]> Raphaƫl G. Git Repositories - youtubedl/blob - youtube_dl/extractor/twitch.py
1981b4d4a8064541c0fcf2faa04d863c72eebb14
[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_HTTPError,
11 compat_parse_qs,
12 compat_str,
13 compat_urllib_parse_urlencode,
14 compat_urllib_parse_urlparse,
15 )
16 from ..utils import (
17 clean_html,
18 ExtractorError,
19 int_or_none,
20 js_to_json,
21 orderedSet,
22 parse_duration,
23 parse_iso8601,
24 update_url_query,
25 urlencode_postdata,
26 urljoin,
27 )
28
29
30 class TwitchBaseIE(InfoExtractor):
31 _VALID_URL_BASE = r'https?://(?:(?:www|go)\.)?twitch\.tv'
32
33 _API_BASE = 'https://api.twitch.tv'
34 _USHER_BASE = 'https://usher.ttvnw.net'
35 _LOGIN_URL = 'https://www.twitch.tv/login'
36 _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6'
37 _NETRC_MACHINE = 'twitch'
38
39 def _handle_error(self, response):
40 if not isinstance(response, dict):
41 return
42 error = response.get('error')
43 if error:
44 raise ExtractorError(
45 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
46 expected=True)
47
48 def _call_api(self, path, item_id, note):
49 response = self._download_json(
50 '%s/%s' % (self._API_BASE, path), item_id, note,
51 headers={'Client-ID': self._CLIENT_ID})
52 self._handle_error(response)
53 return response
54
55 def _real_initialize(self):
56 self._login()
57
58 def _login(self):
59 (username, password) = self._get_login_info()
60 if username is None:
61 return
62
63 def fail(message):
64 raise ExtractorError(
65 'Unable to login. Twitch said: %s' % message, expected=True)
66
67 def login_step(page, urlh, note, data):
68 form = self._hidden_inputs(page)
69 form.update(data)
70
71 page_url = urlh.geturl()
72 post_url = self._search_regex(
73 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
74 'post url', default=page_url, group='url')
75 post_url = urljoin(page_url, post_url)
76
77 headers = {'Referer': page_url}
78
79 try:
80 response = self._download_json(
81 post_url, None, note,
82 data=urlencode_postdata(form),
83 headers=headers)
84 except ExtractorError as e:
85 if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400:
86 response = self._parse_json(
87 e.cause.read().decode('utf-8'), None)
88 fail(response.get('message') or response['errors'][0])
89 raise
90
91 if 'Authenticated successfully' in response.get('message', ''):
92 return None, None
93
94 redirect_url = urljoin(
95 post_url,
96 response.get('redirect') or response['redirect_path'])
97 return self._download_webpage_handle(
98 redirect_url, None, 'Downloading login redirect page',
99 headers=headers)
100
101 login_page, handle = self._download_webpage_handle(
102 self._LOGIN_URL, None, 'Downloading login page')
103
104 # Some TOR nodes and public proxies are blocked completely
105 if 'blacklist_message' in login_page:
106 fail(clean_html(login_page))
107
108 redirect_page, handle = login_step(
109 login_page, handle, 'Logging in', {
110 'username': username,
111 'password': password,
112 })
113
114 # Successful login
115 if not redirect_page:
116 return
117
118 if re.search(r'(?i)<form[^>]+id="two-factor-submit"', redirect_page) is not None:
119 # TODO: Add mechanism to request an SMS or phone call
120 tfa_token = self._get_tfa_info('two-factor authentication token')
121 login_step(redirect_page, handle, 'Submitting TFA token', {
122 'authy_token': tfa_token,
123 'remember_2fa': 'true',
124 })
125
126 def _prefer_source(self, formats):
127 try:
