X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fgeneric.py;h=23891325677e40165252788bba7a47e4cee4f91c;hb=0f2a2ba14b2cbf4bd8bec0ce32f8d7c27a733c68;hp=e1d6a2a013ea599711245ea5b2382f6f02871700;hpb=71791f414ca93d819b8585882ec1eeb006dbb63b;p=youtube-dl diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py index e1d6a2a01..238913256 100644 --- a/youtube_dl/extractor/generic.py +++ b/youtube_dl/extractor/generic.py @@ -1,74 +1,200 @@ # encoding: utf-8 +from __future__ import unicode_literals + import os import re from .common import InfoExtractor +from .youtube import YoutubeIE from ..utils import ( compat_urllib_error, compat_urllib_parse, compat_urllib_request, compat_urlparse, + compat_xml_parse_error, ExtractorError, + HEADRequest, + parse_xml, smuggle_url, unescapeHTML, + unified_strdate, + url_basename, ) from .brightcove import BrightcoveIE +from .ooyala import OoyalaIE +from .rutv import RUTVIE class GenericIE(InfoExtractor): - IE_DESC = u'Generic downloader that works on some sites' + IE_DESC = 'Generic downloader that works on some sites' _VALID_URL = r'.*' - IE_NAME = u'generic' + IE_NAME = 'generic' _TESTS = [ { - u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html', - u'file': u'13601338388002.mp4', - u'md5': u'6e15c93721d7ec9e9ca3fdbf07982cfd', - u'info_dict': { - u"uploader": u"www.hodiho.fr", - u"title": u"R\u00e9gis plante sa Jeep" - } - }, - # embedded vimeo video - { - u'add_ie': ['Vimeo'], - u'url': u'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references', - u'file': u'22444065.mp4', - u'md5': u'2903896e23df39722c33f015af0666e2', - u'info_dict': { - u'title': u'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011', - u"uploader_id": u"skillsmatter", - u"uploader": u"Skills Matter", + 'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html', + 'file': '13601338388002.mp4', + 'md5': '6e15c93721d7ec9e9ca3fdbf07982cfd', + 'info_dict': { + 'uploader': 'www.hodiho.fr', + 'title': 'R\u00e9gis plante sa Jeep', } }, # bandcamp page with custom domain { - u'add_ie': ['Bandcamp'], - u'url': u'http://bronyrock.com/track/the-pony-mash', - u'file': u'3235767654.mp3', - u'info_dict': { - u'title': u'The Pony Mash', - u'uploader': u'M_Pallante', + 'add_ie': ['Bandcamp'], + 'url': 'http://bronyrock.com/track/the-pony-mash', + 'file': '3235767654.mp3', + 'info_dict': { + 'title': 'The Pony Mash', + 'uploader': 'M_Pallante', }, - u'skip': u'There is a limit of 200 free downloads / month for the test song', + 'skip': 'There is a limit of 200 free downloads / month for the test song', }, # embedded brightcove video # it also tests brightcove videos that need to set the 'Referer' in the # http requests { - u'add_ie': ['Brightcove'], - u'url': u'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/', - u'info_dict': { - u'id': u'2765128793001', - u'ext': u'mp4', - u'title': u'Le cours de bourse : l’analyse technique', - u'description': u'md5:7e9ad046e968cb2d1114004aba466fd9', - u'uploader': u'BFM BUSINESS', + 'add_ie': ['Brightcove'], + 'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/', + 'info_dict': { + 'id': '2765128793001', + 'ext': 'mp4', + 'title': 'Le cours de bourse : l’analyse technique', + 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9', + 'uploader': 'BFM BUSINESS', + }, + 'params': { + 'skip_download': True, + }, + }, + { + # https://github.com/rg3/youtube-dl/issues/2253 + 'url': 'http://bcove.me/i6nfkrc3', + 'file': '3101154703001.mp4', + 'md5': '0ba9446db037002366bab3b3eb30c88c', + 'info_dict': { + 'title': 'Still no power', + 'uploader': 'thestar.com', + 'description': 'Mississauga resident David Farmer is still out of power as a result of the ice storm a month ago. To keep the house warm, Farmer cuts wood from his property for a wood burning stove