[ie/facebook:ads] Add extractor (#8870)

Closes #8083
Authored by: kclauhk

yt_dlp/extractor/_extractors.py

@@ -588,6 +588,7 @@ from .facebook import (
     FacebookPluginsVideoIE,
     FacebookRedirectURLIE,
     FacebookReelIE,
+    FacebookAdsIE,
 )
 from .fancode import (
     FancodeVodIE,

yt_dlp/extractor/facebook.py

@@ -20,6 +20,7 @@ from ..utils import (
     get_element_by_id,
     get_first,
     int_or_none,
+    join_nonempty,
     js_to_json,
     merge_dicts,
     parse_count,
@@ -907,3 +908,114 @@ class FacebookReelIE(InfoExtractor):
         video_id = self._match_id(url)
         return self.url_result(
             f'https://m.facebook.com/watch/?v={video_id}&_rdr', FacebookIE, video_id)
+
+
+class FacebookAdsIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/ads/library/?\?(?:[^#]+&)?id=(?P<id>\d+)'
+    IE_NAME = 'facebook:ads'
+
+    _TESTS = [{
+        'url': 'https://www.facebook.com/ads/library/?id=899206155126718',
+        'info_dict': {
+            'id': '899206155126718',
+            'ext': 'mp4',
+            'title': 'video by Kandao',
+            'uploader': 'Kandao',
+            'uploader_id': '774114102743284',
+            'uploader_url': r're:^https?://.*',
+            'timestamp': 1702548330,
+            'thumbnail': r're:^https?://.*',
+            'upload_date': '20231214',
+            'like_count': int,
+        }
+    }, {
+        'url': 'https://www.facebook.com/ads/library/?id=893637265423481',
+        'info_dict': {
+            'id': '893637265423481',
+            'title': 'Jusqu\u2019\u00e0 -25% sur une s\u00e9lection de vins p\u00e9tillants italiens ',
+            'uploader': 'Eataly Paris Marais',
+            'uploader_id': '2086668958314152',
+            'uploader_url': r're:^https?://.*',
+            'timestamp': 1703571529,
+            'upload_date': '20231226',
+            'like_count': int,
+        },
+        'playlist_count': 3,
+    }, {
+        'url': 'https://es-la.facebook.com/ads/library/?id=901230958115569',
+        'only_matching': True,
+    }, {
+        'url': 'https://m.facebook.com/ads/library/?id=901230958115569',
+        'only_matching': True,
+    }]
+
+    _FORMATS_MAP = {
+        'watermarked_video_sd_url': ('sd-wmk', 'SD, watermarked'),
+        'video_sd_url': ('sd', None),
+        'watermarked_video_hd_url': ('hd-wmk', 'HD, watermarked'),
+        'video_hd_url': ('hd', None),
+    }
+
+    def _extract_formats(self, video_dict):
+        formats = []
+        for format_key, format_url in traverse_obj(video_dict, (
+            {dict.items}, lambda _, v: v[0] in self._FORMATS_MAP and url_or_none(v[1])
+        )):
+            formats.append({
+                'format_id': self._FORMATS_MAP[format_key][0],
+                'format_note': self._FORMATS_MAP[format_key][1],
+                'url': format_url,
+                'ext': 'mp4',
+                'quality': qualities(tuple(self._FORMATS_MAP))(format_key),
+            })
+        return formats
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+
+        post_data = [self._parse_json(j, video_id, fatal=False)
+                     for j in re.findall(r's\.handle\(({.*})\);requireLazy\(', webpage)]
+        data = traverse_obj(post_data, (
+            ..., 'require', ..., ..., ..., 'props', 'deeplinkAdCard', 'snapshot', {dict}), get_all=False)
+        if not data:
+            raise ExtractorError('Unable to extract ad data')
+
+        title = data.get('title')
+        if not title or title == '{{product.name}}':
+            title = join_nonempty('display_format', 'page_name', delim=' by ', from_dict=data)
+
+        info_dict = traverse_obj(data, {
+            'description': ('link_description', {str}, {lambda x: x if x != '{{product.description}}' else None}),
+            'uploader': ('page_name', {str}),
+            'uploader_id': ('page_id', {str_or_none}),
+            'uploader_url': ('page_profile_uri', {url_or_none}),
+            'timestamp': ('creation_time', {int_or_none}),
+            'like_count': ('page_like_count', {int_or_none}),
+        })
+
+        entries = []
+        for idx, entry in enumerate(traverse_obj(
+            data, (('videos', 'cards'), lambda _, v: any([url_or_none(v[f]) for f in self._FORMATS_MAP]))), 1
+        ):
+            entries.append({
+                'id': f'{video_id}_{idx}',
+                'title': entry.get('title') or title,
+                'description': entry.get('link_description') or info_dict.get('description'),
+                'thumbnail': url_or_none(entry.get('video_preview_image_url')),
+                'formats': self._extract_formats(entry),
+            })
+
+        if len(entries) == 1:
+            info_dict.update(entries[0])
+
+        elif len(entries) > 1:
+            info_dict.update({
+                'title': entries[0]['title'],
+                'entries': entries,
+                '_type': 'playlist',
+            })
+
+        info_dict['id'] = video_id
+
+        return info_dict

