yt_dlp/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     try_get,
  26 )
  27
  28
  29 class VikiBaseIE(InfoExtractor):
  30     _VALID_URL_BASE = r'https?://(?:www\.)?viki\.(?:com|net|mx|jp|fr)/'
  31     _API_QUERY_TEMPLATE = '/v4/%sapp=%s&t=%s&site=www.viki.com'
  32     _API_URL_TEMPLATE = 'https://api.viki.io%s&sig=%s'
  33
  34     _APP = '100005a'
  35     _APP_VERSION = '6.0.0'
  36     _APP_SECRET = 'MM_d*yP@`&1@]@!AVrXf_o-HVEnoTnm$O-ti4[G~$JDI/Dc-&piU&z&5.;:}95=Iad'
  37
  38     _GEO_BYPASS = False
  39     _NETRC_MACHINE = 'viki'
  40
  41     _token = None
  42
  43     _ERRORS = {
  44         'geo': 'Sorry, this content is not available in your region.',
  45         'upcoming': 'Sorry, this content is not yet available.',
  46         'paywall': 'Sorry, this content is only available to Viki Pass Plus subscribers',
  47     }
  48
  49     def _prepare_call(self, path, timestamp=None, post_data=None):
  50         path += '?' if '?' not in path else '&'
  51         if not timestamp:
  52             timestamp = int(time.time())
  53         query = self._API_QUERY_TEMPLATE % (path, self._APP, timestamp)
  54         if self._token:
  55             query += '&token=%s' % self._token
  56         sig = hmac.new(
  57             self._APP_SECRET.encode('ascii'),
  58             query.encode('ascii'),
  59             hashlib.sha1
  60         ).hexdigest()
  61         url = self._API_URL_TEMPLATE % (query, sig)
  62         return sanitized_Request(
  63             url, json.dumps(post_data).encode('utf-8')) if post_data else url
  64
  65     def _call_api(self, path, video_id, note, timestamp=None, post_data=None):
  66         resp = self._download_json(
  67             self._prepare_call(path, timestamp, post_data),
  68             video_id, note,
  69             headers={
  70                 'x-client-user-agent': std_headers['User-Agent'],
  71                 'x-viki-as-id': self._APP,
  72                 'x-viki-app-ver': self._APP_VERSION,
  73             })
  74
  75         error = resp.get('error')
  76         if error:
  77             if error == 'invalid timestamp':
  78                 resp = self._download_json(
  79                     self._prepare_call(path, int(resp['current_timestamp']), post_data),
  80                     video_id, '%s (retry)' % note,
  81                     headers={
  82                         'x-client-user-agent': std_headers['User-Agent'],
  83                         'x-viki-as-id': self._APP,
  84                         'x-viki-app-ver': self._APP_VERSION,
  85                     })
  86                 error = resp.get('error')
  87             if error:
  88                 self._raise_error(resp['error'])
  89
  90         return resp
  91
  92     def _raise_error(self, error):
  93         raise ExtractorError(
  94             '%s returned error: %s' % (self.IE_NAME, error),
  95             expected=True)
  96
  97     def _check_errors(self, data):
  98         for reason, status in (data.get('blocking') or {}).items():
  99             if status and reason in self._ERRORS:
 100                 message = self._ERRORS[reason]
 101                 if reason == 'geo':
 102                     self.raise_geo_restricted(msg=message)
 103                 elif reason == 'paywall':
 104                     self.raise_login_required(message)
 105                 raise ExtractorError('%s said: %s' % (
 106                     self.IE_NAME, message), expected=True)
 107
 108     def _real_initialize(self):
 109         self._login()
 110
 111     def _login(self):
 112         username, password = self._get_login_info()
 113         if username is None:
 114             return
 115
 116         login_form = {
 117             'login_id': username,
 118             'password': password,
 119         }
 120
 121         login = self._call_api(
 122             'sessions.json', None,
 123             'Logging in', post_data=login_form)
 124
 125         self._token = login.get('token')
 126         if not self._token:
 127             self.report_warning('Unable to get session token, login has probably failed')
 128
 129     @staticmethod
 130     def dict_selection(dict_obj, preferred_key, allow_fallback=True):
 131         if preferred_key in dict_obj:
 132             return dict_obj.get(preferred_key)
 133
 134         if not allow_fallback:
 135             return
 136
 137         filtered_dict = list(filter(None, [dict_obj.get(k) for k in dict_obj.keys()]))
 138         return filtered_dict[0] if filtered_dict else None
 139
 140
 141 class VikiIE(VikiBaseIE):
 142     IE_NAME = 'viki'
 143     _VALID_URL = r'%s(?:videos|player)/(?P<id>[0-9]+v)' % VikiBaseIE._VALID_URL_BASE
 144     _TESTS = [{
 145         'url': 'http://www.viki.com/videos/1023585v-heirs-episode-14',
