Raphaël G. Git Repositories - youtubedl/blob - 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 ..utils import (
   8     ExtractorError,
   9     find_xpath_attr,
  10     unified_strdate,
  11     determine_ext,
  12     get_element_by_id,
  13     get_element_by_attribute,
  14     int_or_none,
  15 )
  16
  17 # There are different sources of video in arte.tv, the extraction process
  18 # is different for each one. The videos usually expire in 7 days, so we can't
  19 # add tests.
  20
  21
  22 class ArteTvIE(InfoExtractor):
  23     _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
  24     IE_NAME = 'arte.tv'
  25
  26     def _real_extract(self, url):
  27         mobj = re.match(self._VALID_URL, url)
  28         lang = mobj.group('lang')
  29         video_id = mobj.group('id')
  30
  31         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  32         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  33         ref_xml_doc = self._download_xml(
  34             ref_xml_url, video_id, note='Downloading metadata')
  35         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  36         config_xml_url = config_node.attrib['ref']
  37         config = self._download_xml(
  38             config_xml_url, video_id, note='Downloading configuration')
  39
  40         formats = [{
  41             'forma_id': q.attrib['quality'],
  42             # The playpath starts at 'mp4:', if we don't manually
  43             # split the url, rtmpdump will incorrectly parse them
  44             'url': q.text.split('mp4:', 1)[0],
  45             'play_path': 'mp4:' + q.text.split('mp4:', 1)[1],
  46             'ext': 'flv',
  47             'quality': 2 if q.attrib['quality'] == 'hd' else 1,
  48         } for q in config.findall('./urls/url')]
  49         self._sort_formats(formats)
  50
  51         title = config.find('.//name').text
  52         thumbnail = config.find('.//firstThumbnailUrl').text
  53         return {
  54             'id': video_id,
  55             'title': title,
  56             'thumbnail': thumbnail,
  57             'formats': formats,
  58         }
  59
  60
  61 class ArteTVPlus7IE(InfoExtractor):
  62     IE_NAME = 'arte.tv:+7'
  63     _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
  64
  65     @classmethod
  66     def _extract_url_info(cls, url):
  67         mobj = re.match(cls._VALID_URL, url)
  68         lang = mobj.group('lang')
  69         # This is not a real id, it can be for example AJT for the news
  70         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
  71         video_id = mobj.group('id')
  72         return video_id, lang
  73
  74     def _real_extract(self, url):
  75         video_id, lang = self._extract_url_info(url)
  76         webpage = self._download_webpage(url, video_id)
  77         return self._extract_from_webpage(webpage, video_id, lang)
  78
  79     def _extract_from_webpage(self, webpage, video_id, lang):
  80         json_url = self._html_search_regex(
  81             [r'arte_vp_url=["\'](.*?)["\']', r'data-url=["\']([^"]+)["\']'],
  82             webpage, 'json vp url')
  83         return self._extract_from_json_url(json_url, video_id, lang)
  84
  85     def _extract_from_json_url(self, json_url, video_id, lang):
  86         info = self._download_json(json_url, video_id)
  87         player_info = info['videoJsonPlayer']
  88
  89         upload_date_str = player_info.get('shootingDate')
  90         if not upload_date_str:
  91             upload_date_str = player_info.get('VDA', '').split(' ')[0]
  92
  93         title = player_info['VTI'].strip()
  94         subtitle = player_info.get('VSU', '').strip()
  95         if subtitle:
  96             title += ' - %s' % subtitle
  97
  98         info_dict = {
  99             'id': player_info['VID'],
 100             'title': title,
 101             'description': player_info.get('VDE'),
 102             'upload_date': unified_strdate(upload_date_str),
 103             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
 104         }
 105
 106         all_formats = []
 107         for format_id, format_dict in player_info['VSR'].items():
 108             fmt = dict(format_dict)
 109             fmt['format_id'] = format_id
 110             all_formats.append(fmt)
 111         # Some formats use the m3u8 protocol
 112         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
 113         def _match_lang(f):
 114             if f.get('versionCode') is None:
 115                 return True
 116             # Return true if that format is in the language of the url
 117             if lang == 'fr':
 118                 l = 'F'
 119             elif lang == 'de':
 120                 l = 'A'
 121             else:
 122                 l = lang
 123             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 124             return any(re.match(r, f['versionCode']) for r in regexes)
 125         # Some formats may not be in the same language as the url
 126         # TODO: Might want not to drop videos that does not match requested language
 127         # but to process those formats with lower precedence
 128         formats = filter(_match_lang, all_formats)
 129         formats = list(formats)  # in python3 filter returns an iterator
 130         if not formats:
 131             # Some videos are only available in the 'Originalversion'
 132             # they aren't tagged as being in French or German
 133             # Sometimes there are neither videos of requested lang code
 134             # nor original version videos available
 135             # For such cases we just take all_formats as is
 136             formats = all_formats
 137             if not formats:
 138                 raise ExtractorError('The formats list is empty')
