_ 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
  23 class ArteTvIE(InfoExtractor):
  24     _VALID_URL = r'(?:http://)?videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
  25     IE_NAME = 'arte.tv'
  26
  27     def _real_extract(self, url):
  28         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
  29         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
  30         ref_xml_doc = self._download_xml(
  31             ref_xml_url, video_id, note='Downloading metadata')
  32         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
  33         config_xml_url = config_node.attrib['ref']
  34         config_xml = self._download_webpage(
  35             config_xml_url, video_id, note='Downloading configuration')
  36
  37         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
  38         def _key(m):
  39             quality = m.group('quality')
  40             if quality == 'hd':
  41                 return 2
  42             else:
  43                 return 1
  44         # We pick the best quality
  45         video_urls = sorted(video_urls, key=_key)
  46         video_url = list(video_urls)[-1].group('url')
  47
  48         title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
  49         thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
  50                                             config_xml, 'thumbnail')
  51         return {'id': video_id,
  52                 'title': title,
  53                 'thumbnail': thumbnail,
  54                 'url': video_url,
  55                 'ext': 'flv',
  56         }
  57
  58
  59 class ArteTVPlus7IE(InfoExtractor):
  60     IE_NAME = 'arte.tv:+7'
  61     _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
  62
  63     @classmethod
  64     def _extract_url_info(cls, url):
  65         mobj = re.match(cls._VALID_URL, url)
  66         lang = mobj.group('lang')
  67         # This is not a real id, it can be for example AJT for the news
  68         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
  69         video_id = mobj.group('id')
  70         return video_id, lang
  71
  72     def _real_extract(self, url):
  73         video_id, lang = self._extract_url_info(url)
  74         webpage = self._download_webpage(url, video_id)
  75         return self._extract_from_webpage(webpage, video_id, lang)
  76
  77     def _extract_from_webpage(self, webpage, video_id, lang):
  78         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
  79         return self._extract_from_json_url(json_url, video_id, lang)
  80
  81     def _extract_from_json_url(self, json_url, video_id, lang):
  82         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
  83         self.report_extraction(video_id)
  84         info = json.loads(json_info)
  85         player_info = info['videoJsonPlayer']
  86
  87         info_dict = {
  88             'id': player_info['VID'],
  89             'title': player_info['VTI'],
  90             'description': player_info.get('VDE'),
  91             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
  92             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
  93         }
  94
  95         all_formats = player_info['VSR'].values()
  96         # Some formats use the m3u8 protocol
  97         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
  98         def _match_lang(f):
  99             if f.get('versionCode') is None:
 100                 return True
 101             # Return true if that format is in the language of the url
 102             if lang == 'fr':
 103                 l = 'F'
 104             elif lang == 'de':
 105                 l = 'A'
 106             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
 107             return any(re.match(r, f['versionCode']) for r in regexes)
 108         # Some formats may not be in the same language as the url
 109         formats = filter(_match_lang, all_formats)
 110         formats = list(formats) # in python3 filter returns an iterator
 111         if not formats:
 112             # Some videos are only available in the 'Originalversion'
 113             # they aren't tagged as being in French or German
 114             if all(f['versionCode'] == 'VO' for f in all_formats):
 115                 formats = all_formats
 116             else:
 117                 raise ExtractorError(u'The formats list is empty')
 118
 119         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 120             def sort_key(f):
 121                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 122         else:
 123             def sort_key(f):
 124                 return (
 125                     # Sort first by quality
 126                     int(f.get('height',-1)),
 127                     int(f.get('bitrate',-1)),
 128                     # The original version with subtitles has lower relevance
 129                     re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
 130                     # The version with sourds/mal subtitles has also lower relevance
 131                     re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
 132                     # Prefer http downloads over m3u8
 133                     0 if f['url'].endswith('m3u8') else 1,
 134                 )
 135         formats = sorted(formats, key=sort_key)
 136         def _format(format_info):
 137             quality = ''
 138             height = format_info.get('height')
 139             if height is not None:
 140                 quality = compat_str(height)
 141             bitrate = format_info.get('bitrate')
 142             if bitrate is not None:
 143                 quality += '-%d' % bitrate
 144             if format_info.get('versionCode') is not None:
 145                 format_id = '%s-%s' % (quality, format_info['versionCode'])
 146             else:
 147                 format_id = quality
 148             info = {
 149                 'format_id': format_id,
 150                 'format_note': format_info.get('versionLibelle'),
 151                 'width': format_info.get('width'),
 152                 'height': height,
 153             }
 154             if format_info['mediaType'] == 'rtmp':
 155                 info['url'] = format_info['streamer']
 156                 info['play_path'] = 'mp4:' + format_info['url']
 157                 info['ext'] = 'flv'
 158             else:
 159                 info['url'] = format_info['url']
 160                 info['ext'] = determine_ext(info['url'])
 161             return info
 162         info_dict['formats'] = [_format(f) for f in formats]
 163
 164         return info_dict
 165
 166
 167 # It also uses the arte_vp_url url from the webpage to extract the information
 168 class ArteTVCreativeIE(ArteTVPlus7IE):
 169     IE_NAME = 'arte.tv:creative'
 170     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 171
 172     _TEST = {
 173         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 174         'info_dict': {
 175             'id': '050489-002',
 176             'ext': 'mp4',
 177             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 178         },
 179     }
 180
 181
 182 class ArteTVFutureIE(ArteTVPlus7IE):
 183     IE_NAME = 'arte.tv:future'
 184     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 185
 186     _TEST = {
 187         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 188         'info_dict': {
 189             'id': '050940-003',
 190             'ext': 'mp4',
 191             'title': 'Les champignons au secours de la planète',
 192         },
 193     }
 194
 195     def _real_extract(self, url):
 196         anchor_id, lang = self._extract_url_info(url)
 197         webpage = self._download_webpage(url, anchor_id)
 198         row = get_element_by_id(anchor_id, webpage)
 199         return self._extract_from_webpage(row, anchor_id, lang)
 200
 201
 202 class ArteTVDDCIE(ArteTVPlus7IE):
 203     IE_NAME = 'arte.tv:ddc'
 204     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 205
 206     def _real_extract(self, url):
 207         video_id, lang = self._extract_url_info(url)
 208         if lang == 'folge':
 209             lang = 'de'
 210         elif lang == 'emission':
 211             lang = 'fr'
 212         webpage = self._download_webpage(url, video_id)
 213         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 214         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 215         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 216         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 217         return self._extract_from_json_url(json_url, video_id, lang)
 218
 219
 220 class ArteTVConcertIE(ArteTVPlus7IE):
 221     IE_NAME = 'arte.tv:concert'
 222     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
 223
 224     _TEST = {
 225         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
 226         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
 227         'info_dict': {
 228             'id': '186',
 229             'ext': 'mp4',
 230             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
 231             'upload_date': '20140128',
 232             'description': 'md5:486eb08f991552ade77439fe6d82c305',
 233         },
 234     }