 146         'info_dict': {
 147             'id': '1023585v',
 148             'ext': 'mp4',
 149             'title': 'Heirs - Episode 14',
 150             'uploader': 'SBS Contents Hub',
 151             'timestamp': 1385047627,
 152             'upload_date': '20131121',
 153             'age_limit': 13,
 154             'duration': 3570,
 155             'episode_number': 14,
 156         },
 157         'params': {
 158             'format': 'bestvideo',
 159         },
 160         'skip': 'Blocked in the US',
 161         'expected_warnings': ['Unknown MIME type image/jpeg in DASH manifest'],
 162     }, {
 163         # clip
 164         'url': 'http://www.viki.com/videos/1067139v-the-avengers-age-of-ultron-press-conference',
 165         'md5': '86c0b5dbd4d83a6611a79987cc7a1989',
 166         'info_dict': {
 167             'id': '1067139v',
 168             'ext': 'mp4',
 169             'title': "'The Avengers: Age of Ultron' Press Conference",
 170             'description': 'md5:d70b2f9428f5488321bfe1db10d612ea',
 171             'duration': 352,
 172             'timestamp': 1430380829,
 173             'upload_date': '20150430',
 174             'uploader': 'Arirang TV',
 175             'like_count': int,
 176             'age_limit': 0,
 177         },
 178         'skip': 'Sorry. There was an error loading this video',
 179     }, {
 180         'url': 'http://www.viki.com/videos/1048879v-ankhon-dekhi',
 181         'info_dict': {
 182             'id': '1048879v',
 183             'ext': 'mp4',
 184             'title': 'Ankhon Dekhi',
 185             'duration': 6512,
 186             'timestamp': 1408532356,
 187             'upload_date': '20140820',
 188             'uploader': 'Spuul',
 189             'like_count': int,
 190             'age_limit': 13,
 191         },
 192         'skip': 'Blocked in the US',
 193     }, {
 194         # episode
 195         'url': 'http://www.viki.com/videos/44699v-boys-over-flowers-episode-1',
 196         'md5': '0a53dc252e6e690feccd756861495a8c',
 197         'info_dict': {
 198             'id': '44699v',
 199             'ext': 'mp4',
 200             'title': 'Boys Over Flowers - Episode 1',
 201             'description': 'md5:b89cf50038b480b88b5b3c93589a9076',
 202             'duration': 4172,
 203             'timestamp': 1270496524,
 204             'upload_date': '20100405',
 205             'uploader': 'group8',
 206             'like_count': int,
 207             'age_limit': 13,
 208             'episode_number': 1,
 209         },
 210         'params': {
 211             'format': 'bestvideo',
 212         },
 213         'expected_warnings': ['Unknown MIME type image/jpeg in DASH manifest'],
 214     }, {
 215         # youtube external
 216         'url': 'http://www.viki.com/videos/50562v-poor-nastya-complete-episode-1',
 217         'md5': '63f8600c1da6f01b7640eee7eca4f1da',
 218         'info_dict': {
 219             'id': '50562v',
 220             'ext': 'webm',
 221             'title': 'Poor Nastya [COMPLETE] - Episode 1',
 222             'description': '',
 223             'duration': 606,
 224             'timestamp': 1274949505,
 225             'upload_date': '20101213',
 226             'uploader': 'ad14065n',
 227             'uploader_id': 'ad14065n',
 228             'like_count': int,
 229             'age_limit': 13,
 230         },
 231         'skip': 'Page not found!',
 232     }, {
 233         'url': 'http://www.viki.com/player/44699v',
 234         'only_matching': True,
 235     }, {
 236         # non-English description
 237         'url': 'http://www.viki.com/videos/158036v-love-in-magic',
 238         'md5': '41faaba0de90483fb4848952af7c7d0d',
 239         'info_dict': {
 240             'id': '158036v',
 241             'ext': 'mp4',
 242             'uploader': 'I Planet Entertainment',
 243             'upload_date': '20111122',
 244             'timestamp': 1321985454,
 245             'description': 'md5:44b1e46619df3a072294645c770cef36',
 246             'title': 'Love In Magic',
 247             'age_limit': 13,
 248         },
 249         'params': {
 250             'format': 'bestvideo',
 251         },
 252         'expected_warnings': ['Unknown MIME type image/jpeg in DASH manifest'],
 253     }]
 254
 255     def _real_extract(self, url):
 256         video_id = self._match_id(url)
 257
 258         resp = self._download_json(
 259             'https://www.viki.com/api/videos/' + video_id,
 260             video_id, 'Downloading video JSON', headers={
 261                 'x-client-user-agent': std_headers['User-Agent'],
 262                 'x-viki-as-id': self._APP,
 263                 'x-viki-app-ver': self._APP_VERSION,
 264             })
 265         video = resp['video']
 266
 267         self._check_errors(video)
