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authordirkf <fieldhouse@gmx.net>2025-01-06 01:22:16 +0000
committerdirkf <fieldhouse@gmx.net>2025-01-06 01:22:16 +0000
commitab7c61ca29ed1d1216d463d01794eb112a9144d5 (patch)
tree9e090669b482ff1a4b97c3248a18b129d149ae69
parent176fc2cb003b7a74f5781a64a4e1ce32e740c149 (diff)
[YouTube] Apply code style changes, trailing commas, etc
-rw-r--r--youtube_dl/extractor/youtube.py60
1 files changed, 31 insertions, 29 deletions
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 7ea30fd40..e9603d155 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -9,6 +9,7 @@ import json
import os.path
import random
import re
+import string
import time
import traceback
@@ -67,6 +68,7 @@ from ..utils import (
class YoutubeBaseInfoExtractor(InfoExtractor):
"""Provide base functions for Youtube extractors"""
+
_LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
_TWOFACTOR_URL = 'https://accounts.google.com/signin/challenge'
@@ -138,7 +140,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
[2, 1, None, 1,
'https://accounts.google.com/ServiceLogin?passive=true&continue=https%3A%2F%2Fwww.youtube.com%2Fsignin%3Fnext%3D%252F%26action_handle_signin%3Dtrue%26hl%3Den%26app%3Ddesktop%26feature%3Dsign_in_button&hl=en&service=youtube&uilel=3&requestPath=%2FServiceLogin&Page=PasswordSeparationSignIn',
None, [], 4],
- 1, [None, None, []], None, None, None, True
+ 1, [None, None, []], None, None, None, True,
],
username,
]
@@ -160,7 +162,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
None, 1, None, [1, None, None, None, [password, None, True]],
[
None, None, [2, 1, None, 1, 'https://accounts.google.com/ServiceLogin?passive=true&continue=https%3A%2F%2Fwww.youtube.com%2Fsignin%3Fnext%3D%252F%26action_handle_signin%3Dtrue%26hl%3Den%26app%3Ddesktop%26feature%3Dsign_in_button&hl=en&service=youtube&uilel=3&requestPath=%2FServiceLogin&Page=PasswordSeparationSignIn', None, [], 4],
- 1, [None, None, []], None, None, None, True
+ 1, [None, None, []], None, None, None, True,
]]
challenge_results = req(
@@ -213,7 +215,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
user_hash, None, 2, None,
[
9, None, None, None, None, None, None, None,
- [None, tfa_code, True, 2]
+ [None, tfa_code, True, 2],
]]
tfa_results = req(
@@ -284,7 +286,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
'client': {
'clientName': 'WEB',
'clientVersion': '2.20201021.03.00',
- }
+ },
},
}
@@ -385,7 +387,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
'client': {
'clientName': 'WEB',
'clientVersion': '2.20201021.03.00',
- }
+ },
},
'query': query,
}
@@ -462,7 +464,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
# (HTML, videodetails, metadata, renderers)
'name': ('content', 'author', (('ownerChannelName', None), 'title'), ['text']),
'url': ('href', 'ownerProfileUrl', 'vanityChannelUrl',
- ['navigationEndpoint', 'browseEndpoint', 'canonicalBaseUrl'])
+ ['navigationEndpoint', 'browseEndpoint', 'canonicalBaseUrl']),
}
if any((videodetails, metadata, renderers)):
result = (
@@ -671,7 +673,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/8KVIDEO',
'description': '',
'uploader': '8KVIDEO',
- 'title': 'UHDTV TEST 8K VIDEO.mp4'
+ 'title': 'UHDTV TEST 8K VIDEO.mp4',
},
'params': {
'youtube_include_dash_manifest': True,
@@ -711,7 +713,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'uploader_url': r're:https?://(?:www\.)?youtube\.com/@theamazingatheist',
'title': 'Burning Everyone\'s Koran',
'description': 'SUBSCRIBE: http://www.youtube.com/saturninefilms \r\n\r\nEven Obama has taken a stand against freedom on this issue: http://www.huffingtonpost.com/2010/09/09/obama-gma-interview-quran_n_710282.html',
- }
+ },
},
# Age-gated videos
{
@@ -839,7 +841,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
},
'expected_warnings': [
'DASH manifest missing',
- ]
