X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/87a0165ca7e39af4dacb7ec637063b2cd35ae40b..20205e79eb4216762c923790b27dbe38f945293a:/youtube_dl/extractor/twitch.py diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py index 77414a2..eadc48c 100644 --- a/youtube_dl/extractor/twitch.py +++ b/youtube_dl/extractor/twitch.py @@ -1,38 +1,46 @@ # coding: utf-8 from __future__ import unicode_literals +import collections import itertools -import re +import json import random +import re from .common import InfoExtractor from ..compat import ( - compat_HTTPError, + compat_kwargs, compat_parse_qs, compat_str, + compat_urlparse, compat_urllib_parse_urlencode, compat_urllib_parse_urlparse, - compat_urlparse, ) from ..utils import ( clean_html, ExtractorError, + float_or_none, int_or_none, - js_to_json, - orderedSet, parse_duration, parse_iso8601, - urlencode_postdata, + qualities, + str_or_none, + try_get, + unified_timestamp, + update_url_query, + url_or_none, + urljoin, ) class TwitchBaseIE(InfoExtractor): - _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv' + _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv' _API_BASE = 'https://api.twitch.tv' _USHER_BASE = 'https://usher.ttvnw.net' - _LOGIN_URL = 'http://www.twitch.tv/login' - _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6' + _LOGIN_FORM_URL = 'https://www.twitch.tv/login' + _LOGIN_POST_URL = 'https://passport.twitch.tv/login' + _CLIENT_ID = 'kimne78kx3ncx6brgo4mv6wki5h1ko' _NETRC_MACHINE = 'twitch' def _handle_error(self, response): @@ -44,10 +52,19 @@ class TwitchBaseIE(InfoExtractor): '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')), expected=True) - def _call_api(self, path, item_id, note): + def _call_api(self, path, item_id, *args, **kwargs): + headers = kwargs.get('headers', {}).copy() + headers.update({ + 'Accept': 'application/vnd.twitchtv.v5+json; charset=UTF-8', + 'Client-ID': self._CLIENT_ID, + }) + kwargs.update({ + 'headers': headers, + 'expected_status': (400, 410), + }) response = self._download_json( - '%s/%s' % (self._API_BASE, path), item_id, note, - headers={'Client-ID': self._CLIENT_ID}) + '%s/%s' % (self._API_BASE, path), item_id, + *args, **compat_kwargs(kwargs)) self._handle_error(response) return response @@ -55,7 +72,7 @@ class TwitchBaseIE(InfoExtractor): self._login() def _login(self): - (username, password) = self._get_login_info() + username, password = self._get_login_info() if username is None: return @@ -63,149 +80,97 @@ class TwitchBaseIE(InfoExtractor): raise ExtractorError( 'Unable to login. Twitch said: %s' % message, expected=True) - login_page, handle = self._download_webpage_handle( - self._LOGIN_URL, None, 'Downloading login page') + def login_step(page, urlh, note, data): + form = self._hidden_inputs(page) + form.update(data) - # Some TOR nodes and public proxies are blocked completely - if 'blacklist_message' in login_page: - fail(clean_html(login_page)) + page_url = urlh.geturl() + post_url = self._search_regex( + r']+action=(["\'])(?P.+?)\1', page, + 'post url', default=self._LOGIN_POST_URL, group='url') + post_url = urljoin(page_url, post_url) - login_form = self._hidden_inputs(login_page) + headers = { + 'Referer': page_url, + 'Origin': page_url, + 'Content-Type': 'text/plain;charset=UTF-8', + } - login_form.update({ - 'username': username, - 'password': password, - }) + response = self._download_json( + post_url, None, note, data=json.dumps(form).encode(), + headers=headers, expected_status=400) + error = response.get('error_description') or response.get('error_code') + if error: + fail(error) + + if 'Authenticated successfully' in response.get('message', ''): + return None, None + + redirect_url = urljoin( + post_url, + response.get('redirect') or response['redirect_path']) + return self._download_webpage_handle( + redirect_url, None, 'Downloading login redirect page', + headers=headers) - redirect_url = handle.geturl() + login_page, handle = self._download_webpage_handle( + self._LOGIN_FORM_URL, None, 'Downloading login page') - post_url = self._search_regex( - r']+action=(["\'])(?P.+?)