yt_dlp/extractor/crunchyroll.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import base64
   5 import re
   6 import json
   7 import zlib
   8
   9 from hashlib import sha1
  10 from math import pow, sqrt, floor
  11 from .common import InfoExtractor
  12 from .vrv import VRVBaseIE
  13 from ..compat import (
  14     compat_b64decode,
  15     compat_etree_Element,
  16     compat_etree_fromstring,
  17     compat_str,
  18     compat_urllib_parse_urlencode,
  19     compat_urllib_request,
  20     compat_urlparse,
  21 )
  22 from ..utils import (
  23     ExtractorError,
  24     bytes_to_intlist,
  25     extract_attributes,
  26     float_or_none,
  27     format_field,
  28     intlist_to_bytes,
  29     int_or_none,
  30     join_nonempty,
  31     lowercase_escape,
  32     merge_dicts,
  33     qualities,
  34     remove_end,
  35     sanitized_Request,
  36     traverse_obj,
  37     try_get,
  38     xpath_text,
  39 )
  40 from ..aes import (
  41     aes_cbc_decrypt,
  42 )
  43
  44
  45 class CrunchyrollBaseIE(InfoExtractor):
  46     _LOGIN_URL = 'https://www.crunchyroll.com/welcome/login'
  47     _API_BASE = 'https://api.crunchyroll.com'
  48     _NETRC_MACHINE = 'crunchyroll'
  49
  50     def _call_rpc_api(self, method, video_id, note=None, data=None):
  51         data = data or {}
  52         data['req'] = 'RpcApi' + method
  53         data = compat_urllib_parse_urlencode(data).encode('utf-8')
  54         return self._download_xml(
  55             'https://www.crunchyroll.com/xml/',
  56             video_id, note, fatal=False, data=data, headers={
  57                 'Content-Type': 'application/x-www-form-urlencoded',
  58             })
  59
  60     def _perform_login(self, username, password):
  61         if self._get_cookies(self._LOGIN_URL).get('etp_rt'):
  62             return
  63
  64         upsell_response = self._download_json(
  65             f'{self._API_BASE}/get_upsell_data.0.json', None, 'Getting session id',
  66             query={
  67                 'sess_id': 1,
  68                 'device_id': 'whatvalueshouldbeforweb',
  69                 'device_type': 'com.crunchyroll.static',
  70                 'access_token': 'giKq5eY27ny3cqz',
  71                 'referer': self._LOGIN_URL
  72             })
  73         if upsell_response['code'] != 'ok':
  74             raise ExtractorError('Could not get session id')
  75         session_id = upsell_response['data']['session_id']
  76
  77         login_response = self._download_json(
  78             f'{self._API_BASE}/login.1.json', None, 'Logging in',
  79             data=compat_urllib_parse_urlencode({
  80                 'account': username,
  81                 'password': password,
  82                 'session_id': session_id
  83             }).encode('ascii'))
  84         if login_response['code'] != 'ok':
  85             raise ExtractorError('Login failed. Server message: %s' % login_response['message'], expected=True)
  86         if not self._get_cookies(self._LOGIN_URL).get('etp_rt'):
  87             raise ExtractorError('Login succeeded but did not set etp_rt cookie')
  88
  89     @staticmethod
  90     def _add_skip_wall(url):
  91         parsed_url = compat_urlparse.urlparse(url)
  92         qs = compat_urlparse.parse_qs(parsed_url.query)
  93         # Always force skip_wall to bypass maturity wall, namely 18+ confirmation message:
  94         # > This content may be inappropriate for some people.
  95         # > Are you sure you want to continue?
  96         # since it's not disabled by default in crunchyroll account's settings.
  97         # See https://github.com/ytdl-org/youtube-dl/issues/7202.
