]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/twitch.py
debian/control: Add recommends on aria2 | wget | curl to use external downloaders.
[youtubedl] / youtube_dl / extractor / twitch.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import itertools
5 import re
6 import random
7
8 from .common import InfoExtractor
9 from ..compat import (
10 compat_str,
11 compat_urllib_parse,
12 compat_urllib_request,
13 )
14 from ..utils import (
15 ExtractorError,
16 parse_iso8601,
17 )
18
19
20 class TwitchBaseIE(InfoExtractor):
21 _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
22
23 _API_BASE = 'https://api.twitch.tv'
24 _USHER_BASE = 'http://usher.twitch.tv'
25 _LOGIN_URL = 'https://secure.twitch.tv/user/login'
26
27 def _handle_error(self, response):
28 if not isinstance(response, dict):
29 return
30 error = response.get('error')
31 if error:
32 raise ExtractorError(
33 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
34 expected=True)
35
36 def _download_json(self, url, video_id, note='Downloading JSON metadata'):
37 response = super(TwitchBaseIE, self)._download_json(url, video_id, note)
38 self._handle_error(response)
39 return response
40
41 def _real_initialize(self):
42 self._login()
43
44 def _login(self):
45 (username, password) = self._get_login_info()
46 if username is None:
47 return
48
49 login_page = self._download_webpage(
50 self._LOGIN_URL, None, 'Downloading login page')
51
52 authenticity_token = self._search_regex(
53 r'<input name="authenticity_token" type="hidden" value="([^"]+)"',
54 login_page, 'authenticity token')
55
56 login_form = {
57 'utf8': '✓'.encode('utf-8'),
58 'authenticity_token': authenticity_token,
59 'redirect_on_login': '',
60 'embed_form': 'false',
61 'mp_source_action': '',
62 'follow': '',
63 'user[login]': username,
64 'user[password]': password,
65 }
66
67 request = compat_urllib_request.Request(
68 self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
69 request.add_header('Referer', self._LOGIN_URL)
70 response = self._download_webpage(
71 request, None, 'Logging in as %s' % username)
72
73 m = re.search(
74 r"id=([\"'])login_error_message\1[^>]*>(?P<msg>[^<]+)", response)
75 if m:
76 raise ExtractorError(
77 'Unable to login: %s' % m.group('msg').strip(), expected=True)
78
79
80 class TwitchItemBaseIE(TwitchBaseIE):
81 def _download_info(self, item, item_id):
82 return self._extract_info(self._download_json(
83 '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
84 'Downloading %s info JSON' % self._ITEM_TYPE))
85
86 def _extract_media(self, item_id):
87 info = self._download_info(self._ITEM_SHORTCUT, item_id)
88 response = self._download_json(
89 '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
90 'Downloading %s playlist JSON' % self._ITEM_TYPE)
91 entries = []
92 chunks = response['chunks']
93 qualities = list(chunks.keys())
94 for num, fragment in enumerate(zip(*chunks.values()), start=1):
95 formats = []
96 for fmt_num, fragment_fmt in enumerate(fragment):
97 format_id = qualities[fmt_num]
98 fmt = {
99 'url': fragment_fmt['url'],
100 'format_id': format_id,
101 'quality': 1 if format_id == 'live' else 0,
102 }
103 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
104 if m:
105 fmt['height'] = int(m.group('height'))
106 formats.append(fmt)
107 self._sort_formats(formats)
108 entry = dict(info)
109 entry['id'] = '%s_%d' % (entry['id'], num)
110 entry['title'] = '%s part %d' % (entry['title'], num)
111 entry['formats'] = formats
112 entries.append(entry)
113 return self.playlist_result(entries, info['id'], info['title'])
114
115 def _extract_info(self, info):
116 return {
117 'id': info['_id'],
118 'title': info['title'],
119 'description': info['description'],
120 'duration': info['length'],
121 'thumbnail': info['preview'],
