_ Git - youtube-dl/blob - youtube_dl/extractor/crunchyroll.py

   1 # encoding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import json
   6 import base64
   7 import zlib
   8 import xml.etree.ElementTree
   9
  10 from hashlib import sha1
  11 from math import pow, sqrt, floor
  12 from .common import InfoExtractor
  13 from ..compat import (
  14     compat_urllib_parse,
  15     compat_urllib_parse_unquote,
  16     compat_urllib_request,
  17     compat_urlparse,
  18 )
  19 from ..utils import (
  20     ExtractorError,
  21     bytes_to_intlist,
  22     intlist_to_bytes,
  23     int_or_none,
  24     remove_end,
  25     unified_strdate,
  26     urlencode_postdata,
  27     xpath_text,
  28 )
  29 from ..aes import (
  30     aes_cbc_decrypt,
  31 )
  32
  33
  34 class CrunchyrollBaseIE(InfoExtractor):
  35     def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None):
  36         request = (url_or_request if isinstance(url_or_request, compat_urllib_request.Request)
  37                    else compat_urllib_request.Request(url_or_request))
  38         # Accept-Language must be set explicitly to accept any language to avoid issues
  39         # similar to https://github.com/rg3/youtube-dl/issues/6797.
  40         # Along with IP address Crunchyroll uses Accept-Language to guess whether georestriction
  41         # should be imposed or not (from what I can see it just takes the first language
  42         # ignoring the priority and requires it to correspond the IP). By the way this causes
  43         # Crunchyroll to not work in georestriction cases in some browsers that don't place
  44         # the locale lang first in header. However allowing any language seems to workaround the issue.
  45         request.add_header('Accept-Language', '*')
  46         return super(CrunchyrollBaseIE, self)._download_webpage(
  47             request, video_id, note, errnote, fatal, tries, timeout, encoding)
  48
  49
  50 class CrunchyrollIE(CrunchyrollBaseIE):
  51     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.(?:com|fr)/(?:media(?:-|/\?id=)|[^/]*/[^/?&]*?)(?P<video_id>[0-9]+))(?:[/?&]|$)'
  52     _NETRC_MACHINE = 'crunchyroll'
  53     _TESTS = [{
  54         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
  55         'info_dict': {
  56             'id': '645513',
  57             'ext': 'flv',
  58             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
  59             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
  60             'thumbnail': 'http://img1.ak.crunchyroll.com/i/spire1-tmb/20c6b5e10f1a47b10516877d3c039cae1380951166_full.jpg',
  61             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
  62             'upload_date': '20131013',
  63             'url': 're:(?!.*&amp)',
  64         },
  65         'params': {
  66             # rtmp
  67             'skip_download': True,
  68         },
  69     }, {
  70         'url': 'http://www.crunchyroll.com/media-589804/culture-japan-1',
  71         'info_dict': {
  72             'id': '589804',
  73             'ext': 'flv',
  74             'title': 'Culture Japan Episode 1 – Rebuilding Japan after the 3.11',
  75             'description': 'md5:fe2743efedb49d279552926d0bd0cd9e',
  76             'thumbnail': 're:^https?://.*\.jpg$',
  77             'uploader': 'Danny Choo Network',
  78             'upload_date': '20120213',
  79         },
  80         'params': {
  81             # rtmp
  82             'skip_download': True,
  83         },
  84
  85     }, {
  86         'url': 'http://www.crunchyroll.fr/girl-friend-beta/episode-11-goodbye-la-mode-661697',
  87         'only_matching': True,
  88     }]
  89
  90     _FORMAT_IDS = {
  91         '360': ('60', '106'),
  92         '480': ('61', '106'),
  93         '720': ('62', '106'),
  94         '1080': ('80', '108'),
  95     }
  96
  97     def _login(self):
  98         (username, password) = self._get_login_info()
  99         if username is None:
 100             return
 101         self.report_login()
 102         login_url = 'https://www.crunchyroll.com/?a=formhandler'
 103         data = urlencode_postdata({
 104             'formname': 'RpcApiUser_Login',
 105             'name': username,
 106             'password': password,
 107         })
 108         login_request = compat_urllib_request.Request(login_url, data)
 109         login_request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 110         self._download_webpage(login_request, None, False, 'Wrong login info')
