From 7129bed51bc2452524c69874023ee4a0b351a046 Mon Sep 17 00:00:00 2001 From: Philipp Hagemeister Date: Sat, 13 Dec 2014 12:35:13 +0100 Subject: [PATCH] [keek] Modernize and extract uploader --- youtube_dl/extractor/keek.py | 28 ++++++++++++++++++---------- 1 file changed, 18 insertions(+), 10 deletions(-) diff --git a/youtube_dl/extractor/keek.py b/youtube_dl/extractor/keek.py index 5d679e88d..e94e2f8ad 100644 --- a/youtube_dl/extractor/keek.py +++ b/youtube_dl/extractor/keek.py @@ -6,29 +6,36 @@ class KeekIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?keek\.com/(?:!|\w+/keeks/)(?P\w+)' + _VALID_URL = r'https?://(?:www\.)?keek\.com/(?:!|\w+/keeks/)(?P\w+)' IE_NAME = 'keek' _TEST = { 'url': 'https://www.keek.com/ytdl/keeks/NODfbab', - 'file': 'NODfbab.mp4', - 'md5': '9b0636f8c0f7614afa4ea5e4c6e57e83', + 'md5': '09c5c109067536c1cec8bac8c21fea05', 'info_dict': { - 'uploader': 'ytdl', + 'id': 'NODfbab', + 'ext': 'mp4', + 'uploader': 'youtube-dl project', + 'uploader_id': 'ytdl', 'title': 'test chars: "\'/\\\u00e4<>This is a test video for youtube-dl.For more information, contact phihag@phihag.de .', }, } def _real_extract(self, url): - m = re.match(self._VALID_URL, url) - video_id = m.group('videoID') + video_id = self._match_id(url) video_url = 'http://cdn.keek.com/keek/video/%s' % video_id thumbnail = 'http://cdn.keek.com/keek/thumbnail/%s/w100/h75' % video_id webpage = self._download_webpage(url, video_id) - uploader = self._html_search_regex( - r'
[\S\s]+?

(?P.+?)

', - webpage, 'uploader', fatal=False) + raw_desc = self._html_search_meta('description', webpage) + if raw_desc: + uploader = self._html_search_regex( + r'Watch (.*?)\s+\(', raw_desc, 'uploader', fatal=False) + uploader_id = self._html_search_regex( + r'Watch .*?\(@(.+?)\)', raw_desc, 'uploader_id', fatal=False) + else: + uploader = None + uploader_id = None return { 'id': video_id, @@ -36,5 +43,6 @@ def _real_extract(self, url): 'ext': 'mp4', 'title': self._og_search_title(webpage), 'thumbnail': thumbnail, - 'uploader': uploader + 'uploader': uploader, + 'uploader_id': uploader_id, }