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