Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/bilibili.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import hashlib
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_parse_qs,
  10     compat_urlparse,
  11 )
  12 from ..utils import (
  13     ExtractorError,
  14     int_or_none,
  15     float_or_none,
  16     parse_iso8601,
  17     smuggle_url,
  18     strip_jsonp,
  19     unified_timestamp,
  20     unsmuggle_url,
  21     urlencode_postdata,
  22 )
  23
  24
  25 class BiliBiliIE(InfoExtractor):
  26     _VALID_URL = r'https?://(?:www\.|bangumi\.|)bilibili\.(?:tv|com)/(?:video/av|anime/(?P<anime_id>\d+)/play#)(?P<id>\d+)'
  27
  28     _TESTS = [{
  29         'url': 'http://www.bilibili.tv/video/av1074402/',
  30         'md5': '9fa226fe2b8a9a4d5a69b4c6a183417e',
  31         'info_dict': {
  32             'id': '1074402',
  33             'ext': 'mp4',
  34             'title': '【金坷垃】金泡沫',
  35             'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
  36             'duration': 308.315,
  37             'timestamp': 1398012660,
  38             'upload_date': '20140420',
  39             'thumbnail': r're:^https?://.+\.jpg',
  40             'uploader': '菊子桑',
  41             'uploader_id': '156160',
  42         },
  43     }, {
  44         # Tested in BiliBiliBangumiIE
  45         'url': 'http://bangumi.bilibili.com/anime/1869/play#40062',
  46         'only_matching': True,
  47     }, {
  48         'url': 'http://bangumi.bilibili.com/anime/5802/play#100643',
  49         'md5': '3f721ad1e75030cc06faf73587cfec57',
  50         'info_dict': {
  51             'id': '100643',
  52             'ext': 'mp4',
  53             'title': 'CHAOS;CHILD',
  54             'description': '如果你是神明，并且能够让妄想成为现实。那你会进行怎么样的妄想？是淫靡的世界？独裁社会？毁灭性的制裁？还是……2015年，涩谷。从6年前发生的大灾害“涩谷地震”之后复兴了的这个街区里新设立的私立高中...',
  55         },
  56         'skip': 'Geo-restricted to China',
  57     }, {
  58         # Title with double quotes
  59         'url': 'http://www.bilibili.com/video/av8903802/',
  60         'info_dict': {
  61             'id': '8903802',
  62             'ext': 'mp4',
  63             'title': '阿滴英文｜英文歌分享#6 "Closer',
  64             'description': '滴妹今天唱Closer給你聽! 有史以来，被推最多次也是最久的歌曲，其实歌词跟我原本想像差蛮多的，不过还是好听！ 微博@阿滴英文',
  65             'uploader': '阿滴英文',
  66             'uploader_id': '65880958',
  67             'timestamp': 1488382620,
  68             'upload_date': '20170301',
  69         },
  70         'params': {
  71             'skip_download': True,  # Test metadata only
  72         },
  73     }]
  74
  75     _APP_KEY = '84956560bc028eb7'
  76     _BILIBILI_KEY = '94aba54af9065f71de72f5508f1cd42e'
  77
  78     def _report_error(self, result):
  79         if 'message' in result:
  80             raise ExtractorError('%s said: %s' % (self.IE_NAME, result['message']), expected=True)
  81         elif 'code' in result:
  82             raise ExtractorError('%s returns error %d' % (self.IE_NAME, result['code']), expected=True)
  83         else:
  84             raise ExtractorError('Can\'t extract Bangumi episode ID')
  85
  86     def _real_extract(self, url):
  87         url, smuggled_data = unsmuggle_url(url, {})
  88
  89         mobj = re.match(self._VALID_URL, url)
  90         video_id = mobj.group('id')
  91         anime_id = mobj.group('anime_id')
  92         webpage = self._download_webpage(url, video_id)
  93
  94         if 'anime/' not in url:
  95             cid = compat_parse_qs(self._search_regex(
  96                 [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
  97                  r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
  98                 webpage, 'player parameters'))['cid'][0]
  99         else:
 100             if 'no_bangumi_tip' not in smuggled_data:
 101                 self.to_screen('Downloading episode %s. To download all videos in anime %s, re-run youtube-dl with %s' % (
 102                     video_id, anime_id, compat_urlparse.urljoin(url, '//bangumi.bilibili.com/anime/%s' % anime_id)))
 103             headers = {
 104                 'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
 105             }
 106             headers.update(self.geo_verification_headers())
 107
 108             js = self._download_json(
 109                 'http://bangumi.bilibili.com/web_api/get_source', video_id,
 110                 data=urlencode_postdata({'episode_id': video_id}),
 111                 headers=headers)
 112             if 'result' not in js:
 113                 self._report_error(js)
 114             cid = js['result']['cid']
 115
 116         payload = 'appkey=%s&cid=%s&otype=json&quality=2&type=mp4' % (self._APP_KEY, cid)
 117         sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
 118
 119         video_info = self._download_json(
 120             'http://interface.bilibili.com/playurl?%s&sign=%s' % (payload, sign),
 121             video_id, note='Downloading video info page',
 122             headers=self.geo_verification_headers())
 123
 124         if 'durl' not in video_info:
 125             self._report_error(video_info)
 126
 127         entries = []
 128
 129         for idx, durl in enumerate(video_info['durl']):
 130             formats = [{
 131                 'url': durl['url'],
