X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Ftwitch.py;h=69882da6337bc6f20d0578f6d612e799429db3e8;hb=5c2266df4b9aeb7881ed8c026a038e2a25e43734;hp=f70002f68495e3e910586fc7ddeccc74523ab665;hpb=d63528c8c7965545969e77b983e0e42d2e41e31f;p=youtube-dl diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py index f70002f68..69882da63 100644 --- a/youtube_dl/extractor/twitch.py +++ b/youtube_dl/extractor/twitch.py @@ -3,23 +3,33 @@ 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_request, + compat_urllib_parse_urlparse, + compat_urlparse, ) from ..utils import ( + encode_dict, ExtractorError, + int_or_none, + parse_duration, parse_iso8601, + sanitized_Request, ) class TwitchBaseIE(InfoExtractor): - _VALID_URL_BASE = r'http://(?:www\.)?twitch\.tv' + _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv' _API_BASE = 'https://api.twitch.tv' - _LOGIN_URL = 'https://secure.twitch.tv/user/login' + _USHER_BASE = 'http://usher.twitch.tv' + _LOGIN_URL = 'http://www.twitch.tv/login' + _NETRC_MACHINE = 'twitch' def _handle_error(self, response): if not isinstance(response, dict): @@ -31,7 +41,15 @@ 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 = sanitized_Request(url, headers=headers) + response = super(TwitchBaseIE, self)._download_json(request, video_id, note) self._handle_error(response) return response @@ -43,35 +61,48 @@ class TwitchBaseIE(InfoExtractor): if username is None: return - login_page = self._download_webpage( + login_page, handle = self._download_webpage_handle( self._LOGIN_URL, None, 'Downloading login page') - authenticity_token = self._search_regex( - r']+action=(["\'])(?P.+?)\1', login_page, + 'post url', default=redirect_url, group='url') + + if not post_url.startswith('http'): + post_url = compat_urlparse.urljoin(redirect_url, post_url) + + request = sanitized_Request( + post_url, compat_urllib_parse.urlencode(encode_dict(login_form)).encode('utf-8')) + request.add_header('Referer', redirect_url) response = self._download_webpage( request, None, 'Logging in as %s' % username) - m = re.search( - r"id=([\"'])login_error_message\1[^>]*>(?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): @@ -112,14 +143,14 @@ class TwitchItemBaseIE(TwitchBaseIE): def _extract_info(self, info): return { 'id': info['_id'], - 'title': info['title'], - 'description': info['description'], - 'duration': info['length'], - 'thumbnail': info['preview'], - 'uploader': info['channel']['display_name'], - 'uploader_id': info['channel']['name'], - 'timestamp': parse_iso8601(info['recorded_at']), - 'view_count': info['views'], + 'title': info.get('title') or 'Untitled Broadcast', + 'description': info.get('description'), + 'duration': int_or_none(info.get('length')), + 'thumbnail': info.get('preview'), + 'uploader': info.get('channel', {}).get('display_name'), + 'uploader_id': info.get('channel', {}).get('name'), + 'timestamp': parse_iso8601(info.get('recorded_at')), + 'view_count': int_or_none(info.get('views')), } def _real_extract(self, url): @@ -128,7 +159,7 @@ 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' @@ -144,57 +175,97 @@ 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' - _TEST = { + _TESTS = [{ 'url': 'http://www.twitch.tv/acracingleague/c/5285812', 'info_dict': { 'id': 'c5285812', 'title': 'ACRL Off Season - Sports Cars @ Nordschleife', }, 'playlist_mincount': 3, - } + }, { + 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361', + 'only_matching': True, + }] 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', + _TESTS = [{ + '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$', - 'duration': 6951, - 'timestamp': 1419028564, - 'upload_date': '20141219', - 'uploader': 'KSPTV', - 'uploader_id': 'ksptv', + 'duration': 17208, + 'timestamp': 1435131709, + 'upload_date': '20150624', + 'uploader': 'Riot Games', + 'uploader_id': 'riotgames', 'view_count': int, + 'start_time': 310, }, 'params': { # m3u8 download 'skip_download': True, }, - } + }, { + # Untitled broadcast (title is None) + 'url': 'http://www.twitch.tv/belkao_o/v/11230755', + 'info_dict': { + 'id': 'v11230755', + 'ext': 'mp4', + 'title': 'Untitled Broadcast', + 'thumbnail': 're:^https?://.