diff options
Diffstat (limited to 'youtube_dl/extractor/generic.py')
| -rw-r--r-- | youtube_dl/extractor/generic.py | 49 |
1 files changed, 40 insertions, 9 deletions
diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py index 8065b5611..4d649fe71 100644 --- a/youtube_dl/extractor/generic.py +++ b/youtube_dl/extractor/generic.py @@ -102,6 +102,20 @@ class GenericIE(InfoExtractor): 'title': '2cc213299525360.mov', # that's what we get }, }, + # second style of embedded ooyala videos + { + 'url': 'http://www.smh.com.au/tv/business/show/financial-review-sunday/behind-the-scenes-financial-review-sunday--4350201.html', + 'info_dict': { + 'id': '13djJjYjptA1XpPx8r9kuzPyj3UZH0Uk', + 'ext': 'mp4', + 'title': 'Behind-the-scenes: Financial Review Sunday ', + 'description': 'Step inside Channel Nine studios for an exclusive tour of its upcoming financial business show.', + }, + 'params': { + # m3u8 download + 'skip_download': True, + }, + }, # google redirect { 'url': 'http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&ved=0CCUQtwIwAA&url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DcmQHVoWB5FY&ei=F-sNU-LLCaXk4QT52ICQBQ&usg=AFQjCNEw4hL29zgOohLXvpJ-Bdh2bils1Q&bvm=bv.61965928,d.bGE', @@ -171,7 +185,18 @@ class GenericIE(InfoExtractor): 'uploader': 'Ze Frank', 'description': 'md5:ddb2a40ecd6b6a147e400e535874947b', } - } + }, + # nowvideo embed hidden behind percent encoding + { + 'url': 'http://www.waoanime.tv/the-super-dimension-fortress-macross-episode-1/', + 'md5': '2baf4ddd70f697d94b1c18cf796d5107', + 'info_dict': { + 'id': '06e53103ca9aa', + 'ext': 'flv', + 'title': 'Macross Episode 001 Watch Macross Episode 001 onl', + 'description': 'No description', + }, + }, ] def report_download_webpage(self, video_id): @@ -323,6 +348,11 @@ class GenericIE(InfoExtractor): except compat_xml_parse_error: pass + # Sometimes embedded video player is hidden behind percent encoding + # (e.g. https://github.com/rg3/youtube-dl/issues/2448) + # Unescaping the whole page allows to handle those cases in a generic way + webpage = compat_urllib_parse.unquote(webpage) + # it's tempting to parse this further, but you would # have to take into account all the variations like # Video Title - Site Name @@ -424,9 +454,10 @@ class GenericIE(InfoExtractor): return self.url_result(mobj.group('url')) # Look for Ooyala videos - mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage) + mobj = (re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=(?P<ec>[^"&]+)', webpage) or + re.search(r'OO.Player.create\([\'"].*?[\'"],\s*[\'"](?P<ec>.{32})[\'"]', webpage)) if mobj is not None: - return OoyalaIE._build_url_result(mobj.group(1)) + return OoyalaIE._build_url_result(mobj.group('ec')) # Look for Aparat videos mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage) @@ -488,6 +519,12 @@ class GenericIE(InfoExtractor): if rutv_url: return self.url_result(rutv_url, 'RUTV') + # Look for embedded TED player + mobj = re.search( + r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage) + if mobj is not None: + return self.url_result(mobj.group('url'), 'TED') + # Start with something easy: JW Player in SWFObject mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage) if mobj is None: @@ -500,12 +537,6 @@ class GenericIE(InfoExtractor): # Broaden the search a little bit: JWPlayer JS loader mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage) - # Look for embedded TED player - mobj = re.search( - r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage) - if mobj is not None: - return self.url_result(mobj.group('url'), 'TED') - if mobj is None: # Try to find twitter cards info mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage) |