youtube_dl/extractor/arte.py

   1 # encoding: utf-8
   2 import re
   3 import json
   4 import xml.etree.ElementTree
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     ExtractorError,
   9     find_xpath_attr,
  10     unified_strdate,
  11     determine_ext,
  12     get_element_by_id,
  13 )
  14
  15 # There are different sources of video in arte.tv, the extraction process
  16 # is different for each one. The videos usually expire in 7 days, so we can't
  17 # add tests.
  18
  19 class ArteTvIE(InfoExtractor):
  20     _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
  21     _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
  22     _LIVE_URL = r'index-[0-9]+\.html$'
  23
  24     IE_NAME = u'arte.tv'
  25
  26     @classmethod
  27     def suitable(cls, url):
  28         return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_URL))
  29
  30     # TODO implement Live Stream
  31     # from ..utils import compat_urllib_parse
  32     # def extractLiveStream(self, url):
  33     #     video_lang = url.split('/')[-4]
  34     #     info = self.grep_webpage(
  35     #         url,
  36     #         r'src="(.*?/videothek_js.*?\.js)',
  37     #         0,
  38     #         [
  39     #             (1, 'url', u'Invalid URL: %s' % url)
  40     #         ]
  41     #     )
  42     #     http_host = url.split('/')[2]
  43     #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
  44     #     info = self.grep_webpage(
  45     #         next_url,
  46     #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
  47     #             '(http://.*?\.swf).*?' +
  48     #             '(rtmp://.*?)\'',
  49     #         re.DOTALL,
  50     #         [
  51     #             (1, 'path',   u'could not extract video path: %s' % url),
  52     #             (2, 'player', u'could not extract video player: %s' % url),
  53     #             (3, 'url',    u'could not extract video url: %s' % url)
  54     #         ]
  55     #     )
  56     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
  57
  58     def _real_extract(self, url):
  59         mobj = re.match(self._VIDEOS_URL, url)
  60         if mobj is not None:
  61             id = mobj.group('id')
  62             lang = mobj.group('lang')
  63             return self._extract_video(url, id, lang)
  64
  65         mobj = re.match(self._LIVEWEB_URL, url)
  66         if mobj is not None:
  67             name = mobj.group('name')
  68             lang = mobj.group('lang')
  69             return self._extract_liveweb(url, name, lang)
  70
  71         if re.search(self._LIVE_URL, video_id) is not None:
  72             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
  73             # self.extractLiveStream(url)
  74             # return
  75
  76     def _extract_video(self, url, video_id, lang):
  77         """Extract from videos.arte.tv"""
  78         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  79         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  80         ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
  81         ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
  82         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  83         config_xml_url = config_node.attrib['ref']
  84         config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
  85
  86         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  87         def _key(m):
  88             quality = m.group('quality')
  89             if quality == 'hd':
  90                 return 2
  91             else:
  92                 return 1
  93         # We pick the best quality
  94         video_urls = sorted(video_urls, key=_key)
  95         video_url = list(video_urls)[-1].group('url')
  96
  97         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
  98         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
  99                                             config_xml, 'thumbnail')
 100         return {'id': video_id,
 101                 'title': title,
 102                 'thumbnail': thumbnail,
 103                 'url': video_url,
 104                 'ext': 'flv',
 105                 }
 106
 107     def _extract_liveweb(self, url, name, lang):
 108         """Extract form http://liveweb.arte.tv/"""
 109         webpage = self._download_webpage(url, name)
 110         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
 111         config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 112                                             video_id, u'Downloading information')
 113         config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
 114         event_doc = config_doc.find('event')
 115         url_node = event_doc.find('video').find('urlHd')
 116         if url_node is None:
 117             url_node = video_doc.find('urlSd')
 118
 119         return {'id': video_id,
 120                 'title': event_doc.find('name%s' % lang.capitalize()).text,
 121                 'url': url_node.text.replace('MP4', 'mp4'),
 122                 'ext': 'flv',
 123                 'thumbnail': self._og_search_thumbnail(webpage),
 124                 }
 125
 126
 127 class ArteTVPlus7IE(InfoExtractor):
 128     IE_NAME = u'arte.tv:+7'
 129     _VALID_URL = r'https?://www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
 130
 131     @classmethod
 132     def _extract_url_info(cls, url):
