youtube_dl/extractor/arte.py

   1 import re
   2 import json
   3 import xml.etree.ElementTree
   4
   5 from .common import InfoExtractor
   6 from ..utils import (
   7     ExtractorError,
   8     find_xpath_attr,
   9     unified_strdate,
  10 )
  11
  12 class ArteTvIE(InfoExtractor):
  13     """
  14     There are two sources of video in arte.tv: videos.arte.tv and
  15     www.arte.tv/guide, the extraction process is different for each one.
  16     The videos expire in 7 days, so we can't add tests.
  17     """
  18     _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
  19     _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
  20     _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
  21     _LIVE_URL = r'index-[0-9]+\.html$'
  22
  23     IE_NAME = u'arte.tv'
  24
  25     @classmethod
  26     def suitable(cls, url):
  27         return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL, cls._LIVEWEB_URL))
  28
  29     # TODO implement Live Stream
  30     # from ..utils import compat_urllib_parse
  31     # def extractLiveStream(self, url):
  32     #     video_lang = url.split('/')[-4]
  33     #     info = self.grep_webpage(
  34     #         url,
  35     #         r'src="(.*?/videothek_js.*?\.js)',
  36     #         0,
  37     #         [
  38     #             (1, 'url', u'Invalid URL: %s' % url)
  39     #         ]
  40     #     )
  41     #     http_host = url.split('/')[2]
  42     #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
  43     #     info = self.grep_webpage(
  44     #         next_url,
  45     #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
  46     #             '(http://.*?\.swf).*?' +
  47     #             '(rtmp://.*?)\'',
  48     #         re.DOTALL,
  49     #         [
  50     #             (1, 'path',   u'could not extract video path: %s' % url),
  51     #             (2, 'player', u'could not extract video player: %s' % url),
  52     #             (3, 'url',    u'could not extract video url: %s' % url)
  53     #         ]
  54     #     )
  55     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
  56
  57     def _real_extract(self, url):
  58         mobj = re.match(self._EMISSION_URL, url)
  59         if mobj is not None:
  60             lang = mobj.group('lang')
  61             # This is not a real id, it can be for example AJT for the news
  62             # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
  63             video_id = mobj.group('id')
  64             return self._extract_emission(url, video_id, lang)
  65
  66         mobj = re.match(self._VIDEOS_URL, url)
  67         if mobj is not None:
  68             id = mobj.group('id')
  69             lang = mobj.group('lang')
  70             return self._extract_video(url, id, lang)
  71
  72         mobj = re.match(self._LIVEWEB_URL, url)
  73         if mobj is not None:
  74             name = mobj.group('name')
  75             lang = mobj.group('lang')
  76             return self._extract_liveweb(url, name, lang)
  77
  78         if re.search(self._LIVE_URL, video_id) is not None:
  79             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
  80             # self.extractLiveStream(url)
  81             # return
  82
  83     def _extract_emission(self, url, video_id, lang):
  84         """Extract from www.arte.tv/guide"""
  85         webpage = self._download_webpage(url, video_id)
  86         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
  87
  88         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
  89         self.report_extraction(video_id)
  90         info = json.loads(json_info)
  91         player_info = info['videoJsonPlayer']
  92
  93         info_dict = {'id': player_info['VID'],
  94                      'title': player_info['VTI'],
  95                      'description': player_info.get('VDE'),
  96                      'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]),
  97                      'thumbnail': player_info['programImage'],
  98                      'ext': 'flv',
  99                      }
 100
 101         formats = player_info['VSR'].values()
 102         def _match_lang(f):
 103             # Return true if that format is in the language of the url
 104             if lang == 'fr':
 105                 l = 'F'
 106             elif lang == 'de':
 107                 l = 'A'
 108             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 109             return any(re.match(r, f['versionCode']) for r in regexes)
 110         # Some formats may not be in the same language as the url
 111         formats = filter(_match_lang, formats)
 112         # We order the formats by quality
 113         formats = sorted(formats, key=lambda f: int(f['height']))
 114         # Prefer videos without subtitles in the same language
 115         formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f['versionCode']) is None)
 116         # Pick the best quality
 117         format_info = formats[-1]
 118         if format_info['mediaType'] == u'rtmp':
 119             info_dict['url'] = format_info['streamer']
 120             info_dict['play_path'] = 'mp4:' + format_info['url']
 121         else:
 122             info_dict['url'] = format_info['url']
 123
 124         return info_dict
 125
 126     def _extract_video(self, url, video_id, lang):
 127         """Extract from videos.arte.tv"""
 128         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
 129         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
 130         ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
 131         ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
 132         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
 133         config_xml_url = config_node.attrib['ref']
 134         config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
 135
 136         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
 137         def _key(m):
 138             quality = m.group('quality')
 139             if quality == 'hd':
 140                 return 2
 141             else:
 142                 return 1
 143         # We pick the best quality
 144         video_urls = sorted(video_urls, key=_key)
 145         video_url = list(video_urls)[-1].group('url')
 146
 147         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
 148         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
 149                                             config_xml, 'thumbnail')
 150         return {'id': video_id,
 151                 'title': title,
 152                 'thumbnail': thumbnail,
 153                 'url': video_url,
 154                 'ext': 'flv',
 155                 }
 156
 157     def _extract_liveweb(self, url, name, lang):
 158         """Extract form http://liveweb.arte.tv/"""
 159         webpage = self._download_webpage(url, name)
 160         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
 161         config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 162                                             video_id, u'Downloading information')
 163         config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
 164         event_doc = config_doc.find('event')
 165         url_node = event_doc.find('video').find('urlHd')
 166         if url_node is None:
 167             url_node = video_doc.find('urlSd')
 168
 169         return {'id': video_id,
 170                 'title': event_doc.find('name%s' % lang.capitalize()).text,
 171                 'url': url_node.text.replace('MP4', 'mp4'),
 172                 'ext': 'flv',
 173                 'thumbnail': self._og_search_thumbnail(webpage),
 174                 }