2 from __future__ import unicode_literals
8 from .common import InfoExtractor
12 compat_urllib_parse_urlencode,
13 compat_urllib_parse_urlparse,
26 class TwitchBaseIE(InfoExtractor):
27 _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
29 _API_BASE = 'https://api.twitch.tv'
30 _USHER_BASE = 'http://usher.twitch.tv'
31 _LOGIN_URL = 'http://www.twitch.tv/login'
32 _NETRC_MACHINE = 'twitch'
34 def _handle_error(self, response):
35 if not isinstance(response, dict):
37 error = response.get('error')
40 '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
43 def _download_json(self, url, video_id, note='Downloading JSON metadata'):
45 'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
46 'X-Requested-With': 'XMLHttpRequest',
48 for cookie in self._downloader.cookiejar:
49 if cookie.name == 'api_token':
50 headers['Twitch-Api-Token'] = cookie.value
51 request = sanitized_Request(url, headers=headers)
52 response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
53 self._handle_error(response)
56 def _real_initialize(self):
60 (username, password) = self._get_login_info()
64 login_page, handle = self._download_webpage_handle(
65 self._LOGIN_URL, None, 'Downloading login page')
67 login_form = self._hidden_inputs(login_page)
74 redirect_url = handle.geturl()
76 post_url = self._search_regex(
77 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', login_page,
78 'post url', default=redirect_url, group='url')
80 if not post_url.startswith('http'):
81 post_url = compat_urlparse.urljoin(redirect_url, post_url)
83 request = sanitized_Request(
84 post_url, compat_urllib_parse_urlencode(login_form).encode('utf-8'))
85 request.add_header('Referer', redirect_url)
86 response = self._download_webpage(
87 request, None, 'Logging in as %s' % username)
89 error_message = self._search_regex(
90 r'<div[^>]+class="subwindow_notice"[^>]*>([^<]+)</div>',
91 response, 'error message', default=None)
94 'Unable to login. Twitch said: %s' % error_message, expected=True)
96 if '>Reset your password<' in response:
97 self.report_warning('Twitch asks you to reset your password, go to https://secure.twitch.tv/reset/submit')
99 def _prefer_source(self, formats):
101 source = next(f for f in formats if f['format_id'] == 'Source')
102 source['preference'] = 10
103 except StopIteration:
104 pass # No Source stream present
105 self._sort_formats(formats)
108 class TwitchItemBaseIE(TwitchBaseIE):
109 def _download_info(self, item, item_id):
110 return self._extract_info(self._download_json(
111 '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
112 'Downloading %s info JSON' % self._ITEM_TYPE))
114 def _extract_media(self, item_id):
115 info = self._download_info(self._ITEM_SHORTCUT, item_id)
116 response = self._download_json(
117 '%s/api/videos/%s%s' % (self._API_BASE, self._ITEM_SHORTCUT, item_id), item_id,
118 'Downloading %s playlist JSON' % self._ITEM_TYPE)
120 chunks = response['chunks']
121 qualities = list(chunks.keys())
122 for num, fragment in enumerate(zip(*chunks.values()), start=1):
124 for fmt_num, fragment_fmt in enumerate(fragment):
125 format_id = qualities[fmt_num]
127 'url': fragment_fmt['url'],
128 'format_id': format_id,
129 'quality': 1 if format_id == 'live' else 0,
131 m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
133 fmt['height'] = int(m.group('height'))
135 self._sort_formats(formats)
137 entry['id'] = '%s_%d' % (entry['id'], num)
138 entry['title'] = '%s part %d' % (entry['title'], num)
139 entry['formats'] = formats
140 entries.append(entry)
141 return self.playlist_result(entries, info['id'], info['title'])
143 def _extract_info(self, info):
146 'title': info.get('title') or 'Untitled Broadcast',
147 'description': info.get('description'),
148 'duration': int_or_none(info.get('length')),
