youtube_dlc/extractor/viki.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import base64
   5 import hashlib
   6 import hmac
   7 import itertools
   8 import json
   9 import re
  10 import time
  11
  12 from .common import InfoExtractor
  13 from ..compat import (
  14     compat_parse_qs,
  15     compat_urllib_parse_urlparse,
  16 )
  17 from ..utils import (
  18     ExtractorError,
  19     int_or_none,
  20     HEADRequest,
  21     parse_age_limit,
  22     parse_iso8601,
  23     sanitized_Request,
  24     std_headers,
  25 )
  26
  27
  28 class VikiBaseIE(InfoExtractor):
  29     _VALID_URL_BASE = r'https?://(?:www\.)?viki\.(?:com|net|mx|jp|fr)/'
  30     _API_QUERY_TEMPLATE = '/v4/%sapp=%s&t=%s&site=www.viki.com'
  31     _API_URL_TEMPLATE = 'https://api.viki.io%s&sig=%s'
  32
  33     _APP = '100005a'
  34     _APP_VERSION = '6.0.0'
  35     _APP_SECRET = 'MM_d*yP@`&1@]@!AVrXf_o-HVEnoTnm$O-ti4[G~$JDI/Dc-&piU&z&5.;:}95=Iad'
  36
  37     _GEO_BYPASS = False
  38     _NETRC_MACHINE = 'viki'
  39
  40     _token = None
  41
  42     _ERRORS = {
  43         'geo': 'Sorry, this content is not available in your region.',
  44         'upcoming': 'Sorry, this content is not yet available.',
  45         # 'paywall': 'paywall',
  46     }
  47
  48     def _prepare_call(self, path, timestamp=None, post_data=None):
  49         path += '?' if '?' not in path else '&'
  50         if not timestamp:
  51             timestamp = int(time.time())
  52         query = self._API_QUERY_TEMPLATE % (path, self._APP, timestamp)
  53         if self._token:
  54             query += '&token=%s' % self._token
  55         sig = hmac.new(
  56             self._APP_SECRET.encode('ascii'),
  57             query.encode('ascii'),
  58             hashlib.sha1
  59         ).hexdigest()
  60         url = self._API_URL_TEMPLATE % (query, sig)
  61         return sanitized_Request(
  62             url, json.dumps(post_data).encode('utf-8')) if post_data else url
  63
  64     def _call_api(self, path, video_id, note, timestamp=None, post_data=None):
  65         resp = self._download_json(
  66             self._prepare_call(path, timestamp, post_data),
  67             video_id, note,
  68             headers={
  69                 'x-client-user-agent': std_headers['User-Agent'],
  70                 'x-viki-as-id': self._APP,
  71                 'x-viki-app-ver': self._APP_VERSION,
  72             })
  73
  74         error = resp.get('error')
  75         if error:
  76             if error == 'invalid timestamp':
  77                 resp = self._download_json(
  78                     self._prepare_call(path, int(resp['current_timestamp']), post_data),
  79                     video_id, '%s (retry)' % note,
  80                     headers={
  81                         'x-client-user-agent': std_headers['User-Agent'],
  82                         'x-viki-as-id': self._APP,
  83                         'x-viki-app-ver': self._APP_VERSION,
  84                     })
  85                 error = resp.get('error')
  86             if error:
  87                 self._raise_error(resp['error'])
  88
  89         return resp
  90
  91     def _raise_error(self, error):
  92         raise ExtractorError(
  93             '%s returned error: %s' % (self.IE_NAME, error),
  94             expected=True)
  95
  96     def _check_errors(self, data):
  97         for reason, status in data.get('blocking', {}).items():
  98             if status and reason in self._ERRORS:
  99                 message = self._ERRORS[reason]
 100                 if reason == 'geo':
 101                     self.raise_geo_restricted(msg=message)
 102                 raise ExtractorError('%s said: %s' % (
 103                     self.IE_NAME, message), expected=True)
 104
 105     def _real_initialize(self):
 106         self._login()
 107
 108     def _login(self):
 109         username, password = self._get_login_info()
 110         if username is None:
 111             return
 112
 113         login_form = {
 114             'login_id': username,
 115             'password': password,
 116         }
 117
 118         login = self._call_api(
 119             'sessions.json', None,
 120             'Logging in', post_data=login_form)
 121
 122         self._token = login.get('token')
 123         if not self._token:
 124             self.report_warning('Unable to get session token, login has probably failed')
 125
 126     @staticmethod
 127     def dict_selection(dict_obj, preferred_key, allow_fallback=True):
 128         if preferred_key in dict_obj:
 129             return dict_obj.get(preferred_key)
 130
 131         if not allow_fallback:
 132             return
 133
