youtube_dl/extractor/arte.py

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import json
   6
   7 from .common import InfoExtractor
   8 from ..utils import (
   9     ExtractorError,
  10     find_xpath_attr,
  11     unified_strdate,
  12     determine_ext,
  13     get_element_by_id,
  14     compat_str,
  15     get_element_by_attribute,
  16 )
  17
  18 # There are different sources of video in arte.tv, the extraction process
  19 # is different for each one. The videos usually expire in 7 days, so we can't
  20 # add tests.
  21
  22 class ArteTvIE(InfoExtractor):
  23     _VIDEOS_URL = r'(?:http://)?videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
  24     _LIVEWEB_URL = r'(?:http://)?liveweb\.arte\.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
  25     _LIVE_URL = r'index-[0-9]+\.html$'
  26
  27     IE_NAME = 'arte.tv'
  28
  29     @classmethod
  30     def suitable(cls, url):
  31         return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_URL))
  32
  33     # TODO implement Live Stream
  34     # from ..utils import compat_urllib_parse
  35     # def extractLiveStream(self, url):
  36     #     video_lang = url.split('/')[-4]
  37     #     info = self.grep_webpage(
  38     #         url,
  39     #         r'src="(.*?/videothek_js.*?\.js)',
  40     #         0,
  41     #         [
  42     #             (1, 'url', 'Invalid URL: %s' % url)
  43     #         ]
  44     #     )
  45     #     http_host = url.split('/')[2]
  46     #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
  47     #     info = self.grep_webpage(
  48     #         next_url,
  49     #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
  50     #             '(http://.*?\.swf).*?' +
  51     #             '(rtmp://.*?)\'',
  52     #         re.DOTALL,
  53     #         [
  54     #             (1, 'path',   'could not extract video path: %s' % url),
  55     #             (2, 'player', 'could not extract video player: %s' % url),
  56     #             (3, 'url',    'could not extract video url: %s' % url)
  57     #         ]
  58     #     )
  59     #     video_url = '%s/%s' % (info.get('url'), info.get('path'))
  60
  61     def _real_extract(self, url):
  62         mobj = re.match(self._VIDEOS_URL, url)
  63         if mobj is not None:
  64             id = mobj.group('id')
  65             lang = mobj.group('lang')
  66             return self._extract_video(url, id, lang)
  67
  68         mobj = re.match(self._LIVEWEB_URL, url)
  69         if mobj is not None:
  70             name = mobj.group('name')
  71             lang = mobj.group('lang')
  72             return self._extract_liveweb(url, name, lang)
  73
  74         if re.search(self._LIVE_URL, url) is not None:
  75             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
  76             # self.extractLiveStream(url)
  77             # return
  78
  79     def _extract_video(self, url, video_id, lang):
  80         """Extract from videos.arte.tv"""
  81         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  82         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  83         ref_xml_doc = self._download_xml(ref_xml_url, video_id, note=u'Downloading metadata')
  84         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  85         config_xml_url = config_node.attrib['ref']
  86         config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
  87
  88         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  89         def _key(m):
  90             quality = m.group('quality')
  91             if quality == 'hd':
  92                 return 2
  93             else:
  94                 return 1
  95         # We pick the best quality
  96         video_urls = sorted(video_urls, key=_key)
  97         video_url = list(video_urls)[-1].group('url')
  98
  99         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
 100         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
 101                                             config_xml, 'thumbnail')
 102         return {'id': video_id,
 103                 'title': title,
 104                 'thumbnail': thumbnail,
 105                 'url': video_url,
 106                 'ext': 'flv',
 107                 }
 108
 109     def _extract_liveweb(self, url, name, lang):
 110         """Extract form http://liveweb.arte.tv/"""
 111         webpage = self._download_webpage(url, name)
 112         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, 'event id')
 113         config_doc = self._download_xml('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 114                                             video_id, 'Downloading information')
 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 = event_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 = '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         return self._extract_from_json_url(json_url, video_id, lang)
 149
 150     def _extract_from_json_url(self, json_url, video_id, lang):
 151         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
 152         self.report_extraction(video_id)
 153         info = json.loads(json_info)
 154         player_info = info['videoJsonPlayer']
 155
 156         info_dict = {
 157             'id': player_info['VID'],
 158             'title': player_info['VTI'],
 159             'description': player_info.get('VDE'),
 160             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
