yt_dlp/extractor/niconico.py

   1 import datetime
   2 import functools
   3 import itertools
   4 import json
   5 import re
   6 import time
   7
   8 from urllib.parse import urlparse
   9
  10 from .common import InfoExtractor, SearchInfoExtractor
  11 from ..compat import (
  12     compat_HTTPError,
  13 )
  14 from ..dependencies import websockets
  15 from ..utils import (
  16     ExtractorError,
  17     OnDemandPagedList,
  18     WebSocketsWrapper,
  19     bug_reports_message,
  20     clean_html,
  21     float_or_none,
  22     int_or_none,
  23     join_nonempty,
  24     parse_duration,
  25     parse_filesize,
  26     parse_iso8601,
  27     parse_resolution,
  28     qualities,
  29     remove_start,
  30     str_or_none,
  31     traverse_obj,
  32     try_get,
  33     unescapeHTML,
  34     update_url_query,
  35     url_or_none,
  36     urlencode_postdata,
  37     urljoin,
  38 )
  39
  40
  41 class NiconicoIE(InfoExtractor):
  42     IE_NAME = 'niconico'
  43     IE_DESC = 'ニコニコ動画'
  44
  45     _TESTS = [{
  46         'url': 'http://www.nicovideo.jp/watch/sm22312215',
  47         'md5': 'd1a75c0823e2f629128c43e1212760f9',
  48         'info_dict': {
  49             'id': 'sm22312215',
  50             'ext': 'mp4',
  51             'title': 'Big Buck Bunny',
  52             'thumbnail': r're:https?://.*',
  53             'uploader': 'takuya0301',
  54             'uploader_id': '2698420',
  55             'upload_date': '20131123',
  56             'timestamp': int,  # timestamp is unstable
  57             'description': '(c) copyright 2008, Blender Foundation / www.bigbuckbunny.org',
  58             'duration': 33,
  59             'view_count': int,
  60             'comment_count': int,
  61         },
  62         'skip': 'Requires an account',
  63     }, {
  64         # File downloaded with and without credentials are different, so omit
  65         # the md5 field
  66         'url': 'http://www.nicovideo.jp/watch/nm14296458',
  67         'info_dict': {
  68             'id': 'nm14296458',
  69             'ext': 'swf',
  70             'title': '【鏡音リン】Dance on media【オリジナル】take2!',
  71             'description': 'md5:689f066d74610b3b22e0f1739add0f58',
  72             'thumbnail': r're:https?://.*',
  73             'uploader': 'りょうた',
  74             'uploader_id': '18822557',
  75             'upload_date': '20110429',
  76             'timestamp': 1304065916,
  77             'duration': 209,
  78         },
  79         'skip': 'Requires an account',
  80     }, {
  81         # 'video exists but is marked as "deleted"
  82         # md5 is unstable
  83         'url': 'http://www.nicovideo.jp/watch/sm10000',
  84         'info_dict': {
  85             'id': 'sm10000',
  86             'ext': 'unknown_video',
  87             'description': 'deleted',
  88             'title': 'ドラえもんエターナル第3話「決戦第3新東京市」＜前編＞',
  89             'thumbnail': r're:https?://.*',
  90             'upload_date': '20071224',
  91             'timestamp': int,  # timestamp field has different value if logged in
  92             'duration': 304,
  93             'view_count': int,
  94         },
  95         'skip': 'Requires an account',
  96     }, {
  97         'url': 'http://www.nicovideo.jp/watch/so22543406',
  98         'info_dict': {
  99             'id': '1388129933',
 100             'ext': 'mp4',
 101             'title': '【第1回】RADIOアニメロミックス ラブライブ！～のぞえりRadio Garden～',
 102             'description': 'md5:b27d224bb0ff53d3c8269e9f8b561cf1',
 103             'thumbnail': r're:https?://.*',
 104             'timestamp': 1388851200,
 105             'upload_date': '20140104',
 106             'uploader': 'アニメロチャンネル',
 107             'uploader_id': '312',
 108         },
 109         'skip': 'The viewing period of the video you were searching for has expired.',
 110     }, {
 111         # video not available via `getflv`; "old" HTML5 video
 112         'url': 'http://www.nicovideo.jp/watch/sm1151009',
 113         'md5': '8fa81c364eb619d4085354eab075598a',
 114         'info_dict': {
 115             'id': 'sm1151009',
 116             'ext': 'mp4',
 117             'title': 'マスターシステム本体内蔵のスペハリのメインテーマ（ＰＳＧ版）',
 118             'description': 'md5:6ee077e0581ff5019773e2e714cdd0b7',
 119             'thumbnail': r're:https?://.*',
 120             'duration': 184,
 121             'timestamp': 1190868283,
 122             'upload_date': '20070927',
 123             'uploader': 'denden2',
 124             'uploader_id': '1392194',
 125             'view_count': int,
 126             'comment_count': int,
 127         },
 128         'skip': 'Requires an account',
 129     }, {
 130         # "New" HTML5 video
 131         # md5 is unstable
 132         'url': 'http://www.nicovideo.jp/watch/sm31464864',
 133         'info_dict': {
 134             'id': 'sm31464864',
 135             'ext': 'mp4',
 136             'title': '新作TVアニメ「戦姫絶唱シンフォギアAXZ」PV 最高画質',
 137             'description': 'md5:e52974af9a96e739196b2c1ca72b5feb',
 138             'timestamp': 1498514060,
 139             'upload_date': '20170626',
 140             'uploader': 'ゲスト',
 141             'uploader_id': '40826363',
 142             'thumbnail': r're:https?://.*',
 143             'duration': 198,
 144             'view_count': int,
 145             'comment_count': int,
 146         },
