_ 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 ..utils import (
  14     ExtractorError,
  15     compat_urllib_parse,
  16     compat_urllib_request,
  17     bytes_to_intlist,
  18     intlist_to_bytes,
  19     unified_strdate,
  20     clean_html,
  21     urlencode_postdata,
  22 )
  23 from ..aes import (
  24     aes_cbc_decrypt,
  25     inc,
  26 )
  27
  28
  29 class CrunchyrollIE(InfoExtractor):
  30     _VALID_URL = r'https?://(?:(?P<prefix>www|m)\.)?(?P<url>crunchyroll\.com/(?:[^/]*/[^/?&]*?|media/\?id=)(?P<video_id>[0-9]+))(?:[/?&]|$)'
  31     _TEST = {
  32         'url': 'http://www.crunchyroll.com/wanna-be-the-strongest-in-the-world/episode-1-an-idol-wrestler-is-born-645513',
  33         #'md5': 'b1639fd6ddfaa43788c85f6d1dddd412',
  34         'info_dict': {
  35             'id': '645513',
  36             'ext': 'flv',
  37             'title': 'Wanna be the Strongest in the World Episode 1 – An Idol-Wrestler is Born!',
  38             'description': 'md5:2d17137920c64f2f49981a7797d275ef',
  39             'thumbnail': 'http://img1.ak.crunchyroll.com/i/spire1-tmb/20c6b5e10f1a47b10516877d3c039cae1380951166_full.jpg',
  40             'uploader': 'Yomiuri Telecasting Corporation (YTV)',
  41             'upload_date': '20131013',
  42         },
  43         'params': {
  44             # rtmp
  45             'skip_download': True,
  46         },
  47     }
  48
  49     _FORMAT_IDS = {
  50         '360': ('60', '106'),
  51         '480': ('61', '106'),
  52         '720': ('62', '106'),
  53         '1080': ('80', '108'),
  54     }
  55
  56     def _login(self):
  57         (username, password) = self._get_login_info()
  58         if username is None:
  59             return
  60         self.report_login()
  61         login_url = 'https://www.crunchyroll.com/?a=formhandler'
  62         data = urlencode_postdata({
  63             'formname': 'RpcApiUser_Login',
  64             'name': username,
  65             'password': password,
  66         })
  67         login_request = compat_urllib_request.Request(login_url, data)
  68         login_request.add_header('Content-Type', 'application/x-www-form-urlencoded')
  69         self._download_webpage(login_request, None, False, 'Wrong login info')
  70
  71
  72     def _real_initialize(self):
  73         self._login()
  74
  75
  76     def _decrypt_subtitles(self, data, iv, id):
  77         data = bytes_to_intlist(data)
  78         iv = bytes_to_intlist(iv)
  79         id = int(id)
  80
  81         def obfuscate_key_aux(count, modulo, start):
  82             output = list(start)
  83             for _ in range(count):
  84                 output.append(output[-1] + output[-2])
  85             # cut off start values
  86             output = output[2:]
  87             output = list(map(lambda x: x % modulo + 33, output))
  88             return output
  89
  90         def obfuscate_key(key):
  91             num1 = int(floor(pow(2, 25) * sqrt(6.9)))
  92             num2 = (num1 ^ key) << 5
  93             num3 = key ^ num1
  94             num4 = num3 ^ (num3 >> 3) ^ num2
  95             prefix = intlist_to_bytes(obfuscate_key_aux(20, 97, (1, 2)))
  96             shaHash = bytes_to_intlist(sha1(prefix + str(num4).encode('ascii')).digest())
  97             # Extend 160 Bit hash to 256 Bit
  98             return shaHash + [0] * 12
  99
 100         key = obfuscate_key(id)
 101         class Counter:
 102             __value = iv
 103             def next_value(self):
 104                 temp = self.__value
 105                 self.__value = inc(self.__value)
 106                 return temp
 107         decrypted_data = intlist_to_bytes(aes_cbc_decrypt(data, key, iv))
 108         return zlib.decompress(decrypted_data)
 109
 110     def _convert_subtitles_to_srt(self, subtitles):
 111         output = ''
 112         for i, (start, end, text) in enumerate(re.findall(r'<event [^>]*?start="([^"]+)" [^>]*?end="([^"]+)" [^>]*?text="([^"]+)"[^>]*?>', subtitles), 1):
 113             start = start.replace('.', ',')
 114             end = end.replace('.', ',')
 115             text = clean_html(text)
 116             text = text.replace('\\N', '\n')
 117             if not text:
 118                 continue
 119             output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
 120         return output
 121
 122     def _convert_subtitles_to_ass(self, subtitles):