downstairs.', + }, + 'add_ie': ['Brightcove'], + }, + # Direct link to a video + { + 'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4', + 'md5': '67d406c2bcb6af27fa886f31aa934bbe', + 'info_dict': { + 'id': 'trailer', + 'ext': 'mp4', + 'title': 'trailer', + 'upload_date': '20100513', + } + }, + # ooyala video + { + 'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219', + 'md5': '5644c6ca5d5782c1d0d350dad9bd840c', + 'info_dict': { + 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ', + 'ext': 'mp4', + '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', + 'info_dict': { + 'id': 'cmQHVoWB5FY', + 'ext': 'mp4', + 'upload_date': '20130224', + 'uploader_id': 'TheVerge', + 'description': 'Chris Ziegler takes a look at the Alcatel OneTouch Fire and the ZTE Open; two of the first Firefox OS handsets to be officially announced.', + 'uploader': 'The Verge', + 'title': 'First Firefox OS phones side-by-side', + }, + 'params': { + 'skip_download': False, + } + }, + # embed.ly video + { + 'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/', + 'info_dict': { + 'id': '9ODmcdjQcHQ', + 'ext': 'mp4', + 'title': 'Tested: Grinding Coffee at 2000 Frames Per Second', + 'upload_date': '20140225', + 'description': 'md5:06a40fbf30b220468f1e0957c0f558ff', + 'uploader': 'Tested', + 'uploader_id': 'testedcom', + }, + # No need to test YoutubeIE here + 'params': { + 'skip_download': True, + }, + }, + # funnyordie embed + { + 'url': 'http://www.theguardian.com/world/2014/mar/11/obama-zach-galifianakis-between-two-ferns', + 'md5': '7cf780be104d40fea7bae52eed4a470e', + 'info_dict': { + 'id': '18e820ec3f', + 'ext': 'mp4', + 'title': 'Between Two Ferns with Zach Galifianakis: President Barack Obama', + 'description': 'Episode 18: President Barack Obama sits down with Zach Galifianakis for his most memorable interview yet.', }, - u'params': { - u'skip_download': True, + }, + # RUTV embed + { + 'url': 'http://www.rg.ru/2014/03/15/reg-dfo/anklav-anons.html', + 'info_dict': { + 'id': '776940', + 'ext': 'mp4', + 'title': 'Охотское море стало целиком российским', + 'description': 'md5:5ed62483b14663e2a95ebbe115eb8f43', + }, + 'params': { + # m3u8 download + 'skip_download': True, + }, + }, + # Embedded TED video + { + 'url': 'http://en.support.wordpress.com/videos/ted-talks/', + 'md5': 'deeeabcc1085eb2ba205474e7235a3d5', + 'info_dict': { + 'id': '981', + 'ext': 'mp4', + 'title': 'My web playroom', + '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', }, }, ] @@ -76,32 +202,34 @@ class GenericIE(InfoExtractor): def report_download_webpage(self, video_id): """Report webpage download.""" if not self._downloader.params.get('test', False): - self._downloader.report_warning(u'Falling back on generic information extractor.') + self._downloader.report_warning('Falling back on generic information extractor.') super(GenericIE, self).report_download_webpage(video_id) def report_following_redirect(self, new_url): """Report information extraction.""" - self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url) + self._downloader.to_screen('[redirect] Following redirect to %s' % new_url) - def _test_redirect(self, url): + def _send_head(self, url): """Check if it is a redirect, like url shorteners, in case return the new url.""" - class HeadRequest(compat_urllib_request.Request): - def get_method(self): - return "HEAD" class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler): """ Subclass the HTTPRedirectHandler to make it use our - HeadRequest also on the redirected URL + HEADRequest also on the redirected URL """ def redirect_request(self, req, fp, code, msg, headers, newurl): if code in (301, 302, 303, 307): newurl = newurl.replace(' ', '%20') newheaders = dict((k,v) for k,v in req.headers.items() if k.lower() not in ("content-length", "content-type")) - return HeadRequest(newurl, + try: + # This function was deprecated in python 3.3 and removed in 3.4 + origin_req_host = req.get_origin_req_host() + except AttributeError: + origin_req_host = req.origin_req_host + return HEADRequest(newurl, headers=newheaders, - origin_req_host=req.get_origin_req_host(), + origin_req_host=origin_req_host, unverifiable=True) else: raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp) @@ -128,87 +256,287 @@ class GenericIE(InfoExtractor): compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]: opener.add_handler(handler()) - response = opener.open(HeadRequest(url)) + response = opener.open(HEADRequest(url)) if response is None: - raise ExtractorError(u'Invalid URL protocol') - new_url = response.geturl() + raise ExtractorError('Invalid URL protocol') + return response - if url == new_url: - return False + def _extract_rss(self, url, video_id, doc): + playlist_title = doc.find('./channel/title').text + playlist_desc_el = doc.find('./channel/description') + playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text - self.report_following_redirect(new_url) - return new_url + entries = [{ + '_type': 'url', + 'url': e.find('link').text, + 'title': e.find('title').text, + } for e in doc.findall('./channel/item')] + + return { + '_type': 'playlist', + 'id': url, + 'title': playlist_title, + 'description': playlist_desc, + 'entries': entries, + } def _real_extract(self, url): parsed_url = compat_urlparse.urlparse(url) if not parsed_url.scheme: - self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http') - return self.url_result('http://' + url) + default_search = self._downloader.params.get('default_search') + if default_search is None: + default_search = 'auto' + + if default_search == 'auto': + if '/' in url: + self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http') + return self.url_result('http://' + url) + else: + return self.url_result('ytsearch:' + url) + else: + assert ':' in default_search + return self.url_result(default_search + url) + video_id = os.path.splitext(url.rstrip('/').split('/')[-1])[0] + + self.to_screen('%s: Requesting header' % video_id) try: - new_url = self._test_redirect(url) - if new_url: - return [self.url_result(new_url)] + response = self._send_head(url) + + # Check for redirect + new_url = response.geturl() + if url != new_url: + self.report_following_redirect(new_url) + return self.url_result(new_url) + + # Check for direct link to a video + content_type = response.headers.get('Content-Type', '') + m = re.match(r'^(?Paudio|video|application(?=/ogg$))/(?P.+)$', content_type) + if m: + upload_date = response.headers.get('Last-Modified') + if upload_date: + upload_date = unified_strdate(upload_date) + return { + 'id': video_id, + 'title': os.path.splitext(url_basename(url))[0], + 'formats': [{ + 'format_id': m.group('format_id'), + 'url': url, + 'vcodec': 'none' if m.group('type') == 'audio' else None + }], + 'upload_date': upload_date, + } + except compat_urllib_error.HTTPError: # This may be a stupid server that doesn't like HEAD, our UA, or so pass - video_id = url.split('/')[-1] try: webpage = self._download_webpage(url, video_id) except ValueError: # since this is the last-resort InfoExtractor, if # this error is thrown, it'll be thrown here - raise ExtractorError(u'Failed to download URL: %s' % url) + raise ExtractorError('Failed to download URL: %s' % url) self.report_extraction(video_id) + # Is it an RSS feed? + try: + doc = parse_xml(webpage) + if doc.tag == 'rss': + return self._extract_rss(url, video_id, doc) + 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 # Site Name | Video Title # Video Title - Tagline | Site Name # and so on and so forth; it's just not practical - video_title = self._html_search_regex(r'(.