diff options
Diffstat (limited to 'youtube_dl/extractor/thisoldhouse.py')
| -rw-r--r-- | youtube_dl/extractor/thisoldhouse.py | 19 | 
1 files changed, 14 insertions, 5 deletions
| diff --git a/youtube_dl/extractor/thisoldhouse.py b/youtube_dl/extractor/thisoldhouse.py index 197258df1..6ab147ad7 100644 --- a/youtube_dl/extractor/thisoldhouse.py +++ b/youtube_dl/extractor/thisoldhouse.py @@ -2,13 +2,15 @@  from __future__ import unicode_literals  from .common import InfoExtractor +from ..compat import compat_str +from ..utils import try_get  class ThisOldHouseIE(InfoExtractor):      _VALID_URL = r'https?://(?:www\.)?thisoldhouse\.com/(?:watch|how-to|tv-episode)/(?P<id>[^/?#]+)'      _TESTS = [{          'url': 'https://www.thisoldhouse.com/how-to/how-to-build-storage-bench', -        'md5': '946f05bbaa12a33f9ae35580d2dfcfe3', +        'md5': '568acf9ca25a639f0c4ff905826b662f',          'info_dict': {              'id': '2REGtUDQ',              'ext': 'mp4', @@ -28,8 +30,15 @@ class ThisOldHouseIE(InfoExtractor):      def _real_extract(self, url):          display_id = self._match_id(url)          webpage = self._download_webpage(url, display_id) -        drupal_settings = self._parse_json(self._search_regex( -            r'jQuery\.extend\(Drupal\.settings\s*,\s*({.+?})\);', -            webpage, 'drupal settings'), display_id) -        video_id = drupal_settings['jwplatform']['video_id'] +        video_id = self._search_regex( +            (r'data-mid=(["\'])(?P<id>(?:(?!\1).)+)\1', +             r'id=(["\'])inline-video-player-(?P<id>(?:(?!\1).)+)\1'), +            webpage, 'video id', default=None, group='id') +        if not video_id: +            drupal_settings = self._parse_json(self._search_regex( +                r'jQuery\.extend\(Drupal\.settings\s*,\s*({.+?})\);', +                webpage, 'drupal settings'), display_id) +            video_id = try_get( +                drupal_settings, lambda x: x['jwplatform']['video_id'], +                compat_str) or list(drupal_settings['comScore'])[0]          return self.url_result('jwplatform:' + video_id, 'JWPlatform', video_id) | 