 111
 112     def _real_initialize(self):
 113         self._login()
 114
 115     def _decrypt_subtitles(self, data, iv, id):
 116         data = bytes_to_intlist(base64.b64decode(data.encode('utf-8')))
 117         iv = bytes_to_intlist(base64.b64decode(iv.encode('utf-8')))
 118         id = int(id)
 119
 120         def obfuscate_key_aux(count, modulo, start):
 121             output = list(start)
 122             for _ in range(count):
 123                 output.append(output[-1] + output[-2])
 124             # cut off start values
 125             output = output[2:]
 126             output = list(map(lambda x: x % modulo + 33, output))
 127             return output
 128
 129         def obfuscate_key(key):
 130             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
 131             num2 = (num1 ^ key) << 5
 132             num3 = key ^ num1
 133             num4 = num3 ^ (num3 >> 3) ^ num2
 134             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
 135             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
 136             # Extend 160 Bit hash to 256 Bit
 137             return shaHash + [0] * 12
 138
 139         key = obfuscate_key(id)
 140
 141         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
 142         return zlib.decompress(decrypted_data)
 143
 144     def _convert_subtitles_to_srt(self, sub_root):
 145         output = ''
 146
 147         for i, event in enumerate(sub_root.findall('./events/event'), 1):
 148             start = event.attrib['start'].replace('.', ',')
 149             end = event.attrib['end'].replace('.', ',')
 150             text = event.attrib['text'].replace('\\N', '\n')
 151             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
 152         return output
 153
 154     def _convert_subtitles_to_ass(self, sub_root):
 155         output = ''
 156
 157         def ass_bool(strvalue):
 158             assvalue = '0'
 159             if strvalue == '1':
 160                 assvalue = '-1'
 161             return assvalue
 162
 163         output = '[Script Info]\n'
 164         output += 'Title: %s\n' % sub_root.attrib["title"]
 165         output += 'ScriptType: v4.00+\n'
 166         output += 'WrapStyle: %s\n' % sub_root.attrib["wrap_style"]
 167         output += 'PlayResX: %s\n' % sub_root.attrib["play_res_x"]
 168         output += 'PlayResY: %s\n' % sub_root.attrib["play_res_y"]
 169         output += """ScaledBorderAndShadow: yes
 170
 171 [V4+ Styles]
 172 Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
 173 """
 174         for style in sub_root.findall('./styles/style'):
 175             output += 'Style: ' + style.attrib["name"]
 176             output += ',' + style.attrib["font_name"]
 177             output += ',' + style.attrib["font_size"]
 178             output += ',' + style.attrib["primary_colour"]
 179             output += ',' + style.attrib["secondary_colour"]
 180             output += ',' + style.attrib["outline_colour"]
 181             output += ',' + style.attrib["back_colour"]
 182             output += ',' + ass_bool(style.attrib["bold"])
 183             output += ',' + ass_bool(style.attrib["italic"])
 184             output += ',' + ass_bool(style.attrib["underline"])
 185             output += ',' + ass_bool(style.attrib["strikeout"])
 186             output += ',' + style.attrib["scale_x"]
 187             output += ',' + style.attrib["scale_y"]
 188             output += ',' + style.attrib["spacing"]
 189             output += ',' + style.attrib["angle"]
 190             output += ',' + style.attrib["border_style"]
 191             output += ',' + style.attrib["outline"]
 192             output += ',' + style.attrib["shadow"]
 193             output += ',' + style.attrib["alignment"]
 194             output += ',' + style.attrib["margin_l"]
 195             output += ',' + style.attrib["margin_r"]
 196             output += ',' + style.attrib["margin_v"]
 197             output += ',' + style.attrib["encoding"]
 198             output += '\n'
 199
 200         output += """
 201 [Events]
 202 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
 203 """
 204         for event in sub_root.findall('./events/event'):
 205             output += 'Dialogue: 0'
 206             output += ',' + event.attrib["start"]
 207             output += ',' + event.attrib["end"]
 208             output += ',' + event.attrib["style"]
 209             output += ',' + event.attrib["name"]
