youtube_dl/extractor/arte.py

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..compat import (
   8     compat_parse_qs,
   9     compat_urllib_parse_urlparse,
  10 )
  11 from ..utils import (
  12     find_xpath_attr,
  13     unified_strdate,
  14     get_element_by_attribute,
  15     int_or_none,
  16     NO_DEFAULT,
  17     qualities,
  18 )
  19
  20 # There are different sources of video in arte.tv, the extraction process
  21 # is different for each one. The videos usually expire in 7 days, so we can't
  22 # add tests.
  23
  24
  25 class ArteTvIE(InfoExtractor):
  26     _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de|en|es)/.*-(?P<id>.*?)\.html'
  27     IE_NAME = 'arte.tv'
  28
  29     def _real_extract(self, url):
  30         mobj = re.match(self._VALID_URL, url)
  31         lang = mobj.group('lang')
  32         video_id = mobj.group('id')
  33
  34         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  35         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  36         ref_xml_doc = self._download_xml(
  37             ref_xml_url, video_id, note='Downloading metadata')
  38         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  39         config_xml_url = config_node.attrib['ref']
  40         config = self._download_xml(
  41             config_xml_url, video_id, note='Downloading configuration')
  42
  43         formats = [{
  44             'format_id': q.attrib['quality'],
  45             # The playpath starts at 'mp4:', if we don't manually
  46             # split the url, rtmpdump will incorrectly parse them
  47             'url': q.text.split('mp4:', 1)[0],
  48             'play_path': 'mp4:' + q.text.split('mp4:', 1)[1],
  49             'ext': 'flv',
  50             'quality': 2 if q.attrib['quality'] == 'hd' else 1,
  51         } for q in config.findall('./urls/url')]
  52         self._sort_formats(formats)
  53
  54         title = config.find('.//name').text
  55         thumbnail = config.find('.//firstThumbnailUrl').text
  56         return {
  57             'id': video_id,
  58             'title': title,
  59             'thumbnail': thumbnail,
  60             'formats': formats,
  61         }
  62
  63
  64 class ArteTVPlus7IE(InfoExtractor):
  65     IE_NAME = 'arte.tv:+7'
  66     _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de|en|es)/(?:(?:sendungen|emissions)/)?(?P<id>[^/]+)/(?P<name>[^/?#&+])'
  67
  68     @classmethod
  69     def _extract_url_info(cls, url):
  70         mobj = re.match(cls._VALID_URL, url)
  71         lang = mobj.group('lang')
  72         query = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
  73         if 'vid' in query:
  74             video_id = query['vid'][0]
  75         else:
  76             # This is not a real id, it can be for example AJT for the news
  77             # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
  78             video_id = mobj.group('id')
  79         return video_id, lang
  80
  81     def _real_extract(self, url):
  82         video_id, lang = self._extract_url_info(url)
  83         webpage = self._download_webpage(url, video_id)
  84         return self._extract_from_webpage(webpage, video_id, lang)
  85
  86     def _extract_from_webpage(self, webpage, video_id, lang):
  87         patterns_templates = (r'arte_vp_url=["\'](.*?%s.*?)["\']', r'data-url=["\']([^"]+%s[^"]+)["\']')
  88         ids = (video_id, '')
  89         # some pages contain multiple videos (like
  90         # http://www.arte.tv/guide/de/sendungen/XEN/xenius/?vid=055918-015_PLUS7-D),
  91         # so we first try to look for json URLs that contain the video id from
  92         # the 'vid' parameter.
  93         patterns = [t % re.escape(_id) for _id in ids for t in patterns_templates]
  94         json_url = self._html_search_regex(
  95             patterns, webpage, 'json vp url', default=None)
  96         if not json_url:
  97             def find_iframe_url(webpage, default=NO_DEFAULT):
  98                 return self._html_search_regex(
  99                     r'<iframe[^>]+src=(["\'])(?P<url>.+\bjson_url=.+?)\1',
 100                     webpage, 'iframe url', group='url', default=default)
 101
 102             iframe_url = find_iframe_url(webpage, None)
 103             if not iframe_url:
 104                 embed_url = self._html_search_regex(
 105                     r'arte_vp_url_oembed=\'([^\']+?)\'', webpage, 'embed url', default=None)
 106                 if embed_url:
 107                     player = self._download_json(
 108                         embed_url, video_id, 'Downloading player page')
 109                     iframe_url = find_iframe_url(player['html'])
 110             # en and es URLs produce react-based pages with different layout (e.g.
