Implement _search_json_ld()

youtube_dl/extractor/megacartoons.py

@@ -2,7 +2,16 @@
 from __future__ import unicode_literals
 
 import json
-from ..utils import url_or_none
+import re
+
+from ..utils import (
+    bug_reports_message,
+    JSON_LD_RE,
+    merge_dicts,
+    NO_DEFAULT,
+    RegexNotFoundError,
+    url_or_none,
+)
 
 from .common import InfoExtractor
 
@@ -14,39 +23,81 @@ class MegaCartoonsIE(InfoExtractor):
         'md5': '4ba9be574f9a17abe0c074e2f955fded',
         'info_dict': {
             'id': 'help-wanted',
-            'title': 'Help Wanted',
             'ext': 'mp4',
+            'title': 'Help Wanted - SpongeBob SquarePants',
+            'upload_date': '20200223',
+            'timestamp': 1582416000,
             'thumbnail': r're:^https?://.*\.jpg$',
             'description': 'md5:2c909daa6c6cb16b2d4d791dd1a31632'
         }
     }
 
+    # adapted from common.py pending yt-dlp back-port
+    def _search_json_ld(self, html, video_id, expected_type=None, **kwargs):
+        json_ld_list = list(re.finditer(JSON_LD_RE, html))
+        default = kwargs.get('default', NO_DEFAULT)
+        fatal = kwargs.get('fatal', True) if default is NO_DEFAULT else False
+        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:
+            # handle initial '@graph' with one level of children
+            if len(json_ld) > 0 and '@graph' in json_ld[0] and '@context' in json_ld[0]:
+                # should always be hit here
+                context = json_ld[0]['@context']
+                json_ld_g = json_ld[0]['@graph'] or []
+                for item in json_ld_g:
+                    item.setdefault('@context', context)
+                json_ld = json_ld_g + json_ld[1:]
+            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.report_warning('unable to extract JSON-LD %s' % bug_reports_message())
+            return {}
+
     def _real_extract(self, url):
         # ID is equal to the episode name
         video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
 
-        # Try to find a good title or fallback to the ID
-        title = self._og_search_title(webpage) or video_id
+        info = self._search_json_ld(webpage, video_id, fatal=False) or {}
 
+        info.update({
+            'id': video_id,
+            # Try to find a good title or fallback to the ID
+            'title': info.get('title') or self._og_search_title(webpage) or video_id.replace('-', ' ').capitalize(),
+        })
+
+        if 'url' not in info or 'thumbnail' not in info:
             # Video data is stored in a json -> extract it from the raw html
             url_json = json.loads(self._html_search_regex(r'<div.*data-item=["/\'](?P<videourls>{.*})["/\'].*>', webpage, 'videourls'))
 
             video_url = url_or_none(url_json.get('sources')[0].get('src') or self._og_search_video_url(webpage))   # Get the video url
             video_thumbnail = url_or_none(url_json.get('splash') or self._og_search_thumbnail(webpage))            # Get the thumbnail
+            info = merge_dicts(info, {
+                'url': video_url,
+                'thumbnail': video_thumbnail,
+            })
 
         # Find the <article> class in the html
         article = self._search_regex(
             r'(?s)<article\b[^>]*?\bclass\s*=\s*[^>]*?\bpost\b[^>]*>(.+?)</article\b', webpage, 'post', default='')
 
         # Extract the actual description from it
-        video_description = (self._html_search_regex(r'(?s)<p>\s*([^<]+)\s*</p>', article, 'videodescription', fatal=False)
+        info['description'] = (
+            self._html_search_regex(r'(?s)<p>\s*([^<]+)\s*</p>', article, 'videodescription', fatal=False)
             or self._og_search_description(webpage))
 
-        return {
-            'id': video_id,
-            'title': title,
-            'url': video_url,
-            'thumbnail': video_thumbnail,
-            'description': video_description,
-        }
+        return info