From e2037b3f7d610beefa1bb3590a4552477035f696 Mon Sep 17 00:00:00 2001 From: Philipp Hagemeister Date: Mon, 15 Sep 2014 23:33:47 +0200 Subject: [PATCH] [einthusan] Add description and beautify --- youtube_dl/extractor/einthusan.py | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-) diff --git a/youtube_dl/extractor/einthusan.py b/youtube_dl/extractor/einthusan.py index 97c5d7b72..5dfea0d39 100644 --- a/youtube_dl/extractor/einthusan.py +++ b/youtube_dl/extractor/einthusan.py @@ -7,7 +7,7 @@ class EinthusanIE(InfoExtractor): - _VALID_URL = r'http://(?:www\.)?einthusan\.com/movies/watch.php\?(.*)?id=(?P[0-9]+).*?' + _VALID_URL = r'https?://(?:www\.)?einthusan\.com/movies/watch.php\?([^#]*?)id=(?P[0-9]+)' _TESTS = [ { 'url': 'http://www.einthusan.com/movies/watch.php?id=2447', @@ -17,6 +17,7 @@ class EinthusanIE(InfoExtractor): 'ext': 'mp4', 'title': 'Ek Villain', 'thumbnail': 're:^https?://.*\.jpg$', + 'description': 'md5:9d29fc91a7abadd4591fb862fa560d93', } }, { @@ -27,6 +28,7 @@ class EinthusanIE(InfoExtractor): 'ext': 'mp4', 'title': 'Soodhu Kavvuum', 'thumbnail': 're:^https?://.*\.jpg$', + 'description': 'md5:05d8a0c0281a4240d86d76e14f2f4d51', } }, ] @@ -36,19 +38,24 @@ def _real_extract(self, url): video_id = mobj.group('id') webpage = self._download_webpage(url, video_id) - video_title = self._html_search_regex(r'''

(.*?)

''', webpage, 'title') + video_title = self._html_search_regex( + r'

(.*?)

', webpage, 'title') video_url = self._html_search_regex( - r'''(?s)jwplayer\("mediaplayer"\)\.setup\({.*?'file': '([^']+)'.*?}\);''', webpage, 'video url') + r'''(?s)jwplayer\("mediaplayer"\)\.setup\({.*?'file': '([^']+)'.*?}\);''', + webpage, 'video url') - thumb_rel_url = self._html_search_regex( - r'''''', webpage, "thumbnail url") - thumb_abs_url = re.sub('\.\.', 'http://www.einthusan.com', thumb_rel_url) + description = self._html_search_meta('description', webpage) + thumbnail = self._html_search_regex( + r'''''', + webpage, "thumbnail url", fatal=False) + if thumbnail is not None: + thumbnail = thumbnail.replace('..', 'http://www.einthusan.com') return { 'id': video_id, - 'ext': 'mp4', 'title': video_title, 'url': video_url, - 'thumbnail': thumb_abs_url, + 'thumbnail': thumbnail, + 'description': description, }