 147         'skip': 'Requires an account',
 148     }, {
 149         # Video without owner
 150         'url': 'http://www.nicovideo.jp/watch/sm18238488',
 151         'md5': 'd265680a1f92bdcbbd2a507fc9e78a9e',
 152         'info_dict': {
 153             'id': 'sm18238488',
 154             'ext': 'mp4',
 155             'title': '【実写版】ミュータントタートルズ',
 156             'description': 'md5:15df8988e47a86f9e978af2064bf6d8e',
 157             'timestamp': 1341160408,
 158             'upload_date': '20120701',
 159             'uploader': None,
 160             'uploader_id': None,
 161             'thumbnail': r're:https?://.*',
 162             'duration': 5271,
 163             'view_count': int,
 164             'comment_count': int,
 165         },
 166         'skip': 'Requires an account',
 167     }, {
 168         'url': 'http://sp.nicovideo.jp/watch/sm28964488?ss_pos=1&cp_in=wt_tg',
 169         'only_matching': True,
 170     }, {
 171         'note': 'a video that is only served as an ENCRYPTED HLS.',
 172         'url': 'https://www.nicovideo.jp/watch/so38016254',
 173         'only_matching': True,
 174     }]
 175
 176     _VALID_URL = r'https?://(?:(?:www\.|secure\.|sp\.)?nicovideo\.jp/watch|nico\.ms)/(?P<id>(?:[a-z]{2})?[0-9]+)'
 177     _NETRC_MACHINE = 'niconico'
 178     _COMMENT_API_ENDPOINTS = (
 179         'https://nvcomment.nicovideo.jp/legacy/api.json',
 180         'https://nmsg.nicovideo.jp/api.json',)
 181     _API_HEADERS = {
 182         'X-Frontend-ID': '6',
 183         'X-Frontend-Version': '0',
 184         'X-Niconico-Language': 'en-us',
 185         'Referer': 'https://www.nicovideo.jp/',
 186         'Origin': 'https://www.nicovideo.jp',
 187     }
 188
 189     def _perform_login(self, username, password):
 190         login_ok = True
 191         login_form_strs = {
 192             'mail_tel': username,
 193             'password': password,
 194         }
 195         self._request_webpage(
 196             'https://account.nicovideo.jp/login', None,
 197             note='Acquiring Login session')
 198         page = self._download_webpage(
 199             'https://account.nicovideo.jp/login/redirector?show_button_twitter=1&site=niconico&show_button_facebook=1', None,
 200             note='Logging in', errnote='Unable to log in',
 201             data=urlencode_postdata(login_form_strs),
 202             headers={
 203                 'Referer': 'https://account.nicovideo.jp/login',
 204                 'Content-Type': 'application/x-www-form-urlencoded',
 205             })
 206         if 'oneTimePw' in page:
 207             post_url = self._search_regex(
 208                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page, 'post url', group='url')
 209             page = self._download_webpage(
 210                 urljoin('https://account.nicovideo.jp', post_url), None,
 211                 note='Performing MFA', errnote='Unable to complete MFA',
 212                 data=urlencode_postdata({
 213                     'otp': self._get_tfa_info('6 digits code')
 214                 }), headers={
 215                     'Content-Type': 'application/x-www-form-urlencoded',
 216                 })
 217             if 'oneTimePw' in page or 'formError' in page:
 218                 err_msg = self._html_search_regex(
 219                     r'formError["\']+>(.*?)</div>', page, 'form_error',
 220                     default='There\'s an error but the message can\'t be parsed.',
 221                     flags=re.DOTALL)
 222                 self.report_warning(f'Unable to log in: MFA challenge failed, "{err_msg}"')
 223                 return False
 224         login_ok = 'class="notice error"' not in page
 225         if not login_ok:
 226             self.report_warning('Unable to log in: bad username or password')
 227         return login_ok
 228
 229     def _get_heartbeat_info(self, info_dict):
 230         video_id, video_src_id, audio_src_id = info_dict['url'].split(':')[1].split('/')
 231         dmc_protocol = info_dict['expected_protocol']
 232
 233         api_data = (
 234             info_dict.get('_api_data')
 235             or self._parse_json(
 236                 self._html_search_regex(
 237                     'data-api-data="([^"]+)"',
 238                     self._download_webpage('https://www.nicovideo.jp/watch/' + video_id, video_id),
 239                     'API data', default='{}'),
 240                 video_id))
 241
 242         session_api_data = try_get(api_data, lambda x: x['media']['delivery']['movie']['session'])
 243         session_api_endpoint = try_get(session_api_data, lambda x: x['urls'][0])
 244
 245         def ping():
 246             tracking_id = traverse_obj(api_data, ('media', 'delivery', 'trackingId'))
 247             if tracking_id:
 248                 tracking_url = update_url_query('https://nvapi.nicovideo.jp/v1/2ab0cbaa/watch', {'t': tracking_id})
 249                 watch_request_response = self._download_json(
 250                     tracking_url, video_id,
 251                     note='Acquiring permission for downloading video', fatal=False,
 252                     headers=self._API_HEADERS)
 253                 if traverse_obj(watch_request_response, ('meta', 'status')) != 200:
 254                     self.report_warning('Failed to acquire permission for playing video. Video download may fail.')
