Cool_Tools/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-15 03:27:46 +01:00
Code Releases Activity

[nbc] fix NBCSport VPlayer URL extraction(closes #16640)

Browse Source
This commit is contained in:
Remita Amine 2020-12-29 17:15:13 +01:00
parent 5966095e65
commit ebdcf70b0d
1 changed files with 19 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
youtube_dl/extractor/nbc.py
View File

@ -158,7 +158,8 @@ class NBCIE(AdobePassIE):
class NBCSportsVPlayerIE(InfoExtractor): class NBCSportsVPlayerIE(InfoExtractor):
_VALID_URL = r'https?://vplayer\.nbcsports\.com/(?:[^/]+/)+(?P<id>[0-9a-zA-Z_]+)' _VALID_URL_BASE = r'https?://(?:vplayer\.nbcsports\.com|(?:www\.)?nbcsports\.com/vplayer)/'
_VALID_URL = _VALID_URL_BASE + r'(?:[^/]+/)+(?P<id>[0-9a-zA-Z_]+)'
_TESTS = [{ _TESTS = [{
'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/9CsDKds0kvHI', 'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/9CsDKds0kvHI',
@ -174,12 +175,15 @@ class NBCSportsVPlayerIE(InfoExtractor):
}, { }, {
'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/media/_hqLjQ95yx8Z', 'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/media/_hqLjQ95yx8Z',
'only_matching': True, 'only_matching': True,
}, {
'url': 'https://www.nbcsports.com/vplayer/p/BxmELC/nbcsports/select/PHJSaFWbrTY9?form=html&autoPlay=true',
'only_matching': True,
}] }]
@staticmethod @staticmethod
def _extract_url(webpage): def _extract_url(webpage):
iframe_m = re.search( iframe_m = re.search(
r'<iframe[^>]+src="(?P<url>https?://vplayer\.nbcsports\.com/[^"]+)"', webpage) r'<(?:iframe[^>]+|div[^>]+data-(?:mpx-)?)src="(?P<url>%s[^"]+)"' % NBCSportsVPlayerIE._VALID_URL_BASE, webpage)
if iframe_m: if iframe_m:
return iframe_m.group('url') return iframe_m.group('url')
@ -192,21 +196,29 @@ class NBCSportsVPlayerIE(InfoExtractor):
class NBCSportsIE(InfoExtractor): class NBCSportsIE(InfoExtractor):
# Does not include https because its certificate is invalid _VALID_URL = r'https?://(?:www\.)?nbcsports\.com//?(?!vplayer/)(?:[^/]+/)+(?P<id>[0-9a-z-]+)'
_VALID_URL = r'https?://(?:www\.)?nbcsports\.com//?(?:[^/]+/)+(?P<id>[0-9a-z-]+)'
_TEST = { _TESTS = [{
# iframe src
'url': 'http://www.nbcsports.com//college-basketball/ncaab/tom-izzo-michigan-st-has-so-much-respect-duke', 'url': 'http://www.nbcsports.com//college-basketball/ncaab/tom-izzo-michigan-st-has-so-much-respect-duke',
'info_dict': { 'info_dict': {
'id': 'PHJSaFWbrTY9', 'id': 'PHJSaFWbrTY9',
'ext': 'flv', 'ext': 'mp4',
'title': 'Tom Izzo, Michigan St. has \'so much respect\' for Duke', 'title': 'Tom Izzo, Michigan St. has \'so much respect\' for Duke',
'description': 'md5:ecb459c9d59e0766ac9c7d5d0eda8113', 'description': 'md5:ecb459c9d59e0766ac9c7d5d0eda8113',
'uploader': 'NBCU-SPORTS', 'uploader': 'NBCU-SPORTS',
'upload_date': '20150330', 'upload_date': '20150330',
'timestamp': 1427726529, 'timestamp': 1427726529,
} }
} }, {
# data-mpx-src
'url': 'https://www.nbcsports.com/philadelphia/philadelphia-phillies/bruce-bochy-hector-neris-hes-idiot',
'only_matching': True,
}, {
# data-src
'url': 'https://www.nbcsports.com/boston/video/report-card-pats-secondary-no-match-josh-allen',
'only_matching': True,
}]
def _real_extract(self, url): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)