\1', login_page, - 'post url', default=redirect_url, group='url') + # Some TOR nodes and public proxies are blocked completely + if 'blacklist_message' in login_page: + fail(clean_html(login_page)) - if not post_url.startswith('http'): - post_url = compat_urlparse.urljoin(redirect_url, post_url) + redirect_page, handle = login_step( + login_page, handle, 'Logging in', { + 'username': username, + 'password': password, + 'client_id': self._CLIENT_ID, + }) - headers = {'Referer': redirect_url} + # Successful login + if not redirect_page: + return - try: - response = self._download_json( - post_url, None, 'Logging in as %s' % username, - data=urlencode_postdata(login_form), - headers=headers) - except ExtractorError as e: - if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400: - response = self._parse_json( - e.cause.read().decode('utf-8'), None) - fail(response['message']) - raise - - if response.get('redirect'): - self._download_webpage( - response['redirect'], None, 'Downloading login redirect page', - headers=headers) + if re.search(r'(?i)]+id="two-factor-submit"', redirect_page) is not None: + # TODO: Add mechanism to request an SMS or phone call + tfa_token = self._get_tfa_info('two-factor authentication token') + login_step(redirect_page, handle, 'Submitting TFA token', { + 'authy_token': tfa_token, + 'remember_2fa': 'true', + }) def _prefer_source(self, formats): try: source = next(f for f in formats if f['format_id'] == 'Source') - source['preference'] = 10 + source['quality'] = 10 except StopIteration: - pass # No Source stream present + for f in formats: + if '/chunked/' in f['url']: + f.update({ + 'quality': 10, + 'format_note': 'Source', + }) self._sort_formats(formats) + def _download_access_token(self, channel_name): + return self._call_api( + 'api/channels/%s/access_token' % channel_name, channel_name, + 'Downloading access token JSON') -class TwitchItemBaseIE(TwitchBaseIE): - def _download_info(self, item, item_id): - return self._extract_info(self._call_api( - 'kraken/videos/%s%s' % (item, item_id), item_id, - 'Downloading %s info JSON' % self._ITEM_TYPE)) - - def _extract_media(self, item_id): - info = self._download_info(self._ITEM_SHORTCUT, item_id) - response = self._call_api( - 'api/videos/%s%s' % (self._ITEM_SHORTCUT, item_id), item_id, - 'Downloading %s playlist JSON' % self._ITEM_TYPE) - entries = [] - chunks = response['chunks'] - qualities = list(chunks.keys()) - for num, fragment in enumerate(zip(*chunks.values()), start=1): - formats = [] - for fmt_num, fragment_fmt in enumerate(fragment): - format_id = qualities[fmt_num] - fmt = { - 'url': fragment_fmt['url'], - 'format_id': format_id, - 'quality': 1 if format_id == 'live' else 0, - } - m = re.search(r'^(?P\d+)[Pp]', format_id) - if m: - fmt['height'] = int(m.group('height')) - formats.append(fmt) - self._sort_formats(formats) - entry = dict(info) - entry['id'] = '%s_%d' % (entry['id'], num) - entry['title'] = '%s part %d' % (entry['title'], num) - entry['formats'] = formats - entries.append(entry) - return self.playlist_result(entries, info['id'], info['title']) - - def _extract_info(self, info): - return { - 'id': info['_id'], - '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): - return self._extract_media(self._match_id(url)) - - -class TwitchVideoIE(TwitchItemBaseIE): - IE_NAME = 'twitch:video' - _VALID_URL = r'%s/[^/]+/b/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE - _ITEM_TYPE = 'video' - _ITEM_SHORTCUT = 'a' - - _TEST = { - 'url': 'http://www.twitch.tv/riotgames/b/577357806', - 'info_dict': { - 'id': 'a577357806', - 'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG', - }, - 'playlist_mincount': 12, - 'skip': 'HTTP Error 404: Not Found', - } - - -class TwitchChapterIE(TwitchItemBaseIE): - IE_NAME = 'twitch:chapter' - _VALID_URL = r'%s/[^/]+/c/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE - _ITEM_TYPE = 'chapter' - _ITEM_SHORTCUT = 'c' - - _TESTS = [{ - 'url': 'http://www.twitch.tv/acracingleague/c/5285812', - 'info_dict': { - 'id': 'c5285812', - 'title': 'ACRL Off Season - Sports Cars @ Nordschleife', - }, - 'playlist_mincount': 3, - 'skip': 'HTTP Error 404: Not Found', - }, { - 'url': 'http://www.twitch.tv/tsm_theoddone/c/2349361', - 'only_matching': True, - }] + def _extract_channel_id(self, token, channel_name): + return compat_str(self._parse_json(token, channel_name)['channel_id']) -class TwitchVodIE(TwitchItemBaseIE): +class TwitchVodIE(TwitchBaseIE): IE_NAME = 'twitch:vod' - _VALID_URL = r'%s/[^/]+/v/(?P\d+)' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'''(?x) + https?:// + (?: + (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/| + player\.twitch\.tv/\?.