 123         output = ''
 124
 125         def ass_bool(strvalue):
 126             assvalue = '0'
 127             if strvalue == '1':
 128                 assvalue = '-1'
 129             return assvalue
 130
 131         sub_root = xml.etree.ElementTree.fromstring(subtitles)
 132         if not sub_root:
 133             return output
 134
 135         output = '[Script Info]\n'
 136         output += 'Title: %s\n' % sub_root.attrib["title"]
 137         output += 'ScriptType: v4.00+\n'
 138         output += 'WrapStyle: %s\n' % sub_root.attrib["wrap_style"]
 139         output += 'PlayResX: %s\n' % sub_root.attrib["play_res_x"]
 140         output += 'PlayResY: %s\n' % sub_root.attrib["play_res_y"]
 141         output += """ScaledBorderAndShadow: yes
 142
 143 [V4+ Styles]
 144 Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
 145 """
 146         for style in sub_root.findall('./styles/style'):
 147             output += 'Style: ' + style.attrib["name"]
 148             output += ',' + style.attrib["font_name"]
 149             output += ',' + style.attrib["font_size"]
 150             output += ',' + style.attrib["primary_colour"]
 151             output += ',' + style.attrib["secondary_colour"]
 152             output += ',' + style.attrib["outline_colour"]
 153             output += ',' + style.attrib["back_colour"]
 154             output += ',' + ass_bool(style.attrib["bold"])
 155             output += ',' + ass_bool(style.attrib["italic"])
 156             output += ',' + ass_bool(style.attrib["underline"])
 157             output += ',' + ass_bool(style.attrib["strikeout"])
 158             output += ',' + style.attrib["scale_x"]
 159             output += ',' + style.attrib["scale_y"]
 160             output += ',' + style.attrib["spacing"]
 161             output += ',' + style.attrib["angle"]
 162             output += ',' + style.attrib["border_style"]
 163             output += ',' + style.attrib["outline"]
 164             output += ',' + style.attrib["shadow"]
 165             output += ',' + style.attrib["alignment"]
 166             output += ',' + style.attrib["margin_l"]
 167             output += ',' + style.attrib["margin_r"]
 168             output += ',' + style.attrib["margin_v"]
 169             output += ',' + style.attrib["encoding"]
 170             output += '\n'
 171
 172         output += """
 173 [Events]
 174 Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
 175 """
 176         for event in sub_root.findall('./events/event'):
 177             output += 'Dialogue: 0'
 178             output += ',' + event.attrib["start"]
 179             output += ',' + event.attrib["end"]
 180             output += ',' + event.attrib["style"]
 181             output += ',' + event.attrib["name"]
 182             output += ',' + event.attrib["margin_l"]
 183             output += ',' + event.attrib["margin_r"]
 184             output += ',' + event.attrib["margin_v"]
 185             output += ',' + event.attrib["effect"]
 186             output += ',' + event.attrib["text"]
 187             output += '\n'
 188
 189         return output
 190
 191     def _real_extract(self,url):
 192         mobj = re.match(self._VALID_URL, url)
 193         video_id = mobj.group('video_id')
 194
 195         if mobj.group('prefix') == 'm':
 196             mobile_webpage = self._download_webpage(url, video_id, 'Downloading mobile webpage')
 197             webpage_url = self._search_regex(r'<link rel="canonical" href="([^"]+)" />', mobile_webpage, 'webpage_url')
 198         else:
 199             webpage_url = 'http://www.' + mobj.group('url')
 200
 201         webpage = self._download_webpage(webpage_url, video_id, 'Downloading webpage')
 202         note_m = self._html_search_regex(r'<div class="showmedia-trailer-notice">(.+?)</div>', webpage, 'trailer-notice', default='')
 203         if note_m:
 204             raise ExtractorError(note_m)
 205
 206         mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P<msg>{.+?})\]\)', webpage)
 207         if mobj:
 208             msg = json.loads(mobj.group('msg'))
 209             if msg.get('type') == 'error':
 210                 raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True)
 211
 212         video_title = self._html_search_regex(r'<h1[^>]*>(.+?)</h1>', webpage, 'video_title', flags=re.DOTALL)
 213         video_title = re.sub(r' {2,}', ' ', video_title)
