diff options
author | Filippo Valsorda <[email protected]> | 2018-01-07 15:03:28 +0100 |
---|---|---|
committer | GitHub <[email protected]> | 2018-01-07 15:03:28 +0100 |
commit | 97bc05116e15df3cf1b7b7216c60fad7ee0acc9f (patch) | |
tree | 453059413184d22a44241d251a82196c5d2461eb /youtube_dl/extractor/nbc.py | |
parent | 7608a91ee7b40c74a35c30b44731606382702592 (diff) | |
parent | 0a5b1295b7c1aa6395b65ee137087c540b37b32b (diff) | |
download | youtube-dl-totalwebcasting.tar.gz youtube-dl-totalwebcasting.zip |
Merge branch 'master' into totalwebcastingtotalwebcasting
Diffstat (limited to 'youtube_dl/extractor/nbc.py')
-rw-r--r-- | youtube_dl/extractor/nbc.py | 85 |
1 files changed, 46 insertions, 39 deletions
diff --git a/youtube_dl/extractor/nbc.py b/youtube_dl/extractor/nbc.py index 434a94de4..554dec36e 100644 --- a/youtube_dl/extractor/nbc.py +++ b/youtube_dl/extractor/nbc.py @@ -4,23 +4,24 @@ import re from .common import InfoExtractor from .theplatform import ThePlatformIE +from .adobepass import AdobePassIE from ..utils import ( find_xpath_attr, - lowercase_escape, smuggle_url, unescapeHTML, update_url_query, + int_or_none, ) -class NBCIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?nbc\.com/(?:[^/]+/)+(?P<id>n?\d+)' +class NBCIE(AdobePassIE): + _VALID_URL = r'https?(?P<permalink>://(?:www\.)?nbc\.com/(?:classic-tv/)?[^/]+/video/[^/]+/(?P<id>n?\d+))' _TESTS = [ { - 'url': 'http://www.nbc.com/the-tonight-show/segments/112966', + 'url': 'http://www.nbc.com/the-tonight-show/video/jimmy-fallon-surprises-fans-at-ben-jerrys/2848237', 'info_dict': { - 'id': '112966', + 'id': '2848237', 'ext': 'mp4', 'title': 'Jimmy Fallon Surprises Fans at Ben & Jerry\'s', 'description': 'Jimmy gives out free scoops of his new "Tonight Dough" ice cream flavor by surprising customers at the Ben & Jerry\'s scoop shop.', @@ -34,16 +35,6 @@ class NBCIE(InfoExtractor): }, }, { - 'url': 'http://www.nbc.com/the-tonight-show/episodes/176', - 'info_dict': { - 'id': '176', - 'ext': 'flv', - 'title': 'Ricky Gervais, Steven Van Zandt, ILoveMakonnen', - 'description': 'A brand new episode of The Tonight Show welcomes Ricky Gervais, Steven Van Zandt and ILoveMakonnen.', - }, - 'skip': '404 Not Found', - }, - { 'url': 'http://www.nbc.com/saturday-night-live/video/star-wars-teaser/2832821', 'info_dict': { 'id': '2832821', @@ -61,15 +52,10 @@ class NBCIE(InfoExtractor): 'skip': 'Only works from US', }, { - # This video has expired but with an escaped embedURL - 'url': 'http://www.nbc.com/parenthood/episode-guide/season-5/just-like-at-home/515', - 'only_matching': True, - }, - { # HLS streams requires the 'hdnea3' cookie 'url': 'http://www.nbc.com/Kings/video/goliath/n1806', 'info_dict': { - 'id': 'n1806', + 'id': '101528f5a9e8127b107e98c5e6ce4638', 'ext': 'mp4', 'title': 'Goliath', 'description': 'When an unknown soldier saves the life of the King\'s son in battle, he\'s thrust into the limelight and politics of the kingdom.', @@ -81,26 +67,46 @@ class NBCIE(InfoExtractor): 'skip_download': True, }, 'skip': 'Only works from US', - } + }, + { + 'url': 'https://www.nbc.com/classic-tv/charles-in-charge/video/charles-in-charge-pilot/n3310', + 'only_matching': True, + }, ] def _real_extract(self, url): - video_id = self._match_id(url) - webpage = self._download_webpage(url, video_id) - theplatform_url = unescapeHTML(lowercase_escape(self._html_search_regex( - [ - r'(?:class="video-player video-player-full" data-mpx-url|class="player" src)="(.*?)"', - r'<iframe[^>]+src="((?:https?:)?//player\.theplatform\.com/[^"]+)"', - r'"embedURL"\s*:\s*"([^"]+)"' - ], - webpage, 'theplatform url').replace('_no_endcard', '').replace('\\/', '/'))) - if theplatform_url.startswith('//'): - theplatform_url = 'http:' + theplatform_url + permalink, video_id = re.match(self._VALID_URL, url).groups() + permalink = 'http' + permalink + video_data = self._download_json( + 'https://api.nbc.com/v3/videos', video_id, query={ + 'filter[permalink]': permalink, + })['data'][0]['attributes'] + query = { + 'mbr': 'true', + 'manifest': 'm3u', + } + video_id = video_data['guid'] + title = video_data['title'] + if video_data.get('entitlement') == 'auth': + resource = self._get_mvpd_resource( + 'nbcentertainment', title, video_id, + video_data.get('vChipRating')) + query['auth'] = self._extract_mvpd_auth( + url, video_id, 'nbcentertainment', resource) + theplatform_url = smuggle_url(update_url_query( + 'http://link.theplatform.com/s/NnzsPC/media/guid/2410887629/' + video_id, + query), {'force_smil_url': True}) return { '_type': 'url_transparent', - 'ie_key': 'ThePlatform', - 'url': smuggle_url(theplatform_url, {'source_url': url}), 'id': video_id, + 'title': title, + 'url': theplatform_url, + 'description': video_data.get('description'), + 'keywords': video_data.get('keywords'), + 'season_number': int_or_none(video_data.get('seasonNumber')), + 'episode_number': int_or_none(video_data.get('episodeNumber')), + 'series': video_data.get('showName'), + 'ie_key': 'ThePlatform', } @@ -108,10 +114,10 @@ class NBCSportsVPlayerIE(InfoExtractor): _VALID_URL = r'https?://vplayer\.nbcsports\.com/(?:[^/]+/)+(?P<id>[0-9a-zA-Z_]+)' _TESTS = [{ - 'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_share/select/9CsDKds0kvHI', + 'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/9CsDKds0kvHI', 'info_dict': { 'id': '9CsDKds0kvHI', - 'ext': 'flv', + 'ext': 'mp4', 'description': 'md5:df390f70a9ba7c95ff1daace988f0d8d', 'title': 'Tyler Kalinoski hits buzzer-beater to lift Davidson', 'timestamp': 1426270238, @@ -119,7 +125,7 @@ class NBCSportsVPlayerIE(InfoExtractor): 'uploader': 'NBCU-SPORTS', } }, { - 'url': 'http://vplayer.nbcsports.com/p/BxmELC/nbc_embedshare/select/_hqLjQ95yx8Z', + 'url': 'https://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/media/_hqLjQ95yx8Z', 'only_matching': True, }] @@ -133,7 +139,8 @@ class NBCSportsVPlayerIE(InfoExtractor): def _real_extract(self, url): video_id = self._match_id(url) webpage = self._download_webpage(url, video_id) - theplatform_url = self._og_search_video_url(webpage) + theplatform_url = self._og_search_video_url(webpage).replace( + 'vplayer.nbcsports.com', 'player.theplatform.com') return self.url_result(theplatform_url, 'ThePlatform') |