*?\bvideo=v? + ) + (?P\d+) + ''' _ITEM_TYPE = 'vod' _ITEM_SHORTCUT = 'v' @@ -215,7 +180,7 @@ class TwitchVodIE(TwitchItemBaseIE): 'id': 'v6528877', 'ext': 'mp4', 'title': 'LCK Summer Split - Week 6 Day 1', - 'thumbnail': 're:^https?://.*\.jpg$', + 'thumbnail': r're:^https?://.*\.jpg$', 'duration': 17208, 'timestamp': 1435131709, 'upload_date': '20150624', @@ -235,7 +200,7 @@ class TwitchVodIE(TwitchItemBaseIE): 'id': 'v11230755', 'ext': 'mp4', 'title': 'Untitled Broadcast', - 'thumbnail': 're:^https?://.*\.jpg$', + 'thumbnail': r're:^https?://.*\.jpg$', 'duration': 1638, 'timestamp': 1439746708, 'upload_date': '20150816', @@ -248,28 +213,87 @@ class TwitchVodIE(TwitchItemBaseIE): 'skip_download': True, }, 'skip': 'HTTP Error 404: Not Found', + }, { + 'url': 'http://player.twitch.tv/?t=5m10s&video=v6528877', + 'only_matching': True, + }, { + 'url': 'https://www.twitch.tv/videos/6528877', + 'only_matching': True, + }, { + 'url': 'https://m.twitch.tv/beagsandjam/v/247478721', + 'only_matching': True, + }, { + 'url': 'https://www.twitch.tv/northernlion/video/291940395', + 'only_matching': True, + }, { + 'url': 'https://player.twitch.tv/?video=480452374', + 'only_matching': True, }] + def _download_info(self, item_id): + return self._extract_info( + self._call_api( + 'kraken/videos/%s' % item_id, item_id, + 'Downloading video info JSON')) + + @staticmethod + def _extract_info(info): + status = info.get('status') + if status == 'recording': + is_live = True + elif status == 'recorded': + is_live = False + else: + is_live = None + _QUALITIES = ('small', 'medium', 'large') + quality_key = qualities(_QUALITIES) + thumbnails = [] + preview = info.get('preview') + if isinstance(preview, dict): + for thumbnail_id, thumbnail_url in preview.items(): + thumbnail_url = url_or_none(thumbnail_url) + if not thumbnail_url: + continue + if thumbnail_id not in _QUALITIES: + continue + thumbnails.append({ + 'url': thumbnail_url, + 'preference': quality_key(thumbnail_id), + }) + return { + 'id': info['_id'], + 'title': info.get('title') or 'Untitled Broadcast', + 'description': info.get('description'), + 'duration': int_or_none(info.get('length')), + 'thumbnails': thumbnails, + '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')), + 'is_live': is_live, + } + def _real_extract(self, url): - item_id = self._match_id(url) + vod_id = self._match_id(url) - info = self._download_info(self._ITEM_SHORTCUT, item_id) + info = self._download_info(vod_id) access_token = self._call_api( - 'api/vods/%s/access_token' % item_id, item_id, + 'api/vods/%s/access_token' % vod_id, vod_id, 'Downloading %s access token' % self._ITEM_TYPE) formats = self._extract_m3u8_formats( - '%s/vod/%s?%s' % ( - self._USHER_BASE, item_id, + '%s/vod/%s.m3u8?%s' % ( + self._USHER_BASE, vod_id, compat_urllib_parse_urlencode({ 'allow_source': 'true', 'allow_audio_only': 'true', 'allow_spectre': 'true', 'player': 'twitchweb', + 'playlist_include_framerate': 'true', 'nauth': access_token['token'], 'nauthsig': access_token['sig'], })), - item_id, 'mp4', entry_protocol='m3u8_native') + vod_id, 'mp4', entry_protocol='m3u8_native') self._prefer_source(formats) info['formats'] = formats @@ -279,96 +303,430 @@ class TwitchVodIE(TwitchItemBaseIE): if 't' in query: info['start_time'] = parse_duration(query['t'][0]) + if info.get('timestamp') is not None: + info['subtitles'] = { + 'rechat': [{ + 'url': update_url_query( + 'https://api.twitch.tv/v5/videos/%s/comments' % vod_id, { + 'client_id': self._CLIENT_ID, + }), + 'ext': 'json', + }], + } + return info -class TwitchPlaylistBaseIE(TwitchBaseIE): - _PLAYLIST_PATH = 'kraken/channels/%s/videos/?offset=%d&limit=%d' +def _make_video_result(node): + assert isinstance(node, dict) + video_id = node.get('id') + if not video_id: + return + return { + '_type': 'url_transparent', + 'ie_key': TwitchVodIE.ie_key(), + 'id': video_id, + 'url': 'https://www.twitch.tv/videos/%s' % video_id, + 'title': node.get('title'), + 'thumbnail': node.get('previewThumbnailURL'), + 'duration': float_or_none(node.get('lengthSeconds')), + 'view_count': int_or_none(node.get('viewCount')), + } + + +class TwitchGraphQLBaseIE(TwitchBaseIE): _PAGE_LIMIT = 100 - def _extract_playlist(self, channel_id): - info = self._call_api( - 'kraken/channels/%s' % channel_id, - channel_id, 'Downloading channel info JSON') - channel_name = info.get('display_name') or info.get('name') + def _download_gql(self, video_id, op, variables, sha256_hash, note, fatal=True): + return self._download_json( + 'https://gql.twitch.tv/gql', video_id, note, + data=json.dumps({ + 'operationName': op, + 'variables': variables, + 'extensions': { + 'persistedQuery': { + 'version': 1, + 'sha256Hash': sha256_hash, + } + } + }).encode(), + headers={ + 'Content-Type': 'text/plain;charset=UTF-8', + 'Client-ID': self._CLIENT_ID, + }, fatal=fatal) + + +class TwitchCollectionIE(TwitchGraphQLBaseIE): + _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/collections/(?P[^/]+)' + + _TESTS = [{ + 'url': 'https://www.twitch.tv/collections/wlDCoH0zEBZZbQ', + 'info_dict': { + 'id': 'wlDCoH0zEBZZbQ', + 'title': 'Overthrow Nook, capitalism for children', + }, + 'playlist_mincount': 13, + }] + + _OPERATION_NAME = 'CollectionSideBar' + _SHA256_HASH = '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14' + + def _real_extract(self, url): + collection_id = self._match_id(url) + collection = self._download_gql( + collection_id, self._OPERATION_NAME, + {'collectionID': collection_id}, self._SHA256_HASH, + 'Downloading collection GraphQL')['data']['collection'] + title = collection.get('title') entries = [] + for edge in collection['items']['edges']: + if not isinstance(edge, dict): + continue + node = edge.get('node') + if not isinstance(node, dict): + continue + video = _make_video_result(node) + if video: + entries.append(video) + return self.playlist_result( + entries, playlist_id=collection_id, playlist_title=title) + + +class TwitchPlaylistBaseIE(TwitchGraphQLBaseIE): + def _entries(self, channel_name, *args): + cursor = None + variables_common = self._make_variables(channel_name, *args) + entries_key = '%ss' % self._ENTRY_KIND + for page_num in itertools.count(1): + variables = variables_common.copy() + variables['limit'] = self._PAGE_LIMIT + if cursor: + variables['cursor'] = cursor + page = self._download_gql( + channel_name, self._OPERATION_NAME, variables, + self._SHA256_HASH, + 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num), + fatal=False) + if not page: + break + edges = try_get( + page, lambda x: x['data']['user'][entries_key]['edges'], list) + if not edges: + break + for edge in edges: + if not isinstance(edge, dict): + continue + if edge.get('__typename') != self._EDGE_KIND: + continue + node = edge.get('node') + if not isinstance(node, dict): + continue + if node.get('__typename') != self._NODE_KIND: + continue + entry = self._extract_entry(node) + if entry: + cursor = edge.get('cursor') + yield entry + if not cursor or not isinstance(cursor, compat_str): + break + + # Deprecated kraken v5 API + def _entries_kraken(self, channel_name, broadcast_type, sort): + access_token = self._download_access_token(channel_name) + channel_id = self._extract_channel_id(access_token['token'], channel_name) offset = 0 - limit = self._PAGE_LIMIT - broken_paging_detected = False counter_override = None for counter in itertools.count(1): response = self._call_api( - self._PLAYLIST_PATH % (channel_id, offset, limit), + 'kraken/channels/%s/videos/' % channel_id, channel_id, - 'Downloading %s videos JSON page %s' - % (self._PLAYLIST_TYPE, counter_override or counter)) - page_entries = self._extract_playlist_page(response) - if not page_entries: + 'Downloading video JSON page %s' % (counter_override