  98         qs['skip_wall'] = ['1']
  99         return compat_urlparse.urlunparse(
 100             parsed_url._replace(query=compat_urllib_parse_urlencode(qs, True)))
 101
 102
 103 class CrunchyrollIE(CrunchyrollBaseIE, VRVBaseIE):
 104     IE_NAME = 'crunchyroll'
 105     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.(?:com|fr)/(?:media(?:-|/\?id=)|(?:[^/]*/){1,2}[^/?&]*?)(?P<id>[0-9]+))(?:[/?&]|$)'
 106     _TESTS = [{
 107         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
 108         'info_dict': {
 109             'id': '645513',
 110             'ext': 'mp4',
 111             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
 112             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
 113             'thumbnail': r're:^https?://.*\.jpg$',
 114             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
 115             'upload_date': '20131013',
 116             'url': 're:(?!.*&amp)',
 117         },
 118         'params': {
 119             # rtmp
 120             'skip_download': True,
 121         },
 122         'skip': 'Video gone',
 123     }, {
 124         'url': 'http://www.crunchyroll.com/media-589804/culture-japan-1',
 125         'info_dict': {
 126             'id': '589804',
 127             'ext': 'flv',
 128             'title': 'Culture Japan Episode 1 – Rebuilding Japan after the 3.11',
 129             'description': 'md5:2fbc01f90b87e8e9137296f37b461c12',
 130             'thumbnail': r're:^https?://.*\.jpg$',
 131             'uploader': 'Danny Choo Network',
 132             'upload_date': '20120213',
 133         },
 134         'params': {
 135             # rtmp
 136             'skip_download': True,
 137         },
 138         'skip': 'Video gone',
 139     }, {
 140         'url': 'http://www.crunchyroll.com/rezero-starting-life-in-another-world-/episode-5-the-morning-of-our-promise-is-still-distant-702409',
 141         'info_dict': {
 142             'id': '702409',
 143             'ext': 'mp4',
 144             'title': compat_str,
 145             'description': compat_str,
 146             'thumbnail': r're:^https?://.*\.jpg$',
 147             'uploader': 'Re:Zero Partners',
 148             'timestamp': 1462098900,
 149             'upload_date': '20160501',
 150         },
 151         'params': {
 152             # m3u8 download
 153             'skip_download': True,
 154         },
 155     }, {
 156         'url': 'http://www.crunchyroll.com/konosuba-gods-blessing-on-this-wonderful-world/episode-1-give-me-deliverance-from-this-judicial-injustice-727589',
 157         'info_dict': {
 158             'id': '727589',
 159             'ext': 'mp4',
 160             'title': compat_str,
 161             'description': compat_str,
 162             'thumbnail': r're:^https?://.*\.jpg$',
 163             'uploader': 'Kadokawa Pictures Inc.',
 164             'timestamp': 1484130900,
 165             'upload_date': '20170111',
 166             'series': compat_str,
 167             'season': "KONOSUBA -God's blessing on this wonderful world! 2",
 168             'season_number': 2,
 169             'episode': 'Give Me Deliverance From This Judicial Injustice!',
 170             'episode_number': 1,
 171         },
 172         'params': {
 173             # m3u8 download
 174             'skip_download': True,
 175         },
 176     }, {
 177         'url': 'http://www.crunchyroll.fr/girl-friend-beta/episode-11-goodbye-la-mode-661697',
 178         'only_matching': True,
 179     }, {
 180         # geo-restricted (US), 18+ maturity wall, non-premium available
 181         'url': 'http://www.crunchyroll.com/cosplay-complex-ova/episode-1-the-birth-of-the-cosplay-club-565617',
 182         'only_matching': True,
 183     }, {
 184         # A description with double quotes
 185         'url': 'http://www.crunchyroll.com/11eyes/episode-1-piros-jszaka-red-night-535080',
 186         'info_dict': {
 187             'id': '535080',
 188             'ext': 'mp4',
 189             'title': compat_str,
 190             'description': compat_str,
 191             'uploader': 'Marvelous AQL Inc.',
 192             'timestamp': 1255512600,
 193             'upload_date': '20091014',
 194         },
 195         'params': {
 196             # Just test metadata extraction
 197             'skip_download': True,
 198         },
 199     }, {
 200         # make sure we can extract an uploader name that's not a link
 201         'url': 'http://www.crunchyroll.com/hakuoki-reimeiroku/episode-1-dawn-of-the-divine-warriors-606899',
 202         'info_dict': {
 203             'id': '606899',
 204             'ext': 'mp4',
 205             'title': 'Hakuoki Reimeiroku Episode 1 – Dawn of the Divine Warriors',
 206             'description': 'Ryunosuke was left to die, but Serizawa-san asked him a simple question "Do you want to live?"',
 207             'uploader': 'Geneon Entertainment',
 208             'upload_date': '20120717',
 209         },
 210         'params': {
 211             # just test metadata extraction
 212             'skip_download': True,
 213         },
 214         'skip': 'Video gone',
 215     }, {
 216         # A video with a vastly different season name compared to the series name
 217         'url': 'http://www.crunchyroll.com/nyarko-san-another-crawling-chaos/episode-1-test-590532',
 218         'info_dict': {
 219             'id': '590532',
 220             'ext': 'mp4',
 221             'title': compat_str,
 222             'description': compat_str,
 223             'uploader': 'TV TOKYO',
 224             'timestamp': 1330956000,
 225             'upload_date': '20120305',
 226             'series': 'Nyarko-san: Another Crawling Chaos',
 227             'season': 'Haiyoru! Nyaruani (ONA)',
 228         },
 229         'params': {
 230             # Just test metadata extraction
 231             'skip_download': True,
 232         },
 233     }, {
 234         'url': 'http://www.crunchyroll.com/media-723735',
 235         'only_matching': True,
 236     }, {
 237         'url': 'https://www.crunchyroll.com/en-gb/mob-psycho-100/episode-2-urban-legends-encountering-rumors-780921',
 238         'only_matching': True,
 239     }]
 240
 241     _FORMAT_IDS = {
 242         '360': ('60', '106'),
 243         '480': ('61', '106'),
 244         '720': ('62', '106'),
 245         '1080': ('80', '108'),
 246     }
 247
 248     def _download_webpage(self, url_or_request, *args, **kwargs):
 249         request = (url_or_request if isinstance(url_or_request, compat_urllib_request.Request)
 250                    else sanitized_Request(url_or_request))
 251         # Accept-Language must be set explicitly to accept any language to avoid issues
 252         # similar to https://github.com/ytdl-org/youtube-dl/issues/6797.