 111             # http://www.arte.tv/guide/en/053330-002-A/carnival-italy?zone=world)
 112             if not iframe_url:
 113                 embed_html = self._parse_json(
 114                     self._search_regex(
 115                         r'program\s*:\s*({.+?["\']embed_html["\'].+?}),?\s*\n',
 116                         webpage, 'program'),
 117                     video_id)['embed_html']
 118                 iframe_url = find_iframe_url(embed_html)
 119             json_url = compat_parse_qs(
 120                 compat_urllib_parse_urlparse(iframe_url).query)['json_url'][0]
 121         return self._extract_from_json_url(json_url, video_id, lang)
 122
 123     def _extract_from_json_url(self, json_url, video_id, lang):
 124         info = self._download_json(json_url, video_id)
 125         player_info = info['videoJsonPlayer']
 126
 127         upload_date_str = player_info.get('shootingDate')
 128         if not upload_date_str:
 129             upload_date_str = (player_info.get('VRA') or player_info.get('VDA') or '').split(' ')[0]
 130
 131         title = player_info['VTI'].strip()
 132         subtitle = player_info.get('VSU', '').strip()
 133         if subtitle:
 134             title += ' - %s' % subtitle
 135
 136         info_dict = {
 137             'id': player_info['VID'],
 138             'title': title,
 139             'description': player_info.get('VDE'),
 140             'upload_date': unified_strdate(upload_date_str),
 141             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 142         }
 143         qfunc = qualities(['HQ', 'MQ', 'EQ', 'SQ'])
 144
 145         LANGS = {
 146             'fr': 'F',
 147             'de': 'A',
 148             'en': 'E[ANG]',
 149             'es': 'E[ESP]',
 150         }
 151
 152         formats = []
 153         for format_id, format_dict in player_info['VSR'].items():
 154             f = dict(format_dict)
 155             versionCode = f.get('versionCode')
 156             langcode = LANGS.get(lang, lang)
 157             lang_rexs = [r'VO?%s-' % re.escape(langcode), r'VO?.-ST%s$' % re.escape(langcode)]
 158             lang_pref = None
 159             if versionCode:
 160                 matched_lang_rexs = [r for r in lang_rexs if re.match(r, versionCode)]
 161                 lang_pref = -10 if not matched_lang_rexs else 10 * len(matched_lang_rexs)
 162             source_pref = 0
 163             if versionCode is not None:
 164                 # The original version with subtitles has lower relevance
 165                 if re.match(r'VO-ST(F|A|E)', versionCode):
 166                     source_pref -= 10
 167                 # The version with sourds/mal subtitles has also lower relevance
 168                 elif re.match(r'VO?(F|A|E)-STM\1', versionCode):
 169                     source_pref -= 9
 170             format = {
 171                 'format_id': format_id,
 172                 'preference': -10 if f.get('videoFormat') == 'M3U8' else None,
 173                 'language_preference': lang_pref,
 174                 'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')),
 175                 'width': int_or_none(f.get('width')),
 176                 'height': int_or_none(f.get('height')),
 177                 'tbr': int_or_none(f.get('bitrate')),
 178                 'quality': qfunc(f.get('quality')),
 179                 'source_preference': source_pref,
 180             }
 181
 182             if f.get('mediaType') == 'rtmp':
 183                 format['url'] = f['streamer']
 184                 format['play_path'] = 'mp4:' + f['url']
 185                 format['ext'] = 'flv'
 186             else:
 187                 format['url'] = f['url']
 188
 189             formats.append(format)
 190
 191         self._check_formats(formats, video_id)
 192         self._sort_formats(formats)
 193
 194         info_dict['formats'] = formats
 195         return info_dict
 196
 197
 198 # It also uses the arte_vp_url url from the webpage to extract the information
 199 class ArteTVCreativeIE(ArteTVPlus7IE):
 200     IE_NAME = 'arte.tv:creative'
 201     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de|en|es)/(?:magazine?/)?(?P<id>[^/?#&]+)'
 202
 203     _TESTS = [{
 204         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 205         'info_dict': {
 206             'id': '72176',
 207             'ext': 'mp4',
 208             'title': 'Folge 2 - Corporate Design',
 209             'upload_date': '20131004',