128 source = next(f for f in formats if f['format_id'] == 'Source')
129 source['preference'] = 10
130 except StopIteration:
131 pass # No Source stream present
132 self._sort_formats(formats)
133
134
135 class TwitchItemBaseIE(TwitchBaseIE):
136 def _download_info(self, item, item_id):
137 return self._extract_info(self._call_api(
138 'kraken/videos/%s%s' % (item, item_id), item_id,
139 'Downloading %s info JSON' % self._ITEM_TYPE))
140
141 def _extract_media(self, item_id):
142 info = self._download_info(self._ITEM_SHORTCUT, item_id)
143 response = self._call_api(
144 'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id,
145 'Downloading %s playlist JSON' % self._ITEM_TYPE)
146 entries = []
147 chunks = response['chunks']
148 qualities = list(chunks.keys())
149 for num, fragment in enumerate(zip(*chunks.values()), start=1):
150 formats = []
151 for fmt_num, fragment_fmt in enumerate(fragment):
152 format_id = qualities[fmt_num]
153 fmt = {
154 'url': fragment_fmt['url'],
155 'format_id': format_id,
156 'quality': 1 if format_id == 'live' else 0,
157 }
158 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
159 if m:
160 fmt['height'] = int(m.group('height'))
161 formats.append(fmt)
162 self._sort_formats(formats)
163 entry = dict(info)
164 entry['id'] = '%s_%d' % (entry['id'], num)
165 entry['title'] = '%s part %d' % (entry['title'], num)
166 entry['formats'] = formats
167 entries.append(entry)
168 return self.playlist_result(entries, info['id'], info['title'])
169
170 def _extract_info(self, info):
171 return {
172 'id': info['_id'],
173 'title': info.get('title') or 'Untitled Broadcast',
174 'description': info.get('description'),
175 'duration': int_or_none(info.get('length')),
176 'thumbnail': info.get('preview'),
177 'uploader': info.get('channel', {}).get('display_name'),
178 'uploader_id': info.get('channel', {}).get('name'),
179 'timestamp': parse_iso8601(info.get('recorded_at')),
180 'view_count': int_or_none(info.get('views')),
181 }
182
183 def _real_extract(self, url):
184 return self._extract_media(self._match_id(url))
185
186
187 class TwitchVideoIE(TwitchItemBaseIE):
188 IE_NAME = 'twitch:video'
189 _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
190 _ITEM_TYPE = 'video'
191 _ITEM_SHORTCUT = 'a'
192
193 _TEST = {
194 'url': 'http://www.twitch.tv/riotgames/b/577357806',
195 'info_dict': {
196 'id': 'a577357806',
197 'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
198 },
199 'playlist_mincount': 12,
200 'skip': 'HTTP Error 404: Not Found',
201 }
202
203
204 class TwitchChapterIE(TwitchItemBaseIE):
205 IE_NAME = 'twitch:chapter'
206 _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
207 _ITEM_TYPE = 'chapter'
208 _ITEM_SHORTCUT = 'c'
209
210 _TESTS = [{
211 'url': 'http://www.twitch.tv/acracingleague/c/5285812',
212 'info_dict': {
213 'id': 'c5285812',
214 'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
215 },
216 'playlist_mincount': 3,
217 'skip': 'HTTP Error 404: Not Found',
218 }, {
219 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
220 'only_matching': True,
221 }]
222
223
224 class TwitchVodIE(TwitchItemBaseIE):
225 IE_NAME = 'twitch:vod'
226 _VALID_URL = r'''(?x)
227 https?://
228 (?:
229 (?:(?:www|go)\.)?twitch\.tv/(?:[^/]+/v|videos)/|
230 player\.twitch\.tv/\?.*?\bvideo=v
231 )
232 (?P<id>\d+)
233 '''
234 _ITEM_TYPE = 'vod'
235 _ITEM_SHORTCUT = 'v'
236
237 _TESTS = [{
238 'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
239 'info_dict': {
240 'id': 'v6528877',