 268
 269         title = self.dict_selection(video.get('titles', {}), 'en', allow_fallback=False)
 270         episode_number = int_or_none(video.get('number'))
 271         if not title:
 272             title = 'Episode %d' % episode_number if video.get('type') == 'episode' else video.get('id') or video_id
 273             container_titles = try_get(video, lambda x: x['container']['titles'], dict) or {}
 274             container_title = self.dict_selection(container_titles, 'en')
 275             title = '%s - %s' % (container_title, title)
 276
 277         description = self.dict_selection(video.get('descriptions', {}), 'en')
 278
 279         like_count = int_or_none(try_get(video, lambda x: x['likes']['count']))
 280
 281         thumbnails = []
 282         for thumbnail_id, thumbnail in (video.get('images') or {}).items():
 283             thumbnails.append({
 284                 'id': thumbnail_id,
 285                 'url': thumbnail.get('url'),
 286             })
 287
 288         subtitles = {}
 289         try:
 290             # New way to fetch subtitles
 291             new_video = self._download_json(
 292                 'https://www.viki.com/api/videos/%s' % video_id, video_id,
 293                 'Downloading new video JSON to get subtitles', fatal=False,
 294                 headers={
 295                     'x-client-user-agent': std_headers['User-Agent'],
 296                     'x-viki-as-id': self._APP,
 297                     'x-viki-app-ver': self._APP_VERSION,
 298                 })
 299             for sub in new_video.get('streamSubtitles').get('dash'):
 300                 subtitles[sub.get('srclang')] = [{
 301                     'ext': 'vtt',
 302                     'url': sub.get('src'),
 303                     'completion': sub.get('percentage'),
 304                 }]
 305         except AttributeError:
 306             # fall-back to the old way if there isn't a streamSubtitles attribute
 307             for subtitle_lang, _ in (video.get('subtitle_completions') or {}).items():
 308                 subtitles[subtitle_lang] = [{
 309                     'ext': subtitles_format,
 310                     'url': self._prepare_call(
 311                         'videos/%s/subtitles/%s.%s' % (video_id, subtitle_lang, subtitles_format)),
 312                 } for subtitles_format in ('srt', 'vtt')]
 313
 314         result = {
 315             'id': video_id,
 316             'title': title,
 317             'description': description,
 318             'duration': int_or_none(video.get('duration')),
 319             'timestamp': parse_iso8601(video.get('created_at')),
 320             'uploader': video.get('author'),
 321             'uploader_url': video.get('author_url'),
 322             'like_count': like_count,
 323             'age_limit': parse_age_limit(video.get('rating')),
 324             'thumbnails': thumbnails,
 325             'subtitles': subtitles,
 326             'episode_number': episode_number,
 327         }
 328
 329         formats = []
 330
 331         def add_format(format_id, format_dict, protocol='http'):
 332             # rtmps URLs does not seem to work
 333             if protocol == 'rtmps':
 334                 return
 335             format_url = format_dict.get('url')
 336             if not format_url:
 337                 return
 338             format_drms = format_dict.get('drms')
 339             format_stream_id = format_dict.get('id')
 340             qs = compat_parse_qs(compat_urllib_parse_urlparse(format_url).query)
 341             stream = qs.get('stream', [None])[0]
 342             if stream:
 343                 format_url = base64.b64decode(stream).decode()
 344             if format_id in ('m3u8', 'hls'):
 345                 m3u8_formats = self._extract_m3u8_formats(
 346                     format_url, video_id, 'mp4',
 347                     entry_protocol='m3u8_native',
 348                     m3u8_id='m3u8-%s' % protocol, fatal=False)
 349                 # Despite CODECS metadata in m3u8 all video-only formats
 350                 # are actually video+audio
 351                 for f in m3u8_formats:
 352                     if not self._downloader.params.get('allow_unplayable_formats') and '_drm/index_' in f['url']:
 353                         continue
 354                     if f.get('acodec') == 'none' and f.get('vcodec') != 'none':
 355                         f['acodec'] = None
 356                     formats.append(f)
 357             elif format_id in ('mpd', 'dash'):
 358                 formats.extend(self._extract_mpd_formats(
 359                     format_url, video_id, 'mpd-%s' % protocol, fatal=False))
 360             elif format_url.startswith('rtmp'):
 361                 mobj = re.search(
