_ Git - youtube-dl/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'6e15c93721d7ec9e9ca3fdbf07982cfd',
  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'add_ie': ['Vimeo'],
  37             u'url': u'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
  38             u'file': u'22444065.mp4',
  39             u'md5': u'2903896e23df39722c33f015af0666e2',
  40             u'info_dict': {
  41                 u'title': u'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
  42                 u"uploader_id": u"skillsmatter",
  43                 u"uploader": u"Skills Matter",
  44             }
  45         },
  46         # bandcamp page with custom domain
  47         {
  48             u'add_ie': ['Bandcamp'],
  49             u'url': u'http://bronyrock.com/track/the-pony-mash',
  50             u'file': u'3235767654.mp3',
  51             u'info_dict': {
  52                 u'title': u'The Pony Mash',
  53                 u'uploader': u'M_Pallante',
  54             },
  55             u'skip': u'There is a limit of 200 free downloads / month for the test song',
  56         },
  57         # embedded brightcove video
  58         {
  59             u'add_ie': ['Brightcove'],
  60             u'url': u'http://www.scientificamerican.com/article.cfm?id=soap-bubble-physics',
  61             u'info_dict': {
  62                 u'id': u'2365799484001',
  63                 u'ext': u'mp4',
  64                 u'title': u'Bubble Simulation',
  65                 u'description': u'A visualization from a new computer model of foam behavior.',
  66                 u'uploader': u'Scientific American',
  67             },
  68             u'params': {
  69                 u'skip_download': True,
  70             },
  71         },
  72     ]
  73
  74     def report_download_webpage(self, video_id):
  75         """Report webpage download."""
  76         if not self._downloader.params.get('test', False):
  77             self._downloader.report_warning(u'Falling back on generic information extractor.')
  78         super(GenericIE, self).report_download_webpage(video_id)
  79
  80     def report_following_redirect(self, new_url):
  81         """Report information extraction."""
  82         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
  83
  84     def _test_redirect(self, url):
  85         """Check if it is a redirect, like url shorteners, in case return the new url."""
  86         class HeadRequest(compat_urllib_request.Request):
  87             def get_method(self):
  88                 return "HEAD"
  89
  90         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
  91             """
  92             Subclass the HTTPRedirectHandler to make it use our
  93             HeadRequest also on the redirected URL
  94             """
  95             def redirect_request(self, req, fp, code, msg, headers, newurl):
  96                 if code in (301, 302, 303, 307):
  97                     newurl = newurl.replace(' ', '%20')
  98                     newheaders = dict((k,v) for k,v in req.headers.items()
  99                                       if k.lower() not in ("content-length", "content-type"))
 100                     return HeadRequest(newurl,
 101                                        headers=newheaders,
 102                                        origin_req_host=req.get_origin_req_host(),
 103                                        unverifiable=True)
 104                 else:
 105                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 106
 107         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 108             """
 109             Fallback to GET if HEAD is not allowed (405 HTTP error)
 110             """
 111             def http_error_405(self, req, fp, code, msg, headers):
 112                 fp.read()
 113                 fp.close()
 114
 115                 newheaders = dict((k,v) for k,v in req.headers.items()
 116                                   if k.lower() not in ("content-length", "content-type"))
 117                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 118                                                  headers=newheaders,
 119                                                  origin_req_host=req.get_origin_req_host(),
 120                                                  unverifiable=True))
 121
 122         # Build our opener
 123         opener = compat_urllib_request.OpenerDirector()
 124         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 125                         HTTPMethodFallback, HEADRedirectHandler,
 126                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 127             opener.add_handler(handler())
 128
 129         response = opener.open(HeadRequest(url))
 130         if response is None:
 131             raise ExtractorError(u'Invalid URL protocol')
 132         new_url = response.geturl()
 133
 134         if url == new_url:
 135             return False
 136
 137         self.report_following_redirect(new_url)
 138         return new_url
 139
 140     def _real_extract(self, url):
 141         parsed_url = compat_urlparse.urlparse(url)
 142         if not parsed_url.scheme:
 143             self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 144             return self.url_result('http://' + url)
 145
 146         try:
 147             new_url = self._test_redirect(url)
 148             if new_url:
 149                 return [self.url_result(new_url)]
 150         except compat_urllib_error.HTTPError:
 151             # This may be a stupid server that doesn't like HEAD, our UA, or so
 152             pass
 153
 154         video_id = url.split('/')[-1]
 155         try:
 156             webpage = self._download_webpage(url, video_id)
 157         except ValueError:
 158             # since this is the last-resort InfoExtractor, if
 159             # this error is thrown, it'll be thrown here
 160             raise ExtractorError(u'Failed to download URL: %s' % url)
 161
 162         self.report_extraction(video_id)
 163         # Look for BrightCove:
 164         bc_url = BrightcoveIE._extract_brightcove_url(webpage)
 165         if bc_url is not None:
 166             self.to_screen(u'Brightcove video detected.')
