[generic] Fix RSS itunes thumbnail extraction (#27405)

This commit is contained in:
renalid 2020-12-19 17:18:51 +01:00 committed by GitHub
parent 76fe4ba3b2
commit 71679eaee8
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23

View file

@ -35,6 +35,7 @@ from ..utils import (
unsmuggle_url, unsmuggle_url,
UnsupportedError, UnsupportedError,
url_or_none, url_or_none,
xpath_attr,
xpath_text, xpath_text,
xpath_with_ns, 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 # RSS feed with enclosures and unsupported link URLs
{ {
'url': 'http://www.hellointernet.fm/podcast?format=rss', 'url': 'http://www.hellointernet.fm/podcast?format=rss',
@ -2234,7 +2259,7 @@ class GenericIE(InfoExtractor):
'timestamp': unified_timestamp( 'timestamp': unified_timestamp(
xpath_text(it, 'pubDate', default=None)), xpath_text(it, 'pubDate', default=None)),
'duration': int_or_none(duration) or parse_duration(duration), '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': itunes('title'),
'episode_number': int_or_none(itunes('episode')), 'episode_number': int_or_none(itunes('episode')),
'season_number': int_or_none(itunes('season')), 'season_number': int_or_none(itunes('season')),