X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/32ad8acade51269177cd01c5aa3c7169557eeac8..39393b81acfaf4045fb7f20454a0226f0dc9142e:/youtube_dl/extractor/twitch.py diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py index fefcd28..e211cd4 100644 --- a/youtube_dl/extractor/twitch.py +++ b/youtube_dl/extractor/twitch.py @@ -4,10 +4,11 @@ from __future__ import unicode_literals import itertools import re import random +import json from .common import InfoExtractor from ..compat import ( - compat_HTTPError, + compat_kwargs, compat_parse_qs, compat_str, compat_urllib_parse_urlencode, @@ -17,23 +18,27 @@ from ..utils import ( clean_html, ExtractorError, int_or_none, - js_to_json, orderedSet, parse_duration, parse_iso8601, + qualities, + str_or_none, + try_get, + unified_timestamp, update_url_query, - urlencode_postdata, + url_or_none, urljoin, ) class TwitchBaseIE(InfoExtractor): - _VALID_URL_BASE = r'https?://(?:(?:www|go)\.)?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 = 'https://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): @@ -45,10 +50,19 @@ 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 @@ -56,7 +70,7 @@ 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 @@ -71,41 +85,50 @@ class TwitchBaseIE(InfoExtractor): page_url = urlh.geturl() post_url = self._search_regex( r']+action=(["\'])(?P.+?)\1', page, - 'post url', default=page_url, group='url') + 'post url', default=self._LOGIN_POST_URL, group='url') post_url = urljoin(page_url, post_url) - headers = {'Referer': page_url} - - try: - response = self._download_json( - post_url, None, note, - data=urlencode_postdata(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 - - redirect_url = urljoin(post_url, response['redirect']) + headers = { + 'Referer': page_url, + 'Origin': page_url, + 'Content-Type': 'text/plain;charset=UTF-8', + } + + 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) login_page, handle = self._download_webpage_handle( - self._LOGIN_URL, None, 'Downloading login page') + self._LOGIN_FORM_URL, None, 'Downloading login page') # Some TOR nodes and public proxies are blocked completely if 'blacklist_message' in login_page: fail(clean_html(login_page)) redirect_page, handle = login_step( - login_page, handle, 'Logging in as %s' % username, { + login_page, handle, 'Logging in', { 'username': username, 'password': password, + 'client_id': self._CLIENT_ID, }) + # Successful login + if not redirect_page: + return + 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') @@ -117,9 +140,14 @@ class TwitchBaseIE(InfoExtractor): 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) @@ -159,16 +187,39 @@ class TwitchItemBaseIE(TwitchBaseIE): return self.playlist_result(entries, info['id'], info['title']) def _extract_info(self, 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')), - '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')), 'view_count': int_or_none(info.get('views')), + 'is_live': is_live, } def _real_extract(self, url): @@ -217,8 +268,8 @@ class TwitchVodIE(TwitchItemBaseIE): _VALID_URL = r'''(?x) https?:// (?: - (?:(?:www|go)\.)?twitch\.tv/(?:[^/]+/v|videos)/| - player\.twitch\.tv/\?.*?\bvideo=v + (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/| + player\.twitch\.tv/\?.*?