Raphaël G. Git Repositories - youtubedl/blob - 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     compat_str,
  14 )
  15
  16 # There are different sources of video in arte.tv, the extraction process
  17 # is different for each one. The videos usually expire in 7 days, so we can't
  18 # add tests.
  19
  20 class ArteTvIE(InfoExtractor):
  21     _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
  22     _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
  23     _LIVE_URL = r'index-[0-9]+\.html$'
  24
  25     IE_NAME = u'arte.tv'
  26
  27     @classmethod
  28     def suitable(cls, url):
  29         return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_URL))
  30
  31     # TODO implement Live Stream
  32     # from ..utils import compat_urllib_parse
  33     # def extractLiveStream(self, url):
  34     #     video_lang = url.split('/')[-4]
  35     #     info = self.grep_webpage(
  36     #         url,
  37     #         r'src="(.*?/videothek_js.*?\.js)',
  38     #         0,
  39     #         [
  40     #             (1, 'url', u'Invalid URL: %s' % url)
  41     #         ]
  42     #     )
  43     #     http_host = url.split('/')[2]
  44     #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
  45     #     info = self.grep_webpage(
  46     #         next_url,
  47     #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
  48     #             '(http://.*?\.swf).*?' +
  49     #             '(rtmp://.*?)\'',
  50     #         re.DOTALL,
  51     #         [
  52     #             (1, 'path',   u'could not extract video path: %s' % url),
  53     #             (2, 'player', u'could not extract video player: %s' % url),
  54     #             (3, 'url',    u'could not extract video url: %s' % url)
  55     #         ]
  56     #     )
  57     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
  58
  59     def _real_extract(self, url):
  60         mobj = re.match(self._VIDEOS_URL, url)
  61         if mobj is not None:
  62             id = mobj.group('id')
  63             lang = mobj.group('lang')
  64             return self._extract_video(url, id, lang)
  65
  66         mobj = re.match(self._LIVEWEB_URL, url)
  67         if mobj is not None:
  68             name = mobj.group('name')
  69             lang = mobj.group('lang')
  70             return self._extract_liveweb(url, name, lang)
  71
  72         if re.search(self._LIVE_URL, video_id) is not None:
  73             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
  74             # self.extractLiveStream(url)
  75             # return
  76
  77     def _extract_video(self, url, video_id, lang):
  78         """Extract from videos.arte.tv"""
  79         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  80         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  81         ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
  82         ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
  83         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  84         config_xml_url = config_node.attrib['ref']
  85         config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
  86
  87         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  88         def _key(m):
  89             quality = m.group('quality')
  90             if quality == 'hd':
  91                 return 2
  92             else:
  93                 return 1
  94         # We pick the best quality
  95         video_urls = sorted(video_urls, key=_key)
  96         video_url = list(video_urls)[-1].group('url')
  97
  98         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
  99         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
 100                                             config_xml, 'thumbnail')
 101         return {'id': video_id,
 102                 'title': title,
 103                 'thumbnail': thumbnail,
 104                 'url': video_url,
 105                 'ext': 'flv',
 106                 }
 107
 108     def _extract_liveweb(self, url, name, lang):
 109         """Extract form http://liveweb.arte.tv/"""
 110         webpage = self._download_webpage(url, name)
 111         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
 112         config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 113                                             video_id, u'Downloading information')
 114         config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
 115         event_doc = config_doc.find('event')
 116         url_node = event_doc.find('video').find('urlHd')
 117         if url_node is None:
 118             url_node = video_doc.find('urlSd')
 119
 120         return {'id': video_id,
 121                 'title': event_doc.find('name%s' % lang.capitalize()).text,
 122                 'url': url_node.text.replace('MP4', 'mp4'),
 123                 'ext': 'flv',
 124                 'thumbnail': self._og_search_thumbnail(webpage),
 125                 }
 126
 127
 128 class ArteTVPlus7IE(InfoExtractor):
 129     IE_NAME = u'arte.tv:+7'
 130     _VALID_URL = r'https?://www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
 131
 132     @classmethod
 133     def _extract_url_info(cls, url):
 134         mobj = re.match(cls._VALID_URL, url)
 135         lang = mobj.group('lang')
 136         # This is not a real id, it can be for example AJT for the news
 137         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
 138         video_id = mobj.group('id')