 255
 256         yesno = lambda x: 'yes' if x else 'no'
 257
 258         if dmc_protocol == 'http':
 259             protocol = 'http'
 260             protocol_parameters = {
 261                 'http_output_download_parameters': {
 262                     'use_ssl': yesno(session_api_data['urls'][0]['isSsl']),
 263                     'use_well_known_port': yesno(session_api_data['urls'][0]['isWellKnownPort']),
 264                 }
 265             }
 266         elif dmc_protocol == 'hls':
 267             protocol = 'm3u8'
 268             segment_duration = try_get(self._configuration_arg('segment_duration'), lambda x: int(x[0])) or 6000
 269             parsed_token = self._parse_json(session_api_data['token'], video_id)
 270             encryption = traverse_obj(api_data, ('media', 'delivery', 'encryption'))
 271             protocol_parameters = {
 272                 'hls_parameters': {
 273                     'segment_duration': segment_duration,
 274                     'transfer_preset': '',
 275                     'use_ssl': yesno(session_api_data['urls'][0]['isSsl']),
 276                     'use_well_known_port': yesno(session_api_data['urls'][0]['isWellKnownPort']),
 277                 }
 278             }
 279             if 'hls_encryption' in parsed_token and encryption:
 280                 protocol_parameters['hls_parameters']['encryption'] = {
 281                     parsed_token['hls_encryption']: {
 282                         'encrypted_key': encryption['encryptedKey'],
 283                         'key_uri': encryption['keyUri'],
 284                     }
 285                 }
 286             else:
 287                 protocol = 'm3u8_native'
 288         else:
 289             raise ExtractorError(f'Unsupported DMC protocol: {dmc_protocol}')
 290
 291         session_response = self._download_json(
 292             session_api_endpoint['url'], video_id,
 293             query={'_format': 'json'},
 294             headers={'Content-Type': 'application/json'},
 295             note='Downloading JSON metadata for %s' % info_dict['format_id'],
 296             data=json.dumps({
 297                 'session': {
 298                     'client_info': {
 299                         'player_id': session_api_data.get('playerId'),
 300                     },
 301                     'content_auth': {
 302                         'auth_type': try_get(session_api_data, lambda x: x['authTypes'][session_api_data['protocols'][0]]),
 303                         'content_key_timeout': session_api_data.get('contentKeyTimeout'),
 304                         'service_id': 'nicovideo',
 305                         'service_user_id': session_api_data.get('serviceUserId')
 306                     },
 307                     'content_id': session_api_data.get('contentId'),
 308                     'content_src_id_sets': [{
 309                         'content_src_ids': [{
 310                             'src_id_to_mux': {
 311                                 'audio_src_ids': [audio_src_id],
 312                                 'video_src_ids': [video_src_id],
 313                             }
 314                         }]
 315                     }],
 316                     'content_type': 'movie',
 317                     'content_uri': '',
 318                     'keep_method': {
 319                         'heartbeat': {
 320                             'lifetime': session_api_data.get('heartbeatLifetime')
 321                         }
 322                     },
 323                     'priority': session_api_data['priority'],
 324                     'protocol': {
 325                         'name': 'http',
 326                         'parameters': {
 327                             'http_parameters': {
 328                                 'parameters': protocol_parameters
 329                             }
 330                         }
 331                     },
 332                     'recipe_id': session_api_data.get('recipeId'),
 333                     'session_operation_auth': {
 334                         'session_operation_auth_by_signature': {
 335                             'signature': session_api_data.get('signature'),
 336                             'token': session_api_data.get('token'),
 337                         }
 338                     },
 339                     'timing_constraint': 'unlimited'
 340                 }
 341             }).encode())
 342
 343         info_dict['url'] = session_response['data']['session']['content_uri']
 344         info_dict['protocol'] = protocol
 345
 346         # get heartbeat info
 347         heartbeat_info_dict = {
 348             'url': session_api_endpoint['url'] + '/' + session_response['data']['session']['id'] + '?_format=json&_method=PUT',
 349             'data': json.dumps(session_response['data']),
 350             # interval, convert milliseconds to seconds, then halve to make a buffer.
 351             'interval': float_or_none(session_api_data.get('heartbeatLifetime'), scale=3000),
 352             'ping': ping
 353         }
 354
 355         return info_dict, heartbeat_info_dict
 356
 357     def _extract_format_for_quality(self, video_id, audio_quality, video_quality, dmc_protocol):
 358
 359         if not audio_quality.get('isAvailable') or not video_quality.get('isAvailable'):
 360             return None
 361
 362         def extract_video_quality(video_quality):
 363             return parse_filesize('%sB' % self._search_regex(
 364                 r'\| ([0-9]*\.?[0-9]*[MK])', video_quality, 'vbr', default=''))
 365
 366         format_id = '-'.join(
 367             [remove_start(s['id'], 'archive_') for s in (video_quality, audio_quality)] + [dmc_protocol])
 368
 369         vid_qual_label = traverse_obj(video_quality, ('metadata', 'label'))
 370         vid_quality = traverse_obj(video_quality, ('metadata', 'bitrate'))
 371
 372         return {
 373             'url': 'niconico_dmc:%s/%s/%s' % (video_id, video_quality['id'], audio_quality['id']),
 374             'format_id': format_id,
 375             'format_note': join_nonempty('DMC', vid_qual_label, dmc_protocol.upper(), delim=' '),
 376             'ext': 'mp4',  # Session API are used in HTML5, which always serves mp4
 377             'acodec': 'aac',
 378             'vcodec': 'h264',
