diff --git a/yt_dlp/extractor/tumblr.py b/yt_dlp/extractor/tumblr.py index a26bdcaae..6ad405cf1 100644 --- a/yt_dlp/extractor/tumblr.py +++ b/yt_dlp/extractor/tumblr.py @@ -8,7 +8,7 @@ from ..utils import ( class TumblrIE(InfoExtractor): - _VALID_URL = r'https?://(?P[^/?#&]+)\.tumblr\.com/(?:post|video)/(?P[0-9]+)(?:$|[/?#])' + _VALID_URL = r'https?://(?P[^/?#&]+)\.tumblr\.com/(?:post|video|(?P[a-zA-Z\d-]+))/(?P[0-9]+)(?:$|[/?#])' _NETRC_MACHINE = 'tumblr' _LOGIN_URL = 'https://www.tumblr.com/login' _OAUTH_URL = 'https://www.tumblr.com/api/v2/oauth2/token' @@ -109,14 +109,13 @@ class TumblrIE(InfoExtractor): 'title': 'The Blues Remembers Everything the Country Forgot', 'alt_title': 'The Blues Remembers Everything the Country Forgot', 'description': 'md5:1a6b4097e451216835a24c1023707c79', - 'release_date': '20201224', 'creator': 'md5:c2239ba15430e87c3b971ba450773272', 'uploader': 'Moor Mother - Topic', 'upload_date': '20201223', 'uploader_id': 'UCxrMtFBRkFvQJ_vVM4il08w', 'uploader_url': 'http://www.youtube.com/channel/UCxrMtFBRkFvQJ_vVM4il08w', 'thumbnail': r're:^https?://i.ytimg.com/.*', - 'channel': 'Moor Mother - Topic', + 'channel': 'Moor Mother', 'channel_id': 'UCxrMtFBRkFvQJ_vVM4il08w', 'channel_url': 'https://www.youtube.com/channel/UCxrMtFBRkFvQJ_vVM4il08w', 'channel_follower_count': int, @@ -135,21 +134,6 @@ class TumblrIE(InfoExtractor): 'release_year': 2020, }, 'add_ie': ['Youtube'], - }, { - 'url': 'http://naked-yogi.tumblr.com/post/118312946248/naked-smoking-stretching', - 'md5': 'de07e5211d60d4f3a2c3df757ea9f6ab', - 'info_dict': { - 'id': 'Wmur', - 'ext': 'mp4', - 'title': 'naked smoking & stretching', - 'upload_date': '20150506', - 'timestamp': 1430931613, - 'age_limit': 18, - 'uploader_id': '1638622', - 'uploader': 'naked-yogi', - }, - # 'add_ie': ['Vidme'], - 'skip': 'dead embedded video host' }, { 'url': 'https://prozdvoices.tumblr.com/post/673201091169681408/what-recording-voice-acting-sounds-like', 'md5': 'a0063fc8110e6c9afe44065b4ea68177', @@ -232,6 +216,139 @@ class TumblrIE(InfoExtractor): 'upload_date': '20140429', }, 'add_ie': ['Instagram'], + }, { + 'note': 'new url scheme', + 'url': 'https://www.tumblr.com/catgirldick/706354197596078080?source=share', + 'info_dict': { + 'id': '706354197596078080', + 'ext': 'mp4', + 'title': 'Bocchi in low quality and spinning, nothing just that', + 'description': 'Bocchi in low quality and spinning, nothing just that', + 'tags': [], + 'uploader_id': str, + 'uploader_url': r're:https://[^/]+\.tumblr\.com/', + 'thumbnail': r're:https?://[^/]+\.media\.tumblr\.com/[^?#]+.jpg', + 'repost_count': int, + 'like_count': int, + 'age_limit': 0, + }, + }, { + 'note': 'bandcamp album embed', + 'url': 'https://patricia-taxxon.tumblr.com/post/704473755725004800/patricia-taxxon-agnes-hilda-patricia-taxxon', + 'info_dict': { + 'id': 'agnes-hilda', + 'title': 'Agnes & Hilda', + 'description': 'The inexplicable joy of an artist. Wash paws after listening.', + 'uploader_id': 'patriciataxxon', + }, + 'playlist_count': 8, + }, { + 'note': 'bandcamp track embeds (many)', + 'url': 'https://lyuboserafimov.tumblr.com/post/706659328557514752/dream-sequencer-i-love-you-sarah-connor-dream', + 'info_dict': { + 'id': '706659328557514752', + 'title': 'md5:d20e162d74d4225ef19ef5700760ea4e', + 'description': 'md5:ede184a5af7e79f09e5835d65986dd1f', + 'tags': ['synthwave', 'dream sequencer', 'retrowave', '80s', 'nostalgia', 'Bandcamp', 'My Music'], + 'uploader_id': 'lyuboserafimov', + 'uploader_url': 'https://lyuboserafimov.tumblr.com/', + 'age_limit': 0, + 'like_count': int, + 'repost_count': int, + }, + 'playlist_count': 4, + }, { + 'note': 'soundcloud track embed', + 'url': 'https://selfisekai.tumblr.com/post/706878583155671040/dont-mind-me-im-just-a-lil-yt-dlp-contributor', + 'info_dict': { + 'id': '1413238837', + 'ext': 'mp3', + 'title': '100 gecs - 800db cloud (maiacore nightcore edit)', + 'description': '', + 'genre': 'nightcore', + 'license': 'all-rights-reserved', + 'uploader': 'maiacore', + 'uploader_id': '1109090314', + 'uploader_url': 'https://soundcloud.com/maiacore', + 'duration': 97.045, + 'timestamp': 1672430792, + 'upload_date': '20221230', + 'thumbnail': r're:https?://.+\.jpg', + 'view_count': int, + 'repost_count': int, + 'like_count': int, + 'comment_count': int, + }, + }, { + 'note': 'soundcloud set embed', + 'url': 'https://selfisekai.tumblr.com/post/706885093077237760/songs-with-maia-in-them', + 'info_dict': { + 'id': '1369208083', + 'title': 'songs with maia in them :)', + 'description': '', + }, + 'playlist_mincount': 8, + }, { + 'note': 'dailymotion video embed', + 'url': 'https://www.tumblr.com/selfisekai/706884794734313472?source=share', + 'info_dict': { + 'id': 'x8hczf5', + 'ext': 'mp4', + 'title': 'Trying a viral pizza bagel in NYC', + 'description': 'md5:17e52b32ae21f23940912b3efff17b74', + 'duration': 58, + 'uploader': 'Insider', + 'uploader_id': 'x29n239', + 'tags': ['food', 'video', 'pizza', 'foody', 'bagel', 'bagels', 'feed:ins', 'feed:live', 'jacky barile', 'reels'], + 'timestamp': 1674064861, + 'upload_date': '20230118', + 'age_limit': 0, + 'thumbnail': 're:https?://.+', + 'view_count': int, + 'like_count': int, + }, + }, { + 'note': 'tiktok video embed', + 'url': 'https://selfisekai.tumblr.com/post/706885498468270080', + 'info_dict': { + 'id': '7098761136534867205', + 'ext': 'mp4', + 'title': 'md5:6e62de3b0157d2ed9999c7d61c6485f2', + 'description': 'md5:6e62de3b0157d2ed9999c7d61c6485f2', + 'creator': 'Poznań 🐐', + 'uploader': 'miasto_poznan', + 'uploader_id': '6998015174997607430', + 'uploader_url': 'https://www.tiktok.com/@MS4wLjABAAAAzgdq4BFpzWfUWq4sO84_HGHHdS7cItoTDdAscuSxYIUQKJvZZ9j99wPe0RuqJpaR', + 'duration': 46, + 'timestamp': 1652809127, + 'upload_date': '20220517', + 'thumbnail': r're:https?://[^/]+\.tiktokcdn\.com/.+', + 'artist': 'The King Khan & BBQ Show', + 'album': 'Love You So', + 'track': 'Love You So', + 'view_count': int, + 'like_count': int, + 'comment_count': int, + 'repost_count': int, + }, + }, { + 'note': 'tumblr video AND youtube embed', + 'url': 'https://selfisekai.tumblr.com/post/706895394042511360', + 'info_dict': { + 'id': '706895394042511360', + 'title': str, + 'uploader_id': 'selfisekai', + 'uploader_url': 'https://selfisekai.tumblr.com/', + 'age_limit': 0, + 'tags': [], + 'like_count': int, + 'repost_count': int, + }, + 'playlist_count': 2, + }, { + # twitch_live provider - error when linked account is not live + 'url': 'https://www.tumblr.com/anarcho-skamunist/722224493650722816/hollow-knight-stream-right-now-going-to-fight', + 'only_matching': True, }] _providers = { @@ -239,6 +356,20 @@ class TumblrIE(InfoExtractor): 'vimeo': 'Vimeo', 'vine': 'Vine', 'youtube': 'Youtube', + 'dailymotion': 'Dailymotion', + 'tiktok': 'TikTok', + 'twitch_live': 'TwitchStream', + } + # these are known providers, but we don't know which entity type + # we are supposed to extract, so we use matching by url + _ambiguous_providers = { + 'bandcamp', + 'soundcloud', + } + # known not to be supported + _unsupported_providers = { + # seems like podcasts can't be embedded + 'spotify', } _ACCESS_TOKEN = None @@ -256,23 +387,41 @@ class TumblrIE(InfoExtractor): if not self._ACCESS_TOKEN: return - self._download_json( - self._OAUTH_URL, None, 'Logging in', - data=urlencode_postdata({ - 'password': password, - 'grant_type': 'password', - 'username': username, - }), headers={ - 'Content-Type': 'application/x-www-form-urlencoded', - 'Authorization': f'Bearer {self._ACCESS_TOKEN}', - }, - errnote='Login failed', fatal=False) + data = { + 'password': password, + 'grant_type': 'password', + 'username': username, + } + if