or counter), + query={ + 'offset': offset, + 'limit': self._PAGE_LIMIT, + 'broadcast_type': broadcast_type, + 'sort': sort, + }) + videos = response.get('videos') + if not isinstance(videos, list): break + for video in videos: + if not isinstance(video, dict): + continue + video_url = url_or_none(video.get('url')) + if not video_url: + continue + yield { + '_type': 'url_transparent', + 'ie_key': TwitchVodIE.ie_key(), + 'id': video.get('_id'), + 'url': video_url, + 'title': video.get('title'), + 'description': video.get('description'), + 'timestamp': unified_timestamp(video.get('published_at')), + 'duration': float_or_none(video.get('length')), + 'view_count': int_or_none(video.get('views')), + 'language': video.get('language'), + } + offset += self._PAGE_LIMIT total = int_or_none(response.get('_total')) - # Since the beginning of March 2016 twitch's paging mechanism - # is completely broken on the twitch side. It simply ignores - # a limit and returns the whole offset number of videos. - # Working around by just requesting all videos at once. - # Upd: pagination bug was fixed by twitch on 15.03.2016. - if not broken_paging_detected and total and len(page_entries) > limit: - self.report_warning( - 'Twitch pagination is broken on twitch side, requesting all videos at once', - channel_id) - broken_paging_detected = True - offset = total - counter_override = '(all at once)' - continue - entries.extend(page_entries) - if broken_paging_detected or total and len(page_entries) >= total: + if total and offset >= total: break - offset += limit - return self.playlist_result( - [self.url_result(entry) for entry in orderedSet(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 [] + +class TwitchVideosIE(TwitchPlaylistBaseIE): + _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P[^/]+)/(?:videos|profile)' + + _TESTS = [{ + # All Videos sorted by Date + 'url': 'https://www.twitch.tv/spamfish/videos?filter=all', + 'info_dict': { + 'id': 'spamfish', + 'title': 'spamfish - All Videos sorted by Date', + }, + 'playlist_mincount': 924, + }, { + # All Videos sorted by Popular + 'url': 'https://www.twitch.tv/spamfish/videos?filter=all&sort=views', + 'info_dict': { + 'id': 'spamfish', + 'title': 'spamfish - All Videos sorted by Popular', + }, + 'playlist_mincount': 931, + }, { + # Past Broadcasts sorted by Date + 'url': 'https://www.twitch.tv/spamfish/videos?filter=archives', + 'info_dict': { + 'id': 'spamfish', + 'title': 'spamfish - Past Broadcasts sorted by Date', + }, + 'playlist_mincount': 27, + }, { + # Highlights sorted by Date + 'url': 'https://www.twitch.tv/spamfish/videos?filter=highlights', + 'info_dict': { + 'id': 'spamfish', + 'title': 'spamfish - Highlights sorted by Date', + }, + 'playlist_mincount': 901, + }, { + # Uploads sorted by Date + 'url': 'https://www.twitch.tv/esl_csgo/videos?filter=uploads&sort=time', + 'info_dict': { + 'id': 'esl_csgo', + 'title': 'esl_csgo - Uploads sorted by Date', + }, + 'playlist_mincount': 5, + }, { + # Past Premieres sorted by Date + 'url': 'https://www.twitch.tv/spamfish/videos?filter=past_premieres', + 'info_dict': { + 'id': 'spamfish', + 'title': 'spamfish - Past Premieres sorted by Date', + }, + 'playlist_mincount': 1, + }, { + 'url': 'https://www.twitch.tv/spamfish/videos/all', + 'only_matching': True, + }, { + 'url': 'https://m.twitch.tv/spamfish/videos/all', + 'only_matching': True, + }, { + 'url': 'https://www.twitch.tv/spamfish/videos', + 'only_matching': True, + }] + + Broadcast = collections.namedtuple('Broadcast', ['type', 'label']) + + _DEFAULT_BROADCAST = Broadcast(None, 'All Videos') + _BROADCASTS = { + 'archives': Broadcast('ARCHIVE', 'Past Broadcasts'), + 'highlights': Broadcast('HIGHLIGHT', 'Highlights'), + 'uploads': Broadcast('UPLOAD', 'Uploads'), + 'past_premieres': Broadcast('PAST_PREMIERE', 'Past Premieres'), + 'all': _DEFAULT_BROADCAST, + } + + _DEFAULT_SORTED_BY = 'Date' + _SORTED_BY = { + 'time': _DEFAULT_SORTED_BY, + 'views': 'Popular', + } + + _SHA256_HASH = 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb' + _OPERATION_NAME = 'FilterableVideoTower_Videos' + _ENTRY_KIND = 