 132                 'filesize': int_or_none(durl['size']),
 133             }]
 134             for backup_url in durl.get('backup_url', []):
 135                 formats.append({
 136                     'url': backup_url,
 137                     # backup URLs have lower priorities
 138                     'preference': -2 if 'hd.mp4' in backup_url else -3,
 139                 })
 140
 141             for a_format in formats:
 142                 a_format.setdefault('http_headers', {}).update({
 143                     'Referer': url,
 144                 })
 145
 146             self._sort_formats(formats)
 147
 148             entries.append({
 149                 'id': '%s_part%s' % (video_id, idx),
 150                 'duration': float_or_none(durl.get('length'), 1000),
 151                 'formats': formats,
 152             })
 153
 154         title = self._html_search_regex('<h1[^>]*>([^<]+)</h1>', webpage, 'title')
 155         description = self._html_search_meta('description', webpage)
 156         timestamp = unified_timestamp(self._html_search_regex(
 157             r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', default=None))
 158         thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage)
 159
 160         # TODO 'view_count' requires deobfuscating Javascript
 161         info = {
 162             'id': video_id,
 163             'title': title,
 164             'description': description,
 165             'timestamp': timestamp,
 166             'thumbnail': thumbnail,
 167             'duration': float_or_none(video_info.get('timelength'), scale=1000),
 168         }
 169
 170         uploader_mobj = re.search(
 171             r'<a[^>]+href="(?:https?:)?//space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"',
 172             webpage)
 173         if uploader_mobj:
 174             info.update({
 175                 'uploader': uploader_mobj.group('name'),
 176                 'uploader_id': uploader_mobj.group('id'),
 177             })
 178
 179         for entry in entries:
 180             entry.update(info)
 181
 182         if len(entries) == 1:
 183             return entries[0]
 184         else:
 185             for idx, entry in enumerate(entries):
 186                 entry['id'] = '%s_part%d' % (video_id, (idx + 1))
 187
 188             return {
 189                 '_type': 'multi_video',
 190                 'id': video_id,
 191                 'title': title,
 192                 'description': description,
 193                 'entries': entries,
 194             }
 195
 196
 197 class BiliBiliBangumiIE(InfoExtractor):
 198     _VALID_URL = r'https?://bangumi\.bilibili\.com/anime/(?P<id>\d+)'
 199
 200     IE_NAME = 'bangumi.bilibili.com'
 201     IE_DESC = 'BiliBili番剧'
 202
 203     _TESTS = [{
 204         'url': 'http://bangumi.bilibili.com/anime/1869',
 205         'info_dict': {
 206             'id': '1869',
 207             'title': '混沌武士',
 208             'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
 209         },
 210         'playlist_count': 26,
 211     }, {
 212         'url': 'http://bangumi.bilibili.com/anime/1869',
 213         'info_dict': {
 214             'id': '1869',
 215             'title': '混沌武士',
 216             'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
 217         },
 218         'playlist': [{
 219             'md5': '91da8621454dd58316851c27c68b0c13',
 220             'info_dict': {
 221                 'id': '40062',
 222                 'ext': 'mp4',
 223                 'title': '混沌武士',
 224                 'description': '故事发生在日本的江户时代。风是一个小酒馆的打工女。一日，酒馆里来了一群恶霸，虽然他们的举动令风十分不满，但是毕竟风只是一届女流，无法对他们采取什么行动，只能在心里嘟哝。这时，酒家里又进来了个“不良份子...',
 225                 'timestamp': 1414538739,
 226                 'upload_date': '20141028',
 227                 'episode': '疾风怒涛 Tempestuous Temperaments',
 228                 'episode_number': 1,
 229             },
 230         }],
 231         'params': {
 232             'playlist_items': '1',
 233         },
 234     }]
 235
 236     @classmethod
 237     def suitable(cls, url):
 238         return False if BiliBiliIE.suitable(url) else super(BiliBiliBangumiIE, cls).suitable(url)
 239
 240     def _real_extract(self, url):
 241         bangumi_id = self._match_id(url)
 242
 243         # Sometimes this API returns a JSONP response
 244         season_info = self._download_json(
 245             'http://bangumi.bilibili.com/jsonp/seasoninfo/%s.ver' % bangumi_id,
 246             bangumi_id, transform_source=strip_jsonp)['result']
 247
 248         entries = [{
 249             '_type': 'url_transparent',
 250             'url': smuggle_url(episode['webplay_url'], {'no_bangumi_tip': 1}),
 251             'ie_key': BiliBiliIE.ie_key(),
 252             'timestamp': parse_iso8601(episode.get('update_time'), delimiter=' '),
 253             'episode': episode.get('index_title'),
 254             'episode_number': int_or_none(episode.get('index')),
 255         } for episode in season_info['episodes']]
 256
 257         entries = sorted(entries, key=lambda entry: entry.get('episode_number'))
 258
 259         return self.playlist_result(
 260             entries, bangumi_id,
 261             season_info.get('bangumi_title'), season_info.get('evaluate'))