241 'ext': 'mp4',
242 'title': 'LCK Summer Split - Week 6 Day 1',
243 'thumbnail': r're:^https?://.*\.jpg$',
244 'duration': 17208,
245 'timestamp': 1435131709,
246 'upload_date': '20150624',
247 'uploader': 'Riot Games',
248 'uploader_id': 'riotgames',
249 'view_count': int,
250 'start_time': 310,
251 },
252 'params': {
253 # m3u8 download
254 'skip_download': True,
255 },
256 }, {
257 # Untitled broadcast (title is None)
258 'url': 'http://www.twitch.tv/belkao_o/v/11230755',
259 'info_dict': {
260 'id': 'v11230755',
261 'ext': 'mp4',
262 'title': 'Untitled Broadcast',
263 'thumbnail': r're:^https?://.*\.jpg$',
264 'duration': 1638,
265 'timestamp': 1439746708,
266 'upload_date': '20150816',
267 'uploader': 'BelkAO_o',
268 'uploader_id': 'belkao_o',
269 'view_count': int,
270 },
271 'params': {
272 # m3u8 download
273 'skip_download': True,
274 },
275 'skip': 'HTTP Error 404: Not Found',
276 }, {
277 'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
278 'only_matching': True,
279 }, {
280 'url': 'https://www.twitch.tv/videos/6528877',
281 'only_matching': True,
282 }]
283
284 def _real_extract(self, url):
285 item_id = self._match_id(url)
286
287 info = self._download_info(self._ITEM_SHORTCUT, item_id)
288 access_token = self._call_api(
289 'api/vods/%s/access_token' % item_id, item_id,
290 'Downloading %s access token' % self._ITEM_TYPE)
291
292 formats = self._extract_m3u8_formats(
293 '%s/vod/%s?%s' % (
294 self._USHER_BASE, item_id,
295 compat_urllib_parse_urlencode({
296 'allow_source': 'true',
297 'allow_audio_only': 'true',
298 'allow_spectre': 'true',
299 'player': 'twitchweb',
300 'nauth': access_token['token'],
301 'nauthsig': access_token['sig'],
302 })),
303 item_id, 'mp4', entry_protocol='m3u8_native')
304
305 self._prefer_source(formats)
306 info['formats'] = formats
307
308 parsed_url = compat_urllib_parse_urlparse(url)
309 query = compat_parse_qs(parsed_url.query)
310 if 't' in query:
311 info['start_time'] = parse_duration(query['t'][0])
312
313 if info.get('timestamp') is not None:
314 info['subtitles'] = {
315 'rechat': [{
316 'url': update_url_query(
317 'https://rechat.twitch.tv/rechat-messages', {
318 'video_id': 'v%s' % item_id,
319 'start': info['timestamp'],
320 }),
321 'ext': 'json',
322 }],
323 }
324
325 return info
326
327
328 class TwitchPlaylistBaseIE(TwitchBaseIE):
329 _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
330 _PAGE_LIMIT = 100
331
332 def _extract_playlist(self, channel_id):
333 info = self._call_api(
334 'kraken/channels/%s' % channel_id,
335 channel_id, 'Downloading channel info JSON')
336 channel_name = info.get('display_name') or info.get('name')
337 entries = []
338 offset = 0
339 limit = self._PAGE_LIMIT
340 broken_paging_detected = False
341 counter_override = None
342 for counter in itertools.count(1):
343 response = self._call_api(
344 self._PLAYLIST_PATH % (channel_id, offset, limit),
345 channel_id,
346 'Downloading %s JSON page %s'
347 % (self._PLAYLIST_TYPE, counter_override or counter))
348 page_entries = self._extract_playlist_page(response)
349 if not page_entries:
350 break
351 total = int_or_none(response.get('_total'))
352 # Since the beginning of March 2016 twitch's paging mechanism
353 # is completely broken on the twitch side. It simply ignores
354 # a limit and returns the whole offset number of videos.
355 # Working around by just requesting all videos at once.
356 # Upd: pagination bug was fixed by twitch on 15.03.2016.
