2 from __future__ import unicode_literals
8 from .common import InfoExtractor
13 compat_urllib_parse_urlencode,
14 compat_urllib_parse_urlparse,
30 class TwitchBaseIE(InfoExtractor):
31 _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv'
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'
39 def _handle_error(self, response):
40 if not isinstance(response, dict):
42 error = response.get('error')
45 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
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)
55 def _real_initialize(self):
59 (username, password) = self._get_login_info()
65 'Unable to login. Twitch said: %s' % message, expected=True)
67 def login_step(page, urlh, note, data):
68 form = self._hidden_inputs(page)
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)
77 headers = {'Referer': page_url}
80 response = self._download_json(
82 data=urlencode_postdata(form),
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])
91 if 'Authenticated successfully' in response.get('message', ''):
94 redirect_url = urljoin(
96 response.get('redirect') or response['redirect_path'])
97 return self._download_webpage_handle(
98 redirect_url, None, 'Downloading login redirect page',
101 login_page, handle = self._download_webpage_handle(
102 self._LOGIN_URL, None, 'Downloading login page')
104 # Some TOR nodes and public proxies are blocked completely
105 if 'blacklist_message' in login_page:
106 fail(clean_html(login_page))
108 redirect_page, handle = login_step(
109 login_page, handle, 'Logging in', {
110 'username': username,
111 'password': password,
115 if not redirect_page:
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',
126 def _prefer_source(self, formats):
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)
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))
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)
147 chunks = response['chunks']
148 qualities = list(chunks.keys())
149 for num, fragment in enumerate(zip(*chunks.values()), start=1):
151 for fmt_num, fragment_fmt in enumerate(fragment):
152 format_id = qualities[fmt_num]
154 'url': fragment_fmt['url'],
155 'format_id': format_id,
156 'quality': 1 if format_id == 'live' else 0,
158 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
160 fmt['height'] = int(m.group('height'))
162 self._sort_formats(formats)
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'])
170 def _extract_info(self, info):
171 status = info.get('status')
172 if status == 'recording':
174 elif status == 'recorded':
180 'title': info.get('title') or 'Untitled Broadcast',
181 'description': info.get('description'),
182 'duration': int_or_none(info.get('length')),
183 'thumbnail': info.get('preview'),
184 'uploader': info.get('channel', {}).get('display_name'),
185 'uploader_id': info.get('channel', {}).get('name'),
186 'timestamp': parse_iso8601(info.get('recorded_at')),
187 'view_count': int_or_none(info.get('views')),
191 def _real_extract(self, url):
192 return self._extract_media(self._match_id(url))
195 class TwitchVideoIE(TwitchItemBaseIE):
196 IE_NAME = 'twitch:video'
197 _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
202 'url': 'http://www.twitch.tv/riotgames/b/577357806',
205 'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
207 'playlist_mincount': 12,
208 'skip': 'HTTP Error 404: Not Found',
212 class TwitchChapterIE(TwitchItemBaseIE):
213 IE_NAME = 'twitch:chapter'
214 _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
215 _ITEM_TYPE = 'chapter'
219 'url': 'http://www.twitch.tv/acracingleague/c/5285812',
222 'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
224 'playlist_mincount': 3,
225 'skip': 'HTTP Error 404: Not Found',
227 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
228 'only_matching': True,
232 class TwitchVodIE(TwitchItemBaseIE):
233 IE_NAME = 'twitch:vod'
234 _VALID_URL = r'''(?x)
237 (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v|videos)/|
238 player\.twitch\.tv/\?.*?\bvideo=v
246 'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
250 'title': 'LCK Summer Split - Week 6 Day 1',
251 'thumbnail': r're:^https?://.*\.jpg$',
253 'timestamp': 1435131709,
254 'upload_date': '20150624',
255 'uploader': 'Riot Games',
256 'uploader_id': 'riotgames',
262 'skip_download': True,
265 # Untitled broadcast (title is None)
266 'url': 'http://www.twitch.tv/belkao_o/v/11230755',
270 'title': 'Untitled Broadcast',
271 'thumbnail': r're:^https?://.*\.jpg$',
273 'timestamp': 1439746708,
274 'upload_date': '20150816',
275 'uploader': 'BelkAO_o',
276 'uploader_id': 'belkao_o',
281 'skip_download': True,
283 'skip': 'HTTP Error 404: Not Found',
285 'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877',
286 'only_matching': True,
288 'url': 'https://www.twitch.tv/videos/6528877',
289 'only_matching': True,
291 'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
