Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/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_urllib_parse_urlencode,
  17     compat_urllib_request,
  18     compat_urlparse,
  19 )
  20 from ..utils import (
  21     ExtractorError,
  22     bytes_to_intlist,
  23     extract_attributes,
  24     float_or_none,
  25     intlist_to_bytes,
  26     int_or_none,
  27     lowercase_escape,
  28     remove_end,
  29     sanitized_Request,
  30     unified_strdate,
  31     urlencode_postdata,
  32     xpath_text,
  33 )
  34 from ..aes import (
  35     aes_cbc_decrypt,
  36 )
  37
  38
  39 class CrunchyrollBaseIE(InfoExtractor):
  40     _LOGIN_URL = 'https://www.crunchyroll.com/login'
  41     _LOGIN_FORM = 'login_form'
  42     _NETRC_MACHINE = 'crunchyroll'
  43
  44     def _call_rpc_api(self, method, video_id, note=None, data=None):
  45         data = data or {}
  46         data['req'] = 'RpcApi' + method
  47         data = compat_urllib_parse_urlencode(data).encode('utf-8')
  48         return self._download_xml(
  49             'https://www.crunchyroll.com/xml/',
  50             video_id, note, fatal=False, data=data, headers={
  51                 'Content-Type': 'application/x-www-form-urlencoded',
  52             })
  53
  54     def _login(self):
  55         username, password = self._get_login_info()
  56         if username is None:
  57             return
  58
  59         login_page = self._download_webpage(
  60             self._LOGIN_URL, None, 'Downloading login page')
  61
  62         def is_logged(webpage):
  63             return 'href="/logout"' in webpage
  64
  65         # Already logged in
  66         if is_logged(login_page):
  67             return
  68
  69         login_form_str = self._search_regex(
  70             r'(?P<form><form[^>]+?id=(["\'])%s\2[^>]*>)' % self._LOGIN_FORM,
  71             login_page, 'login form', group='form')
  72
  73         post_url = extract_attributes(login_form_str).get('action')
  74         if not post_url:
  75             post_url = self._LOGIN_URL
  76         elif not post_url.startswith('http'):
  77             post_url = compat_urlparse.urljoin(self._LOGIN_URL, post_url)
  78
  79         login_form = self._form_hidden_inputs(self._LOGIN_FORM, login_page)
  80
  81         login_form.update({
  82             'login_form[name]': username,
  83             'login_form[password]': password,
  84         })
  85
  86         response = self._download_webpage(
  87             post_url, None, 'Logging in', 'Wrong login info',
  88             data=urlencode_postdata(login_form),
  89             headers={'Content-Type': 'application/x-www-form-urlencoded'})
  90
  91         # Successful login
  92         if is_logged(response):
  93             return
  94
  95         error = self._html_search_regex(
  96             '(?s)<ul[^>]+class=["\']messages["\'][^>]*>(.+?)</ul>',
  97             response, 'error message', default=None)
  98         if error:
  99             raise ExtractorError('Unable to login: %s' % error, expected=True)
 100
 101         raise ExtractorError('Unable to log in')
 102
 103     def _real_initialize(self):
 104         self._login()
 105
 106     @staticmethod
 107     def _add_skip_wall(url):
 108         parsed_url = compat_urlparse.urlparse(url)
 109         qs = compat_urlparse.parse_qs(parsed_url.query)
 110         # Always force skip_wall to bypass maturity wall, namely 18+ confirmation message:
 111         # > This content may be inappropriate for some people.
 112         # > Are you sure you want to continue?
 113         # since it's not disabled by default in crunchyroll account's settings.
 114         # See https://github.com/ytdl-org/youtube-dl/issues/7202.
