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     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         formats = filter(_match_lang, all_formats)
 113         formats = list(formats) # in python3 filter returns an iterator
 114         if not formats:
 115             # Some videos are only available in the 'Originalversion'
 116             # they aren't tagged as being in French or German
 117             if all(f['versionCode'] == 'VO' or f['versionCode'] == 'VA' for f in all_formats):
 118                 formats = all_formats
 119             else:
 120                 raise ExtractorError(u'The formats list is empty')
 121
 122         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
 123             def sort_key(f):
 124                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
 125         else:
 126             def sort_key(f):
 127                 versionCode = f.get('versionCode')
 128                 if versionCode is None:
 129                     versionCode = ''
 130                 return (
 131                     # Sort first by quality
 132                     int(f.get('height', -1)),
 133                     int(f.get('bitrate', -1)),
 134                     # The original version with subtitles has lower relevance
 135                     re.match(r'VO-ST(F|A)', versionCode) is None,
 136                     # The version with sourds/mal subtitles has also lower relevance
 137                     re.match(r'VO?(F|A)-STM\1', versionCode) is None,
 138                     # Prefer http downloads over m3u8
 139                     0 if f['url'].endswith('m3u8') else 1,
 140                 )
 141         formats = sorted(formats, key=sort_key)
 142         def _format(format_info):
 143             quality = ''
 144             height = format_info.get('height')
 145             if height is not None:
 146                 quality = compat_str(height)
 147             bitrate = format_info.get('bitrate')
 148             if bitrate is not None:
 149                 quality += '-%d' % bitrate
 150             if format_info.get('versionCode') is not None:
 151                 format_id = '%s-%s' % (quality, format_info['versionCode'])
 152             else:
 153                 format_id = quality
 154             info = {
 155                 'format_id': format_id,
 156                 'format_note': format_info.get('versionLibelle'),
 157                 'width': format_info.get('width'),
 158                 'height': height,
 159             }
 160             if format_info['mediaType'] == 'rtmp':
 161                 info['url'] = format_info['streamer']
 162                 info['play_path'] = 'mp4:' + format_info['url']
 163                 info['ext'] = 'flv'
 164             else:
 165                 info['url'] = format_info['url']
 166                 info['ext'] = determine_ext(info['url'])
 167             return info
 168         info_dict['formats'] = [_format(f) for f in formats]
 169
 170         return info_dict
 171
 172
 173 # It also uses the arte_vp_url url from the webpage to extract the information
 174 class ArteTVCreativeIE(ArteTVPlus7IE):
 175     IE_NAME = 'arte.tv:creative'
 176     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 177
 178     _TEST = {
 179         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
 180         'info_dict': {
 181             'id': '050489-002',
 182             'ext': 'mp4',
 183             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
 184         },
 185     }
 186
 187
 188 class ArteTVFutureIE(ArteTVPlus7IE):
 189     IE_NAME = 'arte.tv:future'
 190     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
 191
 192     _TEST = {
 193         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
 194         'info_dict': {
 195             'id': '5201',
 196             'ext': 'mp4',
 197             'title': 'Les champignons au secours de la planète',
 198             'upload_date': '20131101',
 199         },
 200     }
 201
 202     def _real_extract(self, url):
 203         anchor_id, lang = self._extract_url_info(url)
 204         webpage = self._download_webpage(url, anchor_id)
 205         row = get_element_by_id(anchor_id, webpage)
 206         return self._extract_from_webpage(row, anchor_id, lang)
 207
 208
 209 class ArteTVDDCIE(ArteTVPlus7IE):
 210     IE_NAME = 'arte.tv:ddc'
 211     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
 212
 213     def _real_extract(self, url):
 214         video_id, lang = self._extract_url_info(url)
 215         if lang == 'folge':
 216             lang = 'de'
 217         elif lang == 'emission':
 218             lang = 'fr'
 219         webpage = self._download_webpage(url, video_id)
 220         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
 221         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
 222         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
 223         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
 224         return self._extract_from_json_url(json_url, video_id, lang)
 225
 226
 227 class ArteTVConcertIE(ArteTVPlus7IE):
 228     IE_NAME = 'arte.tv:concert'
 229     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
 230
 231     _TEST = {
 232         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
 233         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
 234         'info_dict': {
 235             'id': '186',
 236             'ext': 'mp4',
 237             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
 238             'upload_date': '20140128',
 239             'description': 'md5:486eb08f991552ade77439fe6d82c305',
 240         },
 241     }
 242
 243
 244 class ArteTVEmbedIE(ArteTVPlus7IE):
 245     IE_NAME = 'arte.tv:embed'
 246     _VALID_URL = r'''(?x)
 247         http://www\.arte\.tv
 248         /playerv2/embed\.php\?json_url=
 249         (?P<json_url>
 250             http://arte\.tv/papi/tvguide/videos/stream/player/
 251             (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
 252         )
 253     '''
 254
 255     def _real_extract(self, url):
 256         mobj = re.match(self._VALID_URL, url)
 257         video_id = mobj.group('id')
 258         lang = mobj.group('lang')
 259         json_url = mobj.group('json_url')
 260         return self._extract_from_json_url(json_url, video_id, lang)