292 'only_matching': True,
295 def _real_extract(self, url):
296 item_id = self._match_id(url)
298 info = self._download_info(self._ITEM_SHORTCUT, item_id)
299 access_token = self._call_api(
300 'api/vods/%s/access_token' % item_id, item_id,
301 'Downloading %s access token' % self._ITEM_TYPE)
303 formats = self._extract_m3u8_formats(
305 self._USHER_BASE, item_id,
306 compat_urllib_parse_urlencode({
307 'allow_source': 'true',
308 'allow_audio_only': 'true',
309 'allow_spectre': 'true',
310 'player': 'twitchweb',
311 'nauth': access_token['token'],
312 'nauthsig': access_token['sig'],
314 item_id, 'mp4', entry_protocol='m3u8_native')
316 self._prefer_source(formats)
317 info['formats'] = formats
319 parsed_url = compat_urllib_parse_urlparse(url)
320 query = compat_parse_qs(parsed_url.query)
322 info['start_time'] = parse_duration(query['t'][0])
324 if info.get('timestamp') is not None:
325 info['subtitles'] = {
327 'url': update_url_query(
328 'https://rechat.twitch.tv/rechat-messages', {
329 'video_id': 'v%s' % item_id,
330 'start': info['timestamp'],
339 class TwitchPlaylistBaseIE(TwitchBaseIE):
340 _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d'
343 def _extract_playlist(self, channel_id):
344 info = self._call_api(
345 'kraken/channels/%s' % channel_id,
346 channel_id, 'Downloading channel info JSON')
347 channel_name = info.get('display_name') or info.get('name')
350 limit = self._PAGE_LIMIT
351 broken_paging_detected = False
352 counter_override = None
353 for counter in itertools.count(1):
354 response = self._call_api(
355 self._PLAYLIST_PATH % (channel_id, offset, limit),
357 'Downloading %s JSON page %s'
358 % (self._PLAYLIST_TYPE, counter_override or counter))
359 page_entries = self._extract_playlist_page(response)
362 total = int_or_none(response.get('_total'))
363 # Since the beginning of March 2016 twitch's paging mechanism
364 # is completely broken on the twitch side. It simply ignores
365 # a limit and returns the whole offset number of videos.
366 # Working around by just requesting all videos at once.
367 # Upd: pagination bug was fixed by twitch on 15.03.2016.
368 if not broken_paging_detected and total and len(page_entries) > limit:
370 'Twitch pagination is broken on twitch side, requesting all videos at once',
372 broken_paging_detected = True
374 counter_override = '(all at once)'
376 entries.extend(page_entries)
377 if broken_paging_detected or total and len(page_entries) >= total:
380 return self.playlist_result(
381 [self._make_url_result(entry) for entry in orderedSet(entries)],
382 channel_id, channel_name)
384 def _make_url_result(self, url):
386 video_id = 'v%s' % TwitchVodIE._match_id(url)
387 return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id)
388 except AssertionError:
389 return self.url_result(url)
391 def _extract_playlist_page(self, response):
392 videos = response.get('videos')
393 return [video['url'] for video in videos] if videos else []
395 def _real_extract(self, url):
396 return self._extract_playlist(self._match_id(url))
399 class TwitchProfileIE(TwitchPlaylistBaseIE):
400 IE_NAME = 'twitch:profile'
401 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
402 _PLAYLIST_TYPE = 'profile'
405 'url': 'http://www.twitch.tv/vanillatv/profile',
408 'title': 'VanillaTV',
410 'playlist_mincount': 412,
412 'url': 'http://m.twitch.tv/vanillatv/profile',
413 'only_matching': True,
417 class TwitchVideosBaseIE(TwitchPlaylistBaseIE):
418 _VALID_URL_VIDEOS_BASE = r'%s/(?P<id>[^/]+)/videos' % TwitchBaseIE._VALID_URL_BASE
419 _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcast_type='
422 class TwitchAllVideosIE(TwitchVideosBaseIE):
423 IE_NAME = 'twitch:videos:all'
424 _VALID_URL = r'%s/all' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
425 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight'
426 _PLAYLIST_TYPE = 'all videos'
429 'url': 'https://www.twitch.tv/spamfish/videos/all',
434 'playlist_mincount': 869,
436 'url': 'https://m.twitch.tv/spamfish/videos/all',
437 'only_matching': True,
441 class TwitchUploadsIE(TwitchVideosBaseIE):
442 IE_NAME = 'twitch:videos:uploads'
443 _VALID_URL = r'%s/uploads' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
444 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload'
445 _PLAYLIST_TYPE = 'uploads'
448 'url': 'https://www.twitch.tv/spamfish/videos/uploads',
453 'playlist_mincount': 0,
455 'url': 'https://m.twitch.tv/spamfish/videos/uploads',
456 'only_matching': True,
460 class TwitchPastBroadcastsIE(TwitchVideosBaseIE):
461 IE_NAME = 'twitch:videos:past-broadcasts'
462 _VALID_URL = r'%s/past-broadcasts' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
463 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive'
464 _PLAYLIST_TYPE = 'past broadcasts'
467 'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts',
472 'playlist_mincount': 0,