 115         qs['skip_wall'] = ['1']
 116         return compat_urlparse.urlunparse(
 117             parsed_url._replace(query=compat_urllib_parse_urlencode(qs, True)))
 118
 119
 120 class CrunchyrollIE(CrunchyrollBaseIE, VRVIE):
 121     IE_NAME = 'crunchyroll'
 122     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.(?:com|fr)/(?:media(?:-|/\?id=)|(?:[^/]*/){1,2}[^/?&]*?)(?P<video_id>[0-9]+))(?:[/?&]|$)'
 123     _TESTS = [{
 124         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
 125         'info_dict': {
 126             'id': '645513',
 127             'ext': 'mp4',
 128             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
 129             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
 130             'thumbnail': r're:^https?://.*\.jpg$',
 131             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
 132             'upload_date': '20131013',
 133             'url': 're:(?!.*&amp)',
 134         },
 135         'params': {
 136             # rtmp
 137             'skip_download': True,
 138         },
 139     }, {
 140         'url': 'http://www.crunchyroll.com/media-589804/culture-japan-1',
 141         'info_dict': {
 142             'id': '589804',
 143             'ext': 'flv',
 144             'title': 'Culture Japan Episode 1 – Rebuilding Japan after the 3.11',
 145             'description': 'md5:2fbc01f90b87e8e9137296f37b461c12',
 146             'thumbnail': r're:^https?://.*\.jpg$',
 147             'uploader': 'Danny Choo Network',
 148             'upload_date': '20120213',
 149         },
 150         'params': {
 151             # rtmp
 152             'skip_download': True,
 153         },
 154         'skip': 'Video gone',
 155     }, {
 156         'url': 'http://www.crunchyroll.com/rezero-starting-life-in-another-world-/episode-5-the-morning-of-our-promise-is-still-distant-702409',
 157         'info_dict': {
 158             'id': '702409',
 159             'ext': 'mp4',
 160             'title': 'Re:ZERO -Starting Life in Another World- Episode 5 – The Morning of Our Promise Is Still Distant',
 161             'description': 'md5:97664de1ab24bbf77a9c01918cb7dca9',
 162             'thumbnail': r're:^https?://.*\.jpg$',
 163             'uploader': 'TV TOKYO',
 164             'upload_date': '20160508',
 165         },
 166         'params': {
 167             # m3u8 download
 168             'skip_download': True,
 169         },
 170     }, {
 171         'url': 'http://www.crunchyroll.com/konosuba-gods-blessing-on-this-wonderful-world/episode-1-give-me-deliverance-from-this-judicial-injustice-727589',
 172         'info_dict': {
 173             'id': '727589',
 174             'ext': 'mp4',
 175             'title': "KONOSUBA -God's blessing on this wonderful world! 2 Episode 1 – Give Me Deliverance From This Judicial Injustice!",
 176             'description': 'md5:cbcf05e528124b0f3a0a419fc805ea7d',
 177             'thumbnail': r're:^https?://.*\.jpg$',
 178             'uploader': 'Kadokawa Pictures Inc.',
 179             'upload_date': '20170118',
 180             'series': "KONOSUBA -God's blessing on this wonderful world!",
 181             'season': "KONOSUBA -God's blessing on this wonderful world! 2",
 182             'season_number': 2,
 183             'episode': 'Give Me Deliverance From This Judicial Injustice!',
 184             'episode_number': 1,
 185         },
 186         'params': {
 187             # m3u8 download
 188             'skip_download': True,
 189         },
 190     }, {
 191         'url': 'http://www.crunchyroll.fr/girl-friend-beta/episode-11-goodbye-la-mode-661697',
 192         'only_matching': True,
 193     }, {
 194         # geo-restricted (US), 18+ maturity wall, non-premium available