 253         # Along with IP address Crunchyroll uses Accept-Language to guess whether georestriction
 254         # should be imposed or not (from what I can see it just takes the first language
 255         # ignoring the priority and requires it to correspond the IP). By the way this causes
 256         # Crunchyroll to not work in georestriction cases in some browsers that don't place
 257         # the locale lang first in header. However allowing any language seems to workaround the issue.
 258         request.add_header('Accept-Language', '*')
 259         return super(CrunchyrollBaseIE, self)._download_webpage(request, *args, **kwargs)
 260
 261     def _decrypt_subtitles(self, data, iv, id):
 262         data = bytes_to_intlist(compat_b64decode(data))
 263         iv = bytes_to_intlist(compat_b64decode(iv))
 264         id = int(id)
 265
 266         def obfuscate_key_aux(count, modulo, start):
 267             output = list(start)
 268             for _ in range(count):
 269                 output.append(output[-1] + output[-2])
 270             # cut off start values
 271             output = output[2:]
 272             output = list(map(lambda x: x % modulo + 33, output))
 273             return output
 274
 275         def obfuscate_key(key):
 276             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
 277             num2 = (num1 ^ key) << 5
 278             num3 = key ^ num1
 279             num4 = num3 ^ (num3 >> 3) ^ num2
 280             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
 281             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
 282             # Extend 160 Bit hash to 256 Bit
 283             return shaHash + [0] * 12
 284
 285         key = obfuscate_key(id)
 286
 287         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
 288         return zlib.decompress(decrypted_data)
 289
 290     def _convert_subtitles_to_srt(self, sub_root):
 291         output = ''
 292
 293         for i, event in enumerate(sub_root.findall('./events/event'), 1):
 294             start = event.attrib['start'].replace('.', ',')
 295             end = event.attrib['end'].replace('.', ',')
 296             text = event.attrib['text'].replace('\\N', '\n')
 297             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
 298         return output
 299
 300     def _convert_subtitles_to_ass(self, sub_root):
 301         output = ''
 302
 303         def ass_bool(strvalue):
 304             assvalue = '0'
 305             if strvalue == '1':
 306                 assvalue = '-1'
 307             return assvalue
 308
 309         output = '[Script Info]\n'
 310         output += 'Title: %s\n' % sub_root.attrib['title']
 311         output += 'ScriptType: v4.00+\n'
 312         output += 'WrapStyle: %s\n' % sub_root.attrib['wrap_style']
 313         output += 'PlayResX: %s\n' % sub_root.attrib['play_res_x']
 314         output += 'PlayResY: %s\n' % sub_root.attrib['play_res_y']
 315         output += """
 316 [V4+ Styles]
 317 Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
 318 """
 319         for style in sub_root.findall('./styles/style'):
 320             output += 'Style: ' + style.attrib['name']
 321             output += ',' + style.attrib['font_name']
 322             output += ',' + style.attrib['font_size']
 323             output += ',' + style.attrib['primary_colour']
 324             output += ',' + style.attrib['secondary_colour']
 325             output += ',' + style.attrib['outline_colour']
 326             output += ',' + style.attrib['back_colour']
 327             output += ',' + ass_bool(style.attrib['bold'])
 328             output += ',' + ass_bool(style.attrib['italic'])
 329             output += ',' + ass_bool(style.attrib['underline'])
 330             output += ',' + ass_bool(style.attrib['strikeout'])
 331             output += ',' + style.attrib['scale_x']
 332             output += ',' + style.attrib['scale_y']
 333             output += ',' + style.attrib['spacing']
 334             output += ',' + style.attrib['angle']
 335             output += ',' + style.attrib['border_style']
 336             output += ',' + style.attrib['outline']
 337             output += ',' + style.attrib['shadow']
 338             output += ',' + style.attrib['alignment']
 339             output += ',' + style.attrib['margin_l']
 340             output += ',' + style.attrib['margin_r']
 341             output += ',' + style.attrib['margin_v']
 342             output += ',' + style.attrib['encoding']
 343             output += '\n'
 344
 345         output += """
 346 [Events]
 347 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
 348 """
 349         for event in sub_root.findall('./events/event'):
 350             output += 'Dialogue: 0'
 351             output += ',' + event.attrib['start']
 352             output += ',' + event.attrib['end']
 353             output += ',' + event.attrib['style']
 354             output += ',' + event.attrib['name']
 355             output += ',' + event.attrib['margin_l']