 167             return self.url_result(bc_url, 'Brightcove')
 168
 169         # Look for embedded Vimeo player
 170         mobj = re.search(
 171             r'<iframe[^>]+?src="(https?://player.vimeo.com/video/.+?)"', webpage)
 172         if mobj:
 173             player_url = unescapeHTML(mobj.group(1))
 174             surl = smuggle_url(player_url, {'Referer': url})
 175             return self.url_result(surl, 'Vimeo')
 176
 177         # Look for embedded YouTube player
 178         mobj = re.search(
 179             r'<iframe[^>]+?src=(["\'])(?P<url>https?://(?:www\.)?youtube.com/embed/.+?)\1', webpage)
 180         if mobj:
 181             surl = unescapeHTML(mobj.group(u'url'))
 182             return self.url_result(surl, 'Youtube')
 183
 184         # Look for Bandcamp pages with custom domain
 185         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 186         if mobj is not None:
 187             burl = unescapeHTML(mobj.group(1))
 188             return self.url_result(burl, 'Bandcamp')
 189
 190         # Start with something easy: JW Player in SWFObject
 191         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 192         if mobj is None:
 193             # Broaden the search a little bit
 194             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 195         if mobj is None:
 196             # Broaden the search a little bit: JWPlayer JS loader
 197             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"&]*)', webpage)
 198         if mobj is None:
 199             # Try to find twitter cards info
 200             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 201         if mobj is None:
 202             # We look for Open Graph info:
 203             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 204             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 205             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 206             if m_video_type is not None:
 207                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 208         if mobj is None:
 209             # HTML5 video
 210             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
 211         if mobj is None:
 212             raise ExtractorError(u'Unsupported URL: %s' % url)
 213
 214         # It's possible that one of the regexes
 215         # matched, but returned an empty group:
 216         if mobj.group(1) is None:
 217             raise ExtractorError(u'Did not find a valid video URL at %s' % url)
 218
 219         video_url = mobj.group(1)
 220         video_url = compat_urlparse.urljoin(url, video_url)
 221         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 222
 223         # here's a fun little line of code for you:
 224         video_extension = os.path.splitext(video_id)[1][1:]
 225         video_id = os.path.splitext(video_id)[0]
 226
 227         # it's tempting to parse this further, but you would
 228         # have to take into account all the variations like
 229         #   Video Title - Site Name
 230         #   Site Name | Video Title
 231         #   Video Title - Tagline | Site Name
 232         # and so on and so forth; it's just not practical
 233         video_title = self._html_search_regex(r'<title>(.*)</title>',
 234             webpage, u'video title', default=u'video', flags=re.DOTALL)
 235
 236         # video uploader is domain name
 237         video_uploader = self._search_regex(r'(?:https?://)?([^/]*)/.*',
 238             url, u'video uploader')
 239
 240         return [{
 241             'id':       video_id,
 242             'url':      video_url,
 243             'uploader': video_uploader,
 244             'upload_date':  None,
 245             'title':    video_title,
 246             'ext':      video_extension,
 247         }]