'video' + _EDGE_KIND = 'VideoEdge' + _NODE_KIND = 'Video' + + @classmethod + def suitable(cls, url): + return (False + if any(ie.suitable(url) for ie in ( + TwitchVideosClipsIE, + TwitchVideosCollectionsIE)) + else super(TwitchVideosIE, cls).suitable(url)) + + @staticmethod + def _make_variables(channel_name, broadcast_type, sort): + return { + 'channelOwnerLogin': channel_name, + 'broadcastType': broadcast_type, + 'videoSort': sort.upper(), + } + + @staticmethod + def _extract_entry(node): + return _make_video_result(node) def _real_extract(self, url): - return self._extract_playlist(self._match_id(url)) + channel_name = self._match_id(url) + qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) + filter = qs.get('filter', ['all'])[0] + sort = qs.get('sort', ['time'])[0] + broadcast = self._BROADCASTS.get(filter, self._DEFAULT_BROADCAST) + return self.playlist_result( + self._entries(channel_name, broadcast.type, sort), + playlist_id=channel_name, + playlist_title='%s - %s sorted by %s' + % (channel_name, broadcast.label, + self._SORTED_BY.get(sort, self._DEFAULT_SORTED_BY))) -class TwitchProfileIE(TwitchPlaylistBaseIE): - IE_NAME = 'twitch:profile' - _VALID_URL = r'%s/(?P[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE - _PLAYLIST_TYPE = 'profile' +class TwitchVideosClipsIE(TwitchPlaylistBaseIE): + _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P[^/]+)/(?:clips|videos/*?\?.*?\bfilter=clips)' - _TEST = { - 'url': 'http://www.twitch.tv/vanillatv/profile', + _TESTS = [{ + # Clips + 'url': 'https://www.twitch.tv/vanillatv/clips?filter=clips&range=all', 'info_dict': { 'id': 'vanillatv', - 'title': 'VanillaTV', + 'title': 'vanillatv - Clips Top All', }, - 'playlist_mincount': 412, + 'playlist_mincount': 1, + }, { + 'url': 'https://www.twitch.tv/dota2ruhub/videos?filter=clips&range=7d', + 'only_matching': True, + }] + + Clip = collections.namedtuple('Clip', ['filter', 'label']) + + _DEFAULT_CLIP = Clip('LAST_WEEK', 'Top 7D') + _RANGE = { + '24hr': Clip('LAST_DAY', 'Top 24H'), + '7d': _DEFAULT_CLIP, + '30d': Clip('LAST_MONTH', 'Top 30D'), + 'all': Clip('ALL_TIME', 'Top All'), } + # NB: values other than 20 result in skipped videos + _PAGE_LIMIT = 20 + + _SHA256_HASH = 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777' + _OPERATION_NAME = 'ClipsCards__User' + _ENTRY_KIND = 'clip' + _EDGE_KIND = 'ClipEdge' + _NODE_KIND = 'Clip' + + @staticmethod + def _make_variables(channel_name, filter): + return { + 'login': channel_name, + 'criteria': { + 'filter': filter, + }, + } + + @staticmethod + def _extract_entry(node): + assert isinstance(node, dict) + clip_url = url_or_none(node.get('url')) + if not clip_url: + return + return { + '_type': 'url_transparent', + 'ie_key': TwitchClipsIE.ie_key(), + 'id': node.get('id'), + 'url': clip_url, + 'title': node.get('title'), + 'thumbnail': node.get('thumbnailURL'), + 'duration': float_or_none(node.get('durationSeconds')), + 'timestamp': unified_timestamp(node.get('createdAt')), + 'view_count': int_or_none(node.get('viewCount')), + 'language': node.get('language'), + } + + def _real_extract(self, url): + channel_name = self._match_id(url) + qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query) + range = qs.get('range', ['7d'])[0] + clip = self._RANGE.get(range, self._DEFAULT_CLIP) + return self.playlist_result( + self._entries(channel_name, clip.filter), + playlist_id=channel_name, + playlist_title='%s - Clips %s' % (channel_name, clip.label)) -class TwitchPastBroadcastsIE(TwitchPlaylistBaseIE): - IE_NAME = 'twitch:past_broadcasts' - _VALID_URL = r'%s/(?P[^/]+)/profile/past_broadcasts/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE - _PLAYLIST_PATH = TwitchPlaylistBaseIE._PLAYLIST_PATH + '&broadcasts=true' - _PLAYLIST_TYPE = 'past broadcasts' - _TEST = { - 'url': 'http://www.twitch.tv/spamfish/profile/past_broadcasts', +class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE): + _VALID_URL = r'https?://(?:(?:www|go|m)\.)?twitch\.tv/(?P[^/]+)/videos/*?\?.*?