 356             output += ',' + event.attrib['margin_r']
 357             output += ',' + event.attrib['margin_v']
 358             output += ',' + event.attrib['effect']
 359             output += ',' + event.attrib['text']
 360             output += '\n'
 361
 362         return output
 363
 364     def _extract_subtitles(self, subtitle):
 365         sub_root = compat_etree_fromstring(subtitle)
 366         return [{
 367             'ext': 'srt',
 368             'data': self._convert_subtitles_to_srt(sub_root),
 369         }, {
 370             'ext': 'ass',
 371             'data': self._convert_subtitles_to_ass(sub_root),
 372         }]
 373
 374     def _get_subtitles(self, video_id, webpage):
 375         subtitles = {}
 376         for sub_id, sub_name in re.findall(r'\bssid=([0-9]+)"[^>]+?\btitle="([^"]+)', webpage):
 377             sub_doc = self._call_rpc_api(
 378                 'Subtitle_GetXml', video_id,
 379                 'Downloading subtitles for ' + sub_name, data={
 380                     'subtitle_script_id': sub_id,
 381                 })
 382             if not isinstance(sub_doc, compat_etree_Element):
 383                 continue
 384             sid = sub_doc.get('id')
 385             iv = xpath_text(sub_doc, 'iv', 'subtitle iv')
 386             data = xpath_text(sub_doc, 'data', 'subtitle data')
 387             if not sid or not iv or not data:
 388                 continue
 389             subtitle = self._decrypt_subtitles(data, iv, sid).decode('utf-8')
 390             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
 391             if not lang_code:
 392                 continue
 393             subtitles[lang_code] = self._extract_subtitles(subtitle)
 394         return subtitles
 395
 396     def _real_extract(self, url):
 397         mobj = self._match_valid_url(url)
 398         video_id = mobj.group('id')
 399
 400         if mobj.group('prefix') == 'm':
 401             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
 402             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
 403         else:
 404             webpage_url = 'http://www.' + mobj.group('url')
 405
 406         webpage = self._download_webpage(
 407             self._add_skip_wall(webpage_url), video_id,
 408             headers=self.geo_verification_headers())
 409         note_m = self._html_search_regex(
 410             r'<div class="showmedia-trailer-notice">(.+?)</div>',
 411             webpage, 'trailer-notice', default='')
 412         if note_m:
 413             raise ExtractorError(note_m, expected=True)
 414
 415         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
 416         if mobj:
 417             msg = json.loads(mobj.group('msg'))
 418             if msg.get('type') == 'error':
 419                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
 420
 421         if 'To view this, please log in to verify you are 18 or older.' in webpage:
 422             self.raise_login_required()
 423
 424         media = self._parse_json(self._search_regex(
 425             r'vilos\.config\.media\s*=\s*({.+?});',
 426             webpage, 'vilos media', default='{}'), video_id)
 427         media_metadata = media.get('metadata') or {}
 428
 429         language = self._search_regex(
 430             r'(?:vilos\.config\.player\.language|LOCALE)\s*=\s*(["\'])(?P<lang>(?:(?!\1).)+)\1',
 431             webpage, 'language', default=None, group='lang')
 432
 433         video_title = self._html_search_regex(
 434             (r'(?s)<h1[^>]*>((?:(?!<h1).)*?<(?:span[^>]+itemprop=["\']title["\']|meta[^>]+itemprop=["\']position["\'])[^>]*>(?:(?!<h1).)+?)</h1>',
 435              r'<title>(.+?),\s+-\s+.+? Crunchyroll'),
 436             webpage, 'video_title', default=None)
 437         if not video_title:
 438             video_title = re.sub(r'^Watch\s+', '', self._og_search_description(webpage))
 439         video_title = re.sub(r' {2,}', ' ', video_title)
 440         video_description = (self._parse_json(self._html_search_regex(
 441             r'<script[^>]*>\s*.+?\[media_id=%s\].+?({.+?"description"\s*:.+?})\);' % video_id,
 442             webpage, 'description', default='{}'), video_id) or media_metadata).get('description')
 443
 444         thumbnails = []
 445         thumbnail_url = (self._parse_json(self._html_search_regex(
 446             r'<script type="application\/ld\+json">\n\s*(.+?)<\/script>',
 447             webpage, 'thumbnail_url', default='{}'), video_id)).get('image')
 448         if thumbnail_url:
 449             thumbnails.append({
 450                 'url': thumbnail_url,
 451                 'width': 1920,
 452                 'height': 1080
 453             })
 454
 455         if video_description:
 456             video_description = lowercase_escape(video_description.replace(r'\r\n', '\n'))
 457         video_uploader = self._html_search_regex(
 458             # try looking for both an uploader that's a link and one that's not