 210         },
 211     }, {
 212         'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion',
 213         'info_dict': {
 214             'id': '160676',
 215             'ext': 'mp4',
 216             'title': 'Monty Python live (mostly)',
 217             'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n',
 218             'upload_date': '20140805',
 219         }
 220     }]
 221
 222
 223 class ArteTVFutureIE(ArteTVPlus7IE):
 224     IE_NAME = 'arte.tv:future'
 225     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de|en|es)/(?P<id>[^/?#&]+)'
 226
 227     _TESTS = [{
 228         'url': 'http://future.arte.tv/fr/info-sciences/les-ecrevisses-aussi-sont-anxieuses',
 229         'info_dict': {
 230             'id': '050940-028-A',
 231             'ext': 'mp4',
 232             'title': 'Les écrevisses aussi peuvent être anxieuses',
 233         },
 234     }, {
 235         'url': 'http://future.arte.tv/fr/la-science-est-elle-responsable',
 236         'only_matching': True,
 237     }]
 238
 239
 240 class ArteTVDDCIE(ArteTVPlus7IE):
 241     IE_NAME = 'arte.tv:ddc'
 242     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>[^/?#&]+)'
 243
 244     def _real_extract(self, url):
 245         video_id, lang = self._extract_url_info(url)
 246         if lang == 'folge':
 247             lang = 'de'
 248         elif lang == 'emission':
 249             lang = 'fr'
 250         webpage = self._download_webpage(url, video_id)
 251         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 252         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 253         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 254         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 255         return self._extract_from_json_url(json_url, video_id, lang)
 256
 257
 258 class ArteTVConcertIE(ArteTVPlus7IE):
 259     IE_NAME = 'arte.tv:concert'
 260     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>fr|de|en|es)/(?P<id>[^/?#&]+)'
 261
 262     _TEST = {
 263         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
 264         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
 265         'info_dict': {
 266             'id': '186',
 267             'ext': 'mp4',
 268             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
 269             'upload_date': '20140128',
 270             'description': 'md5:486eb08f991552ade77439fe6d82c305',
 271         },
 272     }
 273
 274
 275 class ArteTVCinemaIE(ArteTVPlus7IE):
 276     IE_NAME = 'arte.tv:cinema'
 277     _VALID_URL = r'https?://cinema\.arte\.tv/(?P<lang>fr|de|en|es)/(?P<id>.+)'
 278
 279     _TEST = {
 280         'url': 'http://cinema.arte.tv/de/node/38291',
 281         'md5': '6b275511a5107c60bacbeeda368c3aa1',
 282         'info_dict': {
 283             'id': '055876-000_PWA12025-D',
 284             'ext': 'mp4',
 285             'title': 'Tod auf dem Nil',
 286             'upload_date': '20160122',
 287             'description': 'md5:7f749bbb77d800ef2be11d54529b96bc',
 288         },
 289     }
 290
 291
 292 class ArteTVMagazineIE(ArteTVPlus7IE):
 293     IE_NAME = 'arte.tv:magazine'
 294     _VALID_URL = r'https?://(?:www\.)?arte\.tv/magazine/[^/]+/(?P<lang>fr|de|en|es)/(?P<id>[^/?#&]+)'
 295
 296     _TESTS = [{
 297         'url': 'http://www.arte.tv/magazine/trepalium/fr/entretien-avec-le-realisateur-vincent-lannoo-trepalium',
 298         'md5': '66a093339c1278bb3719157ef07107b2',
 299         'info_dict': {
 300             'id': '065965-000-A',
 301             'ext': 'mp4',
 302             'title': 'Trepalium - Extrait Ep.01',
 303         },
 304     }, {
 305         'url': 'http://www.arte.tv/magazine/metropolis/de/frank-woeste-german-paris-metropolis',
 306         'only_matching': True,
 307     }]
 308
 309
 310 class ArteTVEmbedIE(ArteTVPlus7IE):
 311     IE_NAME = 'arte.tv:embed'
 312     _VALID_URL = r'''(?x)
 313         http://www\.arte\.tv
 314         /playerv2/embed\.php\?json_url=
 315         (?P<json_url>
 316             http://arte\.tv/papi/tvguide/videos/stream/player/
 317             (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
 318         )
 319     '''
 320
 321     def _real_extract(self, url):
 322         mobj = re.match(self._VALID_URL, url)
 323         video_id = mobj.group('id')
 324         lang = mobj.group('lang')
 325         json_url = mobj.group('json_url')
 326         return self._extract_from_json_url(json_url, video_id, lang)