kumi/yt-dlp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[globo] Expand valid URL (#2732)

Browse source 
Closes #2730 
Authored by: Bricio
This commit is contained in:
Bricio 2022-02-11 15:08:55 -03:00 committed by GitHub
parent cd170e8184
commit 17b183886f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 10 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
yt_dlp/extractor/globo.py
View file

@ -186,6 +186,7 @@ class GloboArticleIE(InfoExtractor):
r'\bvideosIDs\s*:\s*["\']?(\d{7,})', r'\bvideosIDs\s*:\s*["\']?(\d{7,})',
r'\bdata-id=["\'](\d{7,})', r'\bdata-id=["\'](\d{7,})',
r'<div[^>]+\bid=["\'](\d{7,})', r'<div[^>]+\bid=["\'](\d{7,})',
r'<bs-player[^>]+\bvideoid=["\'](\d{8,})',
] ]
_TESTS = [{ _TESTS = [{
@ -213,6 +214,14 @@ class GloboArticleIE(InfoExtractor):
}, { }, {
'url': 'http://oglobo.globo.com/rio/a-amizade-entre-um-entregador-de-farmacia-um-piano-19946271', 'url': 'http://oglobo.globo.com/rio/a-amizade-entre-um-entregador-de-farmacia-um-piano-19946271',
'only_matching': True, 'only_matching': True,
}, {
'url': 'https://ge.globo.com/video/ta-na-area-como-foi-assistir-ao-jogo-do-palmeiras-que-a-globo-nao-passou-10287094.ghtml',
'info_dict': {
'id': 'ta-na-area-como-foi-assistir-ao-jogo-do-palmeiras-que-a-globo-nao-passou-10287094',
'title': 'Tá na Área: como foi assistir ao jogo do Palmeiras que a Globo não passou',
'description': 'md5:2d089d036c4c9675117d3a56f8c61739',
},
'playlist_count': 1,
}] }]
@classmethod @classmethod
@ -228,6 +237,6 @@ class GloboArticleIE(InfoExtractor):
entries = [ entries = [
self.url_result('globo:%s' % video_id, GloboIE.ie_key()) self.url_result('globo:%s' % video_id, GloboIE.ie_key())
for video_id in orderedSet(video_ids)] for video_id in orderedSet(video_ids)]
title = self._og_search_title(webpage, fatal=False) title = self._og_search_title(webpage)
description = self._html_search_meta('description', webpage) description = self._html_search_meta('description', webpage)
return self.playlist_result(entries, display_id, title, description) return self.playlist_result(entries, display_id, title, description)