_ 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     """Generic last-resort information extractor."""
  15
  16     _VALID_URL = r'.*'
  17     IE_NAME = u'generic'
  18     _TEST = {
  19         u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  20         u'file': u'13601338388002.mp4',
  21         u'md5': u'85b90ccc9d73b4acd9138d3af4c27f89',
  22         u'info_dict': {
  23             u"uploader": u"www.hodiho.fr",
  24             u"title": u"R\u00e9gis plante sa Jeep"
  25         }
  26     }
  27
  28     def report_download_webpage(self, video_id):
  29         """Report webpage download."""
  30         if not self._downloader.params.get('test', False):
  31             self._downloader.report_warning(u'Falling back on generic information extractor.')
  32         super(GenericIE, self).report_download_webpage(video_id)
  33
  34     def report_following_redirect(self, new_url):
  35         """Report information extraction."""
  36         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
  37
  38     def _test_redirect(self, url):
  39         """Check if it is a redirect, like url shorteners, in case return the new url."""
  40         class HeadRequest(compat_urllib_request.Request):
  41             def get_method(self):
  42                 return "HEAD"
  43
  44         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
  45             """
  46             Subclass the HTTPRedirectHandler to make it use our
  47             HeadRequest also on the redirected URL
  48             """
  49             def redirect_request(self, req, fp, code, msg, headers, newurl):
  50                 if code in (301, 302, 303, 307):
  51                     newurl = newurl.replace(' ', '%20')
  52                     newheaders = dict((k,v) for k,v in req.headers.items()
  53                                       if k.lower() not in ("content-length", "content-type"))
  54                     return HeadRequest(newurl,
  55                                        headers=newheaders,
  56                                        origin_req_host=req.get_origin_req_host(),
  57                                        unverifiable=True)
  58                 else:
  59                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
  60
  61         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
  62             """
  63             Fallback to GET if HEAD is not allowed (405 HTTP error)
  64             """
  65             def http_error_405(self, req, fp, code, msg, headers):
  66                 fp.read()
  67                 fp.close()
  68
  69                 newheaders = dict((k,v) for k,v in req.headers.items()
  70                                   if k.lower() not in ("content-length", "content-type"))
  71                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
  72                                                  headers=newheaders,
  73                                                  origin_req_host=req.get_origin_req_host(),
  74                                                  unverifiable=True))
  75
  76         # Build our opener
  77         opener = compat_urllib_request.OpenerDirector()
  78         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
  79                         HTTPMethodFallback, HEADRedirectHandler,
  80                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
  81             opener.add_handler(handler())
  82
  83         response = opener.open(HeadRequest(url))
  84         if response is None:
  85             raise ExtractorError(u'Invalid URL protocol')
  86         new_url = response.geturl()
  87
  88         if url == new_url:
  89             return False
  90
  91         self.report_following_redirect(new_url)
  92         return new_url
  93
  94     def _real_extract(self, url):
  95         new_url = self._test_redirect(url)
  96         if new_url: return [self.url_result(new_url)]
  97
  98         video_id = url.split('/')[-1]
  99         try:
 100             webpage = self._download_webpage(url, video_id)
 101         except ValueError:
 102             # since this is the last-resort InfoExtractor, if
 103             # this error is thrown, it'll be thrown here
 104             raise ExtractorError(u'Invalid URL: %s' % url)
 105
 106         self.report_extraction(video_id)
 107         # Start with something easy: JW Player in SWFObject
 108         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 109         if mobj is None:
 110             # Broaden the search a little bit
 111             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 112         if mobj is None:
 113             # Broaden the search a little bit: JWPlayer JS loader
 114             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"&]*)', webpage)
 115         if mobj is None:
 116             # Try to find twitter cards info
 117             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 118         if mobj is None:
 119             # We look for Open Graph info:
 120             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 121             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 122             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 123             if m_video_type is not None:
 124                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 125         if mobj is None:
 126             raise ExtractorError(u'Invalid URL: %s' % url)
 127
 128         # It's possible that one of the regexes
 129         # matched, but returned an empty group:
 130         if mobj.group(1) is None:
 131             raise ExtractorError(u'Invalid URL: %s' % url)
 132
 133         video_url = compat_urllib_parse.unquote(mobj.group(1))
 134         video_id = os.path.basename(video_url)
 135
 136         # here's a fun little line of code for you:
 137         video_extension = os.path.splitext(video_id)[1][1:]
 138         video_id = os.path.splitext(video_id)[0]
 139
 140         # it's tempting to parse this further, but you would
 141         # have to take into account all the variations like
 142         #   Video Title - Site Name
 143         #   Site Name | Video Title
 144         #   Video Title - Tagline | Site Name
 145         # and so on and so forth; it's just not practical
 146         video_title = self._html_search_regex(r'<title>(.*)</title>',
 147             webpage, u'video title', default=u'video')
 148
 149         # video uploader is domain name
 150         video_uploader = self._search_regex(r'(?:https?://)?([^/]*)/.*',
 151             url, u'video uploader')
 152
 153         return [{
 154             'id':       video_id,
 155             'url':      video_url,
 156             'uploader': video_uploader,
 157             'upload_date':  None,
 158             'title':    video_title,
 159             'ext':      video_extension,
 160         }]