yt_dlp/extractor/facebook.py

@@ -20,7 +20,6 @@ from ..utils import (
     get_element_by_id,
     get_first,
     int_or_none,
-    join_nonempty,
     js_to_json,
     merge_dicts,
     parse_count,
@@ -903,7 +902,7 @@ class FacebookAdsIE(InfoExtractor):
             'ext': 'mp4',
             'title': 'video by Kandao',
             'uploader': 'Kandao',
-            'uploader_id': '774114102743284',
+            'uploader_id': 774114102743284,
             'uploader_url': r're:^https?://.*',
             'timestamp': 1702548330,
             'thumbnail': r're:^https?://.*',
@@ -916,7 +915,7 @@ class FacebookAdsIE(InfoExtractor):
             'id': '893637265423481',
             'title': 'Jusqu\u2019\u00e0 -25% sur une s\u00e9lection de vins p\u00e9tillants italiens ',
             'uploader': 'Eataly Paris Marais',
-            'uploader_id': '2086668958314152',
+            'uploader_id': 2086668958314152,
             'uploader_url': r're:^https?://.*',
             'timestamp': 1703571529,
             'upload_date': '20231226',
@@ -925,73 +924,71 @@ class FacebookAdsIE(InfoExtractor):
         'playlist_count': 3,
     }]
 
-    _FORMATS_MAP = {
+    def _extract_from_url(self, url, video_id):
-        'watermarked_video_sd_url': ('sd-wmk', 'SD, watermarked'),
+        webpage = self._download_webpage(url, video_id)
-        'video_sd_url': ('sd', None),
-        'watermarked_video_hd_url': ('hd-wmk', 'HD, watermarked'),
-        'video_hd_url': ('hd', None),
-    }
 
-    def _extract_formats(self, video_dict):
+        def extract_metadata(webpage):
-        formats = []
+            def extract_format(video_dict):
-        for format_key, format_url in traverse_obj(video_dict, (
+                formats = []
-            {dict.items}, lambda _, v: v[0] in self._FORMATS_MAP and url_or_none(v[1])
+                for i, url in enumerate(
-        )):
+                    [url_or_none(video_dict.get('watermarked_video_sd_url')), url_or_none(video_dict.get('video_sd_url')),
-            formats.append({
+                     url_or_none(video_dict.get('watermarked_video_hd_url')), url_or_none(video_dict.get('video_hd_url'))]
-                'format_id': self._FORMATS_MAP[format_key][0],
+                ):
-                'format_note': self._FORMATS_MAP[format_key][1],
+                    if url:
-                'url': format_url,
+                        formats.append({
-                'ext': 'mp4',
+                            'format_id': ['sd-wmk', 'sd', 'hd-wmk', 'hd'][i],
-                'quality': qualities(tuple(self._FORMATS_MAP))(format_key),
+                            'format_note': ['SD, watermarked', None, 'HD, watermarked', None][i],
-            })
+                            'url': url,
-        return formats
+                            'ext': 'mp4',
+                            'preference': i,
+                        })
+                return formats
+
+            post_data = [self._parse_json(j, video_id, fatal=False) for j in re.findall(r's.handle\(({.*})\);requireLazy\(', webpage)]
+            ad_data = traverse_obj(post_data, (..., 'require', ..., ..., ..., 'props', 'deeplinkAdCard', 'snapshot'), {dict})
+            info_dict = {}
+            if ad_data and ad_data[0]:
+                data = ad_data[0]
+                title = f"{data['display_format']} by {data['page_name']}" if not data['title'] or data['title'] == '{{product.name}}' else data['title']
+                description = None if data['link_description'] == '{{product.description}}' else data['link_description']
+                info_dict = {
+                    'description': description,
+                    'uploader': data['page_name'],
+                    'uploader_id': data['page_id'],
+                    'uploader_url': data['page_profile_uri'],
+                    'timestamp': data['creation_time'],
+                    'like_count': data['page_like_count'],
+                }
+                entries = []
+                for group in [data['videos'], data['cards']]:
+                    for entry in group:
+                        if entry.get('watermarked_video_sd_url') or entry.get('video_sd_url') or entry.get('watermarked_video_hd_url') or entry.get('video_hd_url'):
+                            entries.append({
+                                'id': f'{video_id}_%s' % str(len(entries) + 1),
+                                'title': entry.get('title') or title,
+                                'description': entry.get('link_description') or description,
+                                'thumbnail': entry.get('video_preview_image_url'),
+                                'formats': extract_format(entry),
+                            })
+                if len(entries) == 1:
+                    info_dict.update(entries[0])
+                    info_dict['id'] = video_id
+                elif len(entries) > 1:
+                    info_dict.update({
+                        'title': entries[0]['title'],
+                        'entries': entries,
+                        '_type': 'playlist',
+                    })
+            return info_dict
+
+        info_dict = {
+            'id': video_id,
+            'title': 'Ad Library',
+        }
+        info_dict.update(extract_metadata(webpage))
+
+        return info_dict
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
+        return self._extract_from_url(f'https://www.facebook.com/ads/library/?id={video_id}', 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