 459             [r'<a[^>]+href="/publisher/[^"]+"[^>]*>([^<]+)</a>', r'<div>\s*Publisher:\s*<span>\s*(.+?)\s*</span>\s*</div>'],
 460             webpage, 'video_uploader', default=False)
 461
 462         requested_languages = self._configuration_arg('language')
 463         requested_hardsubs = [('' if val == 'none' else val) for val in self._configuration_arg('hardsub')]
 464         language_preference = qualities((requested_languages or [language or ''])[::-1])
 465         hardsub_preference = qualities((requested_hardsubs or ['', language or ''])[::-1])
 466
 467         formats = []
 468         for stream in media.get('streams', []):
 469             audio_lang = stream.get('audio_lang') or ''
 470             hardsub_lang = stream.get('hardsub_lang') or ''
 471             if (requested_languages and audio_lang.lower() not in requested_languages
 472                     or requested_hardsubs and hardsub_lang.lower() not in requested_hardsubs):
 473                 continue
 474             vrv_formats = self._extract_vrv_formats(
 475                 stream.get('url'), video_id, stream.get('format'),
 476                 audio_lang, hardsub_lang)
 477             for f in vrv_formats:
 478                 f['language_preference'] = language_preference(audio_lang)
 479                 f['quality'] = hardsub_preference(hardsub_lang)
 480             formats.extend(vrv_formats)
 481         if not formats:
 482             available_fmts = []
 483             for a, fmt in re.findall(r'(<a[^>]+token=["\']showmedia\.([0-9]{3,4})p["\'][^>]+>)', webpage):
 484                 attrs = extract_attributes(a)
 485                 href = attrs.get('href')
 486                 if href and '/freetrial' in href:
 487                     continue
 488                 available_fmts.append(fmt)
 489             if not available_fmts:
 490                 for p in (r'token=["\']showmedia\.([0-9]{3,4})p"', r'showmedia\.([0-9]{3,4})p'):
 491                     available_fmts = re.findall(p, webpage)
 492                     if available_fmts:
 493                         break
 494             if not available_fmts:
 495                 available_fmts = self._FORMAT_IDS.keys()
 496             video_encode_ids = []
 497
 498             for fmt in available_fmts:
 499                 stream_quality, stream_format = self._FORMAT_IDS[fmt]
 500                 video_format = fmt + 'p'
 501                 stream_infos = []
 502                 streamdata = self._call_rpc_api(
 503                     'VideoPlayer_GetStandardConfig', video_id,
 504                     'Downloading media info for %s' % video_format, data={
 505                         'media_id': video_id,
 506                         'video_format': stream_format,
 507                         'video_quality': stream_quality,
 508                         'current_page': url,
 509                     })
 510                 if isinstance(streamdata, compat_etree_Element):
 511                     stream_info = streamdata.find('./{default}preload/stream_info')
 512                     if stream_info is not None:
 513                         stream_infos.append(stream_info)
 514                 stream_info = self._call_rpc_api(
 515                     'VideoEncode_GetStreamInfo', video_id,
 516                     'Downloading stream info for %s' % video_format, data={
 517                         'media_id': video_id,
 518                         'video_format': stream_format,
 519                         'video_encode_quality': stream_quality,
 520                     })
 521                 if isinstance(stream_info, compat_etree_Element):
 522                     stream_infos.append(stream_info)
 523                 for stream_info in stream_infos:
 524                     video_encode_id = xpath_text(stream_info, './video_encode_id')
 525                     if video_encode_id in video_encode_ids:
 526                         continue
 527                     video_encode_ids.append(video_encode_id)
 528
 529                     video_file = xpath_text(stream_info, './file')
 530                     if not video_file:
 531                         continue
 532                     if video_file.startswith('http'):
 533                         formats.extend(self._extract_m3u8_formats(
 534                             video_file, video_id, 'mp4', entry_protocol='m3u8_native',
 535                             m3u8_id='hls', fatal=False))
 536                         continue
 537
 538                     video_url = xpath_text(stream_info, './host')
 539                     if not video_url:
 540                         continue
 541                     metadata = stream_info.find('./metadata')
 542                     format_info = {
 543                         'format': video_format,
 544                         'height': int_or_none(xpath_text(metadata, './height')),
 545                         'width': int_or_none(xpath_text(metadata, './width')),
 546                     }
 547
 548                     if '.fplive.net/' in video_url:
