[msn] improve extraction

- add support for YouTube and NBCSports embeds
- add support for aricles with multiple videos
- improve AOL embed support
- improve format extraction
This commit is contained in:
Remita Amine 2019-11-29 17:31:17 +01:00
parent 88a7a9089a
commit c0b1e01330

View File

@ -14,20 +14,27 @@ from ..utils import (
class MSNIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?msn\.com/(?:[^/]+/)+(?P<display_id>[^/]+)/[a-z]{2}-(?P<id>[\da-zA-Z]+)'
_VALID_URL = r'https?://(?:(?:www|preview)\.)?msn\.com/(?:[^/]+/)+(?P<display_id>[^/]+)/[a-z]{2}-(?P<id>[\da-zA-Z]+)'
_TESTS = [{
'url': 'http://www.msn.com/en-ae/foodanddrink/joinourtable/criminal-minds-shemar-moore-shares-a-touching-goodbye-message/vp-BBqQYNE',
'md5': '8442f66c116cbab1ff7098f986983458',
'url': 'https://www.msn.com/en-in/money/video/7-ways-to-get-rid-of-chest-congestion/vi-BBPxU6d',
'md5': '087548191d273c5c55d05028f8d2cbcd',
'info_dict': {
'id': 'BBqQYNE',
'display_id': 'criminal-minds-shemar-moore-shares-a-touching-goodbye-message',
'id': 'BBPxU6d',
'display_id': '7-ways-to-get-rid-of-chest-congestion',
'ext': 'mp4',
'title': 'Criminal Minds - Shemar Moore Shares A Touching Goodbye Message',
'description': 'md5:e8e89b897b222eb33a6b5067a8f1bc25',
'duration': 104,
'uploader': 'CBS Entertainment',
'uploader_id': 'IT0X5aoJ6bJgYerJXSDCgFmYPB1__54v',
'title': 'Seven ways to get rid of chest congestion',
'description': '7 Ways to Get Rid of Chest Congestion',
'duration': 88,
'uploader': 'Health',
'uploader_id': 'BBPrMqa',
},
}, {
# Article, multiple Dailymotion Embeds
'url': 'https://www.msn.com/en-in/money/sports/hottest-football-wags-greatest-footballers-turned-managers-and-more/ar-BBpc7Nl',
'info_dict': {
'id': 'BBpc7Nl',
},
'playlist_mincount': 4,
}, {
'url': 'http://www.msn.com/en-ae/news/offbeat/meet-the-nine-year-old-self-made-millionaire/ar-BBt6ZKf',
'only_matching': True,
@ -43,44 +50,58 @@ class MSNIE(InfoExtractor):
'only_matching': True,
}, {
# Vidible(AOL) Embed
'url': 'https://www.msn.com/en-us/video/animals/yellowstone-park-staffers-catch-deer-engaged-in-behavior-they-cant-explain/vi-AAGfdg1',
'url': 'https://www.msn.com/en-us/money/other/jupiter-is-about-to-come-so-close-you-can-see-its-moons-with-binoculars/vi-AACqsHR',
'only_matching': True,
}, {
# Dailymotion Embed
'url': 'https://www.msn.com/es-ve/entretenimiento/watch/winston-salem-paire-refait-des-siennes-en-perdant-sa-raquette-au-service/vp-AAG704L',
'only_matching': True,
}, {
# YouTube Embed
'url': 'https://www.msn.com/en-in/money/news/meet-vikram-%E2%80%94-chandrayaan-2s-lander/vi-AAGUr0v',
'only_matching': True,
}, {
# NBCSports Embed
'url': 'https://www.msn.com/en-us/money/football_nfl/week-13-preview-redskins-vs-panthers/vi-BBXsCDb',
'only_matching': True,
}]
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id, display_id = mobj.group('id', 'display_id')
display_id, page_id = re.match(self._VALID_URL, url).groups()
webpage = self._download_webpage(url, display_id)
video = self._parse_json(
self._search_regex(
r'data-metadata\s*=\s*(["\'])(?P<data>.+?)\1',
webpage, 'video data', default='{}', group='data'),
display_id, transform_source=unescapeHTML)
entries = []
