X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Ftwitch.py;h=a2b6a35aa3c89c9ff7c40c5c3385a285abc9a0e5;hb=d5d7bdaeb517f389fff5a6557f072f3586e3c440;hp=14314916a5fbee50d9a6617afa54ecb99fd44ae4;hpb=73e449b226ddfe204fb9f9bea5514141624abbf4;p=youtube-dl diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py index 14314916a..a2b6a35aa 100644 --- a/youtube_dl/extractor/twitch.py +++ b/youtube_dl/extractor/twitch.py @@ -3,15 +3,19 @@ from __future__ import unicode_literals import itertools import re +import random from .common import InfoExtractor from ..compat import ( + compat_parse_qs, compat_str, compat_urllib_parse, + compat_urllib_parse_urlparse, compat_urllib_request, ) from ..utils import ( ExtractorError, + parse_duration, parse_iso8601, ) @@ -21,7 +25,9 @@ class TwitchBaseIE(InfoExtractor): _API_BASE = 'https://api.twitch.tv' _USHER_BASE = 'http://usher.twitch.tv' - _LOGIN_URL = 'https://secure.twitch.tv/user/login' + _LOGIN_URL = 'https://secure.twitch.tv/login' + _LOGIN_POST_URL = 'https://passport.twitch.tv/authorize' + _NETRC_MACHINE = 'twitch' def _handle_error(self, response): if not isinstance(response, dict): @@ -33,7 +39,15 @@ class TwitchBaseIE(InfoExtractor): expected=True) def _download_json(self, url, video_id, note='Downloading JSON metadata'): - response = super(TwitchBaseIE, self)._download_json(url, video_id, note) + headers = { + 'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2', + 'X-Requested-With': 'XMLHttpRequest', + } + for cookie in self._downloader.cookiejar: + if cookie.name == 'api_token': + headers['Twitch-Api-Token'] = cookie.value + request = compat_urllib_request.Request(url, headers=headers) + response = super(TwitchBaseIE, self)._download_json(request, video_id, note) self._handle_error(response) return response @@ -48,32 +62,36 @@ class TwitchBaseIE(InfoExtractor): login_page = self._download_webpage( self._LOGIN_URL, None, 'Downloading login page') - authenticity_token = self._search_regex( - r']*>(?P[^<]+)", response) - if m: + error_message = self._search_regex( + r']+class="subwindow_notice"[^>]*>([^<]+)', + response, 'error message', default=None) + if error_message: raise ExtractorError( - 'Unable to login: %s' % m.group('msg').strip(), expected=True) + 'Unable to login. Twitch said: %s' % error_message, expected=True) + + if '>Reset your password<' in response: + self.report_warning('Twitch asks you to reset your password, go to https://secure.twitch.tv/reset/submit') + + def _prefer_source(self, formats): + try: + source = next(f for f in formats if f['format_id'] == 'Source') + source['preference'] = 10 + except StopIteration: + pass # No Source stream present + self._sort_formats(formats) class TwitchItemBaseIE(TwitchBaseIE): @@ -130,7 +148,7 @@ class TwitchItemBaseIE(TwitchBaseIE): class TwitchVideoIE(TwitchItemBaseIE): IE_NAME = 'twitch:video' - _VALID_URL = r'%s/[^/]+/b/(?P[^/]+)' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'%s/[^/]+/b/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE _ITEM_TYPE = 'video' _ITEM_SHORTCUT = 'a' @@ -146,7 +164,7 @@ class TwitchVideoIE(TwitchItemBaseIE): class TwitchChapterIE(TwitchItemBaseIE): IE_NAME = 'twitch:chapter' - _VALID_URL = r'%s/[^/]+/c/(?P[^/]+)' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'%s/[^/]+/c/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE _ITEM_TYPE = 'chapter' _ITEM_SHORTCUT = 'c' @@ -165,23 +183,24 @@ class TwitchChapterIE(TwitchItemBaseIE): class TwitchVodIE(TwitchItemBaseIE): IE_NAME = 'twitch:vod' - _VALID_URL = r'%s/[^/]+/v/(?P[^/]+)' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'%s/[^/]+/v/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE _ITEM_TYPE = 'vod' _ITEM_SHORTCUT = 'v' _TEST = { - 'url': 'http://www.twitch.tv/ksptv/v/3622000', + 'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s', 'info_dict': { - 'id': 'v3622000', + 'id': 'v6528877', 'ext': 'mp4', - 'title': '''KSPTV: Squadcast: "Everyone's on vacation so here's Dahud" Edition!''', + 'title': 'LCK Summer Split - Week 6 Day 1', 'thumbnail': 're:^https?://.