[extractor/common] Extract multiple JSON-LD entries

This commit is contained in:
Sergey M․ 2020-05-02 23:40:30 +07:00
parent e40c758c2a
commit 4433bb0245
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -1182,16 +1182,33 @@ class InfoExtractor(object):
'twitter card player')
def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
json_ld = self._search_regex(
JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs)
json_ld_list = list(re.finditer(JSON_LD_RE, html))
default = kwargs.get('default', NO_DEFAULT)
if not json_ld:
return default if default is not NO_DEFAULT else {}
# JSON-LD may be malformed and thus `fatal` should be respected.
# At the same time `default` may be passed that assumes `fatal=False`
# for _search_regex. Let's simulate the same behavior here as well.
fatal = kwargs.get('fatal', True) if default == NO_DEFAULT else False
return self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
json_ld = []
for mobj in json_ld_list:
json_ld_item = self._parse_json(
mobj.group('json_ld'), video_id, fatal=fatal)
if not json_ld_item:
continue
if isinstance(json_ld_item, dict):
json_ld.append(json_ld_item)
elif isinstance(json_ld_item, (list, tuple)):
json_ld.extend(json_ld_item)
if json_ld:
json_ld = self._json_ld(json_ld, video_id, fatal=fatal, expected_type=expected_type)
if json_ld:
return json_ld
if default is not NO_DEFAULT:
return default
elif fatal:
raise RegexNotFoundError('Unable to extract JSON-LD')
else:
self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
return {}
def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None):
if isinstance(json_ld, compat_str):
@ -1256,10 +1273,10 @@ class InfoExtractor(object):
extract_interaction_statistic(e)
for e in json_ld:
if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
if '@context' in e:
item_type = e.get('@type')
if expected_type is not None and expected_type != item_type:
return info
continue
if item_type in ('TVEpisode', 'Episode'):
episode_name = unescapeHTML(e.get('name'))
info.update({
@ -1293,10 +1310,16 @@ class InfoExtractor(object):
})
elif item_type == 'VideoObject':
extract_video_object(e)
if expected_type is None:
continue
else:
break
video = e.get('video')
if isinstance(video, dict) and video.get('@type') == 'VideoObject':
extract_video_object(video)
if expected_type is None:
continue
else:
break
return dict((k, v) for k, v in info.items() if v is not None)