 379             'abr': float_or_none(traverse_obj(audio_quality, ('metadata', 'bitrate')), 1000),
 380             'vbr': float_or_none(vid_quality if vid_quality > 0 else extract_video_quality(vid_qual_label), 1000),
 381             'height': traverse_obj(video_quality, ('metadata', 'resolution', 'height')),
 382             'width': traverse_obj(video_quality, ('metadata', 'resolution', 'width')),
 383             'quality': -2 if 'low' in video_quality['id'] else None,
 384             'protocol': 'niconico_dmc',
 385             'expected_protocol': dmc_protocol,  # XXX: This is not a documented field
 386             'http_headers': {
 387                 'Origin': 'https://www.nicovideo.jp',
 388                 'Referer': 'https://www.nicovideo.jp/watch/' + video_id,
 389             }
 390         }
 391
 392     def _real_extract(self, url):
 393         video_id = self._match_id(url)
 394
 395         try:
 396             webpage, handle = self._download_webpage_handle(
 397                 'https://www.nicovideo.jp/watch/' + video_id, video_id)
 398             if video_id.startswith('so'):
 399                 video_id = self._match_id(handle.geturl())
 400
 401             api_data = self._parse_json(self._html_search_regex(
 402                 'data-api-data="([^"]+)"', webpage,
 403                 'API data', default='{}'), video_id)
 404         except ExtractorError as e:
 405             try:
 406                 api_data = self._download_json(
 407                     'https://www.nicovideo.jp/api/watch/v3/%s?_frontendId=6&_frontendVersion=0&actionTrackId=AAAAAAAAAA_%d' % (video_id, round(time.time() * 1000)), video_id,
 408                     note='Downloading API JSON', errnote='Unable to fetch data')['data']
 409             except ExtractorError:
 410                 if not isinstance(e.cause, compat_HTTPError):
 411                     raise
 412                 webpage = e.cause.read().decode('utf-8', 'replace')
 413                 error_msg = self._html_search_regex(
 414                     r'(?s)<section\s+class="(?:(?:ErrorMessage|WatchExceptionPage-message)\s*)+">(.+?)</section>',
 415                     webpage, 'error reason', default=None)
 416                 if not error_msg:
 417                     raise
 418                 raise ExtractorError(re.sub(r'\s+', ' ', error_msg), expected=True)
 419
 420         formats = []
 421
 422         def get_video_info(*items, get_first=True, **kwargs):
 423             return traverse_obj(api_data, ('video', *items), get_all=not get_first, **kwargs)
 424
 425         quality_info = api_data['media']['delivery']['movie']
 426         session_api_data = quality_info['session']
 427         for (audio_quality, video_quality, protocol) in itertools.product(quality_info['audios'], quality_info['videos'], session_api_data['protocols']):
 428             fmt = self._extract_format_for_quality(video_id, audio_quality, video_quality, protocol)
 429             if fmt:
 430                 formats.append(fmt)
 431
 432         # Start extracting information
 433         tags = None
 434         if webpage:
 435             # use og:video:tag (not logged in)
 436             og_video_tags = re.finditer(r'<meta\s+property="og:video:tag"\s*content="(.*?)">', webpage)
 437             tags = list(filter(None, (clean_html(x.group(1)) for x in og_video_tags)))
 438             if not tags:
 439                 # use keywords and split with comma (not logged in)
 440                 kwds = self._html_search_meta('keywords', webpage, default=None)
 441                 if kwds:
 442                     tags = [x for x in kwds.split(',') if x]
 443         if not tags:
 444             # find in json (logged in)
 445             tags = traverse_obj(api_data, ('tag', 'items', ..., 'name'))
 446
 447         thumb_prefs = qualities(['url', 'middleUrl', 'largeUrl', 'player', 'ogp'])
 448
 449         return {
 450             'id': video_id,
 451             '_api_data': api_data,
 452             'title': get_video_info(('originalTitle', 'title')) or self._og_search_title(webpage, default=None),
 453             'formats': formats,
 454             'thumbnails': [{
 455                 'id': key,
 456                 'url': url,
 457                 'ext': 'jpg',
 458                 'preference': thumb_prefs(key),
 459                 **parse_resolution(url, lenient=True),
 460             } for key, url in (get_video_info('thumbnail') or {}).items() if url],
 461             'description': clean_html(get_video_info('description')),
 462             'uploader': traverse_obj(api_data, ('owner', 'nickname'), ('channel', 'name'), ('community', 'name')),
 463             'uploader_id': str_or_none(traverse_obj(api_data, ('owner', 'id'), ('channel', 'id'), ('community', 'id'))),
 464             'timestamp': parse_iso8601(get_video_info('registeredAt')) or parse_iso8601(
 465                 self._html_search_meta('video:release_date', webpage, 'date published', default=None)),
 466             'channel': traverse_obj(api_data, ('channel', 'name'), ('community', 'name')),
 467             'channel_id': traverse_obj(api_data, ('channel', 'id'), ('community', 'id')),
 468             'view_count': int_or_none(get_video_info('count', 'view')),
 469             'tags': tags,
 470             'genre': traverse_obj(api_data, ('genre', 'label'), ('genre', 'key')),
 471             'comment_count': get_video_info('count', 'comment', expected_type=int),
 472             'duration': (
 473                 parse_duration(self._html_search_meta('video:duration', webpage, 'video duration', default=None))
 474                 or get_video_info('duration')),
 475             'webpage_url': url_or_none(url) or f'https://www.nicovideo.jp/watch/{video_id}',
 476             'subtitles': self.extract_subtitles(video_id, api_data, session_api_data),
 477         }
 478
 479     def _get_subtitles(self, video_id, api_data, session_api_data):
 480         comment_user_key = traverse_obj(api_data, ('comment', 'keys', 'userKey'))
 481         user_id_str = session_api_data.get('serviceUserId')
 482
 483         thread_ids = traverse_obj(api_data, ('comment', 'threads', lambda _, v: v['isActive']))
 484         legacy_danmaku = self._extract_legacy_comments(video_id, thread_ids, user_id_str, comment_user_key) or []
 485
 486         new_comments = traverse_obj(api_data, ('comment', 'nvComment'))