*\.jpg$', + 'duration': 1638, + 'timestamp': 1439746708, + 'upload_date': '20150816', + 'uploader': 'BelkAO_o', + 'uploader_id': 'belkao_o', + 'view_count': int, + }, + 'params': { + # m3u8 download + 'skip_download': True, + }, + }] def _real_extract(self, url): item_id = self._match_id(url) + info = self._download_info(self._ITEM_SHORTCUT, item_id) access_token = self._download_json( '%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( - 'http://usher.twitch.tv/vod/%s?nauth=%s&nauthsig=%s' - % (item_id, access_token['token'], access_token['sig']), + '%s/vod/%s?%s' % ( + self._USHER_BASE, item_id, + compat_urllib_parse.urlencode({ + 'allow_source': 'true', + 'allow_spectre': 'true', + 'player': 'twitchweb', + 'nauth': access_token['token'], + 'nauthsig': 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 @@ -214,12 +285,18 @@ 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)) @@ -241,7 +318,7 @@ class TwitchProfileIE(TwitchPlaylistBaseIE): class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE): - IE_NAME = 'twitch:profile' + IE_NAME = 'twitch:past_broadcasts' _VALID_URL = r'%s/(?P[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE _PLAYLIST_URL = TwitchPlaylistBaseIE._PLAYLIST_URL + '&broadcasts=true' _PLAYLIST_TYPE = 'past broadcasts' @@ -254,3 +331,126 @@ class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE): }, 'playlist_mincount': 54, } + + +class TwitchBookmarksIE(TwitchPlaylistBaseIE): + IE_NAME = 'twitch:bookmarks' + _VALID_URL = r'%s/(?P[^/]+)/profile/bookmarks/?(?:\#.*)?$' % 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[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE + + _TESTS = [{ + 'url': 'http://www.twitch.tv/shroomztv', + 'info_dict': { + 'id': '12772022048', + 'display_id': 'shroomztv', + 'ext': 'mp4', + 'title': 're:^ShroomzTV [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$', + 'description': 'H1Z1 - lonewolfing with ShroomzTV | A3 Battle Royale later - @ShroomzTV', + 'is_live': True, + 'timestamp': 1421928037, + 'upload_date': '20150122', + 'uploader': 'ShroomzTV', + 'uploader_id': 'shroomztv', + 'view_count': int, + }, + 'params': { + # 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) + + stream = self._download_json( + '%s/kraken/streams/%s' % (self._API_BASE, channel_id), channel_id, + 'Downloading stream JSON').get('stream') + + # Fallback on profile extraction if stream is offline + if not stream: + return self.url_result( + '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': random.randint(1000000, 10000000), + 'player': 'twitchweb', + 'segment_preference': '4', + '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)), + channel_id, 'mp4') + self._prefer_source(formats) + + view_count = stream.get('viewers') + timestamp = parse_iso8601(stream.get('created_at')) + + channel = stream['channel'] + title = self._live_title(channel.get('display_name') or channel.get('name')) + description = channel.get('status') + + thumbnails = [] + for thumbnail_key, thumbnail_url in stream['preview'].items(): + m = re.search(r'(?P\d+)x(?P\d+)\.jpg$', thumbnail_key) + if not m: + continue + thumbnails.append({ + 'url': thumbnail_url, + 'width': int(m.group('width')), + 'height': int(m.group('height')), + }) + + return { + 'id': compat_str(stream['_id']), + 'display_id': channel_id, + 'title': title, + 'description': description, + 'thumbnails': thumbnails, + 'uploader': channel.get('display_name'), + 'uploader_id': channel.get('name'), + 'timestamp': timestamp, + 'view_count': view_count, + 'formats': formats, + 'is_live': True, + }