[tumblr] Fix extractor (#2883)

Authored by: foghawk

yt_dlp/extractor/tumblr.py

@@ -6,6 +6,7 @@
 from ..utils import (
     ExtractorError,
     int_or_none,
+    traverse_obj,
     urlencode_postdata
 )
 
@@ -14,31 +15,130 @@ class TumblrIE(InfoExtractor):
     _VALID_URL = r'https?://(?P<blog_name>[^/?#&]+)\.tumblr\.com/(?:post|video)/(?P<id>[0-9]+)(?:$|[/?#])'
     _NETRC_MACHINE = 'tumblr'
     _LOGIN_URL = 'https://www.tumblr.com/login'
+    _OAUTH_URL = 'https://www.tumblr.com/api/v2/oauth2/token'
     _TESTS = [{
         'url': 'http://tatianamaslanydaily.tumblr.com/post/54196191430/orphan-black-dvd-extra-behind-the-scenes',
         'md5': '479bb068e5b16462f5176a6828829767',
         'info_dict': {
             'id': '54196191430',
             'ext': 'mp4',
-            'title': 'tatiana maslany news, Orphan Black || DVD extra - behind the scenes ↳...',
+            'title': 'md5:dfac39636969fe6bf1caa2d50405f069',
             'description': 'md5:390ab77358960235b6937ab3b8528956',
+            'uploader_id': 'tatianamaslanydaily',
+            'uploader_url': 'https://tatianamaslanydaily.tumblr.com/',
             'thumbnail': r're:^https?://.*\.jpg',
             'duration': 127,
+            'like_count': int,
+            'repost_count': int,
+            'age_limit': 0,
+            'tags': ['Orphan Black', 'Tatiana Maslany', 'Interview', 'Video', 'OB S1 DVD Extras'],
         }
     }, {
+        'note': 'multiple formats',
         'url': 'https://maskofthedragon.tumblr.com/post/626907179849564160/mona-talking-in-english',
         'md5': 'f43ff8a8861712b6cf0e0c2bd84cfc68',
         'info_dict': {
             'id': '626907179849564160',
             'ext': 'mp4',
-            'title': 'Me roast is buggered!, Mona\xa0“talking” in\xa0“english”',
+            'title': 'Mona\xa0“talking” in\xa0“english”',
             'description': 'md5:082a3a621530cb786ad2b7592a6d9e2c',
+            'uploader_id': 'maskofthedragon',
+            'uploader_url': 'https://maskofthedragon.tumblr.com/',
             'thumbnail': r're:^https?://.*\.jpg',
             'duration': 7,
+            'like_count': int,
+            'repost_count': int,
+            'age_limit': 0,
+            'tags': 'count:19',
         },
         'params': {
             'format': 'hd',
         },
+    }, {
+        'note': 'non-iframe video (with related posts)',
+        'url': 'https://shieldfoss.tumblr.com/post/675519763813908480',
+        'md5': '12bdb75661ef443bffe5a4dac1dbf118',
+        'info_dict': {
+            'id': '675519763813908480',
+            'ext': 'mp4',
+            'title': 'Shieldfoss',
+            'uploader_id': 'nerviovago',
+            'uploader_url': 'https://nerviovago.tumblr.com/',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'like_count': int,
+            'repost_count': int,
+            'age_limit': 0,
+            'tags': [],
+        }
+    }, {
+        'note': 'dashboard only (original post)',
+        'url': 'https://jujanon.tumblr.com/post/159704441298/my-baby-eating',
+        'md5': '029f7c91ab386701b211e3d494d2d95e',
+        'info_dict': {
+            'id': '159704441298',
+            'ext': 'mp4',
+            'title': 'md5:ba79365861101f4911452728d2950561',
+            'description': 'md5:773738196cea76b6996ec71e285bdabc',
+            'uploader_id': 'jujanon',
+            'uploader_url': 'https://jujanon.tumblr.com/',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'like_count': int,
+            'repost_count': int,
+            'age_limit': 0,
+            'tags': ['crabs', 'my video', 'my pets'],
+        }
+    }, {
+        'note': 'dashboard only (reblog)',
+        'url': 'https://bartlebyshop.tumblr.com/post/180294460076/duality-of-bird',
+        'md5': '04334e7cadb1af680d162912559f51a5',
+        'info_dict': {
+            'id': '180294460076',
+            'ext': 'mp4',