474 'url': 'https://m.twitch.tv/spamfish/videos/past-broadcasts',
475 'only_matching': True,
479 class TwitchHighlightsIE(TwitchVideosBaseIE):
480 IE_NAME = 'twitch:videos:highlights'
481 _VALID_URL = r'%s/highlights' % TwitchVideosBaseIE._VALID_URL_VIDEOS_BASE
482 _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight'
483 _PLAYLIST_TYPE = 'highlights'
486 'url': 'https://www.twitch.tv/spamfish/videos/highlights',
491 'playlist_mincount': 805,
493 'url': 'https://m.twitch.tv/spamfish/videos/highlights',
494 'only_matching': True,
498 class TwitchStreamIE(TwitchBaseIE):
499 IE_NAME = 'twitch:stream'
500 _VALID_URL = r'''(?x)
503 (?:(?:www|go|m)\.)?twitch\.tv/|
504 player\.twitch\.tv/\?.*?\bchannel=
510 'url': 'http://www.twitch.tv/shroomztv',
513 'display_id': 'shroomztv',
515 'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
516 'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
518 'timestamp': 1421928037,
519 'upload_date': '20150122',
520 'uploader': 'ShroomzTV',
521 'uploader_id': 'shroomztv',
526 'skip_download': True,
529 'url': 'http://www.twitch.tv/miracle_doto#profile-0',
530 'only_matching': True,
532 'url': 'https://player.twitch.tv/?channel=lotsofs',
533 'only_matching': True,
535 'url': 'https://go.twitch.tv/food',
536 'only_matching': True,
538 'url': 'https://m.twitch.tv/food',
539 'only_matching': True,
543 def suitable(cls, url):
545 if any(ie.suitable(url) for ie in (
552 TwitchPastBroadcastsIE,
554 else super(TwitchStreamIE, cls).suitable(url))
556 def _real_extract(self, url):
557 channel_id = self._match_id(url)
559 stream = self._call_api(
560 'kraken/streams/%s?stream_type=all' % channel_id, channel_id,
561 'Downloading stream JSON').get('stream')
564 raise ExtractorError('%s is offline' % channel_id, expected=True)
566 # Channel name may be typed if different case than the original channel name
567 # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
568 # an invalid m3u8 URL. Working around by use of original channel name from stream
569 # JSON and fallback to lowercase if it's not available.
570 channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
572 access_token = self._call_api(
573 'api/channels/%s/access_token' % channel_id, channel_id,
574 'Downloading channel access token')
577 'allow_source': 'true',
578 'allow_audio_only': 'true',
579 'allow_spectre': 'true',
580 'p': random.randint(1000000, 10000000),
581 'player': 'twitchweb',
582 'segment_preference': '4',
583 'sig': access_token['sig'].encode('utf-8'),
584 'token': access_token['token'].encode('utf-8'),
586 formats = self._extract_m3u8_formats(
587 '%s/api/channel/hls/%s.m3u8?%s'
588 % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
590 self._prefer_source(formats)
592 view_count = stream.get('viewers')
593 timestamp = parse_iso8601(stream.get('created_at'))
595 channel = stream['channel']
596 title = self._live_title(channel.get('display_name') or channel.get('name'))
597 description = channel.get('status')
600 for thumbnail_key, thumbnail_url in stream['preview'].items():
601 m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
605 'url': thumbnail_url,
606 'width': int(m.group('width')),
607 'height': int(m.group('height')),
611 'id': compat_str(stream['_id']),
612 'display_id': channel_id,
614 'description': description,
615 'thumbnails': thumbnails,
616 'uploader': channel.get('display_name'),
617 'uploader_id': channel.get('name'),
618 'timestamp': timestamp,
619 'view_count': view_count,
625 class TwitchClipsIE(InfoExtractor):
626 IE_NAME = 'twitch:clips'
627 _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P<id>[^/?#&]+)'
630 'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound',
631 'md5': '761769e1eafce0ffebfb4089cb3847cd',
633 'id': 'AggressiveCobraPoooound',
635 'title': 'EA Play 2016 Live from the Novo Theatre',
636 'thumbnail': r're:^https?://.*\.jpg',
638 'uploader': 'stereotype_',
639 'uploader_id': 'stereotype_',
643 'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy',
644 'only_matching': True,
647 def _real_extract(self, url):
648 video_id = self._match_id(url)
650 webpage = self._download_webpage(url, video_id)
652 clip = self._parse_json(
654 r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'),
655 video_id, transform_source=js_to_json)
657 title = clip.get('title') or clip.get('channel_title') or self._og_search_title(webpage)
660 'url': option['source'],
661 'format_id': option.get('quality'),
662 'height': int_or_none(option.get('quality')),
663 } for option in clip.get('quality_options', []) if option.get('source')]
667 'url': clip['clip_video_url'],
670 self._sort_formats(formats)
675 'thumbnail': self._og_search_thumbnail(webpage),
676 'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'),
677 'uploader': clip.get('curator_login'),
678 'uploader_id': clip.get('curator_display_name'),