[thesun] fix extraction(closes #16966)

This commit is contained in:
Remita Amine 2019-10-28 13:32:35 +01:00
parent 71fa0b04f9
commit 80c2126e80
1 changed files with 10 additions and 4 deletions

View File

@ -3,7 +3,7 @@ from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from .ooyala import OoyalaIE from ..utils import extract_attributes
class TheSunIE(InfoExtractor): class TheSunIE(InfoExtractor):
@ -16,6 +16,7 @@ class TheSunIE(InfoExtractor):
}, },
'playlist_count': 2, 'playlist_count': 2,
} }
BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s'
def _real_extract(self, url): def _real_extract(self, url):
article_id = self._match_id(url) article_id = self._match_id(url)
@ -23,10 +24,15 @@ class TheSunIE(InfoExtractor):
webpage = self._download_webpage(url, article_id) webpage = self._download_webpage(url, article_id)
entries = [] entries = []
for ooyala_id in re.findall( for video in re.findall(
r'<[^>]+\b(?:id\s*=\s*"thesun-ooyala-player-|data-content-id\s*=\s*")([^"]+)', r'<video[^>]+data-video-id-pending=[^>]+>',
webpage): webpage):
entries.append(OoyalaIE._build_url_result(ooyala_id)) attrs = extract_attributes(video)
video_id = attrs['data-video-id-pending']
account_id = attrs.get('data-account', '5067014667001')
entries.append(self.url_result(
self.BRIGHTCOVE_URL_TEMPLATE % (account_id, video_id),
'BrightcoveNew', video_id))
return self.playlist_result( return self.playlist_result(
entries, article_id, self._og_search_title(webpage, fatal=False)) entries, article_id, self._og_search_title(webpage, fatal=False))