commit: 087d8652302270764bf9679c7621db0d20ad5572
parent a4fc1151f125a2dfb89b2d1505838bcbb71aacc9
Author: dirkf <fieldhouse@gmx.net>
Date: Tue, 25 Mar 2025 22:14:08 +0000
[YouTube] Support new player URL patterns
Diffstat:
2 files changed, 11 insertions(+), 7 deletions(-)
diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py
@@ -302,8 +302,8 @@ class TestSignature(unittest.TestCase):
def t_factory(name, sig_func, url_pattern):
def make_tfunc(url, sig_input, expected_sig):
m = url_pattern.match(url)
- assert m, '%r should follow URL format' % url
- test_id = m.group('id')
+ assert m, '{0!r} should follow URL format'.format(url)
+ test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id'))
def test_func(self):
basename = 'player-{0}-{1}.js'.format(name, test_id)
@@ -336,12 +336,16 @@ def n_sig(jscode, sig_input):
make_sig_test = t_factory(
- 'signature', signature, re.compile(r'.*(?:-|/player/)(?P<id>[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$'))
+ 'signature', signature,
+ re.compile(r'''(?x)
+ .+/(?P<h5>html5)?player(?(h5)(?:-en_US)?-|/)(?P<id>[a-zA-Z0-9/._-]+)
+ (?(h5)/(?:watch_as3|html5player))?\.js$
+ '''))
for test_spec in _SIG_TESTS:
make_sig_test(*test_spec)
make_nsig_test = t_factory(
- 'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_-]+)/.+.js$'))
+ 'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_/.-]+)\.js$'))
for test_spec in _NSIG_TESTS:
make_nsig_test(*test_spec)
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
@@ -692,9 +692,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'invidious': '|'.join(_INVIDIOUS_SITES),
}
_PLAYER_INFO_RE = (
- r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
- r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
- r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
+ r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player',
+ r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
+ r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$',
)
_SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt')