_ Git - youtube-dl/blob - youtube_dl/extractor/generic.py

   1 import os
   2 import re
   3
   4 from .common import InfoExtractor
   5 from ..utils import (
   6     compat_urllib_error,
   7     compat_urllib_parse,
   8     compat_urllib_request,
   9
  10     ExtractorError,
  11 )
  12
  13 class GenericIE(InfoExtractor):
  14     IE_DESC = u'Generic downloader that works on some sites'
  15     _VALID_URL = r'.*'
  16     IE_NAME = u'generic'
  17     _TEST = {
  18         u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  19         u'file': u'13601338388002.mp4',
  20         u'md5': u'85b90ccc9d73b4acd9138d3af4c27f89',
  21         u'info_dict': {
  22             u"uploader": u"www.hodiho.fr",
  23             u"title": u"R\u00e9gis plante sa Jeep"
  24         }
  25     }
  26
  27     def report_download_webpage(self, video_id):
  28         """Report webpage download."""
  29         if not self._downloader.params.get('test', False):
  30             self._downloader.report_warning(u'Falling back on generic information extractor.')
  31         super(GenericIE, self).report_download_webpage(video_id)
  32
  33     def report_following_redirect(self, new_url):
  34         """Report information extraction."""
  35         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
  36
  37     def _test_redirect(self, url):
  38         """Check if it is a redirect, like url shorteners, in case return the new url."""
  39         class HeadRequest(compat_urllib_request.Request):
  40             def get_method(self):
  41                 return "HEAD"
  42
  43         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
  44             """
  45             Subclass the HTTPRedirectHandler to make it use our
  46             HeadRequest also on the redirected URL
  47             """
  48             def redirect_request(self, req, fp, code, msg, headers, newurl):
  49                 if code in (301, 302, 303, 307):
  50                     newurl = newurl.replace(' ', '%20')
  51                     newheaders = dict((k,v) for k,v in req.headers.items()
  52                                       if k.lower() not in ("content-length", "content-type"))
  53                     return HeadRequest(newurl,
  54                                        headers=newheaders,
  55                                        origin_req_host=req.get_origin_req_host(),
  56                                        unverifiable=True)
  57                 else:
  58                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
  59
  60         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
  61             """
  62             Fallback to GET if HEAD is not allowed (405 HTTP error)
  63             """
  64             def http_error_405(self, req, fp, code, msg, headers):
  65                 fp.read()
  66                 fp.close()
  67
  68                 newheaders = dict((k,v) for k,v in req.headers.items()
  69                                   if k.lower() not in ("content-length", "content-type"))
  70                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
  71                                                  headers=newheaders,
  72                                                  origin_req_host=req.get_origin_req_host(),
  73                                                  unverifiable=True))
  74
  75         # Build our opener
  76         opener = compat_urllib_request.OpenerDirector()
  77         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
  78                         HTTPMethodFallback, HEADRedirectHandler,
  79                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
  80             opener.add_handler(handler())
  81
  82         response = opener.open(HeadRequest(url))
  83         if response is None:
  84             raise ExtractorError(u'Invalid URL protocol')
  85         new_url = response.geturl()
  86
  87         if url == new_url:
  88             return False
  89
  90         self.report_following_redirect(new_url)
  91         return new_url
  92
  93     def _real_extract(self, url):
  94         new_url = self._test_redirect(url)
  95         if new_url: return [self.url_result(new_url)]
  96
  97         video_id = url.split('/')[-1]
  98         try:
  99             webpage = self._download_webpage(url, video_id)
 100         except ValueError:
 101             # since this is the last-resort InfoExtractor, if
 102             # this error is thrown, it'll be thrown here
 103             raise ExtractorError(u'Invalid URL: %s' % url)
 104
 105         self.report_extraction(video_id)
 106         # Start with something easy: JW Player in SWFObject
 107         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 108         if mobj is None:
 109             # Broaden the search a little bit
 110             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 111         if mobj is None:
 112             # Broaden the search a little bit: JWPlayer JS loader
 113             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"&]*)', webpage)
 114         if mobj is None:
 115             # Try to find twitter cards info
 116             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 117         if mobj is None:
 118             # We look for Open Graph info:
 119             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 120             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 121             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 122             if m_video_type is not None:
 123                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 124         if mobj is None:
 125             raise ExtractorError(u'Invalid URL: %s' % url)
 126
 127         # It's possible that one of the regexes
 128         # matched, but returned an empty group:
 129         if mobj.group(1) is None:
 130             raise ExtractorError(u'Invalid URL: %s' % url)
 131
 132         video_url = compat_urllib_parse.unquote(mobj.group(1))
 133         video_id = os.path.basename(video_url)
 134
 135         # here's a fun little line of code for you:
 136         video_extension = os.path.splitext(video_id)[1][1:]
 137         video_id = os.path.splitext(video_id)[0]
 138
 139         # it's tempting to parse this further, but you would
 140         # have to take into account all the variations like
 141         #   Video Title - Site Name
 142         #   Site Name | Video Title
 143         #   Video Title - Tagline | Site Name
 144         # and so on and so forth; it's just not practical
 145         video_title = self._html_search_regex(r'<title>(.*)</title>',
 146             webpage, u'video title', default=u'video', flags=re.DOTALL)
 147
 148         # video uploader is domain name
 149         video_uploader = self._search_regex(r'(?:https?://)?([^/]*)/.*',
 150             url, u'video uploader')
 151
 152         return [{
 153             'id':       video_id,
 154             'url':      video_url,
 155             'uploader': video_uploader,
 156             'upload_date':  None,
 157             'title':    video_title,
 158             'ext':      video_extension,
 159         }]