 362                     r'^(?P<url>rtmp://[^/]+/(?P<app>.+?))/(?P<playpath>mp4:.+)$',
 363                     format_url)
 364                 if not mobj:
 365                     return
 366                 formats.append({
 367                     'format_id': 'rtmp-%s' % format_id,
 368                     'ext': 'flv',
 369                     'url': mobj.group('url'),
 370                     'play_path': mobj.group('playpath'),
 371                     'app': mobj.group('app'),
 372                     'page_url': url,
 373                     'drms': format_drms,
 374                     'stream_id': format_stream_id,
 375                 })
 376             else:
 377                 urlh = self._request_webpage(
 378                     HEADRequest(format_url), video_id, 'Checking file size', fatal=False)
 379                 formats.append({
 380                     'url': format_url,
 381                     'format_id': '%s-%s' % (format_id, protocol),
 382                     'height': int_or_none(self._search_regex(
 383                         r'^(\d+)[pP]$', format_id, 'height', default=None)),
 384                     'drms': format_drms,
 385                     'stream_id': format_stream_id,
 386                     'filesize': int_or_none(urlh.headers.get('Content-Length')),
 387                 })
 388
 389         for format_id, format_dict in (resp.get('streams') or {}).items():
 390             add_format(format_id, format_dict)
 391         if not formats:
 392             streams = self._call_api(
 393                 'videos/%s/streams.json' % video_id, video_id,
 394                 'Downloading video streams JSON')
 395
 396             if 'external' in streams:
 397                 result.update({
 398                     '_type': 'url_transparent',
 399                     'url': streams['external']['url'],
 400                 })
 401                 return result
 402
 403             for format_id, stream_dict in streams.items():
 404                 for protocol, format_dict in stream_dict.items():
 405                     add_format(format_id, format_dict, protocol)
 406         self._sort_formats(formats)
 407
 408         result['formats'] = formats
 409         return result
 410
 411
 412 class VikiChannelIE(VikiBaseIE):
 413     IE_NAME = 'viki:channel'
 414     _VALID_URL = r'%s(?:tv|news|movies|artists)/(?P<id>[0-9]+c)' % VikiBaseIE._VALID_URL_BASE
 415     _TESTS = [{
 416         'url': 'http://www.viki.com/tv/50c-boys-over-flowers',
 417         'info_dict': {
 418             'id': '50c',
 419             'title': 'Boys Over Flowers',
 420             'description': 'md5:804ce6e7837e1fd527ad2f25420f4d59',
 421         },
 422         'playlist_mincount': 71,
 423     }, {
 424         'url': 'http://www.viki.com/tv/1354c-poor-nastya-complete',
 425         'info_dict': {
 426             'id': '1354c',
 427             'title': 'Poor Nastya [COMPLETE]',
 428             'description': 'md5:05bf5471385aa8b21c18ad450e350525',
 429         },
 430         'playlist_count': 127,
 431         'skip': 'Page not found',
 432     }, {
 433         'url': 'http://www.viki.com/news/24569c-showbiz-korea',
 434         'only_matching': True,
 435     }, {
 436         'url': 'http://www.viki.com/movies/22047c-pride-and-prejudice-2005',
 437         'only_matching': True,
 438     }, {
 439         'url': 'http://www.viki.com/artists/2141c-shinee',
 440         'only_matching': True,
 441     }]
 442
 443     _PER_PAGE = 25
 444
 445     def _real_extract(self, url):
 446         channel_id = self._match_id(url)
 447
 448         channel = self._call_api(
 449             'containers/%s.json' % channel_id, channel_id,
 450             'Downloading channel JSON')
 451
 452         self._check_errors(channel)
 453
 454         title = self.dict_selection(channel['titles'], 'en')
 455
 456         description = self.dict_selection(channel['descriptions'], 'en')
 457
 458         entries = []
 459         for video_type in ('episodes', 'clips', 'movies'):
 460             for page_num in itertools.count(1):
 461                 page = self._call_api(
 462                     'containers/%s/%s.json?per_page=%d&sort=number&direction=asc&with_paging=true&page=%d'
 463                     % (channel_id, video_type, self._PER_PAGE, page_num), channel_id,
 464                     'Downloading %s JSON page #%d' % (video_type, page_num))
 465                 for video in page['response']:
 466                     video_id = video['id']
 467                     entries.append(self.url_result(
 468                         'https://www.viki.com/videos/%s' % video_id, 'Viki'))
 469                 if not page['pagination']['next']:
 470                     break
 471
 472         return self.playlist_result(entries, channel_id, title, description)