From 0831d95c46e0a198957d44262bb251113346a6b4 Mon Sep 17 00:00:00 2001 From: coletdjnz Date: Sat, 10 Sep 2022 10:06:48 +1200 Subject: [PATCH] [extractor/BiliIntl] Support uppercase lang in `_VALID_URL` Seen in some rare cases Authored by: coletdjnz --- yt_dlp/extractor/bilibili.py | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/yt_dlp/extractor/bilibili.py b/yt_dlp/extractor/bilibili.py index 7e63dad0f..2c29bf3ce 100644 --- a/yt_dlp/extractor/bilibili.py +++ b/yt_dlp/extractor/bilibili.py @@ -909,7 +909,7 @@ class BiliIntlBaseIE(InfoExtractor): class BiliIntlIE(BiliIntlBaseIE): - _VALID_URL = r'https?://(?:www\.)?bili(?:bili\.tv|intl\.com)/(?:[a-z]{2}/)?(play/(?P\d+)/(?P\d+)|video/(?P\d+))' + _VALID_URL = r'https?://(?:www\.)?bili(?:bili\.tv|intl\.com)/(?:[a-zA-Z]{2}/)?(play/(?P\d+)/(?P\d+)|video/(?P\d+))' _TESTS = [{ # Bstation page 'url': 'https://www.bilibili.tv/en/play/34613/341736', @@ -952,6 +952,10 @@ class BiliIntlIE(BiliIntlBaseIE): # No language in URL 'url': 'https://www.bilibili.tv/video/2019955076', 'only_matching': True, + }, { + # Uppercase language in URL + 'url': 'https://www.bilibili.tv/EN/video/2019955076', + 'only_matching': True, }] def _real_extract(self, url): @@ -975,7 +979,7 @@ class BiliIntlIE(BiliIntlBaseIE): class BiliIntlSeriesIE(BiliIntlBaseIE): - _VALID_URL = r'https?://(?:www\.)?bili(?:bili\.tv|intl\.com)/(?:[a-z]{2}/)?play/(?P\d+)/?(?:[?#]|$)' + _VALID_URL = r'https?://(?:www\.)?bili(?:bili\.tv|intl\.com)/(?:[a-zA-Z]{2}/)?play/(?P\d+)/?(?:[?#]|$)' _TESTS = [{ 'url': 'https://www.bilibili.tv/en/play/34613', 'playlist_mincount': 15, @@ -993,6 +997,9 @@ class BiliIntlSeriesIE(BiliIntlBaseIE): }, { 'url': 'https://www.biliintl.com/en/play/34613', 'only_matching': True, + }, { + 'url': 'https://www.biliintl.com/EN/play/34613', + 'only_matching': True, }] def _entries(self, series_id):