yt_dlp/extractor/crunchyroll.py

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