yt_dlp/extractor/floatplane.py

@@ -11,6 +11,7 @@ from ..utils import (
     join_nonempty,
     parse_codecs,
     parse_iso8601,
+    url_or_none,
     urljoin,
 )
 from ..utils.traversal import traverse_obj
@@ -108,6 +109,64 @@ class FloatplaneIE(InfoExtractor):
             'availability': 'subscriber_only',
         },
         'params': {'skip_download': 'm3u8'},
+    }, {
+        'url': 'https://www.floatplane.com/post/65B5PNoBtf',
+        'info_dict': {
+            'id': '65B5PNoBtf',
+            'description': 'I recorded the inbuilt demo mode for your 90\'s enjoyment, thanks for being Floaties!',
+            'display_id': '65B5PNoBtf',
+            'like_count': int,
+            'release_timestamp': 1701249480,
+            'uploader': 'The Trash Network',
+            'availability': 'subscriber_only',
+            'uploader_id': '61bc20c9a131fb692bf2a513',
+            'uploader_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home',
+            'channel_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home/thedrumthing',
+            'comment_count': int,
+            'title': 'The $50 electronic drum kit.',
+            'channel_id': '64424fe73cd58cbcf8d8e131',
+            'thumbnail': 'https://pbs.floatplane.com/blogPost_thumbnails/65B5PNoBtf/725555379422705_1701247052743.jpeg',
+            'dislike_count': int,
+            'channel': 'The Drum Thing',
+            'release_date': '20231129',
+        },
+        'playlist_count': 2,
+        'playlist': [{
+            'info_dict': {
+                'id': 'ISPJjexylS',
+                'ext': 'mp4',
+                'release_date': '20231129',
+                'release_timestamp': 1701249480,
+                'title': 'The $50 electronic drum kit. .mov',
+                'channel_id': '64424fe73cd58cbcf8d8e131',
+                'thumbnail': 'https://pbs.floatplane.com/video_thumbnails/ISPJjexylS/335202812134041_1701249383392.jpeg',
+                'availability': 'subscriber_only',
+                'uploader': 'The Trash Network',
+                'duration': 622,
+                'channel': 'The Drum Thing',
+                'uploader_id': '61bc20c9a131fb692bf2a513',
+                'channel_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home/thedrumthing',
+                'uploader_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home',
+            },
+        }, {
+            'info_dict': {
+                'id': 'qKfxu6fEpu',
+                'ext': 'aac',
+                'release_date': '20231129',
+                'release_timestamp': 1701249480,
+                'title': 'Roland TD-7 Demo.m4a',
+                'channel_id': '64424fe73cd58cbcf8d8e131',
+                'availability': 'subscriber_only',
+                'uploader': 'The Trash Network',
+                'duration': 114,
+                'channel': 'The Drum Thing',
+                'uploader_id': '61bc20c9a131fb692bf2a513',
+                'channel_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home/thedrumthing',
+                'uploader_url': 'https://www.floatplane.com/channel/TheTrashNetwork/home',
+            },
+        }],
+        'skip': 'requires subscription: "The Trash Network"',
+        'params': {'skip_download': 'm3u8'},
     }]
 
