diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py
index cbca22c91d..cf06dbde46 100644
--- a/test/test_InfoExtractor.py
+++ b/test/test_InfoExtractor.py
@@ -99,10 +99,10 @@ class TestInfoExtractor(unittest.TestCase):
self.assertRaises(RegexNotFoundError, ie._html_search_meta, ('z', 'x'), html, None, fatal=True)
def test_search_json_ld_realworld(self):
- # https://github.com/ytdl-org/youtube-dl/issues/23306
- expect_dict(
- self,
- self.ie._search_json_ld(r'''''', None),
- {
- 'title': '1 On 1 With Kleio',
- 'description': 'Kleio Valentien',
- 'url': 'https://gvideo.eporner.com/xN49A1cT3eB/xN49A1cT3eB.mp4',
- 'timestamp': 1449347075,
- 'duration': 743.0,
- 'view_count': 1120958,
- 'width': 1920,
- 'height': 1080,
- })
+ ''',
+ {
+ 'title': '1 On 1 With Kleio',
+ 'description': 'Kleio Valentien',
+ 'url': 'https://gvideo.eporner.com/xN49A1cT3eB/xN49A1cT3eB.mp4',
+ 'timestamp': 1449347075,
+ 'duration': 743.0,
+ 'view_count': 1120958,
+ 'width': 1920,
+ 'height': 1080,
+ },
+ {},
+ ),
+ (
+ r'''''',
+ {
+ 'timestamp': 1636523400,
+ 'title': 'md5:91fe569e952e4d146485740ae927662b',
+ },
+ {'expected_type': 'NewsArticle'},
+ ),
+ ]
+ for html, expected_dict, search_json_ld_kwargs in _TESTS:
+ expect_dict(
+ self,
+ self.ie._search_json_ld(html, None, **search_json_ld_kwargs),
+ expected_dict
+ )
def test_download_json(self):
uri = encode_data_uri(b'{"foo": "blah"}', 'application/json')
diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py
index ebf2e3ceac..52099b4b42 100644
--- a/yt_dlp/extractor/common.py
+++ b/yt_dlp/extractor/common.py
@@ -1451,8 +1451,13 @@ class InfoExtractor(object):
})
extract_interaction_statistic(e)
- for e in json_ld:
- if '@context' in e:
+ def traverse_json_ld(json_ld, at_top_level=True):
+ for e in json_ld:
+ if at_top_level and '@context' not in e:
+ continue
+ if at_top_level and set(e.keys()) == {'@context', '@graph'}:
+ traverse_json_ld(variadic(e['@graph'], allowed_types=(dict,)), at_top_level=False)
+ break
item_type = e.get('@type')
if expected_type is not None and expected_type != item_type:
continue
@@ -1488,7 +1493,7 @@ class InfoExtractor(object):
info.update({
'timestamp': parse_iso8601(e.get('datePublished')),
'title': unescapeHTML(e.get('headline')),
- 'description': unescapeHTML(e.get('articleBody')),
+ 'description': unescapeHTML(e.get('articleBody') or e.get('description')),
})
elif item_type == 'VideoObject':
extract_video_object(e)
@@ -1503,6 +1508,8 @@ class InfoExtractor(object):
continue
else:
break
+ traverse_json_ld(json_ld)
+
return dict((k, v) for k, v in info.items() if v is not None)
def _search_nextjs_data(self, webpage, video_id, **kw):