357 if not broken_paging_detected and total and len(page_entries) > limit:
358 self.report_warning(
359 'Twitch pagination is broken on twitch side, requesting all videos at once',
360 channel_id)
361 broken_paging_detected = True
362 offset = total
363 counter_override = '(all at once)'
364 continue
365 entries.extend(page_entries)
366 if broken_paging_detected or total and len(page_entries) >= total:
367 break
368 offset += limit
369 return self.playlist_result(
370 [self._make_url_result(entry) for entry in orderedSet(entries)],
371 channel_id, channel_name)
372
373 def _make_url_result(self, url):
374 try:
375 video_id = 'v%s' % TwitchVodIE._match_id(url)
376 return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id)
377 except AssertionError:
378 return self.url_result(url)
379
380 def _extract_playlist_page(self, response):
381 videos = response.get('videos')
382 return [video['url'] for video in videos] if videos else []
383
384 def _real_extract(self, url):
385 return self._extract_playlist(self._match_id(url))
386
387
388 class TwitchProfileIE(TwitchPlaylistBaseIE):
389 IE_NAME = 'twitch:profile'
390 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
391 _PLAYLIST_TYPE = 'profile'
392
393 _TEST = {
394 'url': 'http://www.twitch.tv/vanillatv/profile',
395 'info_dict': {
396 'id': 'vanillatv',
397 'title': 'VanillaTV',
398 },
399 'playlist_mincount': 412,
400 }
401
402
403 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
404 _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
405 _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
406
407
408 class TwitchAllVideosIE(TwitchVideosBaseIE):
409 IE_NAME = 'twitch:videos:all'
410 _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
411 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
412 _PLAYLIST_TYPE = 'all videos'
413
414 _TEST = {
415 'url': 'https://www.twitch.tv/spamfish/videos/all',
416 'info_dict': {
417 'id': 'spamfish',
418 'title': 'Spamfish',
419 },
420 'playlist_mincount': 869,
421 }
422
423
424 class TwitchUploadsIE(TwitchVideosBaseIE):
425 IE_NAME = 'twitch:videos:uploads'
426 _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
427 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
428 _PLAYLIST_TYPE = 'uploads'
429
430 _TEST = {
431 'url': 'https://www.twitch.tv/spamfish/videos/uploads',
432 'info_dict': {
433 'id': 'spamfish',
434 'title': 'Spamfish',
435 },
436 'playlist_mincount': 0,
437 }
438
439
440 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
441 IE_NAME = 'twitch:videos:past-broadcasts'
442 _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
443 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
444 _PLAYLIST_TYPE = 'past broadcasts'
445
446 _TEST = {
447 'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
448 'info_dict': {
449 'id': 'spamfish',
450 'title': 'Spamfish',
451 },
452 'playlist_mincount': 0,
453 }
454
455
456 class TwitchHighlightsIE(TwitchVideosBaseIE):
457 IE_NAME = 'twitch:videos:highlights'
458 _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
459 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
460 _PLAYLIST_TYPE = 'highlights'
461
462 _TEST = {
463 'url': 'https://www.twitch.tv/spamfish/videos/highlights',
464 'info_dict': {
465 'id': 'spamfish',
466 'title': 'Spamfish',
467 },
468 'playlist_mincount': 805,
469 }
470
471
472 class TwitchStreamIE(TwitchBaseIE):
473 IE_NAME = 'twitch:stream'
474 _VALID_URL = r'''(?x)
475 https?://
476 (?:
477 (?:(?:www|go)\.)?twitch\.tv/|
478 player\.twitch\.tv/\?.*?\bchannel=
479 )
480 (?P<id>[^/#?]+)
481 '''
482
483 _TESTS = [{
484 'url': 'http://www.twitch.tv/shroomztv',
485 'info_dict': {
486 'id': '12772022048',
487 'display_id': 'shroomztv',
488 'ext': 'mp4',
489 'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
490 'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
491 'is_live': True,
492 'timestamp': 1421928037,
493 'upload_date': '20150122',
494 'uploader': 'ShroomzTV',
495 'uploader_id': 'shroomztv',
496 'view_count': int,
497 },
498 'params': {
499 # m3u8 download
500 'skip_download': True,
501 },
502 }, {
503 'url': 'http://www.twitch.tv/miracle_doto#profile-0',
504 'only_matching': True,
505 }, {
506 'url': 'https://player.twitch.tv/?channel=lotsofs',
507 'only_matching': True,
508 }, {
509 'url': 'https://go.twitch.tv/food',
510 'only_matching': True,
511 }]
512
513 @classmethod
514 def suitable(cls, url):
515 return (False
516 if any(ie.suitable(url) for ie in (
517 TwitchVideoIE,
518 TwitchChapterIE,
519 TwitchVodIE,
520 TwitchProfileIE,
521 TwitchAllVideosIE,
522 TwitchUploadsIE,
523 TwitchPastBroadcastsIE,
524 TwitchHighlightsIE))
525 else super(TwitchStreamIE, cls).suitable(url))
526
527 def _real_extract(self, url):
528 channel_id = self._match_id(url)
529
530 stream = self._call_api(
531 'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
532 'Downloading stream JSON').get('stream')
533
534 if not stream:
535 raise ExtractorError('%s is offline' % channel_id, expected=True)
536
537 # Channel name may be typed if different case than the original channel name
538 # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
539 # an invalid m3u8 URL. Working around by use of original channel name from stream
540 # JSON and fallback to lowercase if it's not available.
