summary refs log tree commit diff
path: root/tests/rest
diff options
context:
space:
mode:
authorPatrick Cloke <clokep@users.noreply.github.com>2022-06-03 12:09:12 -0400
committerGitHub <noreply@github.com>2022-06-03 12:09:12 -0400
commit01df5bacac3aa0e8356fed889ea0b69c4c044535 (patch)
tree46bae7592ce241b95a9a88ecccb9ff8df3717733 /tests/rest
parentWait for lazy join to complete when getting current state (#12872) (diff)
downloadsynapse-01df5bacac3aa0e8356fed889ea0b69c4c044535.tar.xz
Improve URL previews for some pages (#12951)
* Skip `og` and `meta` tags where the value is empty.
* Fallback to the favicon if there are no other images.
* Ignore tags meant for navigation.
Diffstat (limited to 'tests/rest')
-rw-r--r--tests/rest/media/v1/test_html_preview.py37
1 files changed, 36 insertions, 1 deletions
diff --git a/tests/rest/media/v1/test_html_preview.py b/tests/rest/media/v1/test_html_preview.py
index 62e308814d..ea9e5889bf 100644
--- a/tests/rest/media/v1/test_html_preview.py
+++ b/tests/rest/media/v1/test_html_preview.py
@@ -145,7 +145,7 @@ class SummarizeTestCase(unittest.TestCase):
         )
 
 
-class CalcOgTestCase(unittest.TestCase):
+class OpenGraphFromHtmlTestCase(unittest.TestCase):
     if not lxml:
         skip = "url preview feature requires lxml"
 
@@ -235,6 +235,21 @@ class CalcOgTestCase(unittest.TestCase):
 
         self.assertEqual(og, {"og:title": None, "og:description": "Some text."})
 
+        # Another variant is a title with no content.
+        html = b"""
+        <html>
+        <head><title></title></head>
+        <body>
+        <h1>Title</h1>
+        </body>
+        </html>
+        """
+
+        tree = decode_body(html, "http://example.com/test.html")
+        og = parse_html_to_open_graph(tree)
+
+        self.assertEqual(og, {"og:title": "Title", "og:description": "Title"})
+
     def test_h1_as_title(self) -> None:
         html = b"""
         <html>
@@ -250,6 +265,26 @@ class CalcOgTestCase(unittest.TestCase):
 
         self.assertEqual(og, {"og:title": "Title", "og:description": "Some text."})
 
+    def test_empty_description(self) -> None:
+        """Description tags with empty content should be ignored."""
+        html = b"""
+        <html>
+        <meta property="og:description" content=""/>
+        <meta property="og:description"/>
+        <meta name="description" content=""/>
+        <meta name="description"/>
+        <meta name="description" content="Finally!"/>
+        <body>
+        <h1>Title</h1>
+        </body>
+        </html>
+        """
+
+        tree = decode_body(html, "http://example.com/test.html")
+        og = parse_html_to_open_graph(tree)
+
+        self.assertEqual(og, {"og:title": "Title", "og:description": "Finally!"})
+
     def test_missing_title_and_broken_h1(self) -> None:
         html = b"""
         <html>