 195         'url': 'http://www.crunchyroll.com/cosplay-complex-ova/episode-1-the-birth-of-the-cosplay-club-565617',
 196         'only_matching': True,
 197     }, {
 198         # A description with double quotes
 199         'url': 'http://www.crunchyroll.com/11eyes/episode-1-piros-jszaka-red-night-535080',
 200         'info_dict': {
 201             'id': '535080',
 202             'ext': 'mp4',
 203             'title': '11eyes Episode 1 – Red Night ~ Piros éjszaka',
 204             'description': 'Kakeru and Yuka are thrown into an alternate nightmarish world they call "Red Night".',
 205             'uploader': 'Marvelous AQL Inc.',
 206             'upload_date': '20091021',
 207         },
 208         'params': {
 209             # Just test metadata extraction
 210             'skip_download': True,
 211         },
 212     }, {
 213         # make sure we can extract an uploader name that's not a link
 214         'url': 'http://www.crunchyroll.com/hakuoki-reimeiroku/episode-1-dawn-of-the-divine-warriors-606899',
 215         'info_dict': {
 216             'id': '606899',
 217             'ext': 'mp4',
 218             'title': 'Hakuoki Reimeiroku Episode 1 – Dawn of the Divine Warriors',
 219             'description': 'Ryunosuke was left to die, but Serizawa-san asked him a simple question "Do you want to live?"',
 220             'uploader': 'Geneon Entertainment',
 221             'upload_date': '20120717',
 222         },
 223         'params': {
 224             # just test metadata extraction
 225             'skip_download': True,
 226         },
 227     }, {
 228         # A video with a vastly different season name compared to the series name
 229         'url': 'http://www.crunchyroll.com/nyarko-san-another-crawling-chaos/episode-1-test-590532',
 230         'info_dict': {
 231             'id': '590532',
 232             'ext': 'mp4',
 233             'title': 'Haiyoru! Nyaruani (ONA) Episode 1 – Test',
 234             'description': 'Mahiro and Nyaruko talk about official certification.',
 235             'uploader': 'TV TOKYO',
 236             'upload_date': '20120305',
 237             'series': 'Nyarko-san: Another Crawling Chaos',
 238             'season': 'Haiyoru! Nyaruani (ONA)',
 239         },
 240         'params': {
 241             # Just test metadata extraction
 242             'skip_download': True,
 243         },
 244     }, {
 245         'url': 'http://www.crunchyroll.com/media-723735',
 246         'only_matching': True,
 247     }, {
 248         'url': 'https://www.crunchyroll.com/en-gb/mob-psycho-100/episode-2-urban-legends-encountering-rumors-780921',
 249         'only_matching': True,
 250     }]
 251
 252     _FORMAT_IDS = {
 253         '360': ('60', '106'),
 254         '480': ('61', '106'),
 255         '720': ('62', '106'),
 256         '1080': ('80', '108'),
 257     }
 258
 259     def _download_webpage(self, url_or_request, *args, **kwargs):
 260         request = (url_or_request if isinstance(url_or_request, compat_urllib_request.Request)
 261                    else sanitized_Request(url_or_request))
 262         # Accept-Language must be set explicitly to accept any language to avoid issues
 263         # similar to https://github.com/ytdl-org/youtube-dl/issues/6797.
 264         # Along with IP address Crunchyroll uses Accept-Language to guess whether georestriction
 265         # should be imposed or not (from what I can see it just takes the first language
 266         # ignoring the priority and requires it to correspond the IP). By the way this causes
 267         # Crunchyroll to not work in georestriction cases in some browsers that don't place
 268         # the locale lang first in header. However allowing any language seems to workaround the issue.