self.get_param('twofactor'): + data['tfa_token'] = self.get_param('twofactor') + + def _call_login(): + return self._download_json( + self._OAUTH_URL, None, 'Logging in', + data=urlencode_postdata(data), + headers={ + 'Content-Type': 'application/x-www-form-urlencoded', + 'Authorization': f'Bearer {self._ACCESS_TOKEN}', + }, + errnote='Login failed', fatal=False, + expected_status=lambda s: s == 200 or 400 <= s < 500) + + response = _call_login() + if response.get('error') == 'tfa_required': + data['tfa_token'] = self._get_tfa_info() + response = _call_login() + if response.get('error'): + self.report_warning('API returned error {}: {}'.format( + response.get('error'), response.get('error_description'))) def _real_extract(self, url): - blog, video_id = self._match_valid_url(url).groups() + blog_1, blog_2, video_id = self._match_valid_url(url).groups() + blog = blog_2 or blog_1 - url = f'http://{blog}.tumblr.com/post/{video_id}/' - webpage, urlh = self._download_webpage_handle(url, video_id) + url = f'http://{blog}.tumblr.com/post/{video_id}' + # whatsapp ua makes iab tcf shut the fuck up + webpage, urlh = self._download_webpage_handle(url, video_id, headers={ + 'User-Agent': 'WhatsApp/2.0'}) redirect_url = urlh.url @@ -291,21 +440,70 @@ class TumblrIE(InfoExtractor): video_id, headers={'Authorization': f'Bearer {self._ACCESS_TOKEN}'}, fatal=False), ('response', 'timeline', 'elements', 0)) or {} content_json = traverse_obj(post_json, ('trail', 0, 'content'), ('content')) or [] - video_json = next( - (item for item in content_json if item.get('type') == 'video'), {}) - media_json = video_json.get('media') or {} - if api_only and not media_json.get('url') and not video_json.get('url'): - raise ExtractorError('Failed to find video data for dashboard-only post') - if not media_json.get('url') and video_json.get('url'): - # external video host - return self.url_result( - video_json['url'], - self._providers.get(video_json.get('provider'), 'Generic')) + # the url we're extracting from might be an original post or it might be a reblog. + # if it's a reblog, og:description will be the reblogger's comment, not the uploader's. + # content_json is always the op, so if it exists but has no text, there's no description + if content_json: + description = '\n\n'.join(( + item.get('text') for item in content_json if item.get('type') == 'text')) or None + else: + description = self._og_search_description(webpage, default=None) + uploader_id = traverse_obj(post_json, 'reblogged_root_name', 'blog_name') - video_url = self._og_search_video_url(webpage, default=None) - duration = None + info_dict = { + 'id': video_id, + 'title': post_json.get('summary') or (blog if api_only else self._html_search_regex( + r'(?s)(?P<title>.*?)(?: \| Tumblr)?', webpage, 'title', default=blog)), + 'description': description, + 'uploader_id': uploader_id, + 'uploader_url': f'https://{uploader_id}.tumblr.com/' if uploader_id else None, + 'like_count': post_json.get('like_count'), + 'repost_count': post_json.get('reblog_count'), + 'age_limit': {True: 18, False: 0}.get(post_json.get('is_nsfw')), + 'tags': post_json.get('tags'), + } + + # for tumblr's own video hosting + fallback_format = None formats = [] + video_url = self._og_search_video_url(webpage, default=None) + # for external video hosts + entries = [] + ignored_providers = set() + unknown_providers = set() + + video_jsons = (item for item in content_json if item.get('type') in ('video', 'audio')) + if video_jsons: + for video_json in video_jsons: + media_json = video_json.get('media') or {} + if api_only and not media_json.get('url') and not video_json.get('url'): + raise ExtractorError('Failed to find video data for dashboard-only post') + provider = video_json.get('provider') + is_provider_ambiguous = provider in self._ambiguous_providers + + if provider in ('tumblr', None): + fallback_format = { + 'url': media_json.get('url') or video_url, + 'width': int_or_none( + media_json.get('width') or self._og_search_property('video:width', webpage, default=None)), + 'height': int_or_none( + media_json.get('height') or self._og_search_property('video:height', webpage, default=None)), + } + continue + elif provider in self._unsupported_providers: + ignored_providers.add(provider) + continue + elif provider and not is_provider_ambiguous and provider not in self._providers: + unknown_providers.add(provider) + + if video_json.get('url'): + # external video host + entries.append(self.url_result( + video_json['url'], + self._providers.get(provider, 'Generic') if not is_provider_ambiguous else None)) + + duration = None # iframes can supply duration and sometimes additional formats, so check for one iframe_url = self._search_regex( @@ -344,44 +542,35 @@ class TumblrIE(InfoExtractor): 'quality': quality, } for quality, (video_url, format_id) in enumerate(sources)] - if not media_json.get('url') and not video_url and not iframe_url: - # external video host (but we weren't able to figure it out from the api) - iframe_url = self._search_regex( - r'src=["\'](https?://safe\.txmblr\.com/svc/embed/inline/[^"\']+)["\']', - webpage, 'embed iframe url', default=None) - return self.url_result(iframe_url or redirect_url, 'Generic') + if not formats and fallback_format: + formats.append(fallback_format) - formats = formats or [{ - 'url': media_json.get('url') or video_url, - 'width': int_or_none( - media_json.get('width') or self._og_search_property('video:width', webpage, default=None)), - 'height': int_or_none( - media_json.get('height') or self._og_search_property('video:height', webpage, default=None)), - }] + if formats: + # tumblr's own video is always above embeds + entries = [{ + **info_dict, + 'formats': formats, + 'duration': duration, + 'thumbnail': (traverse_obj(video_json, ('poster', 0, 'url')) + or self._og_search_thumbnail(webpage, default=None)), + }] + entries - # the url we're extracting from might be an original post or it might be a reblog. - # if it's a reblog, og:description will be the reblogger's comment, not the uploader's. - # content_json is always the op, so if it exists but has no text, there's no description - if content_json: - description = '\n\n'.join(( - item.get('text') for item in content_json if item.get('type') == 'text')) or None + if ignored_providers: + if not entries: + raise ExtractorError(f'None of embed providers are supported: {str(", ".join(ignored_providers))}', video_id=video_id, expected=True) + else: + self.report_warning(f'Embeds from these providers are ignored as unsupported: {str(", ".join(ignored_providers))}', video_id) + if unknown_providers: + self.report_warning(f'Unrecognized providers, please report: {str(", ".join(unknown_providers))}', video_id) + + if len(entries) > 1: + return { + **info_dict, + '_type': 'playlist', + 'entries': entries, + } else: - description = self._og_search_description(webpage, default=None) - uploader_id = traverse_obj(post_json, 'reblogged_root_name', 'blog_name') - - return { - 'id': video_id, - 'title': post_json.get('summary') or (blog if api_only else self._html_search_regex( - r'(?s)(?P<title>.*?)(?: \| Tumblr)?', webpage, 'title')), - 'description': description, - 'thumbnail': (traverse_obj(video_json, ('poster', 0, 'url')) - or self._og_search_thumbnail(webpage, default=None)), - 'uploader_id': uploader_id, - 'uploader_url': f'https://{uploader_id}.tumblr.com/' if uploader_id else None, - 'duration': duration, - 'like_count': post_json.get('like_count'), - 'repost_count': post_json.get('reblog_count'), - 'age_limit': {True: 18, False: 0}.get(post_json.get('is_nsfw')), - 'tags': post_json.get('tags'), - 'formats': formats, - } + return { + **info_dict, + **entries[0], + }