 487         new_danmaku = self._extract_new_comments(
 488             new_comments.get('server'), video_id,
 489             new_comments.get('params'), new_comments.get('threadKey'))
 490
 491         if not legacy_danmaku and not new_danmaku:
 492             self.report_warning(f'Failed to get comments. {bug_reports_message()}')
 493             return
 494
 495         return {
 496             'comments': [{
 497                 'ext': 'json',
 498                 'data': json.dumps(legacy_danmaku + new_danmaku),
 499             }],
 500         }
 501
 502     def _extract_legacy_comments(self, video_id, threads, user_id, user_key):
 503         auth_data = {
 504             'user_id': user_id,
 505             'userkey': user_key,
 506         } if user_id and user_key else {'user_id': ''}
 507
 508         api_url = traverse_obj(threads, (..., 'server'), get_all=False)
 509
 510         # Request Start
 511         post_data = [{'ping': {'content': 'rs:0'}}]
 512         for i, thread in enumerate(threads):
 513             thread_id = thread['id']
 514             thread_fork = thread['fork']
 515             # Post Start (2N)
 516             post_data.append({'ping': {'content': f'ps:{i * 2}'}})
 517             post_data.append({'thread': {
 518                 'fork': thread_fork,
 519                 'language': 0,
 520                 'nicoru': 3,
 521                 'scores': 1,
 522                 'thread': thread_id,
 523                 'version': '20090904',
 524                 'with_global': 1,
 525                 **auth_data,
 526             }})
 527             # Post Final (2N)
 528             post_data.append({'ping': {'content': f'pf:{i * 2}'}})
 529
 530             # Post Start (2N+1)
 531             post_data.append({'ping': {'content': f'ps:{i * 2 + 1}'}})
 532             post_data.append({'thread_leaves': {
 533                 # format is '<bottom of minute range>-<top of minute range>:<comments per minute>,<total last comments'
 534                 # unfortunately NND limits (deletes?) comment returns this way, so you're only able to grab the last 1000 per language
 535                 'content': '0-999999:999999,999999,nicoru:999999',
 536                 'fork': thread_fork,
 537                 'language': 0,
 538                 'nicoru': 3,
 539                 'scores': 1,
 540                 'thread': thread_id,
 541                 **auth_data,
 542             }})
 543             # Post Final (2N+1)
 544             post_data.append({'ping': {'content': f'pf:{i * 2 + 1}'}})
 545         # Request Final
 546         post_data.append({'ping': {'content': 'rf:0'}})
 547
 548         return self._download_json(
 549             f'{api_url}/api.json', video_id, data=json.dumps(post_data).encode(), fatal=False,
 550             headers={
 551                 'Referer': f'https://www.nicovideo.jp/watch/{video_id}',
 552                 'Origin': 'https://www.nicovideo.jp',
 553                 'Content-Type': 'text/plain;charset=UTF-8',
 554             },
 555             note='Downloading comments', errnote=f'Failed to access endpoint {api_url}')
 556
 557     def _extract_new_comments(self, endpoint, video_id, params, thread_key):
 558         comments = self._download_json(
 559             f'{endpoint}/v1/threads', video_id, data=json.dumps({
 560                 'additionals': {},
 561                 'params': params,
 562                 'threadKey': thread_key,
 563             }).encode(), fatal=False,
 564             headers={
 565                 'Referer': 'https://www.nicovideo.jp/',
 566                 'Origin': 'https://www.nicovideo.jp',
 567                 'Content-Type': 'text/plain;charset=UTF-8',
 568                 'x-client-os-type': 'others',
 569                 'x-frontend-id': '6',
 570                 'x-frontend-version': '0',
 571             },
 572             note='Downloading comments (new)', errnote='Failed to download comments (new)')
 573         return traverse_obj(comments, ('data', 'threads', ..., 'comments', ...))
 574
 575
 576 class NiconicoPlaylistBaseIE(InfoExtractor):
 577     _PAGE_SIZE = 100
 578
 579     _API_HEADERS = {
 580         'X-Frontend-ID': '6',
 581         'X-Frontend-Version': '0',
 582         'X-Niconico-Language': 'en-us'
 583     }
 584
 585     def _call_api(self, list_id, resource, query):
 586         raise NotImplementedError('Must be implemented in subclasses')
 587
 588     @staticmethod
 589     def _parse_owner(item):
 590         return {
 591             'uploader': traverse_obj(item, ('owner', 'name')),
 592             'uploader_id': traverse_obj(item, ('owner', 'id')),
 593         }
 594
 595     def _fetch_page(self, list_id, page):
 596         page += 1
 597         resp = self._call_api(list_id, 'page %d' % page, {
 598             'page': page,
 599             'pageSize': self._PAGE_SIZE,
 600         })
 601         # this is needed to support both mylist and user
 602         for video in traverse_obj(resp, ('items', ..., ('video', None))) or []:
 603             video_id = video.get('id')
 604             if not video_id:
 605                 # skip {"video": {"id": "blablabla", ...}}
 606                 continue
 607             count = video.get('count') or {}
 608             get_count = lambda x: int_or_none(count.get(x))
 609             yield {
 610                 '_type': 'url',
 611                 'id': video_id,
 612                 'title': video.get('title'),
 613                 'url': f'https://www.nicovideo.jp/watch/{video_id}',
 614                 'description': video.get('shortDescription'),
 615                 'duration': int_or_none(video.get('duration')),
 616                 'view_count': get_count('view'),
 617                 'comment_count': get_count('comment'),
 618                 'thumbnail': traverse_obj(video, ('thumbnail', ('nHdUrl', 'largeUrl', 'listingUrl', 'url'))),
 619                 'ie_key': NiconicoIE.ie_key(),
 620                 **self._parse_owner(video),
 621             }
 622
 623     def _entries(self, list_id):
 624         return OnDemandPagedList(functools.partial(self._fetch_page, list_id), self._PAGE_SIZE)
 625
 626
 627 class NiconicoPlaylistIE(NiconicoPlaylistBaseIE):
 628     IE_NAME = 'niconico:playlist'
