Merge pull request buriy#128 from azmeuk/self-closing

buriy · web-flow · commit da9e285f737f · 2019-12-30T18:37:54.000+07:00
Replaced XHTML output with HTML5 output in summary for empty elements (a, br), issue buriy#125
diff --git a/readability/readability.py b/readability/readability.py
@@ -171,7 +171,7 @@ def get_clean_html(self):
         An internal method, which can be overridden in subclasses, for example,
         to disable or to improve DOM-to-text conversion in .summary() method
         """
-        return clean_attributes(tounicode(self.html))
+        return clean_attributes(tounicode(self.html, method='html'))
 
     def summary(self, html_partial=False):
         """
diff --git a/tests/test_article_only.py b/tests/test_article_only.py
@@ -106,3 +106,8 @@ def test_many_repeated_spaces(self):
         s = doc.summary()
 
         assert 'foo' in s
+
+    def test_not_self_closing(self):
+        sample = '<h2><a href="#"></a>foobar</h2>'
+        doc = Document(sample)
+        assert '<body id="readabilityBody"><h2><a href="#"></a>foobar</h2></body>' == doc.summary()
