diff options
Diffstat (limited to 'youtube_dl/extractor/generic.py')
| -rw-r--r-- | youtube_dl/extractor/generic.py | 95 |
1 files changed, 58 insertions, 37 deletions
diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py index 742bc2856..9b6498894 100644 --- a/youtube_dl/extractor/generic.py +++ b/youtube_dl/extractor/generic.py @@ -28,6 +28,7 @@ from .brightcove import BrightcoveIE from .ooyala import OoyalaIE from .rutv import RUTVIE from .smotri import SmotriIE +from .condenast import CondeNastIE class GenericIE(InfoExtractor): @@ -225,21 +226,6 @@ class GenericIE(InfoExtractor): 'skip_download': 'Requires rtmpdump' } }, - # smotri embed - { - 'url': 'http://rbctv.rbc.ru/archive/news/562949990879132.shtml', - 'md5': 'ec40048448e9284c9a1de77bb188108b', - 'info_dict': { - 'id': 'v27008541fad', - 'ext': 'mp4', - 'title': 'Крым и Севастополь вошли в состав России', - 'description': 'md5:fae01b61f68984c7bd2fa741e11c3175', - 'duration': 900, - 'upload_date': '20140318', - 'uploader': 'rbctv_2012_4', - 'uploader_id': 'rbctv_2012_4', - }, - }, # Condé Nast embed { 'url': 'http://www.wired.com/2014/04/honda-asimo/', @@ -394,6 +380,17 @@ class GenericIE(InfoExtractor): 'uploader': 'education-portal.com', }, }, + { + 'url': 'http://thoughtworks.wistia.com/medias/uxjb0lwrcz', + 'md5': 'baf49c2baa8a7de5f3fc145a8506dcd4', + 'info_dict': { + 'id': 'uxjb0lwrcz', + 'ext': 'mp4', + 'title': 'Conversation about Hexagonal Rails Part 1 - ThoughtWorks', + 'duration': 1715.0, + 'uploader': 'thoughtworks.wistia.com', + }, + }, ] def report_following_redirect(self, new_url): @@ -490,7 +487,8 @@ class GenericIE(InfoExtractor): 'Set --default-search "ytsearch" (or run youtube-dl "ytsearch:%s" ) to search YouTube' ) % (url, url), expected=True) else: - assert ':' in default_search + if ':' not in default_search: + default_search += ':' return self.url_result(default_search + url) url, smuggled_data = unsmuggle_url(url) @@ -623,13 +621,13 @@ class GenericIE(InfoExtractor): if mobj: player_url = unescapeHTML(mobj.group('url')) surl = smuggle_url(player_url, {'Referer': url}) - return self.url_result(surl, 'Vimeo') + return self.url_result(surl) # Look for embedded (swf embed) Vimeo player mobj = re.search( - r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage) + r'<embed[^>]+?src="((?:https?:)?//(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage) if mobj: - return self.url_result(mobj.group(1), 'Vimeo') + return self.url_result(mobj.group(1)) # Look for embedded YouTube player matches = re.findall(r'''(?x) @@ -654,19 +652,32 @@ class GenericIE(InfoExtractor): return _playlist_from_matches( matches, lambda m: unescapeHTML(m[1])) + # Look for embedded Dailymotion playlist player (#3822) + m = re.search( + r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.[a-z]{2,3}/widget/jukebox\?.+?)\1', webpage) + if m: + playlists = re.findall( + r'list\[\]=/playlist/([^/]+)/', unescapeHTML(m.group('url'))) + if playlists: + return _playlist_from_matches( + playlists, lambda p: '//dailymotion.com/playlist/%s' % p) + # Look for embedded Wistia player match = re.search( - r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage) + r'<(?:meta[^>]+?content|iframe[^>]+?src)=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage) if match: + embed_url = self._proto_relative_url( + unescapeHTML(match.group('url'))) return { '_type': 'url_transparent', - 'url': unescapeHTML(match.group('url')), + 'url': embed_url, 'ie_key': 'Wistia', 'uploader': video_uploader, 'title': video_title, 'id': video_id, } - match = re.search(r'(?:id=["\']wistia_|data-wistiaid=["\']|Wistia\.embed\(["\'])(?P<id>[^"\']+)', webpage) + + match = re.search(r'(?:id=["\']wistia_|data-wistia-?id=["\']|Wistia\.embed\(["\'])(?P<id>[^"\']+)', webpage) if match: return { '_type': 'url_transparent', @@ -852,47 +863,57 @@ class GenericIE(InfoExtractor): if mobj is not None: return self.url_result(mobj.group('url'), 'MLB') + mobj = re.search( + r'<iframe[^>]+?src=(["\'])(?P<url>%s)\1' % CondeNastIE.EMBED_URL, + webpage) + if mobj is not None: + return self.url_result(self._proto_relative_url(mobj.group('url'), scheme='http:'), 'CondeNast') + + def check_video(vurl): + vpath = compat_urlparse.urlparse(vurl).path + vext = determine_ext(vpath) + return '.' in vpath and vext not in ('swf', 'png', 'jpg', 'srt', 'sbv', 'sub', 'vtt', 'ttml') + + def filter_video(urls): + return list(filter(check_video, urls)) + # Start with something easy: JW Player in SWFObject - found = re.findall(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage) + found = filter_video(re.findall(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)) if not found: # Look for gorilla-vid style embedding - found = re.findall(r'''(?sx) + found = filter_video(re.findall(r'''(?sx) (?: jw_plugins| JWPlayerOptions| jwplayer\s*\(\s*["'][^'"]+["']\s*\)\s*\.setup ) - .*?file\s*:\s*["\'](.*?)["\']''', webpage) + .*?file\s*:\s*["\'](.*?)["\']''', webpage)) if not found: # Broaden the search a little bit - found = re.findall(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage) + found = filter_video(re.findall(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)) if not found: # Broaden the findall a little bit: JWPlayer JS loader - found = re.findall(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage) + found = filter_video(re.findall( + r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)) if not found: # Flow player - found = re.findall(r'''(?xs) + found = filter_video(re.findall(r'''(?xs) flowplayer\("[^"]+",\s* \{[^}]+?\}\s*, \s*{[^}]+? ["']?clip["']?\s*:\s*\{\s* ["']?url["']?\s*:\s*["']([^"']+)["'] - ''', webpage) + ''', webpage)) if not found: # Try to find twitter cards info - found = re.findall(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage) + found = filter_video(re.findall( + r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)) if not found: # We look for Open Graph info: # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am) m_video_type = re.findall(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage) # We only look in og:video if the MIME type is a video, don't try if it's a Flash player: if m_video_type is not None: - def check_video(vurl): - vpath = compat_urlparse.urlparse(vurl).path - vext = determine_ext(vpath) - return '.' in vpath and vext not in ('swf', 'png', 'jpg') - found = list(filter( - check_video, - re.findall(r'<meta.*?property="og:video".*?content="(.*?)"', webpage))) + found = filter_video(re.findall(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)) if not found: # HTML5 video found = re.findall(r'(?s)<video[^<]*(?:>.*?<source[^>]+)? src="([^"]+)"', webpage) |