 629     _VALID_URL = r'https?://(?:(?:www\.|sp\.)?nicovideo\.jp|nico\.ms)/(?:user/\d+/)?(?:my/)?mylist/(?:#/)?(?P<id>\d+)'
 630
 631     _TESTS = [{
 632         'url': 'http://www.nicovideo.jp/mylist/27411728',
 633         'info_dict': {
 634             'id': '27411728',
 635             'title': 'AKB48のオールナイトニッポン',
 636             'description': 'md5:d89694c5ded4b6c693dea2db6e41aa08',
 637             'uploader': 'のっく',
 638             'uploader_id': '805442',
 639         },
 640         'playlist_mincount': 291,
 641     }, {
 642         'url': 'https://www.nicovideo.jp/user/805442/mylist/27411728',
 643         'only_matching': True,
 644     }, {
 645         'url': 'https://www.nicovideo.jp/my/mylist/#/68048635',
 646         'only_matching': True,
 647     }]
 648
 649     def _call_api(self, list_id, resource, query):
 650         return self._download_json(
 651             f'https://nvapi.nicovideo.jp/v2/mylists/{list_id}', list_id,
 652             f'Downloading {resource}', query=query,
 653             headers=self._API_HEADERS)['data']['mylist']
 654
 655     def _real_extract(self, url):
 656         list_id = self._match_id(url)
 657         mylist = self._call_api(list_id, 'list', {
 658             'pageSize': 1,
 659         })
 660         return self.playlist_result(
 661             self._entries(list_id), list_id,
 662             mylist.get('name'), mylist.get('description'), **self._parse_owner(mylist))
 663
 664
 665 class NiconicoSeriesIE(InfoExtractor):
 666     IE_NAME = 'niconico:series'
 667     _VALID_URL = r'https?://(?:(?:www\.|sp\.)?nicovideo\.jp(?:/user/\d+)?|nico\.ms)/series/(?P<id>\d+)'
 668
 669     _TESTS = [{
 670         'url': 'https://www.nicovideo.jp/user/44113208/series/110226',
 671         'info_dict': {
 672             'id': '110226',
 673             'title': 'ご立派ァ！のシリーズ',
 674         },
 675         'playlist_mincount': 10,
 676     }, {
 677         'url': 'https://www.nicovideo.jp/series/12312/',
 678         'info_dict': {
 679             'id': '12312',
 680             'title': 'バトルスピリッツ　お勧めカード紹介(調整中)',
 681         },
 682         'playlist_mincount': 103,
 683     }, {
 684         'url': 'https://nico.ms/series/203559',
 685         'only_matching': True,
 686     }]
 687
 688     def _real_extract(self, url):
 689         list_id = self._match_id(url)
 690         webpage = self._download_webpage(url, list_id)
 691
 692         title = self._search_regex(
 693             (r'<title>「(.+)（全',
 694              r'<div class="TwitterShareButton"\s+data-text="(.+)\s+https:'),
 695             webpage, 'title', fatal=False)
 696         if title:
 697             title = unescapeHTML(title)
 698         json_data = next(self._yield_json_ld(webpage, None, fatal=False))
 699         return self.playlist_from_matches(
 700             traverse_obj(json_data, ('itemListElement', ..., 'url')), list_id, title, ie=NiconicoIE)
 701
 702
 703 class NiconicoHistoryIE(NiconicoPlaylistBaseIE):
 704     IE_NAME = 'niconico:history'
 705     IE_DESC = 'NicoNico user history or likes. Requires cookies.'
 706     _VALID_URL = r'https?://(?:www\.|sp\.)?nicovideo\.jp/my/(?P<id>history(?:/like)?)'
 707
 708     _TESTS = [{
 709         'note': 'PC page, with /video',
 710         'url': 'https://www.nicovideo.jp/my/history/video',
 711         'only_matching': True,
 712     }, {
 713         'note': 'PC page, without /video',
 714         'url': 'https://www.nicovideo.jp/my/history',
 715         'only_matching': True,
 716     }, {
 717         'note': 'mobile page, with /video',
 718         'url': 'https://sp.nicovideo.jp/my/history/video',
 719         'only_matching': True,
 720     }, {
 721         'note': 'mobile page, without /video',
 722         'url': 'https://sp.nicovideo.jp/my/history',
 723         'only_matching': True,
 724     }, {
 725         'note': 'PC page',
 726         'url': 'https://www.nicovideo.jp/my/history/like',
 727         'only_matching': True,
 728     }, {
 729         'note': 'Mobile page',
 730         'url': 'https://sp.nicovideo.jp/my/history/like',
 731         'only_matching': True,
 732     }]
 733
 734     def _call_api(self, list_id, resource, query):
 735         path = 'likes' if list_id == 'history/like' else 'watch/history'
 736         return self._download_json(
 737             f'https://nvapi.nicovideo.jp/v1/users/me/{path}', list_id,
 738             f'Downloading {resource}', query=query, headers=self._API_HEADERS)['data']
 739
 740     def _real_extract(self, url):
 741         list_id = self._match_id(url)
 742         try:
 743             mylist = self._call_api(list_id, 'list', {'pageSize': 1})
 744         except ExtractorError as e:
 745             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 401:
 746                 self.raise_login_required('You have to be logged in to get your history')
 747             raise
 748         return self.playlist_result(self._entries(list_id), list_id, **self._parse_owner(mylist))
 749
 750
 751 class NicovideoSearchBaseIE(InfoExtractor):
 752     _SEARCH_TYPE = 'search'
 753
 754     def _entries(self, url, item_id, query=None, note='Downloading page %(page)s'):
 755         query = query or {}
 756         pages = [query['page']] if 'page' in query else itertools.count(1)
 757         for page_num in pages:
 758             query['page'] = str(page_num)
 759             webpage = self._download_webpage(url, item_id, query=query, note=note % {'page': page_num})
 760             results = re.findall(r'(?<=data-video-id=)["\']?(?P<videoid>.*?)(?=["\'])', webpage)
 761             for item in results:
 762                 yield self.url_result(f'https://www.nicovideo.jp/watch/{item}', 'Niconico', item)
 763             if not results:
 764                 break
 765
 766     def _search_results(self, query):
 767         return self._entries(
 768             self._proto_relative_url(f'//www.nicovideo.jp/{self._SEARCH_TYPE}/{query}'), query)
 769
 770
 771 class NicovideoSearchIE(NicovideoSearchBaseIE, SearchInfoExtractor):
 772     IE_DESC = 'Nico video search'
 773     IE_NAME = 'nicovideo:search'
 774     _SEARCH_KEY = 'nicosearch'
 775
 776
 777 class NicovideoSearchURLIE(NicovideoSearchBaseIE):
 778     IE_NAME = f'{NicovideoSearchIE.IE_NAME}_url'
 779     IE_DESC = 'Nico video search URLs'
 780     _VALID_URL = r'https?://(?:www\.)?nicovideo\.jp/search/(?P<id>[^?#&]+)?'