 161             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 162         }
 163
 164         all_formats = player_info['VSR'].values()
 165         # Some formats use the m3u8 protocol
 166         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
 167         def _match_lang(f):
 168             if f.get('versionCode') is None:
 169                 return True
 170             # Return true if that format is in the language of the url
 171             if lang == 'fr':
 172                 l = 'F'
 173             elif lang == 'de':
 174                 l = 'A'
 175             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 176             return any(re.match(r, f['versionCode']) for r in regexes)
 177         # Some formats may not be in the same language as the url
 178         formats = filter(_match_lang, all_formats)
 179         formats = list(formats) # in python3 filter returns an iterator
 180         if not formats:
 181             # Some videos are only available in the 'Originalversion'
 182             # they aren't tagged as being in French or German
 183             if all(f['versionCode'] == 'VO' for f in all_formats):
 184                 formats = all_formats
 185             else:
 186                 raise ExtractorError(u'The formats list is empty')
 187
 188         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 189             def sort_key(f):
 190                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 191         else:
 192             def sort_key(f):
 193                 return (
 194                     # Sort first by quality
 195                     int(f.get('height',-1)),
 196                     int(f.get('bitrate',-1)),
 197                     # The original version with subtitles has lower relevance
 198                     re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
 199                     # The version with sourds/mal subtitles has also lower relevance
 200                     re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
 201                 )
 202         formats = sorted(formats, key=sort_key)
 203         def _format(format_info):
 204             quality = ''
 205             height = format_info.get('height')
 206             if height is not None:
 207                 quality = compat_str(height)
 208             bitrate = format_info.get('bitrate')
 209             if bitrate is not None:
 210                 quality += '-%d' % bitrate
 211             if format_info.get('versionCode') is not None:
 212                 format_id = '%s-%s' % (quality, format_info['versionCode'])
 213             else:
 214                 format_id = quality
 215             info = {
 216                 'format_id': format_id,
 217                 'format_note': format_info.get('versionLibelle'),
 218                 'width': format_info.get('width'),
 219                 'height': height,
 220             }
 221             if format_info['mediaType'] == 'rtmp':
 222                 info['url'] = format_info['streamer']
 223                 info['play_path'] = 'mp4:' + format_info['url']
 224                 info['ext'] = 'flv'
 225             else:
 226                 info['url'] = format_info['url']
 227                 info['ext'] = determine_ext(info['url'])
 228             return info
 229         info_dict['formats'] = [_format(f) for f in formats]
 230
 231         return info_dict
 232
 233
 234 # It also uses the arte_vp_url url from the webpage to extract the information
 235 class ArteTVCreativeIE(ArteTVPlus7IE):
 236     IE_NAME = 'arte.tv:creative'
 237     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 238
 239     _TEST = {
 240         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 241         'file': '050489-002.mp4',
 242         'info_dict': {
 243             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 244         },
 245     }
 246
 247
 248 class ArteTVFutureIE(ArteTVPlus7IE):
 249     IE_NAME = 'arte.tv:future'
 250     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 251
 252     _TEST = {
 253         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 254         'file': '050940-003.mp4',
 255         'info_dict': {
 256             'title': 'Les champignons au secours de la planète',
 257         },
 258     }
 259
 260     def _real_extract(self, url):
 261         anchor_id, lang = self._extract_url_info(url)
 262         webpage = self._download_webpage(url, anchor_id)
 263         row = get_element_by_id(anchor_id, webpage)
 264         return self._extract_from_webpage(row, anchor_id, lang)
 265
 266
 267 class ArteTVDDCIE(ArteTVPlus7IE):
 268     IE_NAME = 'arte.tv:ddc'
 269     _VALID_URL = r'http?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 270
 271     def _real_extract(self, url):
 272         video_id, lang = self._extract_url_info(url)
 273         if lang == 'folge':
 274             lang = 'de'
 275         elif lang == 'emission':
 276             lang = 'fr'
 277         webpage = self._download_webpage(url, video_id)
 278         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 279         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 280         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 281         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 282         return self._extract_from_json_url(json_url, video_id, lang)