541 channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
542
543 access_token = self._call_api(
544 'api/channels/%s/access_token' % channel_id, channel_id,
545 'Downloading channel access token')
546
547 query = {
548 'allow_source': 'true',
549 'allow_audio_only': 'true',
550 'allow_spectre': 'true',
551 'p': random.randint(1000000, 10000000),
552 'player': 'twitchweb',
553 'segment_preference': '4',
554 'sig': access_token['sig'].encode('utf-8'),
555 'token': access_token['token'].encode('utf-8'),
556 }
557 formats = self._extract_m3u8_formats(
558 '%s/api/channel/hls/%s.m3u8?%s'
559 % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
560 channel_id, 'mp4')
561 self._prefer_source(formats)
562
563 view_count = stream.get('viewers')
564 timestamp = parse_iso8601(stream.get('created_at'))
565
566 channel = stream['channel']
567 title = self._live_title(channel.get('display_name') or channel.get('name'))
568 description = channel.get('status')
569
570 thumbnails = []
571 for thumbnail_key, thumbnail_url in stream['preview'].items():
572 m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
573 if not m:
574 continue
575 thumbnails.append({
576 'url': thumbnail_url,
577 'width': int(m.group('width')),
578 'height': int(m.group('height')),
579 })
580
581 return {
582 'id': compat_str(stream['_id']),
583 'display_id': channel_id,
584 'title': title,
585 'description': description,
586 'thumbnails': thumbnails,
587 'uploader': channel.get('display_name'),
588 'uploader_id': channel.get('name'),
589 'timestamp': timestamp,
590 'view_count': view_count,
591 'formats': formats,
592 'is_live': True,
593 }
594
595
596 class TwitchClipsIE(InfoExtractor):
597 IE_NAME = 'twitch:clips'
598 _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
599
600 _TESTS = [{
601 'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
602 'md5': '761769e1eafce0ffebfb4089cb3847cd',
603 'info_dict': {
604 'id': 'AggressiveCobraPoooound',
605 'ext': 'mp4',
606 'title': 'EA Play 2016 Live from the Novo Theatre',
607 'thumbnail': r're:^https?://.*\.jpg',
608 'creator': 'EA',
609 'uploader': 'stereotype_',
610 'uploader_id': 'stereotype_',
611 },
612 }, {
613 # multiple formats
614 'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
615 'only_matching': True,
616 }]
617
618 def _real_extract(self, url):
619 video_id = self._match_id(url)
620
621 webpage = self._download_webpage(url, video_id)
622
623 clip = self._parse_json(
624 self._search_regex(
625 r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
626 video_id, transform_source=js_to_json)
627
628 title = clip.get('title') or clip.get('channel_title') or self._og_search_title(webpage)
629
630 formats = [{
631 'url': option['source'],
632 'format_id': option.get('quality'),
633 'height': int_or_none(option.get('quality')),
634 } for option in clip.get('quality_options', []) if option.get('source')]
635
636 if not formats:
637 formats = [{
638 'url': clip['clip_video_url'],
639 }]
640
641 self._sort_formats(formats)
642
643 return {
644 'id': video_id,
645 'title': title,
646 'thumbnail': self._og_search_thumbnail(webpage),
647 'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
648 'uploader': clip.get('curator_login'),
649 'uploader_id': clip.get('curator_display_name'),
650 'formats': formats,
651 }