From 3aa3953d28dae68b87aa83682043b5eec0973ddc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Kacper=20Michaj=C5=82ow?= Date: Sun, 25 Sep 2016 20:26:58 +0200 Subject: [PATCH] [vk] Fix date and view count extraction. --- youtube_dl/extractor/vk.py | 17 +++++------------ youtube_dl/utils.py | 2 ++ 2 files changed, 7 insertions(+), 12 deletions(-) diff --git a/youtube_dl/extractor/vk.py b/youtube_dl/extractor/vk.py index 9f7a593ef..3cfbd97af 100644 --- a/youtube_dl/extractor/vk.py +++ b/youtube_dl/extractor/vk.py @@ -373,19 +373,12 @@ class VKIE(VKBaseIE): title = self._live_title(title) # Extract upload date - upload_date = None - mobj = re.search(r'id="mv_date(?:_views)?_wrap"[^>]*>([a-zA-Z]+ [0-9]+), ([0-9]+) at', info_page) - if mobj is not None: - mobj.group(1) + ' ' + mobj.group(2) - upload_date = unified_strdate(mobj.group(1) + ' ' + mobj.group(2)) + upload_date = unified_strdate(self._html_search_regex( + r'class="mv_info_date[^>]*>([^<]*)<', info_page, 'upload date', default=None)) - view_count = None - views = self._html_search_regex( - r'"mv_views_count_number"[^>]*>(.+?\bviews?)<', - info_page, 'view count', default=None) - if views: - view_count = str_to_int(self._search_regex( - r'([\d,.]+)', views, 'view count', fatal=False)) + view_count = str_to_int(self._html_search_regex( + r'class="mv_views_count[^>]*>([\d,.]+)', + info_page, 'view count', default=None)) formats = [] for format_id, format_url in data.items(): diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index fcbfa0d76..243d09034 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -124,6 +124,8 @@ DATE_FORMATS = ( '%d %b %Y', '%B %d %Y', '%b %d %Y', + '%b %d %Y at %H:%M', + '%b %d %Y at %H:%M:%S', '%b %dst %Y %I:%M', '%b %dnd %Y %I:%M', '%b %dth %Y %I:%M',