 210             output += ',' + event.attrib["margin_l"]
 211             output += ',' + event.attrib["margin_r"]
 212             output += ',' + event.attrib["margin_v"]
 213             output += ',' + event.attrib["effect"]
 214             output += ',' + event.attrib["text"]
 215             output += '\n'
 216
 217         return output
 218
 219     def _extract_subtitles(self, subtitle):
 220         sub_root = xml.etree.ElementTree.fromstring(subtitle)
 221         return [{
 222             'ext': 'srt',
 223             'data': self._convert_subtitles_to_srt(sub_root),
 224         }, {
 225             'ext': 'ass',
 226             'data': self._convert_subtitles_to_ass(sub_root),
 227         }]
 228
 229     def _get_subtitles(self, video_id, webpage):
 230         subtitles = {}
 231         for sub_id, sub_name in re.findall(r'\?ssid=([0-9]+)" title="([^"]+)', webpage):
 232             sub_page = self._download_webpage(
 233                 'http://www.crunchyroll.com/xml/?req=RpcApiSubtitle_GetXml&subtitle_script_id=' + sub_id,
 234                 video_id, note='Downloading subtitles for ' + sub_name)
 235             id = self._search_regex(r'id=\'([0-9]+)', sub_page, 'subtitle_id', fatal=False)
 236             iv = self._search_regex(r'<iv>([^<]+)', sub_page, 'subtitle_iv', fatal=False)
 237             data = self._search_regex(r'<data>([^<]+)', sub_page, 'subtitle_data', fatal=False)
 238             if not id or not iv or not data:
 239                 continue
 240             subtitle = self._decrypt_subtitles(data, iv, id).decode('utf-8')
 241             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
 242             if not lang_code:
 243                 continue
 244             subtitles[lang_code] = self._extract_subtitles(subtitle)
 245         return subtitles
 246
 247     def _real_extract(self, url):
 248         mobj = re.match(self._VALID_URL, url)
 249         video_id = mobj.group('video_id')
 250
 251         if mobj.group('prefix') == 'm':
 252             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
 253             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
 254         else:
 255             webpage_url = 'http://www.' + mobj.group('url')
 256
 257         webpage = self._download_webpage(webpage_url, video_id, 'Downloading webpage')
 258         note_m = self._html_search_regex(
 259             r'<div class="showmedia-trailer-notice">(.+?)</div>',
 260             webpage, 'trailer-notice', default='')
 261         if note_m:
 262             raise ExtractorError(note_m)
 263
 264         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
 265         if mobj:
 266             msg = json.loads(mobj.group('msg'))
 267             if msg.get('type') == 'error':
 268                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
 269
 270         if 'To view this, please log in to verify you are 18 or older.' in webpage:
 271             self.raise_login_required()
 272
 273         video_title = self._html_search_regex(r'<h1[^>]*>(.+?)</h1>', webpage, 'video_title', flags=re.DOTALL)
 274         video_title = re.sub(r' {2,}', ' ', video_title)
 275         video_description = self._html_search_regex(r'"description":"([^"]+)', webpage, 'video_description', default='')
 276         if not video_description:
 277             video_description = None
 278         video_upload_date = self._html_search_regex(
 279             [r'<div>Availability for free users:(.+?)</div>', r'<div>[^<>]+<span>\s*(.+?\d{4})\s*</span></div>'],
 280             webpage, 'video_upload_date', fatal=False, flags=re.DOTALL)
 281         if video_upload_date:
 282             video_upload_date = unified_strdate(video_upload_date)
 283         video_uploader = self._html_search_regex(
 284             r'<a[^>]+href="/publisher/[^"]+"[^>]*>([^<]+)</a>', webpage,
 285             'video_uploader', fatal=False)
 286
 287         playerdata_url = compat_urllib_parse_unquote(self._html_search_regex(r'"config_url":"([^"]+)', webpage, 'playerdata_url'))
 288         playerdata_req = compat_urllib_request.Request(playerdata_url)
 289         playerdata_req.data = compat_urllib_parse.urlencode({'current_page': webpage_url})
 290         playerdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 291         playerdata = self._download_webpage(playerdata_req, video_id, note='Downloading media info')
 292
 293         stream_id = self._search_regex(r'<media_id>([^<]+)', playerdata, 'stream_id')
