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