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     compat_str,
  14     get_element_by_attribute,
  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="(.*?)"', webpage, 'json vp url')
  82         return self._extract_from_json_url(json_url, video_id, lang)
  83
  84     def _extract_from_json_url(self, json_url, video_id, lang):
  85         info = self._download_json(json_url, video_id)
  86         player_info = info['videoJsonPlayer']
  87
  88         info_dict = {
  89             'id': player_info['VID'],
  90             'title': player_info['VTI'],
  91             'description': player_info.get('VDE'),
  92             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
  93             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
  94         }
  95
  96         all_formats = player_info['VSR'].values()
  97         # Some formats use the m3u8 protocol
  98         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
  99         def _match_lang(f):
 100             if f.get('versionCode') is None:
 101                 return True
 102             # Return true if that format is in the language of the url
 103             if lang == 'fr':
 104                 l = 'F'
 105             elif lang == 'de':
 106                 l = 'A'
 107             else:
 108                 l = lang
 109             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 110             return any(re.match(r, f['versionCode']) for r in regexes)
 111         # Some formats may not be in the same language as the url
 112         # TODO: Might want not to drop videos that does not match requested language
 113         # but to process those formats with lower precedence
 114         formats = filter(_match_lang, all_formats)
 115         formats = list(formats)  # in python3 filter returns an iterator
 116         if not formats:
 117             # Some videos are only available in the 'Originalversion'
 118             # they aren't tagged as being in French or German
 119             # Sometimes there are neither videos of requested lang code
 120             # nor original version videos available
 121             # For such cases we just take all_formats as is
 122             formats = all_formats
 123             if not formats:
 124                 raise ExtractorError('The formats list is empty')
 125
 126         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 127             def sort_key(f):
 128                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 129         else:
 130             def sort_key(f):
 131                 versionCode = f.get('versionCode')
 132                 if versionCode is None:
 133                     versionCode = ''
 134                 return (
 135                     # Sort first by quality
 136                     int(f.get('height', -1)),
 137                     int(f.get('bitrate', -1)),
 138                     # The original version with subtitles has lower relevance
 139                     re.match(r'VO-ST(F|A)', versionCode) is None,
 140                     # The version with sourds/mal subtitles has also lower relevance
 141                     re.match(r'VO?(F|A)-STM\1', versionCode) is None,
 142                     # Prefer http downloads over m3u8
 143                     0 if f['url'].endswith('m3u8') else 1,
 144                 )
 145         formats = sorted(formats, key=sort_key)
 146         def _format(format_info):
 147             quality = ''
 148             height = format_info.get('height')
 149             if height is not None:
 150                 quality = compat_str(height)
 151             bitrate = format_info.get('bitrate')
 152             if bitrate is not None:
 153                 quality += '-%d' % bitrate
 154             if format_info.get('versionCode') is not None:
 155                 format_id = '%s-%s' % (quality, format_info['versionCode'])
 156             else:
 157                 format_id = quality
 158             info = {
 159                 'format_id': format_id,
 160                 'format_note': format_info.get('versionLibelle'),
 161                 'width': format_info.get('width'),
 162                 'height': height,
 163             }
 164             if format_info['mediaType'] == 'rtmp':
 165                 info['url'] = format_info['streamer']
 166                 info['play_path'] = 'mp4:' + format_info['url']
 167                 info['ext'] = 'flv'
 168             else:
 169                 info['url'] = format_info['url']
 170                 info['ext'] = determine_ext(info['url'])
 171             return info
 172         info_dict['formats'] = [_format(f) for f in formats]
 173
 174         return info_dict
 175
 176
 177 # It also uses the arte_vp_url url from the webpage to extract the information
 178 class ArteTVCreativeIE(ArteTVPlus7IE):
 179     IE_NAME = 'arte.tv:creative'
 180     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 181
 182     _TEST = {
 183         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 184         'info_dict': {
 185             'id': '050489-002',
 186             'ext': 'mp4',
 187             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 188         },
 189     }
 190
 191
 192 class ArteTVFutureIE(ArteTVPlus7IE):
 193     IE_NAME = 'arte.tv:future'
 194     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 195
 196     _TEST = {
 197         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 198         'info_dict': {
 199             'id': '5201',
 200             'ext': 'mp4',
 201             'title': 'Les champignons au secours de la planète',
 202             'upload_date': '20131101',
 203         },
 204     }
 205
 206     def _real_extract(self, url):
 207         anchor_id, lang = self._extract_url_info(url)
 208         webpage = self._download_webpage(url, anchor_id)
 209         row = get_element_by_id(anchor_id, webpage)
 210         return self._extract_from_webpage(row, anchor_id, lang)
 211
 212
 213 class ArteTVDDCIE(ArteTVPlus7IE):
 214     IE_NAME = 'arte.tv:ddc'
 215     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 216
 217     def _real_extract(self, url):
 218         video_id, lang = self._extract_url_info(url)
 219         if lang == 'folge':
 220             lang = 'de'
 221         elif lang == 'emission':
 222             lang = 'fr'
 223         webpage = self._download_webpage(url, video_id)
 224         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 225         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 226         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 227         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 228         return self._extract_from_json_url(json_url, video_id, lang)
 229
 230
 231 class ArteTVConcertIE(ArteTVPlus7IE):
 232     IE_NAME = 'arte.tv:concert'
 233     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
 234
 235     _TEST = {
 236         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
 237         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
 238         'info_dict': {
 239             'id': '186',
 240             'ext': 'mp4',
 241             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
 242             'upload_date': '20140128',
 243             'description': 'md5:486eb08f991552ade77439fe6d82c305',
 244         },
 245     }
 246
 247
 248 class ArteTVEmbedIE(ArteTVPlus7IE):
 249     IE_NAME = 'arte.tv:embed'
 250     _VALID_URL = r'''(?x)
 251         http://www\.arte\.tv
 252         /playerv2/embed\.php\?json_url=
 253         (?P<json_url>
 254             http://arte\.tv/papi/tvguide/videos/stream/player/
 255             (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
 256         )
 257     '''
 258
 259     def _real_extract(self, url):
 260         mobj = re.match(self._VALID_URL, url)
 261         video_id = mobj.group('id')
 262         lang = mobj.group('lang')
 263         json_url = mobj.group('json_url')
 264         return self._extract_from_json_url(json_url, video_id, lang)