 781     _TESTS = [{
 782         'url': 'http://www.nicovideo.jp/search/sm9',
 783         'info_dict': {
 784             'id': 'sm9',
 785             'title': 'sm9'
 786         },
 787         'playlist_mincount': 40,
 788     }, {
 789         'url': 'https://www.nicovideo.jp/search/sm9?sort=h&order=d&end=2020-12-31&start=2020-01-01',
 790         'info_dict': {
 791             'id': 'sm9',
 792             'title': 'sm9'
 793         },
 794         'playlist_count': 31,
 795     }]
 796
 797     def _real_extract(self, url):
 798         query = self._match_id(url)
 799         return self.playlist_result(self._entries(url, query), query, query)
 800
 801
 802 class NicovideoSearchDateIE(NicovideoSearchBaseIE, SearchInfoExtractor):
 803     IE_DESC = 'Nico video search, newest first'
 804     IE_NAME = f'{NicovideoSearchIE.IE_NAME}:date'
 805     _SEARCH_KEY = 'nicosearchdate'
 806     _TESTS = [{
 807         'url': 'nicosearchdateall:a',
 808         'info_dict': {
 809             'id': 'a',
 810             'title': 'a'
 811         },
 812         'playlist_mincount': 1610,
 813     }]
 814
 815     _START_DATE = datetime.date(2007, 1, 1)
 816     _RESULTS_PER_PAGE = 32
 817     _MAX_PAGES = 50
 818
 819     def _entries(self, url, item_id, start_date=None, end_date=None):
 820         start_date, end_date = start_date or self._START_DATE, end_date or datetime.datetime.now().date()
 821
 822         # If the last page has a full page of videos, we need to break down the query interval further
 823         last_page_len = len(list(self._get_entries_for_date(
 824             url, item_id, start_date, end_date, self._MAX_PAGES,
 825             note=f'Checking number of videos from {start_date} to {end_date}')))
 826         if (last_page_len == self._RESULTS_PER_PAGE and start_date != end_date):
 827             midpoint = start_date + ((end_date - start_date) // 2)
 828             yield from self._entries(url, item_id, midpoint, end_date)
 829             yield from self._entries(url, item_id, start_date, midpoint)
 830         else:
 831             self.to_screen(f'{item_id}: Downloading results from {start_date} to {end_date}')
 832             yield from self._get_entries_for_date(
 833                 url, item_id, start_date, end_date, note='    Downloading page %(page)s')
 834
 835     def _get_entries_for_date(self, url, item_id, start_date, end_date=None, page_num=None, note=None):
 836         query = {
 837             'start': str(start_date),
 838             'end': str(end_date or start_date),
 839             'sort': 'f',
 840             'order': 'd',
 841         }
 842         if page_num:
 843             query['page'] = str(page_num)
 844
 845         yield from super()._entries(url, item_id, query=query, note=note)
 846
 847
 848 class NicovideoTagURLIE(NicovideoSearchBaseIE):
 849     IE_NAME = 'niconico:tag'
 850     IE_DESC = 'NicoNico video tag URLs'
 851     _SEARCH_TYPE = 'tag'
 852     _VALID_URL = r'https?://(?:www\.)?nicovideo\.jp/tag/(?P<id>[^?#&]+)?'
 853     _TESTS = [{
 854         'url': 'https://www.nicovideo.jp/tag/ドキュメンタリー淫夢',
 855         'info_dict': {
 856             'id': 'ドキュメンタリー淫夢',
 857             'title': 'ドキュメンタリー淫夢'
 858         },
 859         'playlist_mincount': 400,
 860     }]
 861
 862     def _real_extract(self, url):
 863         query = self._match_id(url)
 864         return self.playlist_result(self._entries(url, query), query, query)
 865
 866
 867 class NiconicoUserIE(InfoExtractor):
 868     _VALID_URL = r'https?://(?:www\.)?nicovideo\.jp/user/(?P<id>\d+)/?(?:$|[#?])'
 869     _TEST = {
 870         'url': 'https://www.nicovideo.jp/user/419948',
 871         'info_dict': {
 872             'id': '419948',
 873         },
 874         'playlist_mincount': 101,
 875     }
 876     _API_URL = "https://nvapi.nicovideo.jp/v1/users/%s/videos?sortKey=registeredAt&sortOrder=desc&pageSize=%s&page=%s"
 877     _PAGE_SIZE = 100
 878
 879     _API_HEADERS = {
 880         'X-Frontend-ID': '6',
 881         'X-Frontend-Version': '0'
 882     }
 883
 884     def _entries(self, list_id):
 885         total_count = 1
 886         count = page_num = 0
 887         while count < total_count:
 888             json_parsed = self._download_json(
 889                 self._API_URL % (list_id, self._PAGE_SIZE, page_num + 1), list_id,
 890                 headers=self._API_HEADERS,
 891                 note='Downloading JSON metadata%s' % (' page %d' % page_num if page_num else ''))
 892             if not page_num:
 893                 total_count = int_or_none(json_parsed['data'].get('totalCount'))
 894             for entry in json_parsed["data"]["items"]:
 895                 count += 1
 896                 yield self.url_result('https://www.nicovideo.jp/watch/%s' % entry['id'])
 897             page_num += 1
 898
 899     def _real_extract(self, url):
 900         list_id = self._match_id(url)
 901         return self.playlist_result(self._entries(list_id), list_id, ie=NiconicoIE.ie_key())
 902
 903
 904 class NiconicoLiveIE(InfoExtractor):
 905     IE_NAME = 'niconico:live'
 906     IE_DESC = 'ニコニコ生放送'
 907     _VALID_URL = r'https?://(?:sp\.)?live2?\.nicovideo\.jp/(?:watch|gate)/(?P<id>lv\d+)'
 908     _TESTS = [{
 909         'note': 'this test case includes invisible characters for title, pasting them as-is',
 910         'url': 'https://live.nicovideo.jp/watch/lv339533123',
 911         'info_dict': {
 912             'id': 'lv339533123',
 913             'title': '激辛ペヤング食べます‪( ;ᯅ; )‬（歌枠オーディション参加中）',
 914             'view_count': 1526,
 915             'comment_count': 1772,
 916             'description': '初めましてもかって言います❕\nのんびり自由に適当に暮らしてます',
 917             'uploader': 'もか',
 918             'channel': 'ゲストさんのコミュニティ',
 919             'channel_id': 'co5776900',
 920             'channel_url': 'https://com.nicovideo.jp/community/co5776900',
 921             'timestamp': 1670677328,
 922             'is_live': True,
 923         },
 924         'skip': 'livestream',
 925     }, {
