]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/bilibili.py
Fix extraction from youtube.
[youtubedl] / 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': '5f7d29e1a2872f3df0cf76b1f87d3788',
31 'info_dict': {
32 'id': '1074402',
33 'ext': 'flv',
34 'title': '【金坷垃】金泡沫',
35 'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
36 'duration': 308.067,
37 'timestamp': 1398012678,
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 'title': '阿滴英文|英文歌分享#6 "Closer',
63 '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',
64 },
65 'playlist': [{
66 'info_dict': {
67 'id': '8903802_part1',
68 'ext': 'flv',
69 'title': '阿滴英文|英文歌分享#6 "Closer',
70 'description': 'md5:3b1b9e25b78da4ef87e9b548b88ee76a',
71 'uploader': '阿滴英文',
72 'uploader_id': '65880958',
73 'timestamp': 1488382634,
74 'upload_date': '20170301',
75 },
76 'params': {
77 'skip_download': True, # Test metadata only
78 },
79 }, {
80 'info_dict': {
81 'id': '8903802_part2',
82 'ext': 'flv',
83 'title': '阿滴英文|英文歌分享#6 "Closer',
84 'description': 'md5:3b1b9e25b78da4ef87e9b548b88ee76a',
85 'uploader': '阿滴英文',
86 'uploader_id': '65880958',
87 'timestamp': 1488382634,
88 'upload_date': '20170301',
89 },
90 'params': {
91 'skip_download': True, # Test metadata only
92 },
93 }]
94 }]
95
96 _APP_KEY = '84956560bc028eb7'
97 _BILIBILI_KEY = '94aba54af9065f71de72f5508f1cd42e'
98
99 def _report_error(self, result):
100 if 'message' in result:
101 raise ExtractorError('%s said: %s' % (self.IE_NAME, result['message']), expected=True)
102 elif 'code' in result:
103 raise ExtractorError('%s returns error %d' % (self.IE_NAME, result['code']), expected=True)
104 else:
105 raise ExtractorError('Can\'t extract Bangumi episode ID')
106
107 def _real_extract(self, url):
108 url, smuggled_data = unsmuggle_url(url, {})
109
110 mobj = re.match(self._VALID_URL, url)
111 video_id = mobj.group('id')
112 anime_id = mobj.group('anime_id')
113 webpage = self._download_webpage(url, video_id)
114
115 if 'anime/' not in url:
116 cid = self._search_regex(
117 r'\bcid(?:["\']:|=)(\d+)', webpage, 'cid',
118 default=None
119 ) or compat_parse_qs(self._search_regex(
120 [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
121 r'EmbedPlayer\([^)]+,\s*\\"([^"]+)\\"\)',
122 r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
123 webpage, 'player parameters'))['cid'][0]
124 else:
125 if 'no_bangumi_tip' not in smuggled_data:
126 self.to_screen('Downloading episode %s. To download all videos in anime %s, re-run youtube-dl with %s' % (
127 video_id, anime_id, compat_urlparse.urljoin(url, '//bangumi.bilibili.com/anime/%s' % anime_id)))
128 headers = {
129 'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
130 'Referer': url
131 }
132 headers.update(self.geo_verification_headers())
133
134 js = self._download_json(
135 'http://bangumi.bilibili.com/web_api/get_source', video_id,
136 data=urlencode_postdata({'episode_id': video_id}),
137 headers=headers)
138 if 'result' not in js:
139 self._report_error(js)
140 cid = js['result']['cid']
141
142 headers = {
143 'Referer': url
144 }
145 headers.update(self.geo_verification_headers())
146
147 entries = []
148
149 RENDITIONS = ('qn=80&quality=80&type=', 'quality=2&type=mp4')
150 for num, rendition in enumerate(RENDITIONS, start=1):
151 payload = 'appkey=%s&cid=%s&otype=json&%s' % (self._APP_KEY, cid, rendition)
152 sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
153
154 video_info = self._download_json(
155 'http://interface.bilibili.com/v2/playurl?%s&sign=%s' % (payload, sign),
156 video_id, note='Downloading video info page',
157 headers=headers, fatal=num == len(RENDITIONS))
158
159 if not video_info:
160 continue
161
162 if 'durl' not in video_info:
163 if num < len(RENDITIONS):
164 continue
165 self._report_error(video_info)
166
167 for idx, durl in enumerate(video_info['durl']):
168 formats = [{
169 'url': durl['url'],
170 'filesize': int_or_none(durl['size']),
171 }]
172 for backup_url in durl.get('backup_url', []):
173 formats.append({
174 'url': backup_url,