 133         mobj = re.match(cls._VALID_URL, url)
 134         lang = mobj.group('lang')
 135         # This is not a real id, it can be for example AJT for the news
 136         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
 137         video_id = mobj.group('id')
 138         return video_id, lang
 139
 140     def _real_extract(self, url):
 141         video_id, lang = self._extract_url_info(url)
 142         webpage = self._download_webpage(url, video_id)
 143         return self._extract_from_webpage(webpage, video_id, lang)
 144
 145     def _extract_from_webpage(self, webpage, video_id, lang):
 146         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 147
 148         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
 149         self.report_extraction(video_id)
 150         info = json.loads(json_info)
 151         player_info = info['videoJsonPlayer']
 152
 153         info_dict = {
 154             'id': player_info['VID'],
 155             'title': player_info['VTI'],
 156             'description': player_info.get('VDE'),
 157             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
 158             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 159         }
 160
 161         all_formats = player_info['VSR'].values()
 162         # Some formats use the m3u8 protocol
 163         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
 164         def _match_lang(f):
 165             if f.get('versionCode') is None:
 166                 return True
 167             # Return true if that format is in the language of the url
 168             if lang == 'fr':
 169                 l = 'F'
 170             elif lang == 'de':
 171                 l = 'A'
 172             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 173             return any(re.match(r, f['versionCode']) for r in regexes)
 174         # Some formats may not be in the same language as the url
 175         formats = filter(_match_lang, all_formats)
 176         formats = list(formats) # in python3 filter returns an iterator
 177         if not formats:
 178             # Some videos are only available in the 'Originalversion'
 179             # they aren't tagged as being in French or German
 180             if all(f['versionCode'] == 'VO' for f in all_formats):
 181                 formats = all_formats
 182             else:
 183                 raise ExtractorError(u'The formats list is empty')
 184         # We order the formats by quality
 185         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 186             sort_key = lambda f: ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 187         else:
 188             sort_key = lambda f: int(f.get('height',-1))
 189         formats = sorted(formats, key=sort_key)
 190         # Prefer videos without subtitles in the same language
 191         formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f.get('versionCode', '')) is None)
 192         # Pick the best quality
 193         def _format(format_info):
 194             quality = format_info['quality']
 195             m_quality = re.match(r'\w*? - (\d*)p', quality)
 196             if m_quality is not None:
 197                 quality = m_quality.group(1)
 198             if format_info.get('versionCode') is not None:
 199                 format_id = u'%s-%s' % (quality, format_info['versionCode'])
 200             else:
 201                 format_id = quality
 202             info = {
 203                 'format_id': format_id,
 204                 'format_note': format_info.get('versionLibelle'),
 205                 'width': format_info.get('width'),
 206                 'height': format_info.get('height'),
 207             }
 208             if format_info['mediaType'] == u'rtmp':
 209                 info['url'] = format_info['streamer']
 210                 info['play_path'] = 'mp4:' + format_info['url']
 211                 info['ext'] = 'flv'
 212             else:
 213                 info['url'] = format_info['url']
 214                 info['ext'] = determine_ext(info['url'])
 215             return info
 216         info_dict['formats'] = [_format(f) for f in formats]
 217
 218         return info_dict
 219
 220
 221 # It also uses the arte_vp_url url from the webpage to extract the information
 222 class ArteTVCreativeIE(ArteTVPlus7IE):
 223     IE_NAME = u'arte.tv:creative'
 224     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 225
 226     _TEST = {
 227         u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 228         u'file': u'050489-002.mp4',
 229         u'info_dict': {
 230             u'title': u'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 231         },
 232     }
 233
 234
 235 class ArteTVFutureIE(ArteTVPlus7IE):
 236     IE_NAME = u'arte.tv:future'
 237     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 238
 239     _TEST = {
 240         u'url': u'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 241         u'file': u'050940-003.mp4',
 242         u'info_dict': {
 243             u'title': u'Les champignons au secours de la planète',
 244         },
 245     }
 246
 247     def _real_extract(self, url):
 248         anchor_id, lang = self._extract_url_info(url)
 249         webpage = self._download_webpage(url, anchor_id)
 250         row = get_element_by_id(anchor_id, webpage)
 251         return self._extract_from_webpage(row, anchor_id, lang)