microformats2.html_to_activities(): limit to h-entry, h-event, and h-…

…cite items

fixes #192

README.md

@@ -327,6 +327,7 @@ Non-breaking changes:
   * Handle malformed attachments better.
 * microformats2:
   * Don't crash on string `context` fields.
+  * `html_to_activities()`: limit to `h-entry`, `h-event`, and `h-cite` items ([#192](https://github.com/snarfed/granary/issues/192)).
 * The `cache` kwarg to `Source.original_post_discovery()` now has no effect. `webutil.util.follow_redirects()` has its own built in caching now.
 * Added Meetup.com support for publishing RSVPs.
 

granary/microformats2.py

@@ -587,9 +587,11 @@ def html_to_activities(html, url=None, actor=None, id=None):
 
   activities = []
   for item in items:
-    obj = json_to_object(item, actor=actor)
-    obj['content_is_html'] = True
-    activities.append({'object': obj})
+    types = item.get('type', [])
+    if 'h-entry' in types or 'h-event' in types or 'h-cite' in types:
+      obj = json_to_object(item, actor=actor)
+      obj['content_is_html'] = True
+      activities.append({'object': obj})
 
   return activities
 

granary/tests/test_microformats2.py

@@ -905,6 +905,15 @@ def test_html_to_activities_brs_to_newlines(self):
       'displayName': 'foo bar\nbaz \n\n baj',
     }}], activities)
 
+  def test_html_to_activities_filters_items(self):
+    """Check that we omit h-cards inside h-feeds."""
+    self.assert_equals([], microformats2.html_to_activities("""\
+<div class="h-feed">
+  <article class="h-card">
+    <a href="http://foo">bar</a>
+  </article>
+</div>"""))
+
   def test_size_to_bytes(self):
     for input, expected in (
         (None, None),