+ ],
},
# Olympics (https://github.com/ytdl-org/youtube-dl/issues/4431)
{
@@ -1820,8 +1822,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
# cpn generation algorithm is reverse engineered from base.js.
# In fact it works even with dummy cpn.
- CPN_ALPHABET = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_'
- cpn = ''.join(CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(0, 16))
+ CPN_ALPHABET = string.ascii_letters + string.digits + '-_'
+ cpn = ''.join(CPN_ALPHABET[random.randint(0, 256) & 63] for _ in range(16))
# more consistent results setting it to right before the end
qs = parse_qs(playback_url)
@@ -1881,8 +1883,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
if mobj is None:
raise ExtractorError('Invalid URL: %s' % url)
- video_id = mobj.group(2)
- return video_id
+ return mobj.group(2)
def _extract_chapters_from_json(self, data, video_id, duration):
chapters_list = try_get(
@@ -2035,7 +2036,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
headers = {
'X-YouTube-Client-Name': '85',
'X-YouTube-Client-Version': '2.0',
- 'Origin': 'https://www.youtube.com'
+ 'Origin': 'https://www.youtube.com',
}
video_info = self._call_api('player', query, video_id, fatal=False, headers=headers)
@@ -2064,8 +2065,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
return ''.join([r['text'] for r in runs if isinstance(r.get('text'), compat_str)])
search_meta = (
- lambda x: self._html_search_meta(x, webpage, default=None)) \
- if webpage else lambda x: None
+ (lambda x: self._html_search_meta(x, webpage, default=None))
+ if webpage else lambda _: None)
video_details = player_response.get('videoDetails') or {}
microformat = try_get(
@@ -2137,7 +2138,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
def build_fragments(f):
return LazyList({
'url': update_url_query(f['url'], {
- 'range': '{0}-{1}'.format(range_start, min(range_start + CHUNK_SIZE - 1, f['filesize']))
+ 'range': '{0}-{1}'.format(range_start, min(range_start + CHUNK_SIZE - 1, f['filesize'])),
})
} for range_start in range(0, f['filesize'], CHUNK_SIZE))
@@ -2236,7 +2237,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'protocol': 'http_dash_segments',
'fragments': build_fragments(dct),
} if dct['filesize'] else {
- 'downloader_options': {'http_chunk_size': CHUNK_SIZE} # No longer useful?
+ 'downloader_options': {'http_chunk_size': CHUNK_SIZE}, # No longer useful?
})
formats.append(dct)
@@ -2454,7 +2455,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
info['subtitles'] = subtitles
parsed_url = compat_urllib_parse_urlparse(url)
- for component in [parsed_url.fragment, parsed_url.query]:
+ for component in (parsed_url.fragment, parsed_url.query):
query = compat_parse_qs(component)
for k, v in query.items():
for d_k, s_ks in [('start', ('start', 't')), ('end', ('end',))]:
@@ -2684,7 +2685,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
'title': 'Super Cooper Shorts - Shorts',
'uploader': 'Super Cooper Shorts',
'uploader_id': '@SuperCooperShorts',
- }
+ },
}, {
# Channel that does not have a Shorts tab. Test should just download videos on Home tab instead
'url': 'https://www.youtube.com/@emergencyawesome/shorts',
@@ -2738,7 +2739,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
'description': 'md5:609399d937ea957b0f53cbffb747a14c',
'uploader': 'ThirstForScience',
'uploader_id': '@ThirstForScience',
- }
+ },
}, {
'url': 'https://www.youtube.com/c/ChristophLaimer/playlists',
'only_matching': True,
@@ -3037,7 +3038,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
'uploader': '3Blue1Brown',
'uploader_id': '@3blue1brown',
'channel_id': 'UCYO_jab_esuFRV4b17AJtAw',
- }
+ },
}]
@classmethod
@@ -3335,7 +3336,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
'client': {
'clientName': 'WEB',
'clientVersion': client_version,
- }
+ },
}
visitor_data = try_get(context, lambda x: x['client']['visitorData'], compat_str)
@@ -3354,7 +3355,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
headers['x-goog-visitor-id'] = visitor_data
data['continuation'] = continuation['continuation']
data['clickTracking'] = {
- 'clickTrackingParams': continuation['itct']
+ 'clickTrackingParams': continuation['itct'],
}
count = 0
retries = 3
@@ -3613,7 +3614,7 @@ class YoutubePlaylistIE(InfoExtractor):
'uploader': 'milan',
'uploader_id': '@milan5503',
'channel_id': 'UCEI1-PVPcYXjB73Hfelbmaw',
- }
+ },
}, {
'url': 'http://www.youtube.com/embed/_xDOZElKyNU?list=PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
'playlist_mincount': 455,
@@ -3623,7 +3624,7 @@ class YoutubePlaylistIE(InfoExtractor):
'uploader': 'LBK',
'uploader_id': '@music_king',
'channel_id': 'UC21nz3_MesPLqtDqwdvnoxA',
- }
+ },
}, {
'url': 'TLGGrESM50VT6acwMjAyMjAxNw',
'only_matching': True,
@@ -3734,7 +3735,7 @@ class YoutubeSearchIE(SearchInfoExtractor, YoutubeBaseInfoExtractor):
'info_dict': {
'id': 'youtube-dl test video',
'title': 'youtube-dl test video',
- }
+ },
}]
def _get_n_results(self, query, n):
@@ -3754,7 +3755,7 @@ class YoutubeSearchDateIE(YoutubeSearchIE):
'info_dict': {
'id': 'youtube-dl test video',
'title': 'youtube-dl test video',
- }
+ },
}]
@@ -3769,7 +3770,7 @@ class YoutubeSearchURLIE(YoutubeBaseInfoExtractor):
'id': 'youtube-dl test video',
'title': 'youtube-dl test video',
},
- 'params': {'playlistend': 5}
+ 'params': {'playlistend': 5},
}, {
'url': 'https://www.youtube.com/results?q=test&sp=EgQIBBgB',
'only_matching': True,
@@ -3785,6 +3786,7 @@ class YoutubeSearchURLIE(YoutubeBaseInfoExtractor):
class YoutubeFeedsInfoExtractor(YoutubeTabIE):
"""
Base class for feed extractors
+
Subclasses must define the _FEED_NAME property.
"""
_LOGIN_REQUIRED = True