X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/929d540d5979815dd13aa119b6fdb41454907d0c..2dbb05e5969fc6d6b5d51b6bb9576b03fb63d936:/youtube_dl/extractor/twitch.py diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py index dc5ff29..e211cd4 100644 --- a/youtube_dl/extractor/twitch.py +++ b/youtube_dl/extractor/twitch.py @@ -17,12 +17,12 @@ from ..compat import ( from ..utils import ( clean_html, ExtractorError, - float_or_none, int_or_none, orderedSet, parse_duration, parse_iso8601, qualities, + str_or_none, try_get, unified_timestamp, update_url_query, @@ -52,8 +52,14 @@ class TwitchBaseIE(InfoExtractor): def _call_api(self, path, item_id, *args, **kwargs): headers = kwargs.get('headers', {}).copy() - headers['Client-ID'] = self._CLIENT_ID - kwargs['headers'] = headers + 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, *args, **compat_kwargs(kwargs)) @@ -188,12 +194,27 @@ class TwitchItemBaseIE(TwitchBaseIE): 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')), - 'thumbnail': info.get('preview'), + 'thumbnails': thumbnails, 'uploader': info.get('channel', {}).get('display_name'), 'uploader_id': info.get('channel', {}).get('name'), 'timestamp': parse_iso8601(info.get('recorded_at')), @@ -248,7 +269,7 @@ class TwitchVodIE(TwitchItemBaseIE): https?:// (?: (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/| - player\.twitch\.tv/\?.*?\bvideo=v + player\.twitch\.tv/\?.*?\bvideo=v? ) (?P\d+) ''' @@ -306,6 +327,9 @@ class TwitchVodIE(TwitchItemBaseIE): }, { 'url': 'https://www.twitch.tv/northernlion/video/291940395', 'only_matching': True, + }, { + 'url': 'https://player.twitch.tv/?video=480452374', + 'only_matching': True, }] def _real_extract(self, url): @@ -317,13 +341,14 @@ class TwitchVodIE(TwitchItemBaseIE): 'Downloading %s access token' % self._ITEM_TYPE) formats = self._extract_m3u8_formats( - '%s/vod/%s?%s' % ( + '%s/vod/%s.m3u8?%s' % ( self._USHER_BASE, item_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'], })), @@ -341,9 +366,8 @@ class TwitchVodIE(TwitchItemBaseIE): info['subtitles'] = { 'rechat': [{ 'url': update_url_query( - 'https://rechat.twitch.tv/rechat-messages', { - 'video_id': 'v%s' % item_id, - 'start': info['timestamp'], + 'https://api.twitch.tv/v5/videos/%s/comments' % item_id, { + 'client_id': self._CLIENT_ID, }), 'ext': 'json', }], @@ -571,11 +595,19 @@ class TwitchStreamIE(TwitchBaseIE): 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._call_api( + 'api/channels/%s/access_token' % channel_name, channel_name, + 'Downloading access token JSON') + + token = access_token['token'] + channel_id = compat_str(self._parse_json( + token, channel_name)['channel_id']) 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) @@ -584,11 +616,9 @@ 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', @@ -596,13 +626,14 @@ 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) @@ -625,8 +656,8 @@ 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, @@ -641,7 +672,14 @@ class TwitchStreamIE(TwitchBaseIE): class TwitchClipsIE(TwitchBaseIE): IE_NAME = 'twitch:clips' - _VALID_URL = r'https?://(?:clips\.twitch\.tv/(?:[^/]+/)*|(?:www\.)?twitch\.tv/[^/]+/clip/)(?P[^/?#&]+)' + _VALID_URL = r'''(?x) + https?:// + (?: + clips\.twitch\.tv/(?:embed\?.*?\bclip=|(?:[^/]+/)*)| + (?:(?:www|go|m)\.)?twitch\.tv/[^/]+/clip/ + ) + (?P[^/?#&]+) + ''' _TESTS = [{ 'url': 'https://clips.twitch.tv/FaintLightGullWholeWheat', @@ -664,68 +702,95 @@ class TwitchClipsIE(TwitchBaseIE): }, { '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) - status = self._download_json( - 'https://clips.twitch.tv/api/v2/clips/%s/status' % video_id, - video_id) + 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 status['quality_options']: + for option in clip.get('videoQualities', []): if not isinstance(option, dict): continue - source = url_or_none(option.get('source')) + 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('frame_rate')), + 'fps': int_or_none(option.get('frameRate')), }) - self._sort_formats(formats) - info = { + 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': 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), } - - clip = self._call_api( - 'kraken/clips/%s' % video_id, video_id, fatal=False, headers={ - 'Accept': 'application/vnd.twitchtv.v5+json', - }) - - if clip: - quality_key = qualities(('tiny', 'small', 'medium')) - thumbnails = [] - thumbnails_dict = clip.get('thumbnails') - if isinstance(thumbnails_dict, dict): - for thumbnail_id, thumbnail_url in thumbnails_dict.items(): - thumbnails.append({ - 'id': thumbnail_id, - 'url': thumbnail_url, - 'preference': quality_key(thumbnail_id), - }) - - info.update({ - 'id': clip.get('tracking_id') or video_id, - 'title': clip.get('title') or video_id, - 'duration': float_or_none(clip.get('duration')), - 'views': int_or_none(clip.get('views')), - 'timestamp': unified_timestamp(clip.get('created_at')), - 'thumbnails': thumbnails, - 'creator': try_get(clip, lambda x: x['broadcaster']['display_name'], compat_str), - 'uploader': try_get(clip, lambda x: x['curator']['display_name'], compat_str), - 'uploader_id': try_get(clip, lambda x: x['curator']['id'], compat_str), - }) - else: - info.update({ - 'title': video_id, - 'id': video_id, - }) - - return info