149 'thumbnail': info.get('preview'),
150 'uploader': info.get('channel', {}).get('display_name'),
151 'uploader_id': info.get('channel', {}).get('name'),
152 'timestamp': parse_iso8601(info.get('recorded_at')),
153 'view_count': int_or_none(info.get('views')),
156 def _real_extract(self, url):
157 return self._extract_media(self._match_id(url))
160 class TwitchVideoIE(TwitchItemBaseIE):
161 IE_NAME = 'twitch:video'
162 _VALID_URL = r'%s/[^/]+/b/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
167 'url': 'http://www.twitch.tv/riotgames/b/577357806',
170 'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
172 'playlist_mincount': 12,
176 class TwitchChapterIE(TwitchItemBaseIE):
177 IE_NAME = 'twitch:chapter'
178 _VALID_URL = r'%s/[^/]+/c/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
179 _ITEM_TYPE = 'chapter'
183 'url': 'http://www.twitch.tv/acracingleague/c/5285812',
186 'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
188 'playlist_mincount': 3,
190 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361',
191 'only_matching': True,
195 class TwitchVodIE(TwitchItemBaseIE):
196 IE_NAME = 'twitch:vod'
197 _VALID_URL = r'%s/[^/]+/v/(?P<id>\d+)' % TwitchBaseIE._VALID_URL_BASE
202 'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
206 'title': 'LCK Summer Split - Week 6 Day 1',
207 'thumbnail': 're:^https?://.*\.jpg$',
209 'timestamp': 1435131709,
210 'upload_date': '20150624',
211 'uploader': 'Riot Games',
212 'uploader_id': 'riotgames',
218 'skip_download': True,
221 # Untitled broadcast (title is None)
222 'url': 'http://www.twitch.tv/belkao_o/v/11230755',
226 'title': 'Untitled Broadcast',
227 'thumbnail': 're:^https?://.*\.jpg$',
229 'timestamp': 1439746708,
230 'upload_date': '20150816',
231 'uploader': 'BelkAO_o',
232 'uploader_id': 'belkao_o',
237 'skip_download': True,
241 def _real_extract(self, url):
242 item_id = self._match_id(url)
244 info = self._download_info(self._ITEM_SHORTCUT, item_id)
245 access_token = self._download_json(
246 '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id,
247 'Downloading %s access token' % self._ITEM_TYPE)
249 formats = self._extract_m3u8_formats(
251 self._USHER_BASE, item_id,
252 compat_urllib_parse_urlencode({
253 'allow_source': 'true',
254 'allow_audio_only': 'true',
255 'allow_spectre': 'true',
256 'player': 'twitchweb',
257 'nauth': access_token['token'],
258 'nauthsig': access_token['sig'],
262 self._prefer_source(formats)
263 info['formats'] = formats
265 parsed_url = compat_urllib_parse_urlparse(url)
266 query = compat_parse_qs(parsed_url.query)
268 info['start_time'] = parse_duration(query['t'][0])
273 class TwitchPlaylistBaseIE(TwitchBaseIE):
274 _PLAYLIST_URL = '%s/kraken/channels/%%s/videos/?offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
277 def _extract_playlist(self, channel_id):
278 info = self._download_json(
279 '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
280 channel_id, 'Downloading channel info JSON')
281 channel_name = info.get('display_name') or info.get('name')
284 limit = self._PAGE_LIMIT
285 broken_paging_detected = False
286 counter_override = None
287 for counter in itertools.count(1):
288 response = self._download_json(
289 self._PLAYLIST_URL % (channel_id, offset, limit),
291 'Downloading %s videos JSON page %s'
292 % (self._PLAYLIST_TYPE, counter_override or counter))
293 page_entries = self._extract_playlist_page(response)
296 total = int_or_none(response.get('_total'))
297 # Since the beginning of March 2016 twitch's paging mechanism
298 # is completely broken on the twitch side. It simply ignores
299 # a limit and returns the whole offset number of videos.
300 # Working around by just requesting all videos at once.
301 # Upd: pagination bug was fixed by twitch on 15.03.2016.