 926         'url': 'https://live2.nicovideo.jp/watch/lv339533123',
 927         'only_matching': True,
 928     }, {
 929         'url': 'https://sp.live.nicovideo.jp/watch/lv339533123',
 930         'only_matching': True,
 931     }, {
 932         'url': 'https://sp.live2.nicovideo.jp/watch/lv339533123',
 933         'only_matching': True,
 934     }]
 935
 936     _KNOWN_LATENCY = ('high', 'low')
 937
 938     def _real_extract(self, url):
 939         if not websockets:
 940             raise ExtractorError('websockets library is not available. Please install it.', expected=True)
 941         video_id = self._match_id(url)
 942         webpage, urlh = self._download_webpage_handle(f'https://live.nicovideo.jp/watch/{video_id}', video_id)
 943
 944         embedded_data = self._parse_json(unescapeHTML(self._search_regex(
 945             r'<script\s+id="embedded-data"\s*data-props="(.+?)"', webpage, 'embedded data')), video_id)
 946
 947         ws_url = traverse_obj(embedded_data, ('site', 'relive', 'webSocketUrl'))
 948         if not ws_url:
 949             raise ExtractorError('The live hasn\'t started yet or already ended.', expected=True)
 950         ws_url = update_url_query(ws_url, {
 951             'frontend_id': traverse_obj(embedded_data, ('site', 'frontendId')) or '9',
 952         })
 953
 954         hostname = remove_start(urlparse(urlh.geturl()).hostname, 'sp.')
 955         cookies = try_get(urlh.geturl(), self._downloader._calc_cookies)
 956         latency = try_get(self._configuration_arg('latency'), lambda x: x[0])
 957         if latency not in self._KNOWN_LATENCY:
 958             latency = 'high'
 959
 960         ws = WebSocketsWrapper(ws_url, {
 961             'Cookies': str_or_none(cookies) or '',
 962             'Origin': f'https://{hostname}',
 963             'Accept': '*/*',
 964             'User-Agent': self.get_param('http_headers')['User-Agent'],
 965         })
 966
 967         self.write_debug('[debug] Sending HLS server request')
 968         ws.send(json.dumps({
 969             'type': 'startWatching',
 970             'data': {
 971                 'stream': {
 972                     'quality': 'abr',
 973                     'protocol': 'hls+fmp4',
 974                     'latency': latency,
 975                     'chasePlay': False
 976                 },
 977                 'room': {
 978                     'protocol': 'webSocket',
 979                     'commentable': True
 980                 },
 981                 'reconnect': False,
 982             }
 983         }))
 984
 985         while True:
 986             recv = ws.recv()
 987             if not recv:
 988                 continue
 989             data = json.loads(recv)
 990             if not isinstance(data, dict):
 991                 continue
 992             if data.get('type') == 'stream':
 993                 m3u8_url = data['data']['uri']
 994                 qualities = data['data']['availableQualities']
 995                 break
 996             elif data.get('type') == 'disconnect':
 997                 self.write_debug(recv)
 998                 raise ExtractorError('Disconnected at middle of extraction')
 999             elif data.get('type') == 'error':
1000                 self.write_debug(recv)
1001                 message = traverse_obj(data, ('body', 'code')) or recv
1002                 raise ExtractorError(message)
1003             elif self.get_param('verbose', False):
1004                 if len(recv) > 100:
1005                     recv = recv[:100] + '...'
1006                 self.write_debug('Server said: %s' % recv)
1007
1008         title = traverse_obj(embedded_data, ('program', 'title')) or self._html_search_meta(
1009             ('og:title', 'twitter:title'), webpage, 'live title', fatal=False)
1010
1011         raw_thumbs = traverse_obj(embedded_data, ('program', 'thumbnail')) or {}
1012         thumbnails = []
1013         for name, value in raw_thumbs.items():
1014             if not isinstance(value, dict):
1015                 thumbnails.append({
1016                     'id': name,
1017                     'url': value,
1018                     **parse_resolution(value, lenient=True),
1019                 })
1020                 continue
1021
1022             for k, img_url in value.items():
1023                 res = parse_resolution(k, lenient=True) or parse_resolution(img_url, lenient=True)
1024                 width, height = res.get('width'), res.get('height')
1025
1026                 thumbnails.append({
1027                     'id': f'{name}_{width}x{height}',
1028                     'url': img_url,
1029                     **res,
1030                 })
1031
1032         formats = self._extract_m3u8_formats(m3u8_url, video_id, ext='mp4', live=True)
1033         for fmt, q in zip(formats, reversed(qualities[1:])):
1034             fmt.update({
1035                 'format_id': q,
1036                 'protocol': 'niconico_live',
1037                 'ws': ws,
1038                 'video_id': video_id,
1039                 'cookies': cookies,
1040                 'live_latency': latency,
1041                 'origin': hostname,
1042             })
1043
1044         return {
1045             'id': video_id,
1046             'title': title,
1047             **traverse_obj(embedded_data, {
1048                 'view_count': ('program', 'statistics', 'watchCount'),
1049                 'comment_count': ('program', 'statistics', 'commentCount'),
1050                 'uploader': ('program', 'supplier', 'name'),
1051                 'channel': ('socialGroup', 'name'),
1052                 'channel_id': ('socialGroup', 'id'),
1053                 'channel_url': ('socialGroup', 'socialGroupPageUrl'),
1054             }),
1055             'description': clean_html(traverse_obj(embedded_data, ('program', 'description'))),
1056             'timestamp': int_or_none(traverse_obj(embedded_data, ('program', 'openTime'))),
1057             'is_live': True,
1058             'thumbnails': thumbnails,
1059             'formats': formats,
1060         }