 269         request.add_header('Accept-Language', '*')
 270         return super(CrunchyrollBaseIE, self)._download_webpage(request, *args, **kwargs)
 271
 272     def _decrypt_subtitles(self, data, iv, id):
 273         data = bytes_to_intlist(compat_b64decode(data))
 274         iv = bytes_to_intlist(compat_b64decode(iv))
 275         id = int(id)
 276
 277         def obfuscate_key_aux(count, modulo, start):
 278             output = list(start)
 279             for _ in range(count):
 280                 output.append(output[-1] + output[-2])
 281             # cut off start values
 282             output = output[2:]
 283             output = list(map(lambda x: x % modulo + 33, output))
 284             return output
 285
 286         def obfuscate_key(key):
 287             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
 288             num2 = (num1 ^ key) << 5
 289             num3 = key ^ num1
 290             num4 = num3 ^ (num3 >> 3) ^ num2
 291             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
 292             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
 293             # Extend 160 Bit hash to 256 Bit
 294             return shaHash + [0] * 12
 295
 296         key = obfuscate_key(id)
 297
 298         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
 299         return zlib.decompress(decrypted_data)
 300
 301     def _convert_subtitles_to_srt(self, sub_root):
 302         output = ''
 303
 304         for i, event in enumerate(sub_root.findall('./events/event'), 1):
 305             start = event.attrib['start'].replace('.', ',')
 306             end = event.attrib['end'].replace('.', ',')
 307             text = event.attrib['text'].replace('\\N', '\n')
 308             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
 309         return output
 310
 311     def _convert_subtitles_to_ass(self, sub_root):
 312         output = ''
 313
 314         def ass_bool(strvalue):
 315             assvalue = '0'
 316             if strvalue == '1':
 317                 assvalue = '-1'
 318             return assvalue
 319
 320         output = '[Script Info]\n'
 321         output += 'Title: %s\n' % sub_root.attrib['title']
 322         output += 'ScriptType: v4.00+\n'
 323         output += 'WrapStyle: %s\n' % sub_root.attrib['wrap_style']
 324         output += 'PlayResX: %s\n' % sub_root.attrib['play_res_x']
 325         output += 'PlayResY: %s\n' % sub_root.attrib['play_res_y']
 326         output += """
 327 [V4+ Styles]
 328 Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
 329 """
 330         for style in sub_root.findall('./styles/style'):
 331             output += 'Style: ' + style.attrib['name']
 332             output += ',' + style.attrib['font_name']
 333             output += ',' + style.attrib['font_size']
 334             output += ',' + style.attrib['primary_colour']
 335             output += ',' + style.attrib['secondary_colour']
 336             output += ',' + style.attrib['outline_colour']
 337             output += ',' + style.attrib['back_colour']
 338             output += ',' + ass_bool(style.attrib['bold'])
 339             output += ',' + ass_bool(style.attrib['italic'])
 340             output += ',' + ass_bool(style.attrib['underline'])
 341             output += ',' + ass_bool(style.attrib['strikeout'])
 342             output += ',' + style.attrib['scale_x']
 343             output += ',' + style.attrib['scale_y']
 344             output += ',' + style.attrib['spacing']
 345             output += ',' + style.attrib['angle']
 346             output += ',' + style.attrib['border_style']
 347             output += ',' + style.attrib['outline']
 348             output += ',' + style.attrib['shadow']
 349             output += ',' + style.attrib['alignment']
 350             output += ',' + style.attrib['margin_l']
 351             output += ',' + style.attrib['margin_r']
 352             output += ',' + style.attrib['margin_v']
 353             output += ',' + style.attrib['encoding']
 354             output += '\n'
 355
 356         output += """
 357 [Events]
 358 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
 359 """
 360         for event in sub_root.findall('./events/event'):
 361             output += 'Dialogue: 0'
 362             output += ',' + event.attrib['start']
 363             output += ',' + event.attrib['end']
 364             output += ',' + event.attrib['style']
 365             output += ',' + event.attrib['name']
 366             output += ',' + event.attrib['margin_l']
 367             output += ',' + event.attrib['margin_r']