 139
 140         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 141             def sort_key(f):
 142                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 143         else:
 144             def sort_key(f):
 145                 versionCode = f.get('versionCode')
 146                 if versionCode is None:
 147                     versionCode = ''
 148                 return (
 149                     # Sort first by quality
 150                     int(f.get('height', -1)),
 151                     int(f.get('bitrate', -1)),
 152                     # The original version with subtitles has lower relevance
 153                     re.match(r'VO-ST(F|A)', versionCode) is None,
 154                     # The version with sourds/mal subtitles has also lower relevance
 155                     re.match(r'VO?(F|A)-STM\1', versionCode) is None,
 156                     # Prefer http downloads over m3u8
 157                     0 if f['url'].endswith('m3u8') else 1,
 158                 )
 159         formats = sorted(formats, key=sort_key)
 160         def _format(format_info):
 161             info = {
 162                 'format_id': format_info['format_id'],
 163                 'format_note': '%s, %s' % (format_info.get('versionCode'), format_info.get('versionLibelle')),
 164                 'width': int_or_none(format_info.get('width')),
 165                 'height': int_or_none(format_info.get('height')),
 166                 'tbr': int_or_none(format_info.get('bitrate')),
 167             }
 168             if format_info['mediaType'] == 'rtmp':
 169                 info['url'] = format_info['streamer']
 170                 info['play_path'] = 'mp4:' + format_info['url']
 171                 info['ext'] = 'flv'
 172             else:
 173                 info['url'] = format_info['url']
 174                 info['ext'] = determine_ext(info['url'])
 175             return info
 176         info_dict['formats'] = [_format(f) for f in formats]
 177
 178         return info_dict
 179
 180
 181 # It also uses the arte_vp_url url from the webpage to extract the information
 182 class ArteTVCreativeIE(ArteTVPlus7IE):
 183     IE_NAME = 'arte.tv:creative'
 184     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/(?:magazine?/)?(?P<id>[^?#]+)'
 185
 186     _TESTS = [{
 187         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 188         'info_dict': {
 189             'id': '72176',
 190             'ext': 'mp4',
 191             'title': 'Folge 2 - Corporate Design',
 192             'upload_date': '20131004',
 193         },
 194     }, {
 195         'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion',
 196         'info_dict': {
 197             'id': '160676',
 198             'ext': 'mp4',
 199             'title': 'Monty Python live (mostly)',
 200             'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n',
 201             'upload_date': '20140805',
 202         }
 203     }]
 204
 205
 206 class ArteTVFutureIE(ArteTVPlus7IE):
 207     IE_NAME = 'arte.tv:future'
 208     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 209
 210     _TEST = {
 211         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 212         'info_dict': {
 213             'id': '5201',
 214             'ext': 'mp4',
 215             'title': 'Les champignons au secours de la planète',
 216             'upload_date': '20131101',
 217         },
 218     }
 219
 220     def _real_extract(self, url):
 221         anchor_id, lang = self._extract_url_info(url)
 222         webpage = self._download_webpage(url, anchor_id)
 223         row = get_element_by_id(anchor_id, webpage)
 224         return self._extract_from_webpage(row, anchor_id, lang)
 225
 226
 227 class ArteTVDDCIE(ArteTVPlus7IE):
 228     IE_NAME = 'arte.tv:ddc'
 229     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 230
 231     def _real_extract(self, url):
 232         video_id, lang = self._extract_url_info(url)
 233         if lang == 'folge':
 234             lang = 'de'
 235         elif lang == 'emission':
 236             lang = 'fr'
 237         webpage = self._download_webpage(url, video_id)
 238         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 239         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 240         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 241         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 242         return self._extract_from_json_url(json_url, video_id, lang)
 243
 244
 245 class ArteTVConcertIE(ArteTVPlus7IE):
 246     IE_NAME = 'arte.tv:concert'
 247     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
 248
 249     _TEST = {
 250         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
 251         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
 252         'info_dict': {
 253             'id': '186',
 254             'ext': 'mp4',
 255             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
 256             'upload_date': '20140128',
 257             'description': 'md5:486eb08f991552ade77439fe6d82c305',
 258         },
 259     }
 260
 261
 262 class ArteTVEmbedIE(ArteTVPlus7IE):
 263     IE_NAME = 'arte.tv:embed'
 264     _VALID_URL = r'''(?x)
 265         http://www\.arte\.tv
 266         /playerv2/embed\.php\?json_url=
 267         (?P<json_url>
 268             http://arte\.tv/papi/tvguide/videos/stream/player/
 269             (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
 270         )
 271     '''
 272
 273     def _real_extract(self, url):
 274         mobj = re.match(self._VALID_URL, url)
 275         video_id = mobj.group('id')
 276         lang = mobj.group('lang')
 277         json_url = mobj.group('json_url')
 278         return self._extract_from_json_url(json_url, video_id, lang)