*)', - webpage, u'video title', default=u'video', flags=re.DOTALL) + video_title = self._html_search_regex( + r'(?s)(.*?)', webpage, 'video title', + default='video') + + # video uploader is domain name + video_uploader = self._search_regex( + r'^(?:https?://)?([^/]*)/.*', url, 'video uploader') # Look for BrightCove: - bc_url = BrightcoveIE._extract_brightcove_url(webpage) - if bc_url is not None: - self.to_screen(u'Brightcove video detected.') - return self.url_result(bc_url, 'Brightcove') + bc_urls = BrightcoveIE._extract_brightcove_urls(webpage) + if bc_urls: + self.to_screen('Brightcove video detected.') + entries = [{ + '_type': 'url', + 'url': smuggle_url(bc_url, {'Referer': url}), + 'ie_key': 'Brightcove' + } for bc_url in bc_urls] - # Look for embedded Vimeo player + return { + '_type': 'playlist', + 'title': video_title, + 'id': video_id, + 'entries': entries, + } + + # Look for embedded (iframe) Vimeo player mobj = re.search( - r']+?src="(https?://player.vimeo.com/video/.+?)"', webpage) + r']+?src=(["\'])(?P(?:https?:)?//player\.vimeo\.com/video/.+?)\1', webpage) if mobj: - player_url = unescapeHTML(mobj.group(1)) + player_url = unescapeHTML(mobj.group('url')) surl = smuggle_url(player_url, {'Referer': url}) return self.url_result(surl, 'Vimeo') + # Look for embedded (swf embed) Vimeo player + mobj = re.search( + r']+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage) + if mobj: + return self.url_result(mobj.group(1), 'Vimeo') + # Look for embedded YouTube player - matches = re.findall( - r']+?src=(["\'])(?P(?:https?:)?//(?:www\.)?youtube.com/embed/.+?)\1', webpage) + matches = re.findall(r'''(?x) + (?:]+?src=|embedSWF\(\s*) + (["\'])(?P(?:https?:)?//(?:www\.)?youtube\.com/ + (?:embed|v)/.+?) + \1''', webpage) if matches: urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube') for tuppl in matches] return self.playlist_result( urlrs, playlist_id=video_id, playlist_title=video_title) + # Look for embedded Dailymotion player + matches = re.findall( + r']+?src=(["\'])(?P(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage) + if matches: + urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion') + for tuppl in matches] + return self.playlist_result( + urlrs, playlist_id=video_id, playlist_title=video_title) + + # Look for embedded Wistia player + match = re.search( + r']+?src=(["\'])(?P(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage) + if match: + return { + '_type': 'url_transparent', + 'url': unescapeHTML(match.group('url')), + 'ie_key': 'Wistia', + 'uploader': video_uploader, + 'title': video_title, + 'id': video_id, + } + + # Look for embedded blip.tv player + mobj = re.search(r']*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage) + if mobj: + return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV') + mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage) + if mobj: + return self.url_result(mobj.group(1), 'BlipTV') + # Look for Bandcamp pages with custom domain mobj = re.search(r']*?content="(.*?bandcamp\.com.*?)"', webpage) if mobj is not None: burl = unescapeHTML(mobj.group(1)) - return self.url_result(burl, 'Bandcamp') + # Don't set the extractor because it can be a track url or an album + return self.url_result(burl) + + # Look for embedded Vevo player + mobj = re.search( + r']+?src=(["\'])(?P(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage) + if mobj is not None: + return self.url_result(mobj.group('url')) + + # Look for Ooyala videos + mobj = (re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=(?P[^"&]+)', webpage) or + re.search(r'OO.Player.create\([\'"].*?[\'"],\s*[\'"](?P.{32})[\'"]', webpage)) + if mobj is not None: + return OoyalaIE._build_url_result(mobj.group('ec')) + + # Look for Aparat videos + mobj = re.search(r'