diff options
| author | renalid <renalid@gmail.com> | 2020-12-19 17:18:51 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2020-12-19 23:18:51 +0700 |
| commit | 71679eaee857d9dfcf1fdc4b5016abfc1fe8bc45 (patch) | |
| tree | 6805e4ad28d80edf6a4c16b5195951c9d01890e2 | |
| parent | 76fe4ba3b22a689ba7b5c6ea9064796ce1f4af55 (diff) | |
| download | youtube-dl-71679eaee857d9dfcf1fdc4b5016abfc1fe8bc45.tar.xz | |
[generic] Fix RSS itunes thumbnail extraction (#27405)
| -rw-r--r-- | youtube_dl/extractor/generic.py | 27 |
1 files changed, 26 insertions, 1 deletions
diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py index 80ecaf795..ab24572e4 100644 --- a/youtube_dl/extractor/generic.py +++ b/youtube_dl/extractor/generic.py @@ -35,6 +35,7 @@ from ..utils import ( unsmuggle_url, UnsupportedError, url_or_none, + xpath_attr, xpath_text, xpath_with_ns, ) @@ -217,6 +218,30 @@ class GenericIE(InfoExtractor): }, }], }, + # RSS feed with item with description and thumbnails + { + 'url': 'https://anchor.fm/s/dd00e14/podcast/rss', + 'info_dict': { + 'id': 'https://anchor.fm/s/dd00e14/podcast/rss', + 'title': 're:.*100% Hydrogen.*', + 'description': 're:.*In this episode.*', + }, + 'playlist': [{ + 'info_dict': { + 'ext': 'm4a', + 'id': 'c1c879525ce2cb640b344507e682c36d', + 'title': 're:Hydrogen!', + 'description': 're:.*In this episode we are going.*', + 'timestamp': int, + 'upload_date': '20190908', + 'duration': int, + 'thumbnail': r're:^https?://.*\.jpg$', + }, + }], + 'params': { + 'skip_download': True, + }, + }, # RSS feed with enclosures and unsupported link URLs { 'url': 'http://www.hellointernet.fm/podcast?format=rss', @@ -2234,7 +2259,7 @@ class GenericIE(InfoExtractor): 'timestamp': unified_timestamp( xpath_text(it, 'pubDate', default=None)), 'duration': int_or_none(duration) or parse_duration(duration), - 'thumbnail': url_or_none(itunes('image')), + 'thumbnail': url_or_none(xpath_attr(it, xpath_with_ns('./itunes:image', NS_MAP), 'href')), 'episode': itunes('title'), 'episode_number': int_or_none(itunes('episode')), 'season_number': int_or_none(itunes('season')), |