diff options
| author | Dario Guarascio <dario@pink.esync.lan> | 2014-05-15 13:59:27 +0700 |
|---|---|---|
| committer | Dario Guarascio <dario@pink.esync.lan> | 2014-05-15 13:59:27 +0700 |
| commit | ec8deefc275c2f7ddbbd9e78397f2e202c4a621f (patch) | |
| tree | c4b2f9111721d6ed34f8a7b300e0e62b264e2e71 | |
| parent | a2d5a4ee64926c82d0f7d29aebd7654044a61f89 (diff) | |
| download | youtube-dl-ec8deefc275c2f7ddbbd9e78397f2e202c4a621f.tar.xz | |
[youtube] Video categories added to metadata
| -rw-r--r-- | youtube_dl/extractor/youtube.py | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 1f3aa4322..f0a92d182 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1141,6 +1141,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split()) upload_date = unified_strdate(upload_date) + + video_categories = [] + # categories + m_cat_container = get_element_by_id("eow-category", video_webpage) + if m_cat_container: + video_categories = re.findall(r'<a[^<]+>(.*?)</a>', + m_cat_container, re.DOTALL) + # description video_description = get_element_by_id("eow-description", video_webpage) if video_description: @@ -1347,6 +1355,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor): 'title': video_title, 'thumbnail': video_thumbnail, 'description': video_description, + 'categories': video_categories, 'subtitles': video_subtitles, 'duration': video_duration, 'age_limit': 18 if age_gate else 0, |