Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/generic.py

   1 # encoding: utf-8
   2
   3 import os
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     compat_urllib_error,
   9     compat_urllib_parse,
  10     compat_urllib_request,
  11     compat_urlparse,
  12
  13     ExtractorError,
  14     smuggle_url,
  15     unescapeHTML,
  16 )
  17 from .brightcove import BrightcoveIE
  18
  19
  20 class GenericIE(InfoExtractor):
  21     IE_DESC = u'Generic downloader that works on some sites'
  22     _VALID_URL = r'.*'
  23     IE_NAME = u'generic'
  24     _TESTS = [
  25         {
  26             u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  27             u'file': u'13601338388002.mp4',
  28             u'md5': u'85b90ccc9d73b4acd9138d3af4c27f89',
  29             u'info_dict': {
  30                 u"uploader": u"www.hodiho.fr",
  31                 u"title": u"R\u00e9gis plante sa Jeep"
  32             }
  33         },
  34         # embedded vimeo video
  35         {
  36             u'url': u'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
  37             u'file': u'22444065.mp4',
  38             u'md5': u'2903896e23df39722c33f015af0666e2',
  39             u'info_dict': {
  40                 u'title': u'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
  41                 u"uploader_id": u"skillsmatter",
  42                 u"uploader": u"Skills Matter",
  43             }
  44         }
  45     ]
  46
  47     def report_download_webpage(self, video_id):
  48         """Report webpage download."""
  49         if not self._downloader.params.get('test', False):
  50             self._downloader.report_warning(u'Falling back on generic information extractor.')
  51         super(GenericIE, self).report_download_webpage(video_id)
  52
  53     def report_following_redirect(self, new_url):
  54         """Report information extraction."""
  55         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
  56
  57     def _test_redirect(self, url):
  58         """Check if it is a redirect, like url shorteners, in case return the new url."""
  59         class HeadRequest(compat_urllib_request.Request):
  60             def get_method(self):
  61                 return "HEAD"
  62
  63         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
  64             """
  65             Subclass the HTTPRedirectHandler to make it use our
  66             HeadRequest also on the redirected URL
  67             """
  68             def redirect_request(self, req, fp, code, msg, headers, newurl):
  69                 if code in (301, 302, 303, 307):
  70                     newurl = newurl.replace(' ', '%20')
  71                     newheaders = dict((k,v) for k,v in req.headers.items()
  72                                       if k.lower() not in ("content-length", "content-type"))
  73                     return HeadRequest(newurl,
  74                                        headers=newheaders,
  75                                        origin_req_host=req.get_origin_req_host(),
  76                                        unverifiable=True)
  77                 else:
  78                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
  79
  80         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
  81             """
  82             Fallback to GET if HEAD is not allowed (405 HTTP error)
  83             """
  84             def http_error_405(self, req, fp, code, msg, headers):
  85                 fp.read()
  86                 fp.close()
  87
  88                 newheaders = dict((k,v) for k,v in req.headers.items()
  89                                   if k.lower() not in ("content-length", "content-type"))
  90                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
  91                                                  headers=newheaders,
  92                                                  origin_req_host=req.get_origin_req_host(),
  93                                                  unverifiable=True))
  94
  95         # Build our opener
  96         opener = compat_urllib_request.OpenerDirector()
  97         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
  98                         HTTPMethodFallback, HEADRedirectHandler,
  99                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 100             opener.add_handler(handler())
 101
 102         response = opener.open(HeadRequest(url))
 103         if response is None:
 104             raise ExtractorError(u'Invalid URL protocol')
 105         new_url = response.geturl()
 106
 107         if url == new_url:
 108             return False
 109
 110         self.report_following_redirect(new_url)
 111         return new_url
 112
 113     def _real_extract(self, url):
 114         parsed_url = compat_urlparse.urlparse(url)
 115         if not parsed_url.scheme:
 116             self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 117             return self.url_result('http://' + url)
 118
 119         try:
 120             new_url = self._test_redirect(url)
 121             if new_url:
 122                 return [self.url_result(new_url)]
 123         except compat_urllib_error.HTTPError:
 124             # This may be a stupid server that doesn't like HEAD, our UA, or so
 125             pass
 126
 127         video_id = url.split('/')[-1]
 128         try:
 129             webpage = self._download_webpage(url, video_id)
 130         except ValueError:
 131             # since this is the last-resort InfoExtractor, if
 132             # this error is thrown, it'll be thrown here
 133             raise ExtractorError(u'Failed to download URL: %s' % url)
 134
 135         self.report_extraction(video_id)
 136         # Look for BrightCove:
 137         m_brightcove = re.search(r'<object[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1.+?</object>', webpage, re.DOTALL)
 138         if m_brightcove is not None:
 139             self.to_screen(u'Brightcove video detected.')
 140             bc_url = BrightcoveIE._build_brighcove_url(m_brightcove.group())
 141             return self.url_result(bc_url, 'Brightcove')
 142
 143         # Look for embedded Vimeo player
 144         mobj = re.search(
 145             r'<iframe[^>]+?src="(https?://player.vimeo.com/video/.+?)"', webpage)
 146         if mobj:
 147             player_url = unescapeHTML(mobj.group(1))
 148             surl = smuggle_url(player_url, {'Referer': url})
 149             return self.url_result(surl, 'Vimeo')
 150
 151         # Look for embedded YouTube player
 152         mobj = re.search(
 153             r'<iframe[^>]+?src="(https?://(?:www\.)?youtube.com/embed/.+?)"', webpage)
 154         if mobj:
 155             surl = unescapeHTML(mobj.group(1))
 156             return self.url_result(surl, 'Youtube')
 157
 158         # Start with something easy: JW Player in SWFObject
 159         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 160         if mobj is None:
 161             # Broaden the search a little bit
 162             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 163         if mobj is None:
 164             # Broaden the search a little bit: JWPlayer JS loader
 165             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"&]*)', webpage)
 166         if mobj is None:
 167             # Try to find twitter cards info
 168             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 169         if mobj is None:
 170             # We look for Open Graph info:
 171             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 172             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 173             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 174             if m_video_type is not None:
 175                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 176         if mobj is None:
 177             # HTML5 video
 178             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
 179         if mobj is None:
 180             raise ExtractorError(u'Unsupported URL: %s' % url)
 181
 182         # It's possible that one of the regexes
 183         # matched, but returned an empty group:
 184         if mobj.group(1) is None:
 185             raise ExtractorError(u'Did not find a valid video URL at %s' % url)
 186
 187         video_url = mobj.group(1)
 188         video_url = compat_urlparse.urljoin(url, video_url)
 189         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 190
 191         # here's a fun little line of code for you:
 192         video_extension = os.path.splitext(video_id)[1][1:]
 193         video_id = os.path.splitext(video_id)[0]
 194
 195         # it's tempting to parse this further, but you would
 196         # have to take into account all the variations like
 197         #   Video Title - Site Name
 198         #   Site Name | Video Title
 199         #   Video Title - Tagline | Site Name
 200         # and so on and so forth; it's just not practical
 201         video_title = self._html_search_regex(r'<title>(.*)</title>',
 202             webpage, u'video title', default=u'video', flags=re.DOTALL)
 203
 204         # video uploader is domain name
 205         video_uploader = self._search_regex(r'(?:https?://)?([^/]*)/.*',
 206             url, u'video uploader')
 207
 208         return [{
 209             'id':       video_id,
 210             'url':      video_url,
 211             'uploader': video_uploader,
 212             'upload_date':  None,
 213             'title':    video_title,
 214             'ext':      video_extension,
 215         }]