 139         return video_id, lang
 140
 141     def _real_extract(self, url):
 142         video_id, lang = self._extract_url_info(url)
 143         webpage = self._download_webpage(url, video_id)
 144         return self._extract_from_webpage(webpage, video_id, lang)
 145
 146     def _extract_from_webpage(self, webpage, video_id, lang):
 147         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 148
 149         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
 150         self.report_extraction(video_id)
 151         info = json.loads(json_info)
 152         player_info = info['videoJsonPlayer']
 153
 154         info_dict = {
 155             'id': player_info['VID'],
 156             'title': player_info['VTI'],
 157             'description': player_info.get('VDE'),
 158             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
 159             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 160         }
 161
 162         all_formats = player_info['VSR'].values()
 163         # Some formats use the m3u8 protocol
 164         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
 165         def _match_lang(f):
 166             if f.get('versionCode') is None:
 167                 return True
 168             # Return true if that format is in the language of the url
 169             if lang == 'fr':
 170                 l = 'F'
 171             elif lang == 'de':
 172                 l = 'A'
 173             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 174             return any(re.match(r, f['versionCode']) for r in regexes)
 175         # Some formats may not be in the same language as the url
 176         formats = filter(_match_lang, all_formats)
 177         formats = list(formats) # in python3 filter returns an iterator
 178         if not formats:
 179             # Some videos are only available in the 'Originalversion'
 180             # they aren't tagged as being in French or German
 181             if all(f['versionCode'] == 'VO' for f in all_formats):
 182                 formats = all_formats
 183             else:
 184                 raise ExtractorError(u'The formats list is empty')
 185
 186         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 187             def sort_key(f):
 188                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 189         else:
 190             def sort_key(f):
 191                 return (
 192                     # Sort first by quality
 193                     int(f.get('height',-1)),
 194                     int(f.get('bitrate',-1)),
 195                     # The original version with subtitles has lower relevance
 196                     re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
 197                     # The version with sourds/mal subtitles has also lower relevance
 198                     re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
 199                 )
 200         formats = sorted(formats, key=sort_key)
 201         def _format(format_info):
 202             quality = ''
 203             height = format_info.get('height')
 204             if height is not None:
 205                 quality = compat_str(height)
 206             bitrate = format_info.get('bitrate')
 207             if bitrate is not None:
 208                 quality += '-%d' % bitrate
 209             if format_info.get('versionCode') is not None:
 210                 format_id = u'%s-%s' % (quality, format_info['versionCode'])
 211             else:
 212                 format_id = quality
 213             info = {
 214                 'format_id': format_id,
 215                 'format_note': format_info.get('versionLibelle'),
 216                 'width': format_info.get('width'),
 217                 'height': height,
 218             }
 219             if format_info['mediaType'] == u'rtmp':
 220                 info['url'] = format_info['streamer']
 221                 info['play_path'] = 'mp4:' + format_info['url']
 222                 info['ext'] = 'flv'
 223             else:
 224                 info['url'] = format_info['url']
 225                 info['ext'] = determine_ext(info['url'])
 226             return info
 227         info_dict['formats'] = [_format(f) for f in formats]
 228
 229         return info_dict
 230
 231
 232 # It also uses the arte_vp_url url from the webpage to extract the information
 233 class ArteTVCreativeIE(ArteTVPlus7IE):
 234     IE_NAME = u'arte.tv:creative'
 235     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 236
 237     _TEST = {
 238         u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 239         u'file': u'050489-002.mp4',
 240         u'info_dict': {
 241             u'title': u'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 242         },
 243     }
 244
 245
 246 class ArteTVFutureIE(ArteTVPlus7IE):
 247     IE_NAME = u'arte.tv:future'
 248     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 249
 250     _TEST = {
 251         u'url': u'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 252         u'file': u'050940-003.mp4',
 253         u'info_dict': {
 254             u'title': u'Les champignons au secours de la planète',
 255         },
 256     }
 257
 258     def _real_extract(self, url):
 259         anchor_id, lang = self._extract_url_info(url)
 260         webpage = self._download_webpage(url, anchor_id)
 261         row = get_element_by_id(anchor_id, webpage)
 262         return self._extract_from_webpage(row, anchor_id, lang)