302 if not broken_paging_detected and total and len(page_entries) > limit:
304 'Twitch pagination is broken on twitch side, requesting all videos at once',
306 broken_paging_detected = True
308 counter_override = '(all at once)'
310 entries.extend(page_entries)
311 if broken_paging_detected or total and len(page_entries) >= total:
314 return self.playlist_result(
315 [self.url_result(entry) for entry in orderedSet(entries)],
316 channel_id, channel_name)
318 def _extract_playlist_page(self, response):
319 videos = response.get('videos')
320 return [video['url'] for video in videos] if videos else []
322 def _real_extract(self, url):
323 return self._extract_playlist(self._match_id(url))
326 class TwitchProfileIE(TwitchPlaylistBaseIE):
327 IE_NAME = 'twitch:profile'
328 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
329 _PLAYLIST_TYPE = 'profile'
332 'url': 'http://www.twitch.tv/vanillatv/profile',
335 'title': 'VanillaTV',
337 'playlist_mincount': 412,
341 class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE):
342 IE_NAME = 'twitch:past_broadcasts'
343 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
344 _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true'
345 _PLAYLIST_TYPE = 'past broadcasts'
348 'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts',
353 'playlist_mincount': 54,
357 class TwitchBookmarksIE(TwitchPlaylistBaseIE):
358 IE_NAME = 'twitch:bookmarks'
359 _VALID_URL = r'%s/(?P<id>[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
360 _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE
361 _PLAYLIST_TYPE = 'bookmarks'
364 'url': 'http://www.twitch.tv/ognos/profile/bookmarks',
369 'playlist_mincount': 3,
372 def _extract_playlist_page(self, response):
374 for bookmark in response.get('bookmarks', []):
375 video = bookmark.get('video')
378 entries.append(video['url'])
382 class TwitchStreamIE(TwitchBaseIE):
383 IE_NAME = 'twitch:stream'
384 _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
387 'url': 'http://www.twitch.tv/shroomztv',
390 'display_id': 'shroomztv',
392 'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$',
393 'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV',
395 'timestamp': 1421928037,
396 'upload_date': '20150122',
397 'uploader': 'ShroomzTV',
398 'uploader_id': 'shroomztv',
403 'skip_download': True,
406 'url': 'http://www.twitch.tv/miracle_doto#profile-0',
407 'only_matching': True,
410 def _real_extract(self, url):
411 channel_id = self._match_id(url)
413 stream = self._download_json(
414 '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id,
415 'Downloading stream JSON').get('stream')
417 # Fallback on profile extraction if stream is offline
419 return self.url_result(
420 'http://www.twitch.tv/%s/profile' % channel_id,
421 'TwitchProfile', channel_id)
423 # Channel name may be typed if different case than the original channel name
424 # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
425 # an invalid m3u8 URL. Working around by use of original channel name from stream
426 # JSON and fallback to lowercase if it's not available.
427 channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
429 access_token = self._download_json(
430 '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
431 'Downloading channel access token')
434 'allow_source': 'true',
435 'allow_audio_only': 'true',
436 'p': random.randint(1000000, 10000000),
437 'player': 'twitchweb',
438 'segment_preference': '4',
439 'sig': access_token['sig'].encode('utf-8'),
440 'token': access_token['token'].encode('utf-8'),
442 formats = self._extract_m3u8_formats(
443 '%s/api/channel/hls/%s.m3u8?%s'
444 % (self._USHER_BASE, channel_id, compat_urllib_parse_urlencode(query)),
446 self._prefer_source(formats)
448 view_count = stream.get('viewers')
449 timestamp = parse_iso8601(stream.get('created_at'))
451 channel = stream['channel']
452 title = self._live_title(channel.get('display_name') or channel.get('name'))
453 description = channel.get('status')
456 for thumbnail_key, thumbnail_url in stream['preview'].items():
457 m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
461 'url': thumbnail_url,
462 'width': int(m.group('width')),
463 'height': int(m.group('height')),
467 'id': compat_str(stream['_id']),
468 'display_id': channel_id,
470 'description': description,
471 'thumbnails': thumbnails,
472 'uploader': channel.get('display_name'),
473 'uploader_id': channel.get('name'),
474 'timestamp': timestamp,
475 'view_count': view_count,