122 'uploader': info['channel']['display_name'],
123 'uploader_id': info['channel']['name'],
124 'timestamp': parse_iso8601(info['recorded_at']),
125 'view_count': info['views'],
126 }
127
128 def _real_extract(self, url):
129 return self._extract_media(self._match_id(url))
130
131
132 class TwitchVideoIE(TwitchItemBaseIE):
133 IE_NAME = 'twitch:video'
134 _VALID_URL = r'%s/[^/]+/b/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
135 _ITEM_TYPE = 'video'
136 _ITEM_SHORTCUT = 'a'
137
138 _TEST = {
139 'url': 'http://www.twitch.tv/riotgames/b/577357806',
140 'info_dict': {
141 'id': 'a577357806',
142 'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
143 },
144 'playlist_mincount': 12,
145 }
146
147
148 class TwitchChapterIE(TwitchItemBaseIE):
149 IE_NAME = 'twitch:chapter'
150 _VALID_URL = r'%s/[^/]+/c/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
151 _ITEM_TYPE = 'chapter'
152 _ITEM_SHORTCUT = 'c'
153
154 _TESTS = [{
155 'url': 'http://www.twitch.tv/acracingleague/c/5285812',
156 'info_dict': {
157 'id': 'c5285812',
158 'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
159 },
160 'playlist_mincount': 3,
161 }, {
162 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
163 'only_matching': True,
164 }]
165
166
167 class TwitchVodIE(TwitchItemBaseIE):
168 IE_NAME = 'twitch:vod'
169 _VALID_URL = r'%s/[^/]+/v/(?P<id>[^/]+)' % TwitchBaseIE._VALID_URL_BASE
170 _ITEM_TYPE = 'vod'
171 _ITEM_SHORTCUT = 'v'
172
173 _TEST = {
174 'url': 'http://www.twitch.tv/ksptv/v/3622000',
175 'info_dict': {
176 'id': 'v3622000',
177 'ext': 'mp4',
178 'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''',
179 'thumbnail': 're:^https?://.*\.jpg$',
180 'duration': 6951,
181 'timestamp': 1419028564,
182 'upload_date': '20141219',
183 'uploader': 'KSPTV',
184 'uploader_id': 'ksptv',
185 'view_count': int,
186 },
187 'params': {
188 # m3u8 download
189 'skip_download': True,
190 },
191 }
192
193 def _real_extract(self, url):
194 item_id = self._match_id(url)
195 info = self._download_info(self._ITEM_SHORTCUT, item_id)
196 access_token = self._download_json(
197 '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
198 'Downloading %s access token' % self._ITEM_TYPE)
199 formats = self._extract_m3u8_formats(
200 '%s/vod/%s?nauth=%s&nauthsig=%s'
201 % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']),
202 item_id, 'mp4')
203 info['formats'] = formats
204 return info
205
206
207 class TwitchPlaylistBaseIE(TwitchBaseIE):
208 _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
209 _PAGE_LIMIT = 100
210
211 def _extract_playlist(self, channel_id):
212 info = self._download_json(
213 '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
214 channel_id, 'Downloading channel info JSON')
215 channel_name = info.get('display_name') or info.get('name')
216 entries = []
217 offset = 0
218 limit = self._PAGE_LIMIT
219 for counter in itertools.count(1):
220 response = self._download_json(
221 self._PLAYLIST_URL % (channel_id, offset, limit),
222 channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter))
223 page_entries = self._extract_playlist_page(response)
224 if not page_entries:
225 break
226 entries.extend(page_entries)
227 offset += limit
228 return self.playlist_result(
229 [self.url_result(entry) for entry in set(entries)],
230 channel_id, channel_name)
231
232 def _extract_playlist_page(self, response):
233 videos = response.get('videos')
234 return [video['url'] for video in videos] if videos else []
235
236 def _real_extract(self, url):
237 return self._extract_playlist(self._match_id(url))
238
239
240 class TwitchProfileIE(TwitchPlaylistBaseIE):
241 IE_NAME = 'twitch:profile'
242 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
243 _PLAYLIST_TYPE = 'profile'
244
245 _TEST = {
246 'url': 'http://www.twitch.tv/vanillatv/profile',
