diff options
| author | Sergey M․ <dstftw@gmail.com> | 2017-07-14 22:37:04 +0700 |
|---|---|---|
| committer | Sergey M․ <dstftw@gmail.com> | 2017-07-14 22:37:04 +0700 |
| commit | 7d02dcfaa2589453ee3cc6c88ee27f04c252f8a7 (patch) | |
| tree | 60dcf197be51091a0995d5c3ff9889b362deb99f | |
| parent | 00dbdfc1f741b919a0add36394065ce1aeccfda8 (diff) | |
[youtube] Don't capture YouTube Red ad for creator meta field (closes #13621)
| -rw-r--r-- | youtube_dl/extractor/youtube.py | 17 |
1 files changed, 16 insertions, 1 deletions
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 77cd271ef..4597ccb3a 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -673,6 +673,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, }, # video_info is None (https://github.com/rg3/youtube-dl/issues/4421) + # YouTube Red ad is not captured for creator { 'url': '__2ABJjxzNo', 'info_dict': { @@ -1649,7 +1650,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_webpage, 'license', default=None) m_music = re.search( - r'<h4[^>]+class="title"[^>]*>\s*Music\s*</h4>\s*<ul[^>]*>\s*<li>(?P<title>.+?) by (?P<creator>.+?)(?:\(.+?\))?</li', + r'''(?x) + <h4[^>]+class="title"[^>]*>\s*Music\s*</h4>\s* + <ul[^>]*>\s* + <li>(?P<title>.+?) + by (?P<creator>.+?) + (?: + \(.+?\)| + <a[^>]* + (?: + \bhref=["\']/red[^>]*>| # drop possible + >\s*Listen ad-free with YouTube Red # YouTube Red ad + ) + .*? + )?</li + ''', video_webpage) if m_music: video_alt_title = remove_quotes(unescapeHTML(m_music.group('title'))) |