_ Git - youtube-dl/blob - 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     determine_ext,
  11 )
  12
  13 # There are different sources of video in arte.tv, the extraction process
  14 # is different for each one. The videos usually expire in 7 days, so we can't
  15 # add tests.
  16
  17 class ArteTvIE(InfoExtractor):
  18     _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
  19     _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
  20     _LIVE_URL = r'index-[0-9]+\.html$'
  21
  22     IE_NAME = u'arte.tv'
  23
  24     @classmethod
  25     def suitable(cls, url):
  26         return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_URL))
  27
  28     # TODO implement Live Stream
  29     # from ..utils import compat_urllib_parse
  30     # def extractLiveStream(self, url):
  31     #     video_lang = url.split('/')[-4]
  32     #     info = self.grep_webpage(
  33     #         url,
  34     #         r'src="(.*?/videothek_js.*?\.js)',
  35     #         0,
  36     #         [
  37     #             (1, 'url', u'Invalid URL: %s' % url)
  38     #         ]
  39     #     )
  40     #     http_host = url.split('/')[2]
  41     #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
  42     #     info = self.grep_webpage(
  43     #         next_url,
  44     #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
  45     #             '(http://.*?\.swf).*?' +
  46     #             '(rtmp://.*?)\'',
  47     #         re.DOTALL,
  48     #         [
  49     #             (1, 'path',   u'could not extract video path: %s' % url),
  50     #             (2, 'player', u'could not extract video player: %s' % url),
  51     #             (3, 'url',    u'could not extract video url: %s' % url)
  52     #         ]
  53     #     )
  54     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
  55
  56     def _real_extract(self, url):
  57         mobj = re.match(self._VIDEOS_URL, url)
  58         if mobj is not None:
  59             id = mobj.group('id')
  60             lang = mobj.group('lang')
  61             return self._extract_video(url, id, lang)
  62
  63         mobj = re.match(self._LIVEWEB_URL, url)
  64         if mobj is not None:
  65             name = mobj.group('name')
  66             lang = mobj.group('lang')
  67             return self._extract_liveweb(url, name, lang)
  68
  69         if re.search(self._LIVE_URL, video_id) is not None:
  70             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
  71             # self.extractLiveStream(url)
  72             # return
  73
  74     def _extract_video(self, url, video_id, lang):
  75         """Extract from videos.arte.tv"""
  76         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  77         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  78         ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
  79         ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
  80         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  81         config_xml_url = config_node.attrib['ref']
  82         config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
  83
  84         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  85         def _key(m):
  86             quality = m.group('quality')
  87             if quality == 'hd':
  88                 return 2
  89             else:
  90                 return 1
  91         # We pick the best quality
  92         video_urls = sorted(video_urls, key=_key)
  93         video_url = list(video_urls)[-1].group('url')
  94
  95         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
  96         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
  97                                             config_xml, 'thumbnail')
  98         return {'id': video_id,
  99                 'title': title,
 100                 'thumbnail': thumbnail,
 101                 'url': video_url,
 102                 'ext': 'flv',
 103                 }
 104
 105     def _extract_liveweb(self, url, name, lang):
 106         """Extract form http://liveweb.arte.tv/"""
 107         webpage = self._download_webpage(url, name)
 108         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
 109         config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 110                                             video_id, u'Downloading information')
 111         config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
 112         event_doc = config_doc.find('event')
 113         url_node = event_doc.find('video').find('urlHd')
 114         if url_node is None:
 115             url_node = video_doc.find('urlSd')
 116
 117         return {'id': video_id,
 118                 'title': event_doc.find('name%s' % lang.capitalize()).text,
 119                 'url': url_node.text.replace('MP4', 'mp4'),
 120                 'ext': 'flv',
 121                 'thumbnail': self._og_search_thumbnail(webpage),
 122                 }
 123
 124
 125 class ArteTVPlus7IE(InfoExtractor):
 126     IE_NAME = u'arte.tv:+7'
 127     _VALID_URL = r'https?://www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
 128
 129     def _real_extract(self, url):
 130         mobj = re.match(self._VALID_URL, url)
 131         lang = mobj.group('lang')
 132         # This is not a real id, it can be for example AJT for the news
 133         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
 134         video_id = mobj.group('id')
 135
 136         webpage = self._download_webpage(url, video_id)
 137         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 138
 139         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
 140         self.report_extraction(video_id)
 141         info = json.loads(json_info)
 142         player_info = info['videoJsonPlayer']
 143
 144         info_dict = {
 145             'id': player_info['VID'],
 146             'title': player_info['VTI'],
 147             'description': player_info.get('VDE'),
 148             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
 149             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 150         }
 151
 152         formats = player_info['VSR'].values()
 153         def _match_lang(f):
 154             if f.get('versionCode') is None:
 155                 return True
 156             # Return true if that format is in the language of the url
 157             if lang == 'fr':
 158                 l = 'F'
 159             elif lang == 'de':
 160                 l = 'A'
 161             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 162             return any(re.match(r, f['versionCode']) for r in regexes)
 163         # Some formats may not be in the same language as the url
 164         formats = filter(_match_lang, formats)
 165         # Some formats use the m3u8 protocol
 166         formats = filter(lambda f: f.get('videoFormat') != 'M3U8', formats)
 167         # We order the formats by quality
 168         formats = sorted(formats, key=lambda f: int(f.get('height',-1)))
 169         # Prefer videos without subtitles in the same language
 170         formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f.get('versionCode', '')) is None)
 171         # Pick the best quality
 172         def _format(format_info):
 173             info = {
 174                 'width': format_info.get('width'),
 175                 'height': format_info.get('height'),
 176             }
 177             if format_info['mediaType'] == u'rtmp':
 178                 info['url'] = format_info['streamer']
 179                 info['play_path'] = 'mp4:' + format_info['url']
 180                 info['ext'] = 'flv'
 181             else:
 182                 info['url'] = format_info['url']
 183                 info['ext'] = determine_ext(info['url'])
 184             return info
 185         info_dict['formats'] = [_format(f) for f in formats]
 186         # TODO: Remove when #980 has been merged
 187         info_dict.update(info_dict['formats'][-1])
 188
 189         return info_dict
 190
 191
 192 # It also uses the arte_vp_url url from the webpage to extract the information
 193 class ArteTVCreativeIE(ArteTVPlus7IE):
 194     IE_NAME = u'arte.tv:creative'
 195     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 196
 197     _TEST = {
 198         u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 199         u'file': u'050489-002.mp4',
 200         u'info_dict': {
 201             u'title': u'Agentur Amateur #2 - Corporate Design',
 202         },
 203     }
 204