 549                         video_url = re.sub(r'^rtmpe?://', 'http://', video_url.strip())
 550                         parsed_video_url = compat_urlparse.urlparse(video_url)
 551                         direct_video_url = compat_urlparse.urlunparse(parsed_video_url._replace(
 552                             netloc='v.lvlt.crcdn.net',
 553                             path='%s/%s' % (remove_end(parsed_video_url.path, '/'), video_file.split(':')[-1])))
 554                         if self._is_valid_url(direct_video_url, video_id, video_format):
 555                             format_info.update({
 556                                 'format_id': 'http-' + video_format,
 557                                 'url': direct_video_url,
 558                             })
 559                             formats.append(format_info)
 560                             continue
 561
 562                     format_info.update({
 563                         'format_id': 'rtmp-' + video_format,
 564                         'url': video_url,
 565                         'play_path': video_file,
 566                         'ext': 'flv',
 567                     })
 568                     formats.append(format_info)
 569         self._sort_formats(formats)
 570
 571         metadata = self._call_rpc_api(
 572             'VideoPlayer_GetMediaMetadata', video_id,
 573             note='Downloading media info', data={
 574                 'media_id': video_id,
 575             })
 576
 577         subtitles = {}
 578         for subtitle in media.get('subtitles', []):
 579             subtitle_url = subtitle.get('url')
 580             if not subtitle_url:
 581                 continue
 582             subtitles.setdefault(subtitle.get('language', 'enUS'), []).append({
 583                 'url': subtitle_url,
 584                 'ext': subtitle.get('format', 'ass'),
 585             })
 586         if not subtitles:
 587             subtitles = self.extract_subtitles(video_id, webpage)
 588
 589         # webpage provide more accurate data than series_title from XML
 590         series = self._html_search_regex(
 591             r'(?s)<h\d[^>]+\bid=["\']showmedia_about_episode_num[^>]+>(.+?)</h\d',
 592             webpage, 'series', fatal=False)
 593
 594         season = episode = episode_number = duration = None
 595
 596         if isinstance(metadata, compat_etree_Element):
 597             season = xpath_text(metadata, 'series_title')
 598             episode = xpath_text(metadata, 'episode_title')
 599             episode_number = int_or_none(xpath_text(metadata, 'episode_number'))
 600             duration = float_or_none(media_metadata.get('duration'), 1000)
 601
 602         if not episode:
 603             episode = media_metadata.get('title')
 604         if not episode_number:
 605             episode_number = int_or_none(media_metadata.get('episode_number'))
 606         thumbnail_url = try_get(media, lambda x: x['thumbnail']['url'])
 607         if thumbnail_url:
 608             thumbnails.append({
 609                 'url': thumbnail_url,
 610                 'width': 640,
 611                 'height': 360
 612             })
 613
 614         season_number = int_or_none(self._search_regex(
 615             r'(?s)<h\d[^>]+id=["\']showmedia_about_episode_num[^>]+>.+?</h\d>\s*<h4>\s*Season (\d+)',
 616             webpage, 'season number', default=None))
 617
 618         info = self._search_json_ld(webpage, video_id, default={})
 619
 620         return merge_dicts({
 621             'id': video_id,
 622             'title': video_title,
 623             'description': video_description,
 624             'duration': duration,
 625             'thumbnails': thumbnails,
 626             'uploader': video_uploader,
 627             'series': series,
 628             'season': season,
 629             'season_number': season_number,
 630             'episode': episode,
 631             'episode_number': episode_number,
 632             'subtitles': subtitles,
 633             'formats': formats,
 634         }, info)
 635
 636
 637 class CrunchyrollShowPlaylistIE(CrunchyrollBaseIE):
 638     IE_NAME = 'crunchyroll:playlist'
 639     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.com/(?:\w{1,2}/)?(?!(?:news|anime-news|library|forum|launchcalendar|lineup|store|comics|freetrial|login|media-\d+))(?P<id>[\w\-]+))/?(?:\?|$)'
 640
 641     _TESTS = [{
 642         'url': 'https://www.crunchyroll.com/a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 643         'info_dict': {
 644             'id': 'a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 645             'title': 'A Bridge to the Starry Skies - Hoshizora e Kakaru Hashi'
 646         },
 647         'playlist_count': 13,
 648     }, {
 649         # geo-restricted (US), 18+ maturity wall, non-premium available
 650         'url': 'http://www.crunchyroll.com/cosplay-complex-ova',
 651         'info_dict': {
 652             'id': 'cosplay-complex-ova',
