util.parse_mf2: set metaformats on main mf2 item

either representative h-card or h-entry, if available
snarfed · Jul 25, 2024 · ac77c73 · ac77c73
1 parent 3019ba2
commit ac77c73
Show file tree

Hide file tree

Showing 2 changed files with 50 additions and 19 deletions.
diff --git a/tests/test_util.py b/tests/test_util.py
@@ -1718,3 +1718,27 @@ def test_parse_mf2_metaformats_hcard_photo_fallback(self):
 </html>
 """, url='http://xyz', metaformats=True),
     ignore=['debug', 'rels', 'rel-urls'])
+
+  def test_parse_mf2_metaformats_hentry_inside_hfeed(self):
+    self.assert_equals({
+      'items': [{
+        'type': ['h-feed'],
+        'properties': {},
+        'children': [{
+          'type': ['h-entry'],
+          'properties': {
+            'name': ['foo'],
+            'photo': ['http://pic'],
+          },
+        }],
+      }],
+    }, util.parse_mf2("""\
+<html>
+<head><meta property="og:image" content="http://pic" /></head>
+<body>
+<div class="h-feed"><div class="h-entry">foo</div></div>
+</body>
+</html>
+""", url='http://xyz/post', metaformats=True),
+    ignore=['debug', 'rels', 'rel-urls'])
+
diff --git a/util.py b/util.py
@@ -28,6 +28,7 @@
 from cachetools import cached, TTLCache
 from domain2idna import domain2idna
 from flask import abort
+import mf2util
 
 try:
   import ujson
@@ -226,28 +227,28 @@
 METAFORMAT_TO_MF2 = [
     # in priority order, descending
     # OGP
-    ("property", "article:author", "url"),
-    ("property", "article:published_time", "published"),
-    ("property", "article:modified_time", "updated"),
-    ("property", "og:audio", "audio"),
-    ("property", "og:description", "summary"),
-    ("property", "og:image", "photo"),
-    ("property", "og:title", "name"),
-    ("property", "og:video", "video"),
+    ('property', 'article:author', 'url'),
+    ('property', 'article:published_time', 'published'),
+    ('property', 'article:modified_time', 'updated'),
+    ('property', 'og:audio', 'audio'),
+    ('property', 'og:description', 'summary'),
+    ('property', 'og:image', 'photo'),
+    ('property', 'og:title', 'name'),
+    ('property', 'og:video', 'video'),
     # Twitter
-    ("name", "twitter:title", "name"),
-    ("name", "twitter:description", "summary"),
-    ("name", "twitter:image", "photo"),
+    ('name', 'twitter:title', 'name'),
+    ('name', 'twitter:description', 'summary'),
+    ('name', 'twitter:image', 'photo'),
     # HTML standard meta names
     # https://developer.mozilla.org/en-US/docs/Web/HTML/Element/meta/name
-    ("name", "description", "summary"),
+    ('name', 'description', 'summary'),
 ]
 METAFORMAT_URL_PROPERTIES = {
-    "article:author",
-    "og:audio",
-    "og:image",
-    "og:video",
-    "twitter:image",
+    'article:author',
+    'og:audio',
+    'og:image',
+    'og:video',
+    'twitter:image',
 }
 
 
@@ -2094,16 +2095,22 @@ def parse_mf2(input, url=None, id=None, metaformats=None):
       return None
 
   mf2 = mf2py.parse(url=url, doc=input)
+  if urlparse(url).path in ('', '/'):
+    type = 'h-card'
+    mf2_item = mf2util.representative_hcard(mf2, mf2.get('url') or url)
+  else:
+    type = 'h-entry'
+    mf2_item = mf2util.find_first_entry(mf2, ['h-entry'])
 
-  mf2_item = mf2['items'][0] if mf2['items'] else None
+  if not mf2_item and mf2['items']:
+    mf2_item = mf2['items'][0]
 
   mf2_hcard = None
   for item in mf2['items']:
     if 'h-card' in item.get('type', []):
       mf2_hcard = item
 
   if metaformats and url:
-    type = 'h-card' if urlparse(url).path in ('', '/') else 'h-entry'
     if meta_item := parse_metaformats(input, url, type=type):
       if mf2_item:
         # if mf2 item doesn't have a photo, fall back to metaformats