*\.jpg$', - 'duration': 6951, - 'timestamp': 1419028564, - 'upload_date': '20141219', - 'uploader': 'KSPTV', - 'uploader_id': 'ksptv', + 'duration': 17208, + 'timestamp': 1435131709, + 'upload_date': '20150624', + 'uploader': 'Riot Games', + 'uploader_id': 'riotgames', 'view_count': int, + 'start_time': 310, }, 'params': { # m3u8 download @@ -196,10 +215,17 @@ class TwitchVodIE(TwitchItemBaseIE): '%s/api/vods/%s/access_token' % (self._API_BASE, item_id), item_id, 'Downloading %s access token' % self._ITEM_TYPE) formats = self._extract_m3u8_formats( - '%s/vod/%s?nauth=%s&nauthsig=%s' + '%s/vod/%s?nauth=%s&nauthsig=%s&allow_source=true' % (self._USHER_BASE, item_id, access_token['token'], access_token['sig']), item_id, 'mp4') + self._prefer_source(formats) info['formats'] = formats + + parsed_url = compat_urllib_parse_urlparse(url) + query = compat_parse_qs(parsed_url.query) + if 't' in query: + info['start_time'] = parse_duration(query['t'][0]) + return info @@ -219,12 +245,18 @@ class TwitchPlaylistBaseIE(TwitchBaseIE): response = self._download_json( self._PLAYLIST_URL % (channel_id, offset, limit), channel_id, 'Downloading %s videos JSON page %d' % (self._PLAYLIST_TYPE, counter)) - videos = response['videos'] - if not videos: + page_entries = self._extract_playlist_page(response) + if not page_entries: break - entries.extend([self.url_result(video['url']) for video in videos]) + entries.extend(page_entries) offset += limit - return self.playlist_result(entries, channel_id, channel_name) + return self.playlist_result( + [self.url_result(entry) for entry in set(entries)], + channel_id, channel_name) + + def _extract_playlist_page(self, response): + videos = response.get('videos') + return [video['url'] for video in videos] if videos else [] def _real_extract(self, url): return self._extract_playlist(self._match_id(url)) @@ -261,11 +293,36 @@ class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE): } +class TwitchBookmarksIE(TwitchPlaylistBaseIE): + IE_NAME = 'twitch:bookmarks' + _VALID_URL = r'%s/(?P[^/]+)/profile/bookmarks/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE + _PLAYLIST_URL = '%s/api/bookmark/?user=%%s&offset=%%d&limit=%%d' % TwitchBaseIE._API_BASE + _PLAYLIST_TYPE = 'bookmarks' + + _TEST = { + 'url': 'http://www.twitch.tv/ognos/profile/bookmarks', + 'info_dict': { + 'id': 'ognos', + 'title': 'Ognos', + }, + 'playlist_mincount': 3, + } + + def _extract_playlist_page(self, response): + entries = [] + for bookmark in response.get('bookmarks', []): + video = bookmark.get('video') + if not video: + continue + entries.append(video['url']) + return entries + + class TwitchStreamIE(TwitchBaseIE): IE_NAME = 'twitch:stream' - _VALID_URL = r'%s/(?P[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'%s/(?P[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE - _TEST = { + _TESTS = [{ 'url': 'http://www.twitch.tv/shroomztv', 'info_dict': { 'id': '12772022048', @@ -284,7 +341,10 @@ class TwitchStreamIE(TwitchBaseIE): # m3u8 download 'skip_download': True, }, - } + }, { + 'url': 'http://www.twitch.tv/miracle_doto#profile-0', + 'only_matching': True, + }] def _real_extract(self, url): channel_id = self._match_id(url) @@ -299,23 +359,29 @@ class TwitchStreamIE(TwitchBaseIE): 'http://www.twitch.tv/%s/profile' % channel_id, 'TwitchProfile', channel_id) + # Channel name may be typed if different case than the original channel name + # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing + # an invalid m3u8 URL. Working around by use of original channel name from stream + # JSON and fallback to lowercase if it's not available. + channel_id = stream.get('channel', {}).get('name') or channel_id.lower() + access_token = self._download_json( '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id, 'Downloading channel access token') query = { 'allow_source': 'true', - 'p': '9386337', + 'p': random.randint(1000000, 10000000), 'player': 'twitchweb', 'segment_preference': '4', - 'sig': access_token['sig'], - 'token': access_token['token'], + 'sig': access_token['sig'].encode('utf-8'), + 'token': access_token['token'].encode('utf-8'), } - formats = self._extract_m3u8_formats( '%s/api/channel/hls/%s.m3u8?%s' - % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query).encode('utf-8')), + % (self._USHER_BASE, channel_id, compat_urllib_parse.urlencode(query)), channel_id, 'mp4') + self._prefer_source(formats) view_count = stream.get('viewers') timestamp = parse_iso8601(stream.get('created_at')) @@ -347,4 +413,4 @@ class TwitchStreamIE(TwitchBaseIE): 'view_count': view_count, 'formats': formats, 'is_live': True, - } \ No newline at end of file + }