for _, metadata in re.findall(r'data-metadata\s*=\s*(["\'])(?P<data>.+?)\1', webpage):
video = self._parse_json(unescapeHTML(metadata), display_id)
if not video:
error = unescapeHTML(self._search_regex(
r'data-error=(["\'])(?P<error>.+?)\1',
webpage, 'error', group='error'))
raise ExtractorError('%s said: %s' % (self.IE_NAME, error), expected=True)
player_name = video.get('playerName')
if player_name:
provider_id = video.get('providerId')
if provider_id:
player_name = video.get('playerName')
if player_name and provider_id:
entry = None
if player_name == 'AOL':
return self.url_result(
if provider_id.startswith('http'):
provider_id = self._search_regex(
r'https?://delivery\.vidible\.tv/video/redirect/([0-9a-f]{24})',
provider_id, 'vidible id')
entry = self.url_result(
'aol-video:' + provider_id, 'Aol', provider_id)
elif player_name == 'Dailymotion':
return self.url_result(
entry = self.url_result(
'https://www.dailymotion.com/video/' + provider_id,
'Dailymotion', provider_id)
elif player_name == 'YouTube':
entry = self.url_result(
provider_id, 'Youtube', provider_id)
elif player_name == 'NBCSports':
entry = self.url_result(
'http://vplayer.nbcsports.com/p/BxmELC/nbcsports_embed/select/media/' + provider_id,
'NBCSportsVPlayer', provider_id)
if entry:
entries.append(entry)
continue
video_id = video['uuid']
title = video['title']
formats = []
@ -88,23 +109,30 @@ class MSNIE(InfoExtractor):
format_url = file_.get('url')
if not format_url:
continue
if 'm3u8' in format_url:
if 'format=m3u8-aapl' in format_url:
# m3u8_native should not be used here until
# https://github.com/ytdl-org/youtube-dl/issues/9913 is fixed
m3u8_formats = self._extract_m3u8_formats(
formats.extend(self._extract_m3u8_formats(
format_url, display_id, 'mp4',
m3u8_id='hls', fatal=False)
formats.extend(m3u8_formats)
elif determine_ext(format_url) == 'ism':
m3u8_id='hls', fatal=False))
elif 'format=mpd-time-csf' in format_url:
formats.extend(self._extract_mpd_formats(
format_url, display_id, 'dash', fatal=False))
elif '.ism' in format_url:
if format_url.endswith('.ism'):
format_url += '/manifest'
formats.extend(self._extract_ism_formats(
format_url + '/Manifest', display_id, 'mss', fatal=False))
format_url, display_id, 'mss', fatal=False))
else:
format_id = file_.get('formatCode')
formats.append({
'url': format_url,
'ext': 'mp4',
'format_id': 'http',
'format_id': format_id,
'width': int_or_none(file_.get('width')),
'height': int_or_none(file_.get('height')),
'vbr': int_or_none(self._search_regex(r'_(\d+)\.mp4', format_url, 'vbr', default=None)),
'preference': 1 if format_id == '1001' else None,
})
self._sort_formats(formats)
@ -120,7 +148,7 @@ class MSNIE(InfoExtractor):
'url': format_url,
})
return {
entries.append({
'id': video_id,
'display_id': display_id,
'title': title,
@ -132,4 +160,12 @@ class MSNIE(InfoExtractor):
'creator': video.get('creator'),
'subtitles': subtitles,
'formats': formats,
}
})
if not entries:
error = unescapeHTML(self._search_regex(
r'data-error=(["\'])(?P<error>.+?)\1',
webpage, 'error', group='error'))
raise ExtractorError('%s said: %s' % (self.IE_NAME, error), expected=True)
return self.playlist_result(entries, page_id)