 653             'title': 'Cosplay Complex OVA'
 654         },
 655         'playlist_count': 3,
 656         'skip': 'Georestricted',
 657     }, {
 658         # geo-restricted (US), 18+ maturity wall, non-premium will be available since 2015.11.14
 659         'url': 'http://www.crunchyroll.com/ladies-versus-butlers?skip_wall=1',
 660         'only_matching': True,
 661     }, {
 662         'url': 'http://www.crunchyroll.com/fr/ladies-versus-butlers',
 663         'only_matching': True,
 664     }]
 665
 666     def _real_extract(self, url):
 667         show_id = self._match_id(url)
 668
 669         webpage = self._download_webpage(
 670             # https:// gives a 403, but http:// does not
 671             self._add_skip_wall(url).replace('https://', 'http://'), show_id,
 672             headers=self.geo_verification_headers())
 673         title = self._html_search_meta('name', webpage, default=None)
 674
 675         episode_re = r'<li id="showview_videos_media_(\d+)"[^>]+>.*?<a href="([^"]+)"'
 676         season_re = r'<a [^>]+season-dropdown[^>]+>([^<]+)'
 677         paths = re.findall(f'(?s){episode_re}|{season_re}', webpage)
 678
 679         entries, current_season = [], None
 680         for ep_id, ep, season in paths:
 681             if season:
 682                 current_season = season
 683                 continue
 684             entries.append(self.url_result(
 685                 f'http://www.crunchyroll.com{ep}', CrunchyrollIE.ie_key(), ep_id, season=current_season))
 686
 687         return {
 688             '_type': 'playlist',
 689             'id': show_id,
 690             'title': title,
 691             'entries': reversed(entries),
 692         }
 693
 694
 695 class CrunchyrollBetaIE(CrunchyrollBaseIE):
 696     IE_NAME = 'crunchyroll:beta'
 697     _VALID_URL = r'https?://beta\.crunchyroll\.com/(?P<lang>(?:\w{1,2}/)?)watch/(?P<internal_id>\w+)/(?P<id>[\w\-]+)/?(?:\?|$)'
 698     _TESTS = [{
 699         'url': 'https://beta.crunchyroll.com/watch/GY2P1Q98Y/to-the-future',
 700         'info_dict': {
 701             'id': '696363',
 702             'ext': 'mp4',
 703             'timestamp': 1459610100,
 704             'description': 'md5:a022fbec4fbb023d43631032c91ed64b',
 705             'uploader': 'Toei Animation',
 706             'title': 'World Trigger Episode 73 – To the Future',
 707             'upload_date': '20160402',
 708         },
 709         'params': {'skip_download': 'm3u8'},
 710         'expected_warnings': ['Unable to download XML']
 711     }]
 712
 713     def _real_extract(self, url):
 714         lang, internal_id, display_id = self._match_valid_url(url).group('lang', 'internal_id', 'id')
 715         webpage = self._download_webpage(url, display_id)
 716         initial_state = self._parse_json(
 717             self._search_regex(r'__INITIAL_STATE__\s*=\s*({.+?})\s*;', webpage, 'initial state'),
 718             display_id)
 719         episode_data = initial_state['content']['byId'][internal_id]
 720         if not self._get_cookies(url).get('etp_rt'):
 721             video_id = episode_data['external_id'].split('.')[1]
 722             series_id = episode_data['episode_metadata']['series_slug_title']
 723             return self.url_result(f'https://www.crunchyroll.com/{lang}{series_id}/{display_id}-{video_id}',
 724                                    CrunchyrollIE.ie_key(), video_id)
 725
 726         app_config = self._parse_json(
 727             self._search_regex(r'__APP_CONFIG__\s*=\s*({.+?})\s*;', webpage, 'app config'),
 728             display_id)
 729         client_id = app_config['cxApiParams']['accountAuthClientId']
 730         api_domain = app_config['cxApiParams']['apiDomain']
 731         basic_token = str(base64.b64encode(('%s:' % client_id).encode('ascii')), 'ascii')
 732         auth_response = self._download_json(
 733             f'{api_domain}/auth/v1/token', display_id,
 734             note='Authenticating with cookie',
 735             headers={
 736                 'Authorization': 'Basic ' + basic_token
 737             }, data='grant_type=etp_rt_cookie'.encode('ascii'))
 738         policy_response = self._download_json(
 739             f'{api_domain}/index/v2', display_id,
 740             note='Retrieving signed policy',
 741             headers={
 742                 'Authorization': auth_response['token_type'] + ' ' + auth_response['access_token']
 743             })
 744         bucket = policy_response['cms']['bucket']
 745         params = {
 746             'Policy': policy_response['cms']['policy'],
 747             'Signature': policy_response['cms']['signature'],
 748             'Key-Pair-Id': policy_response['cms']['key_pair_id']
 749         }
 750         locale = traverse_obj(initial_state, ('localization', 'locale'))
 751         if locale:
 752             params['locale'] = locale
 753         episode_response = self._download_json(
 754             f'{api_domain}/cms/v2{bucket}/episodes/{internal_id}', display_id,
 755             note='Retrieving episode metadata',