+            'title': 'duality of bird',
+            'description': 'duality of bird',
+            'uploader_id': 'todaysbird',
+            'uploader_url': 'https://todaysbird.tumblr.com/',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'like_count': int,
+            'repost_count': int,
+            'age_limit': 0,
+            'tags': [],
+        }
+    }, {
+        'note': 'dashboard only (external)',
+        'url': 'https://afloweroutofstone.tumblr.com/post/675661759168823296/the-blues-remembers-everything-the-country-forgot',
+        'info_dict': {
+            'id': 'q67_fd7b8SU',
+            'ext': 'mp4',
+            '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_id': 'UCxrMtFBRkFvQJ_vVM4il08w',
+            'channel_url': 'https://www.youtube.com/channel/UCxrMtFBRkFvQJ_vVM4il08w',
+            'channel_follower_count': int,
+            'duration': 181,
+            'view_count': int,
+            'like_count': int,
+            'age_limit': 0,
+            'categories': ['Music'],
+            'tags': 'count:7',
+            'live_status': 'not_live',
+            'playable_in_embed': True,
+            'availability': 'public',
+            'track': 'The Blues Remembers Everything the Country Forgot',
+            'artist': 'md5:c2239ba15430e87c3b971ba450773272',
+            'album': 'Brass',
+            'release_year': 2020,
+        },
+        'add_ie': ['Youtube'],
     }, {
         'url': 'http://naked-yogi.tumblr.com/post/118312946248/naked-smoking-stretching',
         'md5': 'de07e5211d60d4f3a2c3df757ea9f6ab',
@@ -54,17 +154,48 @@ class TumblrIE(InfoExtractor):
         },
         # 'add_ie': ['Vidme'],
         'skip': 'dead embedded video host'
+    }, {
+        'url': 'https://prozdvoices.tumblr.com/post/673201091169681408/what-recording-voice-acting-sounds-like',
+        'md5': 'a0063fc8110e6c9afe44065b4ea68177',
+        'info_dict': {
+            'id': 'eomhW5MLGWA',
+            'ext': 'mp4',
+            'title': 'what recording voice acting sounds like',
+            'description': 'md5:1da3faa22d0e0b1d8b50216c284ee798',
+            'uploader': 'ProZD',
+            'upload_date': '20220112',
+            'uploader_id': 'ProZD',
+            'uploader_url': 'http://www.youtube.com/user/ProZD',
+            'thumbnail': r're:^https?://i.ytimg.com/.*',
+            'channel': 'ProZD',
+            'channel_id': 'UC6MFZAOHXlKK1FI7V0XQVeA',
+            'channel_url': 'https://www.youtube.com/channel/UC6MFZAOHXlKK1FI7V0XQVeA',
+            'channel_follower_count': int,
+            'duration': 20,
+            'view_count': int,
+            'like_count': int,
+            'age_limit': 0,
+            'categories': ['Film & Animation'],
+            'tags': [],
+            'live_status': 'not_live',
+            'playable_in_embed': True,
+            'availability': 'public',
+        },
+        'add_ie': ['Youtube'],
     }, {
         'url': 'https://dominustempori.tumblr.com/post/673572712813297664/youtubes-all-right-for-some-pretty-cool',
-        'md5': '5e45724c70b748f64f5a1731ac72c84a',
+        'md5': '203e9eb8077e3f45bfaeb4c86c1467b8',
         'info_dict': {
             'id': '87816359',
-            'ext': 'mp4',
+            'ext': 'mov',
             'title': 'Harold Ramis',
+            'description': 'md5:be8e68cbf56ce0785c77f0c6c6dfaf2c',
             'uploader': 'Resolution Productions Group',
             'uploader_id': 'resolutionproductions',
             'uploader_url': 'https://vimeo.com/resolutionproductions',
+            'upload_date': '20140227',
             'thumbnail': r're:^https?://i.vimeocdn.com/video/.*',
+            'timestamp': 1393523719,
             'duration': 291,
         },
         'add_ie': ['Vimeo'],
@@ -107,83 +238,96 @@ class TumblrIE(InfoExtractor):
         'add_ie': ['Instagram'],
     }]
 
