diff options
author | Ismael Mejia <iemejia@gmail.com> | 2013-09-06 23:23:23 +0200 |
---|---|---|
committer | Ismael Mejia <iemejia@gmail.com> | 2013-09-06 23:24:41 +0200 |
commit | 72836fcee453386f4f16325c5b8fa4c1ba1bb442 (patch) | |
tree | 58efd36f4a56269a07774969e2ac385aacf8eae6 /youtube_dl/extractor/xhamster.py | |
parent | d6e203b3dcef8f291b57021903e629d3e30e1f0b (diff) | |
parent | a7130543fa0368175740f5fa173ef920671db866 (diff) |
Merge branch 'master' into subtitles_rework
Diffstat (limited to 'youtube_dl/extractor/xhamster.py')
-rw-r--r-- | youtube_dl/extractor/xhamster.py | 18 |
1 files changed, 10 insertions, 8 deletions
diff --git a/youtube_dl/extractor/xhamster.py b/youtube_dl/extractor/xhamster.py index 0f1feeffd..88b8b6be0 100644 --- a/youtube_dl/extractor/xhamster.py +++ b/youtube_dl/extractor/xhamster.py @@ -3,7 +3,8 @@ import re from .common import InfoExtractor from ..utils import ( compat_urllib_parse, - + unescapeHTML, + determine_ext, ExtractorError, ) @@ -36,15 +37,16 @@ class XHamsterIE(InfoExtractor): video_url = compat_urllib_parse.unquote(mobj.group('file')) else: video_url = mobj.group('server')+'/key='+mobj.group('file') - video_extension = video_url.split('.')[-1] video_title = self._html_search_regex(r'<title>(?P<title>.+?) - xHamster\.com</title>', webpage, u'title') - # Can't see the description anywhere in the UI - # video_description = self._html_search_regex(r'<span>Description: </span>(?P<description>[^<]+)', - # webpage, u'description', fatal=False) - # if video_description: video_description = unescapeHTML(video_description) + # Only a few videos have an description + mobj = re.search('<span>Description: </span>(?P<description>[^<]+)', webpage) + if mobj: + video_description = unescapeHTML(mobj.group('description')) + else: + video_description = None mobj = re.search(r'hint=\'(?P<upload_date_Y>[0-9]{4})-(?P<upload_date_m>[0-9]{2})-(?P<upload_date_d>[0-9]{2}) [0-9]{2}:[0-9]{2}:[0-9]{2} [A-Z]{3,4}\'', webpage) if mobj: @@ -62,9 +64,9 @@ class XHamsterIE(InfoExtractor): return [{ 'id': video_id, 'url': video_url, - 'ext': video_extension, + 'ext': determine_ext(video_url), 'title': video_title, - # 'description': video_description, + 'description': video_description, 'upload_date': video_upload_date, 'uploader_id': video_uploader_id, 'thumbnail': video_thumbnail |