 294         video_thumbnail = self._search_regex(r'<episode_image_url>([^<]+)', playerdata, 'thumbnail', fatal=False)
 295
 296         formats = []
 297         for fmt in re.findall(r'showmedia\.([0-9]{3,4})p', webpage):
 298             stream_quality, stream_format = self._FORMAT_IDS[fmt]
 299             video_format = fmt + 'p'
 300             streamdata_req = compat_urllib_request.Request(
 301                 'http://www.crunchyroll.com/xml/?req=RpcApiVideoPlayer_GetStandardConfig&media_id=%s&video_format=%s&video_quality=%s'
 302                 % (stream_id, stream_format, stream_quality),
 303                 compat_urllib_parse.urlencode({'current_page': url}).encode('utf-8'))
 304             streamdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 305             streamdata = self._download_xml(
 306                 streamdata_req, video_id,
 307                 note='Downloading media info for %s' % video_format)
 308             stream_info = streamdata.find('./{default}preload/stream_info')
 309             video_url = stream_info.find('./host').text
 310             video_play_path = stream_info.find('./file').text
 311             metadata = stream_info.find('./metadata')
 312             format_info = {
 313                 'format': video_format,
 314                 'format_id': video_format,
 315                 'height': int_or_none(xpath_text(metadata, './height')),
 316                 'width': int_or_none(xpath_text(metadata, './width')),
 317             }
 318
 319             if '.fplive.net/' in video_url:
 320                 video_url = re.sub(r'^rtmpe?://', 'http://', video_url.strip())
 321                 parsed_video_url = compat_urlparse.urlparse(video_url)
 322                 direct_video_url = compat_urlparse.urlunparse(parsed_video_url._replace(
 323                     netloc='v.lvlt.crcdn.net',
 324                     path='%s/%s' % (remove_end(parsed_video_url.path, '/'), video_play_path.split(':')[-1])))
 325                 if self._is_valid_url(direct_video_url, video_id, video_format):
 326                     format_info.update({
 327                         'url': direct_video_url,
 328                     })
 329                     formats.append(format_info)
 330                     continue
 331
 332             format_info.update({
 333                 'url': video_url,
 334                 'play_path': video_play_path,
 335                 'ext': 'flv',
 336             })
 337             formats.append(format_info)
 338
 339         subtitles = self.extract_subtitles(video_id, webpage)
 340
 341         return {
 342             'id': video_id,
 343             'title': video_title,
 344             'description': video_description,
 345             'thumbnail': video_thumbnail,
 346             'uploader': video_uploader,
 347             'upload_date': video_upload_date,
 348             'subtitles': subtitles,
 349             'formats': formats,
 350         }
 351
 352
 353 class CrunchyrollShowPlaylistIE(CrunchyrollBaseIE):
 354     IE_NAME = "crunchyroll:playlist"
 355     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.com/(?!(?:news|anime-news|library|forum|launchcalendar|lineup|store|comics|freetrial|login))(?P<id>[\w\-]+))/?$'
 356
 357     _TESTS = [{
 358         'url': 'http://www.crunchyroll.com/a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 359         'info_dict': {
 360             'id': 'a-bridge-to-the-starry-skies-hoshizora-e-kakaru-hashi',
 361             'title': 'A Bridge to the Starry Skies - Hoshizora e Kakaru Hashi'
 362         },
 363         'playlist_count': 13,
 364     }]
 365
 366     def _real_extract(self, url):
 367         show_id = self._match_id(url)
 368
 369         webpage = self._download_webpage(url, show_id)
 370         title = self._html_search_regex(
 371             r'(?s)<h1[^>]*>\s*<span itemprop="name">(.*?)</span>',
 372             webpage, 'title')
 373         episode_paths = re.findall(
 374             r'(?s)<li id="showview_videos_media_[0-9]+"[^>]+>.*?<a href="([^"]+)"',
 375             webpage)
 376         entries = [
 377             self.url_result('http://www.crunchyroll.com' + ep, 'Crunchyroll')
 378             for ep in episode_paths
 379         ]
 380         entries.reverse()
 381
 382         return {
 383             '_type': 'playlist',
 384             'id': show_id,
 385             'title': title,
 386             'entries': entries,
 387         }