+    _providers = {
+        'instagram': 'Instagram',
+        'vimeo': 'Vimeo',
+        'vine': 'Vine',
+        'youtube': 'Youtube',
+    }
+
+    _ACCESS_TOKEN = None
+
     def _real_initialize(self):
+        self.get_access_token()
         self._login()
 
+    def get_access_token(self):
+        login_page = self._download_webpage(
+            self._LOGIN_URL, None, 'Downloading login page', fatal=False)
+        if login_page:
+            self._ACCESS_TOKEN = self._search_regex(
+                r'"API_TOKEN":\s*"(\w+)"', login_page, 'API access token', fatal=False)
+        if not self._ACCESS_TOKEN:
+            self.report_warning('Failed to get access token; metadata will be missing and some videos may not work')
+
     def _login(self):
         username, password = self._get_login_info()
-        if username is None:
+        if not username:
             return
 
-        login_page = self._download_webpage(
-            self._LOGIN_URL, None, 'Downloading login page')
+        if not self._ACCESS_TOKEN:
+            return
 
-        login_form = self._hidden_inputs(login_page)
-        login_form.update({
-            'user[email]': username,
-            'user[password]': password
-        })
-
-        response, urlh = self._download_webpage_handle(
-            self._LOGIN_URL, None, 'Logging in',
-            data=urlencode_postdata(login_form), headers={
+        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',
-                'Referer': self._LOGIN_URL,
-            })
-
-        # Successful login
-        if '/dashboard' in urlh.geturl():
-            return
-
-        login_errors = self._parse_json(
-            self._search_regex(
-                r'RegistrationForm\.errors\s*=\s*(\[.+?\])\s*;', response,
-                'login errors', default='[]'),
-            None, fatal=False)
-        if login_errors:
-            raise ExtractorError(
-                'Unable to login: %s' % login_errors[0], expected=True)
-
-        self.report_warning('Login has probably failed')
+                'Authorization': f'Bearer {self._ACCESS_TOKEN}',
+            },
+            errnote='Login failed', fatal=False)
 
     def _real_extract(self, url):
-        m_url = self._match_valid_url(url)
-        video_id = m_url.group('id')
-        blog = m_url.group('blog_name')
+        blog, video_id = self._match_valid_url(url).groups()
 
-        url = 'http://%s.tumblr.com/post/%s/' % (blog, video_id)
+        url = f'http://{blog}.tumblr.com/post/{video_id}/'
         webpage, urlh = self._download_webpage_handle(url, video_id)
 
         redirect_url = urlh.geturl()
-        if 'tumblr.com/safe-mode' in redirect_url or redirect_url.startswith('/safe-mode'):
-            raise ExtractorError(
-                'This Tumblr may contain sensitive media. '
-                'Disable safe mode in your account settings '
-                'at https://www.tumblr.com/settings/account#safe_mode',
-                expected=True)
 
+        api_only = bool(self._search_regex(
+            r'(tumblr.com|^)/(safe-mode|login_required|blog/view)',
+            redirect_url, 'redirect', default=None))
+
+        if api_only and not self._ACCESS_TOKEN:
+            raise ExtractorError('Cannot get data for dashboard-only post without access token')
+
+        post_json = {}
+        if self._ACCESS_TOKEN:
+            post_json = traverse_obj(
+                self._download_json(
+                    f'https://www.tumblr.com/api/v2/blog/{blog}/posts/{video_id}/permalink',
+                    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'))
+
+        video_url = self._og_search_video_url(webpage, default=None)
+        duration = None
+        formats = []
+
+        # iframes can supply duration and sometimes additional formats, so check for one
         iframe_url = self._search_regex(
-            r'src=\'(https?://www\.tumblr\.com/video/[^\']+)\'',
+            fr'src=\'(https?://www\.tumblr\.com/video/{blog}/{video_id}/[^\']+)\'',
             webpage, 'iframe url', default=None)
-        if iframe_url is None:
-            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 iframe_url:
             iframe = self._download_webpage(
                 iframe_url, video_id, 'Downloading iframe page',
                 headers={'Referer': redirect_url})
 
-        duration = None
-        sources = []
-
-        sd_url = self._search_regex(
-            r'<source[^>]+src=(["\'])(?P<url>.+?)\1', iframe,
-            'sd video url', default=None, group='url')
-        if sd_url:
-            sources.append((sd_url, 'sd'))
-
             options = self._parse_json(
                 self._search_regex(
                     r'data-crt-options=(["\'])(?P<options>.+?)\1', iframe,
@@ -191,32 +335,66 @@ def _real_extract(self, url):
                 video_id, fatal=False)
             if options:
                 duration = int_or_none(options.get('duration'))
+
                 hd_url = options.get('hdUrl')
                 if hd_url:
+                    # there are multiple formats; extract them
+                    # ignore other sources of width/height data as they may be wrong
+                    sources = []
+                    sd_url = self._search_regex(
+                        r'<source[^>]+src=(["\'])(?P<url>.+?)\1', iframe,
+                        'sd video url', default=None, group='url')
+                    if sd_url:
+                        sources.append((sd_url, 'sd'))
                     sources.append((hd_url, 'hd'))
 
                     formats = [{
                         'url': video_url,
-            'ext': 'mp4',
                         'format_id': format_id,
                         'height': int_or_none(self._search_regex(
-                r'/(\d{3,4})$', video_url, 'height', default=None)),
+                            r'_(\d+)\.\w+$', video_url, 'height', default=None)),
                         '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')
+
+        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)),
+        }]
         self._sort_formats(formats)
 
-        # The only place where you can get a title, it's not complete,
-        # but searching in other places doesn't work for all videos
-        video_title = self._html_search_regex(
-            r'(?s)<title>(?P<title>.*?)(?: \| Tumblr)?</title>',
-            webpage, 'title')
+        # 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')
 
         return {
             'id': video_id,
-            'title': video_title,
-            'description': self._og_search_description(webpage, default=None),
-            'thumbnail': self._og_search_thumbnail(webpage, default=None),
+            'title': post_json.get('summary') or (blog if api_only else self._html_search_regex(
+                r'(?s)<title>(?P<title>.*?)(?: \| Tumblr)?</title>', 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,
         }