 368             output += ',' + event.attrib['margin_v']
 369             output += ',' + event.attrib['effect']
 370             output += ',' + event.attrib['text']
 371             output += '\n'
 372
 373         return output
 374
 375     def _extract_subtitles(self, subtitle):
 376         sub_root = compat_etree_fromstring(subtitle)
 377         return [{
 378             'ext': 'srt',
 379             'data': self._convert_subtitles_to_srt(sub_root),
 380         }, {
 381             'ext': 'ass',
 382             'data': self._convert_subtitles_to_ass(sub_root),
 383         }]
 384
 385     def _get_subtitles(self, video_id, webpage):
 386         subtitles = {}
 387         for sub_id, sub_name in re.findall(r'\bssid=([0-9]+)"[^>]+?\btitle="([^"]+)', webpage):
 388             sub_doc = self._call_rpc_api(
 389                 'Subtitle_GetXml', video_id,
 390                 'Downloading subtitles for ' + sub_name, data={
 391                     'subtitle_script_id': sub_id,
 392                 })
 393             if not isinstance(sub_doc, compat_etree_Element):
 394                 continue
 395             sid = sub_doc.get('id')
 396             iv = xpath_text(sub_doc, 'iv', 'subtitle iv')
 397             data = xpath_text(sub_doc, 'data', 'subtitle data')
 398             if not sid or not iv or not data:
 399                 continue
 400             subtitle = self._decrypt_subtitles(data, iv, sid).decode('utf-8')
 401             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
 402             if not lang_code:
 403                 continue
 404             subtitles[lang_code] = self._extract_subtitles(subtitle)
 405         return subtitles
 406
 407     def _real_extract(self, url):
 408         mobj = re.match(self._VALID_URL, url)
 409         video_id = mobj.group('video_id')
 410
 411         if mobj.group('prefix') == 'm':
 412             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
 413             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
 414         else:
 415             webpage_url = 'http://www.' + mobj.group('url')
 416
 417         webpage = self._download_webpage(
 418             self._add_skip_wall(webpage_url), video_id,
 419             headers=self.geo_verification_headers())
 420         note_m = self._html_search_regex(
 421             r'<div class="showmedia-trailer-notice">(.+?)</div>',
 422             webpage, 'trailer-notice', default='')
 423         if note_m:
 424             raise ExtractorError(note_m)
 425
 426         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
 427         if mobj:
 428             msg = json.loads(mobj.group('msg'))
 429             if msg.get('type') == 'error':
 430                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
 431
 432         if 'To view this, please log in to verify you are 18 or older.' in webpage:
 433             self.raise_login_required()
 434
 435         media = self._parse_json(self._search_regex(
 436             r'vilos\.config\.media\s*=\s*({.+?});',
 437             webpage, 'vilos media', default='{}'), video_id)
 438         media_metadata = media.get('metadata') or {}
 439
 440         language = self._search_regex(
 441             r'(?:vilos\.config\.player\.language|LOCALE)\s*=\s*(["\'])(?P<lang>(?:(?!\1).)+)\1',
 442             webpage, 'language', default=None, group='lang')
 443
 444         video_title = self._html_search_regex(
 445             r'(?s)<h1[^>]*>((?:(?!<h1).)*?<span[^>]+itemprop=["\']title["\'][^>]*>(?:(?!<h1).)+?)</h1>',
 446             webpage, 'video_title')
 447         video_title = re.sub(r' {2,}', ' ', video_title)
 448         video_description = (self._parse_json(self._html_search_regex(
 449             r'<script[^>]*>\s*.+?\[media_id=%s\].+?({.+?"description"\s*:.+?})\);' % video_id,
 450             webpage, 'description', default='{}'), video_id) or media_metadata).get('description')
 451         if video_description:
 452             video_description = lowercase_escape(video_description.replace(r'\r\n', '\n'))
 453         video_upload_date = self._html_search_regex(
 454             [r'<div>Availability for free users:(.+?)</div>', r'<div>[^<>]+<span>\s*(.+?\d{4})\s*</span></div>'],
 455             webpage, 'video_upload_date', fatal=False, flags=re.DOTALL)
 456         if video_upload_date:
 457             video_upload_date = unified_strdate(video_upload_date)
 458         video_uploader = self._html_search_regex(
 459             # try looking for both an uploader that's a link and one that's not