175 # backup URLs have lower priorities
176 'preference': -2 if 'hd.mp4' in backup_url else -3,
177 })
178
179 for a_format in formats:
180 a_format.setdefault('http_headers', {}).update({
181 'Referer': url,
182 })
183
184 self._sort_formats(formats)
185
186 entries.append({
187 'id': '%s_part%s' % (video_id, idx),
188 'duration': float_or_none(durl.get('length'), 1000),
189 'formats': formats,
190 })
191 break
192
193 title = self._html_search_regex(
194 ('<h1[^>]+\btitle=(["\'])(?P<title>(?:(?!\1).)+)\1',
195 '(?s)<h1[^>]*>(?P<title>.+?)</h1>'), webpage, 'title',
196 group='title')
197 description = self._html_search_meta('description', webpage)
198 timestamp = unified_timestamp(self._html_search_regex(
199 r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time',
200 default=None) or self._html_search_meta(
201 'uploadDate', webpage, 'timestamp', default=None))
202 thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage)
203
204 # TODO 'view_count' requires deobfuscating Javascript
205 info = {
206 'id': video_id,
207 'title': title,
208 'description': description,
209 'timestamp': timestamp,
210 'thumbnail': thumbnail,
211 'duration': float_or_none(video_info.get('timelength'), scale=1000),
212 }
213
214 uploader_mobj = re.search(
215 r'<a[^>]+href="(?:https?:)?//space\.bilibili\.com/(?P<id>\d+)"[^>]*>(?P<name>[^<]+)',
216 webpage)
217 if uploader_mobj:
218 info.update({
219 'uploader': uploader_mobj.group('name'),
220 'uploader_id': uploader_mobj.group('id'),
221 })
222 if not info.get('uploader'):
223 info['uploader'] = self._html_search_meta(
224 'author', webpage, 'uploader', default=None)
225
226 for entry in entries:
227 entry.update(info)
228
229 if len(entries) == 1:
230 return entries[0]
231 else:
232 for idx, entry in enumerate(entries):
233 entry['id'] = '%s_part%d' % (video_id, (idx + 1))
234
235 return {
236 '_type': 'multi_video',
237 'id': video_id,
238 'title': title,
239 'description': description,
240 'entries': entries,
241 }
242
243
244 class BiliBiliBangumiIE(InfoExtractor):
245 _VALID_URL = r'https?://bangumi\.bilibili\.com/anime/(?P<id>\d+)'
246
247 IE_NAME = 'bangumi.bilibili.com'
248 IE_DESC = 'BiliBili番剧'
249
250 _TESTS = [{
251 'url': 'http://bangumi.bilibili.com/anime/1869',
252 'info_dict': {
253 'id': '1869',
254 'title': '混沌武士',
255 'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
256 },
257 'playlist_count': 26,
258 }, {
259 'url': 'http://bangumi.bilibili.com/anime/1869',
260 'info_dict': {
261 'id': '1869',
262 'title': '混沌武士',
263 'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
264 },
265 'playlist': [{
266 'md5': '91da8621454dd58316851c27c68b0c13',
267 'info_dict': {
268 'id': '40062',
269 'ext': 'mp4',
270 'title': '混沌武士',
271 '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...',
272 'timestamp': 1414538739,
273 'upload_date': '20141028',
274 'episode': '疾风怒涛 Tempestuous Temperaments',
275 'episode_number': 1,
276 },
277 }],
278 'params': {
279 'playlist_items': '1',
280 },
281 }]
282
283 @classmethod
284 def suitable(cls, url):
285 return False if BiliBiliIE.suitable(url) else super(BiliBiliBangumiIE, cls).suitable(url)
286
287 def _real_extract(self, url):
288 bangumi_id = self._match_id(url)
289
290 # Sometimes this API returns a JSONP response
291 season_info = self._download_json(
292 'http://bangumi.bilibili.com/jsonp/seasoninfo/%s.ver' % bangumi_id,
293 bangumi_id, transform_source=strip_jsonp)['result']
294
295 entries = [{
296 '_type': 'url_transparent',
297 'url': smuggle_url(episode['webplay_url'], {'no_bangumi_tip': 1}),
298 'ie_key': BiliBiliIE.ie_key(),
299 'timestamp': parse_iso8601(episode.get('update_time'), delimiter=' '),
300 'episode': episode.get('index_title'),
301 'episode_number': int_or_none(episode.get('index')),
302 } for episode in season_info['episodes']]
303
304 entries = sorted(entries, key=lambda entry: entry.get('episode_number'))
305
306 return self.playlist_result(
307 entries, bangumi_id,
308 season_info.get('bangumi_title'), season_info.get('evaluate'))