 134         filtered_dict = list(filter(None, [dict_obj.get(k) for k in dict_obj.keys()]))
 135         return filtered_dict[0] if filtered_dict else None
 136
 137
 138 class VikiIE(VikiBaseIE):
 139     IE_NAME = 'viki'
 140     _VALID_URL = r'%s(?:videos|player)/(?P<id>[0-9]+v)' % VikiBaseIE._VALID_URL_BASE
 141     _TESTS = [{
 142         'url': 'http://www.viki.com/videos/1023585v-heirs-episode-14',
 143         'info_dict': {
 144             'id': '1023585v',
 145             'ext': 'mp4',
 146             'title': 'Heirs Episode 14',
 147             'uploader': 'SBS',
 148             'description': 'md5:c4b17b9626dd4b143dcc4d855ba3474e',
 149             'upload_date': '20131121',
 150             'age_limit': 13,
 151         },
 152         'skip': 'Blocked in the US',
 153     }, {
 154         # clip
 155         'url': 'http://www.viki.com/videos/1067139v-the-avengers-age-of-ultron-press-conference',
 156         'md5': '86c0b5dbd4d83a6611a79987cc7a1989',
 157         'info_dict': {
 158             'id': '1067139v',
 159             'ext': 'mp4',
 160             'title': "'The Avengers: Age of Ultron' Press Conference",
 161             'description': 'md5:d70b2f9428f5488321bfe1db10d612ea',
 162             'duration': 352,
 163             'timestamp': 1430380829,
 164             'upload_date': '20150430',
 165             'uploader': 'Arirang TV',
 166             'like_count': int,
 167             'age_limit': 0,
 168         }
 169     }, {
 170         'url': 'http://www.viki.com/videos/1048879v-ankhon-dekhi',
 171         'info_dict': {
 172             'id': '1048879v',
 173             'ext': 'mp4',
 174             'title': 'Ankhon Dekhi',
 175             'duration': 6512,
 176             'timestamp': 1408532356,
 177             'upload_date': '20140820',
 178             'uploader': 'Spuul',
 179             'like_count': int,
 180             'age_limit': 13,
 181         },
 182         'skip': 'Blocked in the US',
 183     }, {
 184         # episode
 185         'url': 'http://www.viki.com/videos/44699v-boys-over-flowers-episode-1',
 186         'md5': '94e0e34fd58f169f40c184f232356cfe',
 187         'info_dict': {
 188             'id': '44699v',
 189             'ext': 'mp4',
 190             'title': 'Boys Over Flowers - Episode 1',
 191             'description': 'md5:b89cf50038b480b88b5b3c93589a9076',
 192             'duration': 4172,
 193             'timestamp': 1270496524,
 194             'upload_date': '20100405',
 195             'uploader': 'group8',
 196             'like_count': int,
 197             'age_limit': 13,
 198         },
 199         'expected_warnings': ['Unknown MIME type image/jpeg in DASH manifest'],
 200     }, {
 201         # youtube external
 202         'url': 'http://www.viki.com/videos/50562v-poor-nastya-complete-episode-1',
 203         'md5': '63f8600c1da6f01b7640eee7eca4f1da',
 204         'info_dict': {
 205             'id': '50562v',
 206             'ext': 'webm',
 207             'title': 'Poor Nastya [COMPLETE] - Episode 1',
 208             'description': '',
 209             'duration': 606,
 210             'timestamp': 1274949505,
 211             'upload_date': '20101213',
 212             'uploader': 'ad14065n',
 213             'uploader_id': 'ad14065n',
 214             'like_count': int,
 215             'age_limit': 13,
 216         },
 217         'skip': 'Page not found!',
 218     }, {
 219         'url': 'http://www.viki.com/player/44699v',
 220         'only_matching': True,
 221     }, {
 222         # non-English description
 223         'url': 'http://www.viki.com/videos/158036v-love-in-magic',
 224         'md5': 'adf9e321a0ae5d0aace349efaaff7691',
 225         'info_dict': {
 226             'id': '158036v',
 227             'ext': 'mp4',
 228             'uploader': 'I Planet Entertainment',
 229             'upload_date': '20111122',
 230             'timestamp': 1321985454,
 231             'description': 'md5:44b1e46619df3a072294645c770cef36',
 232             'title': 'Love In Magic',
 233             'age_limit': 13,
 234         },
 235     }]
 236
 237     def _real_extract(self, url):
 238         video_id = self._match_id(url)
 239
 240         resp = self._download_json(
 241             'https://www.viki.com/api/videos/' + video_id,
 242             video_id, 'Downloading video JSON', headers={
 243                 'x-client-user-agent': std_headers['User-Agent'],
 244                 'x-viki-as-id': self._APP,
 245                 'x-viki-app-ver': self._APP_VERSION,
 246             })
 247         video = resp['video']
 248
 249         self._check_errors(video)
 250
 251         title = self.dict_selection(video.get('titles', {}), 'en', allow_fallback=False)