 756             query=params)
 757         if episode_response.get('is_premium_only') and not episode_response.get('playback'):
 758             raise ExtractorError('This video is for premium members only.', expected=True)
 759         stream_response = self._download_json(
 760             episode_response['playback'], display_id,
 761             note='Retrieving stream info')
 762
 763         thumbnails = []
 764         for thumbnails_data in traverse_obj(episode_response, ('images', 'thumbnail')):
 765             for thumbnail_data in thumbnails_data:
 766                 thumbnails.append({
 767                     'url': thumbnail_data.get('source'),
 768                     'width': thumbnail_data.get('width'),
 769                     'height': thumbnail_data.get('height'),
 770                 })
 771         subtitles = {}
 772         for lang, subtitle_data in stream_response.get('subtitles').items():
 773             subtitles[lang] = [{
 774                 'url': subtitle_data.get('url'),
 775                 'ext': subtitle_data.get('format')
 776             }]
 777
 778         requested_hardsubs = [('' if val == 'none' else val) for val in (self._configuration_arg('hardsub') or ['none'])]
 779         hardsub_preference = qualities(requested_hardsubs[::-1])
 780         requested_formats = self._configuration_arg('format') or ['adaptive_hls']
 781
 782         formats = []
 783         for stream_type, streams in stream_response.get('streams', {}).items():
 784             if stream_type not in requested_formats:
 785                 continue
 786             for stream in streams.values():
 787                 hardsub_lang = stream.get('hardsub_locale') or ''
 788                 if hardsub_lang.lower() not in requested_hardsubs:
 789                     continue
 790                 format_id = join_nonempty(
 791                     stream_type,
 792                     format_field(stream, 'hardsub_locale', 'hardsub-%s'))
 793                 if not stream.get('url'):
 794                     continue
 795                 if stream_type.split('_')[-1] == 'hls':
 796                     adaptive_formats = self._extract_m3u8_formats(
 797                         stream['url'], display_id, 'mp4', m3u8_id=format_id,
 798                         note='Downloading %s information' % format_id,
 799                         fatal=False)
 800                 elif stream_type.split('_')[-1] == 'dash':
 801                     adaptive_formats = self._extract_mpd_formats(
 802                         stream['url'], display_id, mpd_id=format_id,
 803                         note='Downloading %s information' % format_id,
 804                         fatal=False)
 805                 for f in adaptive_formats:
 806                     if f.get('acodec') != 'none':
 807                         f['language'] = stream_response.get('audio_locale')
 808                     f['quality'] = hardsub_preference(hardsub_lang.lower())
 809                 formats.extend(adaptive_formats)
 810         self._sort_formats(formats)
 811
 812         return {
 813             'id': internal_id,
 814             'title': '%s Episode %s – %s' % (episode_response.get('season_title'), episode_response.get('episode'), episode_response.get('title')),
 815             'description': episode_response.get('description').replace(r'\r\n', '\n'),
 816             'duration': float_or_none(episode_response.get('duration_ms'), 1000),
 817             'thumbnails': thumbnails,
 818             'series': episode_response.get('series_title'),
 819             'series_id': episode_response.get('series_id'),
 820             'season': episode_response.get('season_title'),
 821             'season_id': episode_response.get('season_id'),
 822             'season_number': episode_response.get('season_number'),
 823             'episode': episode_response.get('title'),
 824             'episode_number': episode_response.get('sequence_number'),
 825             'subtitles': subtitles,
 826             'formats': formats
 827         }
 828
 829
 830 class CrunchyrollBetaShowIE(CrunchyrollBaseIE):
 831     IE_NAME = 'crunchyroll:playlist:beta'
 832     _VALID_URL = r'https?://beta\.crunchyroll\.com/(?P<lang>(?:\w{1,2}/)?)series/\w+/(?P<id>[\w\-]+)/?(?:\?|$)'
 833     _TESTS = [{
 834         'url': 'https://beta.crunchyroll.com/series/GY19NQ2QR/Girl-Friend-BETA',
 835         'info_dict': {
 836             'id': 'girl-friend-beta',
 837             'title': 'Girl Friend BETA',
 838         },
 839         'playlist_mincount': 10,
 840     }, {
 841         'url': 'https://beta.crunchyroll.com/it/series/GY19NQ2QR/Girl-Friend-BETA',
 842         'only_matching': True,
 843     }]
 844
 845     def _real_extract(self, url):
 846         lang, series_id = self._match_valid_url(url).group('lang', 'id')
 847         return self.url_result(f'https://www.crunchyroll.com/{lang}{series_id.lower()}',
 848                                CrunchyrollShowPlaylistIE.ie_key(), series_id)