\bfilter=collections' + + _TESTS = [{ + # Collections + 'url': 'https://www.twitch.tv/spamfish/videos?filter=collections', 'info_dict': { 'id': 'spamfish', - 'title': 'Spamfish', + 'title': 'spamfish - Collections', }, - 'playlist_mincount': 54, - } + 'playlist_mincount': 3, + }] + + _SHA256_HASH = '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84' + _OPERATION_NAME = 'ChannelCollectionsContent' + _ENTRY_KIND = 'collection' + _EDGE_KIND = 'CollectionsItemEdge' + _NODE_KIND = 'Collection' + + @staticmethod + def _make_variables(channel_name): + return { + 'ownerLogin': channel_name, + } + + @staticmethod + def _extract_entry(node): + assert isinstance(node, dict) + collection_id = node.get('id') + if not collection_id: + return + return { + '_type': 'url_transparent', + 'ie_key': TwitchCollectionIE.ie_key(), + 'id': collection_id, + 'url': 'https://www.twitch.tv/collections/%s' % collection_id, + 'title': node.get('title'), + 'thumbnail': node.get('thumbnailURL'), + 'duration': float_or_none(node.get('lengthSeconds')), + 'timestamp': unified_timestamp(node.get('updatedAt')), + 'view_count': int_or_none(node.get('viewCount')), + } + + def _real_extract(self, url): + channel_name = self._match_id(url) + return self.playlist_result( + self._entries(channel_name), playlist_id=channel_name, + playlist_title='%s - Collections' % channel_name) class TwitchStreamIE(TwitchBaseIE): IE_NAME = 'twitch:stream' - _VALID_URL = r'%s/(?P[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE + _VALID_URL = r'''(?x) + https?:// + (?: + (?:(?:www|go|m)\.)?twitch\.tv/| + player\.twitch\.tv/\?.*?\bchannel= + ) + (?P[^/#?]+) + ''' _TESTS = [{ 'url': 'http://www.twitch.tv/shroomztv', @@ -392,14 +750,40 @@ class TwitchStreamIE(TwitchBaseIE): }, { 'url': 'http://www.twitch.tv/miracle_doto#profile-0', 'only_matching': True, + }, { + 'url': 'https://player.twitch.tv/?channel=lotsofs', + 'only_matching': True, + }, { + 'url': 'https://go.twitch.tv/food', + 'only_matching': True, + }, { + 'url': 'https://m.twitch.tv/food', + 'only_matching': True, }] + @classmethod + def suitable(cls, url): + return (False + if any(ie.suitable(url) for ie in ( + TwitchVodIE, + TwitchCollectionIE, + TwitchVideosIE, + TwitchVideosClipsIE, + TwitchVideosCollectionsIE, + TwitchClipsIE)) + else super(TwitchStreamIE, cls).suitable(url)) + def _real_extract(self, url): - channel_id = self._match_id(url) + channel_name = self._match_id(url) + + access_token = self._download_access_token(channel_name) + + token = access_token['token'] + channel_id = self._extract_channel_id(token, channel_name) stream = self._call_api( - 'kraken/streams/%s?stream_type=all' % channel_id, channel_id, - 'Downloading stream JSON').get('stream') + 'kraken/streams/%s?stream_type=all' % channel_id, + channel_id, 'Downloading stream JSON').get('stream') if not stream: raise ExtractorError('%s is offline' % channel_id, expected=True) @@ -408,11 +792,9 @@ class TwitchStreamIE(TwitchBaseIE): # (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._call_api( - 'api/channels/%s/access_token' % channel_id, channel_id, - 'Downloading channel access token') + channel_name = try_get( + stream, lambda x: x['channel']['name'], + compat_str) or channel_name.lower() query = { 'allow_source': 'true', @@ -420,13 +802,14 @@ class TwitchStreamIE(TwitchBaseIE): 'allow_spectre': 'true', 'p': random.randint(1000000, 10000000), 'player': 'twitchweb', + 'playlist_include_framerate': 'true', 'segment_preference': '4', 'sig': access_token['sig'].encode('utf-8'), - 'token': access_token['token'].encode('utf-8'), + '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)), + % (self._USHER_BASE, channel_name, compat_urllib_parse_urlencode(query)), channel_id, 'mp4') self._prefer_source(formats) @@ -449,8 +832,8 @@ class TwitchStreamIE(TwitchBaseIE): }) return { - 'id': compat_str(stream['_id']), - 'display_id': channel_id, + 'id': str_or_none(stream.get('_id')) or channel_id, + 'display_id': channel_name, 'title': title, 'description': description, 'thumbnails': thumbnails, @@ -463,59 +846,127 @@ class TwitchStreamIE(TwitchBaseIE): } -class TwitchClipsIE(InfoExtractor): +class TwitchClipsIE(TwitchBaseIE): IE_NAME = 'twitch:clips' - _VALID_URL = r'https?://clips\.twitch\.tv/(?:[^/]+/)*(?P[^/?