 252         if not title:
 253             title = 'Episode %d' % video.get('number') if video.get('type') == 'episode' else video.get('id') or video_id
 254             container_titles = video.get('container', {}).get('titles', {})
 255             container_title = self.dict_selection(container_titles, 'en')
 256             title = '%s - %s' % (container_title, title)
 257
 258         description = self.dict_selection(video.get('descriptions', {}), 'en')
 259
 260         duration = int_or_none(video.get('duration'))
 261         timestamp = parse_iso8601(video.get('created_at'))
 262         uploader = video.get('author')
 263         like_count = int_or_none(video.get('likes', {}).get('count'))
 264         age_limit = parse_age_limit(video.get('rating'))
 265
 266         thumbnails = []
 267         for thumbnail_id, thumbnail in video.get('images', {}).items():
 268             thumbnails.append({
 269                 'id': thumbnail_id,
 270                 'url': thumbnail.get('url'),
 271             })
 272
 273         subtitles = {}
 274         try:
 275             # New way to fetch subtitles
 276             new_video = self._download_json(
 277                 'https://www.viki.com/api/videos/%s' % video_id, video_id,
 278                 'Downloading new video JSON to get subtitles', fatal=False,
 279                 headers={
 280                     'x-client-user-agent': std_headers['User-Agent'],
 281                     'x-viki-as-id': self._APP,
 282                     'x-viki-app-ver': self._APP_VERSION,
 283                 })
 284             for sub in new_video.get('streamSubtitles').get('dash'):
 285                 subtitles[sub.get('srclang')] = [{
 286                     'ext': 'vtt',
 287                     'url': sub.get('src'),
 288                     'completion': sub.get('percentage'),
 289                 }]
 290         except AttributeError:
 291             # fall-back to the old way if there isn't a streamSubtitles attribute
 292             for subtitle_lang, _ in video.get('subtitle_completions', {}).items():
 293                 subtitles[subtitle_lang] = [{
 294                     'ext': subtitles_format,
 295                     'url': self._prepare_call(
 296                         'videos/%s/subtitles/%s.%s' % (video_id, subtitle_lang, subtitles_format)),
 297                 } for subtitles_format in ('srt', 'vtt')]
 298
 299         result = {
 300             'id': video_id,
 301             'title': title,
 302             'description': description,
 303             'duration': duration,
 304             'timestamp': timestamp,
 305             'uploader': uploader,
 306             'like_count': like_count,
 307             'age_limit': age_limit,
 308             'thumbnails': thumbnails,
 309             'subtitles': subtitles,
 310         }
 311
 312         formats = []
 313
 314         def add_format(format_id, format_dict, protocol='http'):
 315             # rtmps URLs does not seem to work
 316             if protocol == 'rtmps':
 317                 return
 318             format_url = format_dict.get('url')
 319             if not format_url:
 320                 return
 321             format_drms = format_dict.get('drms')
 322             format_stream_id = format_dict.get('id')
 323             qs = compat_parse_qs(compat_urllib_parse_urlparse(format_url).query)
 324             stream = qs.get('stream', [None])[0]
 325             if stream:
 326                 format_url = base64.b64decode(stream).decode()
 327             if format_id in ('m3u8', 'hls'):
 328                 m3u8_formats = self._extract_m3u8_formats(
 329                     format_url, video_id, 'mp4',
 330                     entry_protocol='m3u8_native',
 331                     m3u8_id='m3u8-%s' % protocol, fatal=False)
 332                 # Despite CODECS metadata in m3u8 all video-only formats
 333                 # are actually video+audio
 334                 for f in m3u8_formats:
 335                     if not self._downloader.params.get('allow_unplayable_formats') and '_drm/index_' in f['url']:
 336                         continue
 337                     if f.get('acodec') == 'none' and f.get('vcodec') != 'none':
 338                         f['acodec'] = None
 339                     formats.append(f)
 340             elif format_id in ('mpd', 'dash'):
 341                 formats.extend(self._extract_mpd_formats(
 342                     format_url, video_id, 'mpd-%s' % protocol, fatal=False))
 343             elif format_url.startswith('rtmp'):
 344                 mobj = re.search(
 345                     r'^(?P<url>rtmp://[^/]+/(?P<app>.+?))/(?P<playpath>mp4:.+)$',