247 'info_dict': {
248 'id': 'vanillatv',
249 'title': 'VanillaTV',
250 },
251 'playlist_mincount': 412,
252 }
253
254
255 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
256 IE_NAME = 'twitch:past_broadcasts'
257 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
258 _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
259 _PLAYLIST_TYPE = 'past broadcasts'
260
261 _TEST = {
262 'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
263 'info_dict': {
264 'id': 'spamfish',
265 'title': 'Spamfish',
266 },
267 'playlist_mincount': 54,
268 }
269
270
271 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
272 IE_NAME = 'twitch:bookmarks'
273 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
274 _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
275 _PLAYLIST_TYPE = 'bookmarks'
276
277 _TEST = {
278 'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
279 'info_dict': {
280 'id': 'ognos',
281 'title': 'Ognos',
282 },
283 'playlist_mincount': 3,
284 }
285
286 def _extract_playlist_page(self, response):
287 entries = []
288 for bookmark in response.get('bookmarks', []):
289 video = bookmark.get('video')
290 if not video:
291 continue
292 entries.append(video['url'])
293 return entries
294
295
296 class TwitchStreamIE(TwitchBaseIE):
297 IE_NAME = 'twitch:stream'
298 _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
299
300 _TEST = {
301 'url': 'http://www.twitch.tv/shroomztv',
302 'info_dict': {
303 'id': '12772022048',
304 'display_id': 'shroomztv',
305 'ext': 'mp4',
306 'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
307 'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
308 'is_live': True,
309 'timestamp': 1421928037,
310 'upload_date': '20150122',
311 'uploader': 'ShroomzTV',
312 'uploader_id': 'shroomztv',
313 'view_count': int,
314 },
315 'params': {
316 # m3u8 download
317 'skip_download': True,
318 },
319 }
320
321 def _real_extract(self, url):
322 channel_id = self._match_id(url)
323
324 stream = self._download_json(
325 '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
326 'Downloading stream JSON').get('stream')
327
328 # Fallback on profile extraction if stream is offline
329 if not stream:
330 return self.url_result(
331 'http://www.twitch.tv/%s/profile' % channel_id,
332 'TwitchProfile', channel_id)
333
334 access_token = self._download_json(
335 '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
336 'Downloading channel access token')
337
338 query = {
339 'allow_source': 'true',
340 'p': random.randint(1000000, 10000000),
341 'player': 'twitchweb',
342 'segment_preference': '4',
343 'sig': access_token['sig'],
344 'token': access_token['token'],
345 }
346
347 formats = self._extract_m3u8_formats(
348 '%s/api/channel/hls/%s.m3u8?%s'
349 % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query).encode('utf-8')),
350 channel_id, 'mp4')
351
352 view_count = stream.get('viewers')
353 timestamp = parse_iso8601(stream.get('created_at'))
354
355 channel = stream['channel']
356 title = self._live_title(channel.get('display_name') or channel.get('name'))
357 description = channel.get('status')
358
359 thumbnails = []
360 for thumbnail_key, thumbnail_url in stream['preview'].items():
361 m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
362 if not m:
363 continue
364 thumbnails.append({
365 'url': thumbnail_url,
366 'width': int(m.group('width')),
367 'height': int(m.group('height')),
368 })
369
370 return {
371 'id': compat_str(stream['_id']),
372 'display_id': channel_id,
373 'title': title,
374 'description': description,
375 'thumbnails': thumbnails,
376 'uploader': channel.get('display_name'),
377 'uploader_id': channel.get('name'),
378 'timestamp': timestamp,
379 'view_count': view_count,
380 'formats': formats,
381 'is_live': True,
382 }