#&]+)' + _VALID_URL = r'''(?x) + https?:// + (?: + clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)| + (?:(?:www|go|m)\.)?twitch\.tv/[^/]+/clip/ + ) + (?P[^/?#&]+) + ''' _TESTS = [{ - 'url': 'https://clips.twitch.tv/ea/AggressiveCobraPoooound', + 'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat', 'md5': '761769e1eafce0ffebfb4089cb3847cd', 'info_dict': { - 'id': 'AggressiveCobraPoooound', + 'id': '42850523', 'ext': 'mp4', 'title': 'EA Play 2016 Live from the Novo Theatre', - 'thumbnail': 're:^https?://.*\.jpg', + 'thumbnail': r're:^https?://.*\.jpg', + 'timestamp': 1465767393, + 'upload_date': '20160612', 'creator': 'EA', 'uploader': 'stereotype_', - 'uploader_id': 'stereotype_', + 'uploader_id': '43566419', }, }, { # multiple formats 'url': 'https://clips.twitch.tv/rflegendary/UninterestedBeeDAESuppy', 'only_matching': True, + }, { + 'url': 'https://www.twitch.tv/sergeynixon/clip/StormyThankfulSproutFutureMan', + 'only_matching': True, + }, { + 'url': 'https://clips.twitch.tv/embed?clip=InquisitiveBreakableYogurtJebaited', + 'only_matching': True, + }, { + 'url': 'https://m.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank', + 'only_matching': True, + }, { + 'url': 'https://go.twitch.tv/rossbroadcast/clip/ConfidentBraveHumanChefFrank', + 'only_matching': True, }] def _real_extract(self, url): video_id = self._match_id(url) - webpage = self._download_webpage(url, video_id) - - clip = self._parse_json( - self._search_regex( - r'(?s)clipInfo\s*=\s*({.+?});', webpage, 'clip info'), - video_id, transform_source=js_to_json) - - title = clip.get('channel_title') or self._og_search_title(webpage) - - formats = [{ - 'url': option['source'], - 'format_id': option.get('quality'), - 'height': int_or_none(option.get('quality')), - } for option in clip.get('quality_options', []) if option.get('source')] - - if not formats: - formats = [{ - 'url': clip['clip_video_url'], - }] + clip = self._download_json( + 'https://gql.twitch.tv/gql', video_id, data=json.dumps({ + 'query': '''{ + clip(slug: "%s") { + broadcaster { + displayName + } + createdAt + curator { + displayName + id + } + durationSeconds + id + tiny: thumbnailURL(width: 86, height: 45) + small: thumbnailURL(width: 260, height: 147) + medium: thumbnailURL(width: 480, height: 272) + title + videoQualities { + frameRate + quality + sourceURL + } + viewCount + } +}''' % video_id, + }).encode(), headers={ + 'Client-ID': self._CLIENT_ID, + })['data']['clip'] + + if not clip: + raise ExtractorError( + 'This clip is no longer available', expected=True) + formats = [] + for option in clip.get('videoQualities', []): + if not isinstance(option, dict): + continue + source = url_or_none(option.get('sourceURL')) + if not source: + continue + formats.append({ + 'url': source, + 'format_id': option.get('quality'), + 'height': int_or_none(option.get('quality')), + 'fps': int_or_none(option.get('frameRate')), + }) self._sort_formats(formats) + thumbnails = [] + for thumbnail_id in ('tiny', 'small', 'medium'): + thumbnail_url = clip.get(thumbnail_id) + if not thumbnail_url: + continue + thumb = { + 'id': thumbnail_id, + 'url': thumbnail_url, + } + mobj = re.search(r'-(\d+)x(\d+)\.', thumbnail_url) + if mobj: + thumb.update({ + 'height': int(mobj.group(2)), + 'width': int(mobj.group(1)), + }) + thumbnails.append(thumb) + return { - 'id': video_id, - 'title': title, - 'thumbnail': self._og_search_thumbnail(webpage), - 'creator': clip.get('broadcaster_display_name') or clip.get('broadcaster_login'), - 'uploader': clip.get('curator_login'), - 'uploader_id': clip.get('curator_display_name'), + 'id': clip.get('id') or video_id, + 'title': clip.get('title') or video_id, 'formats': formats, + 'duration': int_or_none(clip.get('durationSeconds')), + 'views': int_or_none(clip.get('viewCount')), + 'timestamp': unified_timestamp(clip.get('createdAt')), + 'thumbnails': thumbnails, + 'creator': try_get(clip, lambda x: x['broadcaster']['displayName'], compat_str), + 'uploader': try_get(clip, lambda x: x['curator']['displayName'], compat_str), + 'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str), }