 346                     format_url)
 347                 if not mobj:
 348                     return
 349                 formats.append({
 350                     'format_id': 'rtmp-%s' % format_id,
 351                     'ext': 'flv',
 352                     'url': mobj.group('url'),
 353                     'play_path': mobj.group('playpath'),
 354                     'app': mobj.group('app'),
 355                     'page_url': url,
 356                     'drms': format_drms,
 357                     'stream_id': format_stream_id,
 358                 })
 359             else:
 360                 urlh = self._request_webpage(
 361                     HEADRequest(format_url), video_id, 'Checking file size', fatal=False)
 362                 formats.append({
 363                     'url': format_url,
 364                     'format_id': '%s-%s' % (format_id, protocol),
 365                     'height': int_or_none(self._search_regex(
 366                         r'^(\d+)[pP]$', format_id, 'height', default=None)),
 367                     'drms': format_drms,
 368                     'stream_id': format_stream_id,
 369                     'filesize': int_or_none(urlh.headers.get('Content-Length')),
 370                 })
 371
 372         for format_id, format_dict in (resp.get('streams') or {}).items():
 373             add_format(format_id, format_dict)
 374         if not formats:
 375             streams = self._call_api(
 376                 'videos/%s/streams.json' % video_id, video_id,
 377                 'Downloading video streams JSON')
 378
 379             if 'external' in streams:
 380                 result.update({
 381                     '_type': 'url_transparent',
 382                     'url': streams['external']['url'],
 383                 })
 384                 return result
 385
 386             for format_id, stream_dict in streams.items():
 387                 for protocol, format_dict in stream_dict.items():
 388                     add_format(format_id, format_dict, protocol)
 389         self._sort_formats(formats)
 390
 391         result['formats'] = formats
 392         return result
 393
 394
 395 class VikiChannelIE(VikiBaseIE):
 396     IE_NAME = 'viki:channel'
 397     _VALID_URL = r'%s(?:tv|news|movies|artists)/(?P<id>[0-9]+c)' % VikiBaseIE._VALID_URL_BASE
 398     _TESTS = [{
 399         'url': 'http://www.viki.com/tv/50c-boys-over-flowers',
 400         'info_dict': {
 401             'id': '50c',
 402             'title': 'Boys Over Flowers',
 403             'description': 'md5:ecd3cff47967fe193cff37c0bec52790',
 404         },
 405         'playlist_mincount': 71,
 406     }, {
 407         'url': 'http://www.viki.com/tv/1354c-poor-nastya-complete',
 408         'info_dict': {
 409             'id': '1354c',
 410             'title': 'Poor Nastya [COMPLETE]',
 411             'description': 'md5:05bf5471385aa8b21c18ad450e350525',
 412         },
 413         'playlist_count': 127,
 414     }, {
 415         'url': 'http://www.viki.com/news/24569c-showbiz-korea',
 416         'only_matching': True,
 417     }, {
 418         'url': 'http://www.viki.com/movies/22047c-pride-and-prejudice-2005',
 419         'only_matching': True,
 420     }, {
 421         'url': 'http://www.viki.com/artists/2141c-shinee',
 422         'only_matching': True,
 423     }]
 424
 425     _PER_PAGE = 25
 426
 427     def _real_extract(self, url):
 428         channel_id = self._match_id(url)
 429
 430         channel = self._call_api(
 431             'containers/%s.json' % channel_id, channel_id,
 432             'Downloading channel JSON')
 433
 434         self._check_errors(channel)
 435
 436         title = self.dict_selection(channel['titles'], 'en')
 437
 438         description = self.dict_selection(channel['descriptions'], 'en')
 439
 440         entries = []
 441         for video_type in ('episodes', 'clips', 'movies'):
 442             for page_num in itertools.count(1):
 443                 page = self._call_api(
 444                     'containers/%s/%s.json?per_page=%d&sort=number&direction=asc&with_paging=true&page=%d'
 445                     % (channel_id, video_type, self._PER_PAGE, page_num), channel_id,
 446                     'Downloading %s JSON page #%d' % (video_type, page_num))
 447                 for video in page['response']:
 448                     video_id = video['id']
 449                     entries.append(self.url_result(
 450                         'https://www.viki.com/videos/%s' % video_id, 'Viki'))
 451                 if not page['pagination']['next']:
 452                     break
 453
 454         return self.playlist_result(entries, channel_id, title, description)