     def _real_initialize(self):
@@ -124,6 +183,22 @@ class FloatplaneIE(InfoExtractor):
         if not any(traverse_obj(post_data, ('metadata', ('hasVideo', 'hasAudio')))):
             raise ExtractorError('Post does not contain a video or audio track', expected=True)
 
+        uploader_url = format_field(
+            post_data, [('creator', 'urlname')], 'https://www.floatplane.com/channel/%s/home') or None
+
+        common_info = {
+            'uploader_url': uploader_url,
+            'channel_url': urljoin(f'{uploader_url}/', traverse_obj(post_data, ('channel', 'urlname'))),
+            'availability': self._availability(needs_subscription=True),
+            **traverse_obj(post_data, {
+                'uploader': ('creator', 'title', {str}),
+                'uploader_id': ('creator', 'id', {str}),
+                'channel': ('channel', 'title', {str}),
+                'channel_id': ('channel', 'id', {str}),
+                'release_timestamp': ('releaseDate', {parse_iso8601}),
+            }),
+        }
+
         items = []
         for media in traverse_obj(post_data, (('videoAttachments', 'audioAttachments'), ...)):
             media_id = media['id']
@@ -150,11 +225,11 @@ class FloatplaneIE(InfoExtractor):
             formats = []
             for quality in traverse_obj(stream, ('resource', 'data', 'qualityLevels', ...)):
                 url = urljoin(stream['cdn'], format_path(traverse_obj(
-                    stream, ('resource', 'data', 'qualityLevelParams', quality['name']))))
+                    stream, ('resource', 'data', 'qualityLevelParams', quality['name'], {dict}))))
                 formats.append({
                     **traverse_obj(quality, {
-                        'format_id': 'name',
-                        'format_note': 'label',
+                        'format_id': ('name', {str}),
+                        'format_note': ('label', {str}),
                         'width': ('width', {int}),
                         'height': ('height', {int}),
                     }),
@@ -164,38 +239,28 @@ class FloatplaneIE(InfoExtractor):
                 })
 
             items.append({
+                **common_info,
                 'id': media_id,
                 **traverse_obj(metadata, {
-                    'title': 'title',
+                    'title': ('title', {str}),
                     'duration': ('duration', {int_or_none}),
-                    'thumbnail': ('thumbnail', 'path'),
+                    'thumbnail': ('thumbnail', 'path', {url_or_none}),
                 }),
                 'formats': formats,
             })
 
-        uploader_url = format_field(
-            post_data, [('creator', 'urlname')], 'https://www.floatplane.com/channel/%s/home') or None
-        channel_url = urljoin(f'{uploader_url}/', traverse_obj(post_data, ('channel', 'urlname')))
-
         post_info = {
+            **common_info,
             'id': post_id,
             'display_id': post_id,
             **traverse_obj(post_data, {
-                'title': 'title',
+                'title': ('title', {str}),
                 'description': ('text', {clean_html}),
-                'uploader': ('creator', 'title'),
-                'uploader_id': ('creator', 'id'),
-                'channel': ('channel', 'title'),
-                'channel_id': ('channel', 'id'),
                 'like_count': ('likes', {int_or_none}),
                 'dislike_count': ('dislikes', {int_or_none}),
                 'comment_count': ('comments', {int_or_none}),
-                'release_timestamp': ('releaseDate', {parse_iso8601}),
-                'thumbnail': ('thumbnail', 'path'),
+                'thumbnail': ('thumbnail', 'path', {url_or_none}),
             }),
-            'uploader_url': uploader_url,
-            'channel_url': channel_url,
-            'availability': self._availability(needs_subscription=True),
         }
 
         if len(items) > 1: