_ 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
   9 from hashlib import sha1
  10 from math import pow, sqrt, floor
  11 from .common import InfoExtractor
  12 from ..utils import (
  13     ExtractorError,
  14     compat_urllib_parse,
  15     compat_urllib_request,
  16     bytes_to_intlist,
  17     intlist_to_bytes,
  18     unified_strdate,
  19     clean_html,
  20     urlencode_postdata,
  21 )
  22 from ..aes import (
  23     aes_cbc_decrypt,
  24     inc,
  25 )
  26
  27
  28 class CrunchyrollIE(InfoExtractor):
  29     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.com/(?:[^/]*/[^/?&]*?|media/\?id=)(?P<video_id>[0-9]+))(?:[/?&]|$)'
  30     _TEST = {
  31         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
  32         #'md5': 'b1639fd6ddfaa43788c85f6d1dddd412',
  33         'info_dict': {
  34             'id': '645513',
  35             'ext': 'flv',
  36             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
  37             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
  38             'thumbnail': 'http://img1.ak.crunchyroll.com/i/spire1-tmb/20c6b5e10f1a47b10516877d3c039cae1380951166_full.jpg',
  39             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
  40             'upload_date': '20131013',
  41         },
  42         'params': {
  43             # rtmp
  44             'skip_download': True,
  45         },
  46     }
  47
  48     _FORMAT_IDS = {
  49         '360': ('60', '106'),
  50         '480': ('61', '106'),
  51         '720': ('62', '106'),
  52         '1080': ('80', '108'),
  53     }
  54
  55     def _login(self):
  56         (username, password) = self._get_login_info()
  57         if username is None:
  58             return
  59         self.report_login()
  60         login_url = 'https://www.crunchyroll.com/?a=formhandler'
  61         data = urlencode_postdata({
  62             'formname': 'RpcApiUser_Login',
  63             'name': username,
  64             'password': password,
  65         })
  66         login_request = compat_urllib_request.Request(login_url, data)
  67         login_request.add_header('Content-Type', 'application/x-www-form-urlencoded')
  68         self._download_webpage(login_request, None, False, 'Wrong login info')
  69
  70     def _real_initialize(self):
  71         self._login()
  72
  73     def _decrypt_subtitles(self, data, iv, id):
  74         data = bytes_to_intlist(data)
  75         iv = bytes_to_intlist(iv)
  76         id = int(id)
  77
  78         def obfuscate_key_aux(count, modulo, start):
  79             output = list(start)
  80             for _ in range(count):
  81                 output.append(output[-1] + output[-2])
  82             # cut off start values
  83             output = output[2:]
  84             output = list(map(lambda x: x % modulo + 33, output))
  85             return output
  86
  87         def obfuscate_key(key):
  88             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
  89             num2 = (num1 ^ key) << 5
  90             num3 = key ^ num1
  91             num4 = num3 ^ (num3 >> 3) ^ num2
  92             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
  93             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
  94             # Extend 160 Bit hash to 256 Bit
  95             return shaHash + [0] * 12
  96
  97         key = obfuscate_key(id)
  98         class Counter:
  99             __value = iv
 100             def next_value(self):
 101                 temp = self.__value
 102                 self.__value = inc(self.__value)
 103                 return temp
 104         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
 105         return zlib.decompress(decrypted_data)
 106
 107     def _convert_subtitles_to_srt(self, subtitles):
 108         output = ''
 109         for i, (start, end, text) in enumerate(re.findall(r'<event [^>]*?start="([^"]+)" [^>]*?end="([^"]+)" [^>]*?text="([^"]+)"[^>]*?>', subtitles), 1):
 110             start = start.replace('.', ',')
 111             end = end.replace('.', ',')
 112             text = clean_html(text)
 113             text = text.replace('\\N', '\n')
 114             if not text:
 115                 continue
 116             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
 117         return output
 118
 119     def _real_extract(self,url):
 120         mobj = re.match(self._VALID_URL, url)
 121         video_id = mobj.group('video_id')
 122
 123         if mobj.group('prefix') == 'm':
 124             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
 125             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
 126         else:
 127             webpage_url = 'http://www.' + mobj.group('url')
 128
 129         webpage = self._download_webpage(webpage_url, video_id, 'Downloading webpage')
 130         note_m = self._html_search_regex(r'<div class="showmedia-trailer-notice">(.+?)</div>', webpage, 'trailer-notice', default='')
 131         if note_m:
 132             raise ExtractorError(note_m)
 133
 134         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
 135         if mobj:
 136             msg = json.loads(mobj.group('msg'))
 137             if msg.get('type') == 'error':
 138                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
 139
 140         video_title = self._html_search_regex(r'<h1[^>]*>(.+?)</h1>', webpage, 'video_title', flags=re.DOTALL)
 141         video_title = re.sub(r' {2,}', ' ', video_title)
 142         video_description = self._html_search_regex(r'"description":"([^"]+)', webpage, 'video_description', default='')
 143         if not video_description:
 144             video_description = None
 145         video_upload_date = self._html_search_regex(r'<div>Availability for free users:(.+?)</div>', webpage, 'video_upload_date', fatal=False, flags=re.DOTALL)
 146         if video_upload_date:
 147             video_upload_date = unified_strdate(video_upload_date)
 148         video_uploader = self._html_search_regex(r'<div>\s*Publisher:(.+?)</div>', webpage, 'video_uploader', fatal=False, flags=re.DOTALL)
 149
 150         playerdata_url = compat_urllib_parse.unquote(self._html_search_regex(r'"config_url":"([^"]+)', webpage, 'playerdata_url'))
 151         playerdata_req = compat_urllib_request.Request(playerdata_url)
 152         playerdata_req.data = compat_urllib_parse.urlencode({'current_page': webpage_url})
 153         playerdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 154         playerdata = self._download_webpage(playerdata_req, video_id, note='Downloading media info')
 155
 156         stream_id = self._search_regex(r'<media_id>([^<]+)', playerdata, 'stream_id')
 157         video_thumbnail = self._search_regex(r'<episode_image_url>([^<]+)', playerdata, 'thumbnail', fatal=False)
 158
 159         formats = []
 160         for fmt in re.findall(r'\?p([0-9]{3,4})=1', webpage):
 161             stream_quality, stream_format = self._FORMAT_IDS[fmt]
 162             video_format = fmt+'p'
 163             streamdata_req = compat_urllib_request.Request('http://www.crunchyroll.com/xml/')
 164             # urlencode doesn't work!
 165             streamdata_req.data = 'req=RpcApiVideoEncode%5FGetStreamInfo&video%5Fencode%5Fquality='+stream_quality+'&media%5Fid='+stream_id+'&video%5Fformat='+stream_format
 166             streamdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 167             streamdata_req.add_header('Content-Length', str(len(streamdata_req.data)))
 168             streamdata = self._download_webpage(streamdata_req, video_id, note='Downloading media info for '+video_format)
 169             video_url = self._search_regex(r'<host>([^<]+)', streamdata, 'video_url')
 170             video_play_path = self._search_regex(r'<file>([^<]+)', streamdata, 'video_play_path')
 171             formats.append({
 172                 'url': video_url,
 173                 'play_path':   video_play_path,
 174                 'ext': 'flv',
 175                 'format': video_format,
 176                 'format_id': video_format,
 177             })
 178
 179         subtitles = {}
 180         for sub_id, sub_name in re.findall(r'\?ssid=([0-9]+)" title="([^"]+)', webpage):
 181             sub_page = self._download_webpage('http://www.crunchyroll.com/xml/?req=RpcApiSubtitle_GetXml&subtitle_script_id='+sub_id,\
 182                                               video_id, note='Downloading subtitles for '+sub_name)
 183             id = self._search_regex(r'id=\'([0-9]+)', sub_page, 'subtitle_id', fatal=False)
 184             iv = self._search_regex(r'<iv>([^<]+)', sub_page, 'subtitle_iv', fatal=False)
 185             data = self._search_regex(r'<data>([^<]+)', sub_page, 'subtitle_data', fatal=False)
 186             if not id or not iv or not data:
 187                 continue
 188             id = int(id)
 189             iv = base64.b64decode(iv)
 190             data = base64.b64decode(data)
 191
 192             subtitle = self._decrypt_subtitles(data, iv, id).decode('utf-8')
 193             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
 194             if not lang_code:
 195                 continue
 196             subtitles[lang_code] = self._convert_subtitles_to_srt(subtitle)
 197
 198         return {
 199             'id':          video_id,
 200             'title':       video_title,
 201             'description': video_description,
 202             'thumbnail':   video_thumbnail,
 203             'uploader':    video_uploader,
 204             'upload_date': video_upload_date,
 205             'subtitles':   subtitles,
 206             'formats':     formats,
 207         }