_ Git - youtube-dl/blob - 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('Arte live streams are not yet supported, sorry')
  76             # self.extractLiveStream(url)
  77             # return
  78
  79         raise ExtractorError('No video found')
  80
  81     def _extract_video(self, url, video_id, lang):
  82         """Extract from videos.arte.tv"""
  83         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  84         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  85         ref_xml_doc = self._download_xml(
  86             ref_xml_url, video_id, note='Downloading metadata')
  87         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  88         config_xml_url = config_node.attrib['ref']
  89         config_xml = self._download_webpage(
  90             config_xml_url, video_id, note='Downloading configuration')
  91
  92         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  93         def _key(m):
  94             quality = m.group('quality')
  95             if quality == 'hd':
  96                 return 2
  97             else:
  98                 return 1
  99         # We pick the best quality
 100         video_urls = sorted(video_urls, key=_key)
 101         video_url = list(video_urls)[-1].group('url')
 102
 103         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
 104         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
 105                                             config_xml, 'thumbnail')
 106         return {'id': video_id,
 107                 'title': title,
 108                 'thumbnail': thumbnail,
 109                 'url': video_url,
 110                 'ext': 'flv',
 111                 }
 112
 113     def _extract_liveweb(self, url, name, lang):
 114         """Extract form http://liveweb.arte.tv/"""
 115         webpage = self._download_webpage(url, name)
 116         video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, 'event id')
 117         config_doc = self._download_xml('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
 118                                             video_id, 'Downloading information')
 119         event_doc = config_doc.find('event')
 120         url_node = event_doc.find('video').find('urlHd')
 121         if url_node is None:
 122             url_node = event_doc.find('urlSd')
 123
 124         return {'id': video_id,
 125                 'title': event_doc.find('name%s' % lang.capitalize()).text,
 126                 'url': url_node.text.replace('MP4', 'mp4'),
 127                 'ext': 'flv',
 128                 'thumbnail': self._og_search_thumbnail(webpage),
 129                 }
 130
 131
 132 class ArteTVPlus7IE(InfoExtractor):
 133     IE_NAME = 'arte.tv:+7'
 134     _VALID_URL = r'https?://www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
 135
 136     @classmethod
 137     def _extract_url_info(cls, url):
 138         mobj = re.match(cls._VALID_URL, url)
 139         lang = mobj.group('lang')
 140         # This is not a real id, it can be for example AJT for the news
 141         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
 142         video_id = mobj.group('id')
 143         return video_id, lang
 144
 145     def _real_extract(self, url):
 146         video_id, lang = self._extract_url_info(url)
 147         webpage = self._download_webpage(url, video_id)
 148         return self._extract_from_webpage(webpage, video_id, lang)
 149
 150     def _extract_from_webpage(self, webpage, video_id, lang):
 151         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 152         return self._extract_from_json_url(json_url, video_id, lang)
 153
 154     def _extract_from_json_url(self, json_url, video_id, lang):
 155         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
 156         self.report_extraction(video_id)
 157         info = json.loads(json_info)
 158         player_info = info['videoJsonPlayer']
 159
 160         info_dict = {
 161             'id': player_info['VID'],
 162             'title': player_info['VTI'],
 163             'description': player_info.get('VDE'),
 164             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
 165             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 166         }
 167
 168         all_formats = player_info['VSR'].values()
 169         # Some formats use the m3u8 protocol
 170         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
 171         def _match_lang(f):
 172             if f.get('versionCode') is None:
 173                 return True
 174             # Return true if that format is in the language of the url
 175             if lang == 'fr':
 176                 l = 'F'
 177             elif lang == 'de':
 178                 l = 'A'
 179             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 180             return any(re.match(r, f['versionCode']) for r in regexes)
 181         # Some formats may not be in the same language as the url
 182         formats = filter(_match_lang, all_formats)
 183         formats = list(formats) # in python3 filter returns an iterator
 184         if not formats:
 185             # Some videos are only available in the 'Originalversion'
 186             # they aren't tagged as being in French or German
 187             if all(f['versionCode'] == 'VO' for f in all_formats):
 188                 formats = all_formats
 189             else:
 190                 raise ExtractorError(u'The formats list is empty')
 191
 192         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 193             def sort_key(f):
 194                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 195         else:
 196             def sort_key(f):
 197                 return (
 198                     # Sort first by quality
 199                     int(f.get('height',-1)),
 200                     int(f.get('bitrate',-1)),
 201                     # The original version with subtitles has lower relevance
 202                     re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
 203                     # The version with sourds/mal subtitles has also lower relevance
 204                     re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
 205                 )
 206         formats = sorted(formats, key=sort_key)
 207         def _format(format_info):
 208             quality = ''
 209             height = format_info.get('height')
 210             if height is not None:
 211                 quality = compat_str(height)
 212             bitrate = format_info.get('bitrate')
 213             if bitrate is not None:
 214                 quality += '-%d' % bitrate
 215             if format_info.get('versionCode') is not None:
 216                 format_id = '%s-%s' % (quality, format_info['versionCode'])
 217             else:
 218                 format_id = quality
 219             info = {
 220                 'format_id': format_id,
 221                 'format_note': format_info.get('versionLibelle'),
 222                 'width': format_info.get('width'),
 223                 'height': height,
 224             }
 225             if format_info['mediaType'] == 'rtmp':
 226                 info['url'] = format_info['streamer']
 227                 info['play_path'] = 'mp4:' + format_info['url']
 228                 info['ext'] = 'flv'
 229             else:
 230                 info['url'] = format_info['url']
 231                 info['ext'] = determine_ext(info['url'])
 232             return info
 233         info_dict['formats'] = [_format(f) for f in formats]
 234
 235         return info_dict
 236
 237
 238 # It also uses the arte_vp_url url from the webpage to extract the information
 239 class ArteTVCreativeIE(ArteTVPlus7IE):
 240     IE_NAME = 'arte.tv:creative'
 241     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 242
 243     _TEST = {
 244         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 245         'file': '050489-002.mp4',
 246         'info_dict': {
 247             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 248         },
 249     }
 250
 251
 252 class ArteTVFutureIE(ArteTVPlus7IE):
 253     IE_NAME = 'arte.tv:future'
 254     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 255
 256     _TEST = {
 257         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 258         'file': '050940-003.mp4',
 259         'info_dict': {
 260             'title': 'Les champignons au secours de la planète',
 261         },
 262     }
 263
 264     def _real_extract(self, url):
 265         anchor_id, lang = self._extract_url_info(url)
 266         webpage = self._download_webpage(url, anchor_id)
 267         row = get_element_by_id(anchor_id, webpage)
 268         return self._extract_from_webpage(row, anchor_id, lang)
 269
 270
 271 class ArteTVDDCIE(ArteTVPlus7IE):
 272     IE_NAME = 'arte.tv:ddc'
 273     _VALID_URL = r'http?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 274
 275     def _real_extract(self, url):
 276         video_id, lang = self._extract_url_info(url)
 277         if lang == 'folge':
 278             lang = 'de'
 279         elif lang == 'emission':
 280             lang = 'fr'
 281         webpage = self._download_webpage(url, video_id)
 282         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 283         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 284         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 285         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 286         return self._extract_from_json_url(json_url, video_id, lang)