\bvideo=v? ) (?P\d+) ''' @@ -270,6 +321,15 @@ class TwitchVodIE(TwitchItemBaseIE): }, { '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 _real_extract(self, url): @@ -281,13 +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'], })), @@ -305,9 +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', }], @@ -358,9 +418,16 @@ class TwitchPlaylistBaseIE(TwitchBaseIE): break offset += limit return self.playlist_result( - [self.url_result(entry) for entry in orderedSet(entries)], + [self._make_url_result(entry) for entry in orderedSet(entries)], channel_id, channel_name) + def _make_url_result(self, url): + try: + video_id = 'v%s' % TwitchVodIE._match_id(url) + return self.url_result(url, TwitchVodIE.ie_key(), video_id=video_id) + except AssertionError: + return self.url_result(url) + def _extract_playlist_page(self, response): videos = response.get('videos') return [video['url'] for video in videos] if videos else [] @@ -374,14 +441,17 @@ class TwitchProfileIE(TwitchPlaylistBaseIE): _VALID_URL = r'%s/(?P[^/]+)/profile/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE _PLAYLIST_TYPE = 'profile' - _TEST = { + _TESTS = [{ 'url': 'http://www.twitch.tv/vanillatv/profile', 'info_dict': { 'id': 'vanillatv', 'title': 'VanillaTV', }, 'playlist_mincount': 412, - } + }, { + 'url': 'http://m.twitch.tv/vanillatv/profile', + 'only_matching': True, + }] class TwitchVideosBaseIE(TwitchPlaylistBaseIE): @@ -395,14 +465,17 @@ class TwitchAllVideosIE(TwitchVideosBaseIE): _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive,upload,highlight' _PLAYLIST_TYPE = 'all videos' - _TEST = { + _TESTS = [{ 'url': 'https://www.twitch.tv/spamfish/videos/all', 'info_dict': { 'id': 'spamfish', 'title': 'Spamfish', }, 'playlist_mincount': 869, - } + }, { + 'url': 'https://m.twitch.tv/spamfish/videos/all', + 'only_matching': True, + }] class TwitchUploadsIE(TwitchVideosBaseIE): @@ -411,14 +484,17 @@ class TwitchUploadsIE(TwitchVideosBaseIE): _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'upload' _PLAYLIST_TYPE = 'uploads' - _TEST = { + _TESTS = [{ 'url': 'https://www.twitch.tv/spamfish/videos/uploads', 'info_dict': { 'id': 'spamfish', 'title': 'Spamfish', }, 'playlist_mincount': 0, - } + }, { + 'url': 'https://m.twitch.tv/spamfish/videos/uploads', + 'only_matching': True, + }] class TwitchPastBroadcastsIE(TwitchVideosBaseIE): @@ -427,14 +503,17 @@ class TwitchPastBroadcastsIE(TwitchVideosBaseIE): _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'archive' _PLAYLIST_TYPE = 'past broadcasts' - _TEST = { + _TESTS = [{ 'url': 'https://www.twitch.tv/spamfish/videos/past-broadcasts', 'info_dict': { 'id': 'spamfish', 'title': 'Spamfish', }, 'playlist_mincount': 0, - } + }, { + 'url': 'https://m.twitch.tv/spamfish/videos/past-broadcasts', + 'only_matching': True, + }] class TwitchHighlightsIE(TwitchVideosBaseIE): @@ -443,14 +522,17 @@ class TwitchHighlightsIE(TwitchVideosBaseIE): _PLAYLIST_PATH = TwitchVideosBaseIE._PLAYLIST_PATH + 'highlight' _PLAYLIST_TYPE = 'highlights' - _TEST = { + _TESTS = [{ 'url': 'https://www.twitch.tv/spamfish/videos/highlights', 'info_dict': { 'id': 'spamfish', 'title': 'Spamfish', }, 'playlist_mincount': 805, - } + }, { + 'url': 'https://m.twitch.tv/spamfish/videos/highlights', + 'only_matching': True, + }] class TwitchStreamIE(TwitchBaseIE): @@ -458,7 +540,7 @@ class TwitchStreamIE(TwitchBaseIE): _VALID_URL = r'''(?x) https?:// (?: - (?:(?:www|go)\.)?twitch\.tv/| + (?:(?:www|go|m)\.)?twitch\.tv/| player\.twitch\.tv/\?.*?\bchannel= ) (?P[^/#?]+) @@ -492,6 +574,9 @@ class TwitchStreamIE(TwitchBaseIE): }, { 'url': 'https://go.twitch.tv/food', 'only_matching': True, + }, { + 'url': 'https://m.twitch.tv/food', + 'only_matching': True, }] @classmethod @@ -505,15 +590,24 @@ class TwitchStreamIE(TwitchBaseIE): TwitchAllVideosIE, TwitchUploadsIE, TwitchPastBroadcastsIE, - TwitchHighlightsIE)) + TwitchHighlightsIE, + 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._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) @@ -522,11 +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', @@ -534,13 +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) @@ -563,8 +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, @@ -577,59 +670,127 @@ 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': 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('title') or 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), }