 460             [r'<a[^>]+href="/publisher/[^"]+"[^>]*>([^<]+)</a>', r'<div>\s*Publisher:\s*<span>\s*(.+?)\s*</span>\s*</div>'],
 461             webpage, 'video_uploader', fatal=False)
 462
 463         formats = []
 464         for stream in media.get('streams', []):
 465             audio_lang = stream.get('audio_lang')
 466             hardsub_lang = stream.get('hardsub_lang')
 467             vrv_formats = self._extract_vrv_formats(
 468                 stream.get('url'), video_id, stream.get('format'),
 469                 audio_lang, hardsub_lang)
 470             for f in vrv_formats:
 471                 if not hardsub_lang:
 472                     f['preference'] = 1
 473                 language_preference = 0
 474                 if audio_lang == language:
 475                     language_preference += 1
 476                 if hardsub_lang == language:
 477                     language_preference += 1
 478                 if language_preference:
 479                     f['language_preference'] = language_preference
 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, ('preference', 'language_preference', 'height', 'width', 'tbr', 'fps'))
 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 = thumbnail = 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             thumbnail = xpath_text(metadata, 'episode_image_url')
 602
 603         if not episode:
 604             episode = media_metadata.get('title')
 605         if not episode_number:
 606             episode_number = int_or_none(media_metadata.get('episode_number'))
 607         if not thumbnail:
 608             thumbnail = media_metadata.get('thumbnail', {}).get('url')
 609
 610         season_number = int_or_none(self._search_regex(
 611             r'(?s)<h\d[^>]+id=["\']showmedia_about_episode_num[^>]+>.+?</h\d>\s*<h4>\s*Season (\d+)',
 612             webpage, 'season number', default=None))
 613
 614         return {
 615             'id': video_id,
 616             'title': video_title,
 617             'description': video_description,
 618             'duration': duration,
 619             'thumbnail': thumbnail,
 620             'uploader': video_uploader,
 621             'upload_date': video_upload_date,
 622             'series': series,
 623             'season': season,
 624             'season_number': season_number,
 625             'episode': episode,
 626             'episode_number': episode_number,
 627             'subtitles': subtitles,
 628             'formats': formats,
 629         }
 630
 631
 632 class CrunchyrollShowPlaylistIE(CrunchyrollBaseIE):
 633     IE_NAME = 'crunchyroll:playlist'
 634     _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\-]+))/?(?:\?|$)'
 635
 636     _TESTS = [{
 637         'url': 'http://www.crunchyroll.com/a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 638         'info_dict': {
 639             'id': 'a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 640             'title': 'A Bridge to the Starry Skies - Hoshizora e Kakaru Hashi'
 641         },
 642         'playlist_count': 13,
 643     }, {
 644         # geo-restricted (US), 18+ maturity wall, non-premium available
 645         'url': 'http://www.crunchyroll.com/cosplay-complex-ova',
 646         'info_dict': {
 647             'id': 'cosplay-complex-ova',
 648             'title': 'Cosplay Complex OVA'
 649         },
 650         'playlist_count': 3,
 651         'skip': 'Georestricted',
 652     }, {
 653         # geo-restricted (US), 18+ maturity wall, non-premium will be available since 2015.11.14
 654         'url': 'http://www.crunchyroll.com/ladies-versus-butlers?skip_wall=1',
 655         'only_matching': True,
 656     }]
 657
 658     def _real_extract(self, url):
 659         show_id = self._match_id(url)
 660
 661         webpage = self._download_webpage(
 662             self._add_skip_wall(url), show_id,
 663             headers=self.geo_verification_headers())
 664         title = self._html_search_meta('name', webpage, default=None)
 665
 666         episode_paths = re.findall(
 667             r'(?s)<li id="showview_videos_media_(\d+)"[^>]+>.*?<a href="([^"]+)"',
 668             webpage)
 669         entries = [
 670             self.url_result('http://www.crunchyroll.com' + ep, 'Crunchyroll', ep_id)
 671             for ep_id, ep in episode_paths
 672         ]
 673         entries.reverse()
 674
 675         return {
 676             '_type': 'playlist',
 677             'id': show_id,
 678             'title': title,
 679             'entries': entries,
 680         }