 214         video_description = self._html_search_regex(r'"description":"([^"]+)', webpage, 'video_description', default='')
 215         if not video_description:
 216             video_description = None
 217         video_upload_date = self._html_search_regex(r'<div>Availability for free users:(.+?)</div>', webpage, 'video_upload_date', fatal=False, flags=re.DOTALL)
 218         if video_upload_date:
 219             video_upload_date = unified_strdate(video_upload_date)
 220         video_uploader = self._html_search_regex(r'<div>\s*Publisher:(.+?)</div>', webpage, 'video_uploader', fatal=False, flags=re.DOTALL)
 221
 222         playerdata_url = compat_urllib_parse.unquote(self._html_search_regex(r'"config_url":"([^"]+)', webpage, 'playerdata_url'))
 223         playerdata_req = compat_urllib_request.Request(playerdata_url)
 224         playerdata_req.data = compat_urllib_parse.urlencode({'current_page': webpage_url})
 225         playerdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 226         playerdata = self._download_webpage(playerdata_req, video_id, note='Downloading media info')
 227
 228         stream_id = self._search_regex(r'<media_id>([^<]+)', playerdata, 'stream_id')
 229         video_thumbnail = self._search_regex(r'<episode_image_url>([^<]+)', playerdata, 'thumbnail', fatal=False)
 230
 231         formats = []
 232         for fmt in re.findall(r'\?p([0-9]{3,4})=1', webpage):
 233             stream_quality, stream_format = self._FORMAT_IDS[fmt]
 234             video_format = fmt+'p'
 235             streamdata_req = compat_urllib_request.Request('http://www.crunchyroll.com/xml/')
 236             # urlencode doesn't work!
 237             streamdata_req.data = 'req=RpcApiVideoEncode%5FGetStreamInfo&video%5Fencode%5Fquality='+stream_quality+'&media%5Fid='+stream_id+'&video%5Fformat='+stream_format
 238             streamdata_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 239             streamdata_req.add_header('Content-Length', str(len(streamdata_req.data)))
 240             streamdata = self._download_webpage(streamdata_req, video_id, note='Downloading media info for '+video_format)
 241             video_url = self._search_regex(r'<host>([^<]+)', streamdata, 'video_url')
 242             video_play_path = self._search_regex(r'<file>([^<]+)', streamdata, 'video_play_path')
 243             formats.append({
 244                 'url': video_url,
 245                 'play_path':   video_play_path,
 246                 'ext': 'flv',
 247                 'format': video_format,
 248                 'format_id': video_format,
 249             })
 250
 251         subtitles = {}
 252         sub_format = self._downloader.params.get('subtitlesformat', 'srt')
 253         for sub_id, sub_name in re.findall(r'\?ssid=([0-9]+)" title="([^"]+)', webpage):
 254             sub_page = self._download_webpage('http://www.crunchyroll.com/xml/?req=RpcApiSubtitle_GetXml&subtitle_script_id='+sub_id,\
 255                                               video_id, note='Downloading subtitles for '+sub_name)
 256             id = self._search_regex(r'id=\'([0-9]+)', sub_page, 'subtitle_id', fatal=False)
 257             iv = self._search_regex(r'<iv>([^<]+)', sub_page, 'subtitle_iv', fatal=False)
 258             data = self._search_regex(r'<data>([^<]+)', sub_page, 'subtitle_data', fatal=False)
 259             if not id or not iv or not data:
 260                 continue
 261             id = int(id)
 262             iv = base64.b64decode(iv)
 263             data = base64.b64decode(data)
 264
 265             subtitle = self._decrypt_subtitles(data, iv, id).decode('utf-8')
 266             lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
 267             if not lang_code:
 268                 continue
 269             if sub_format == 'ass':
 270                 subtitles[lang_code] = self._convert_subtitles_to_ass(subtitle)
 271             else:
 272                 subtitles[lang_code] = self._convert_subtitles_to_srt(subtitle)
 273
 274         return {
 275             'id':          video_id,
 276             'title':       video_title,
 277             'description': video_description,
 278             'thumbnail':   video_thumbnail,
 279             'uploader':    video_uploader,
 280             'upload_date': video_upload_date,
 281             'subtitles':   subtitles,
 282             'formats':     formats,
 283         }