Added a feature to allow more images per article (with a test)

readability/readability.py

@@ -482,7 +482,7 @@ class Document:
 
                 #if el.tag == 'div' and counts["img"] >= 1:
                 #    continue
-                if counts["p"] and counts["img"] > counts["p"]:
+                if counts["p"] and counts["img"] > 1+counts["p"]*1.3:
                     reason = "too many images (%s)" % counts["img"]
                     to_remove = True
                 elif counts["li"] > counts["p"] and tag != "ul" and tag != "ol":

tests/test_article_only.py

@@ -37,3 +37,10 @@ class TestArticleOnly(unittest.TestCase):
         res = doc.summary(html_partial=True)
         self.assertEqual('<div><div class="', res[0:17])
 
+    def test_too_many_images_sample_html_partial(self):
+        """Using the too-many-images sample, make sure we still get the article."""
+        sample = load_sample('too-many-images.sample.html')
+        doc = Document(sample)
+        res = doc.summary(html_partial=True)
+        self.assertEqual('<div><div class="post-body', res[0:26])
+