1 pt for 100 inner text chars is computed as float

11 years ago · f858f0dbb0
parent 31b75c1cd8
commit f858f0dbb0
2 changed files with 26 additions and 26 deletions
--- a/readability/scoring.py
+++ b/readability/scoring.py
@ -183,8 +183,8 @@ def score_candidates(nodes):

            # for every 100 characters in this paragraph, add another point
            # up to 3 points
-            length_points = len(inner_text) // 100
-            content_score += min(length_points, 3)
+            length_points = len(inner_text) / 100
+            content_score += min(length_points, 3.0)
            logger.debug("Length/content points: %d : %f", length_points, content_score)

        # add the score to the parent
@ -195,9 +195,10 @@ def score_candidates(nodes):
        logger.debug("Giving grand bonus points: %f", candidates[grand].content_score)

    for candidate in candidates.values():
-        adjustment = 1 - get_link_density(candidate.node)
-        logger.debug("Getting link density adjustment: %f * %f", candidate.content_score, adjustment)
-        candidate.content_score = candidate.content_score * adjustment
+        adjustment = 1.0 - get_link_density(candidate.node)
+        candidate.content_score *= adjustment
+        logger.debug("Link density adjustment for %s %r: %f",
+            candidate.node.tag, candidate.node.attrib, adjustment)

    return candidates

--- a/tests/test_readable.py
+++ b/tests/test_readable.py
@ -233,33 +233,32 @@ class TestScoringNodes(unittest.TestCase):
        self.assertTrue(scores[-1] > 100)

    def test_bonus_score_per_100_chars_in_p(self):
-        """Nodes get 1pt per 100 characters up to 3 max points"""
-        def build_doc(length):
-            div = '<div id="content" class=""><p>{0}</p></div>'
-            document_str = '<html><body>{0}</body></html>'
-            content = 'c' * length
-            test_div = div.format(content)
-            doc = document_fromstring(document_str.format(test_div))
-            test_nodes = []
-            for node in doc.iter('p'):
-                test_nodes.append(node)
-
-            return test_nodes
-
-        test_nodes = build_doc(400)
+        """Nodes get 1 point per 100 characters up to max. 3 points."""
+        def build_candidates(length):
+            html = "<p>%s</p>" % ("c" * length)
+            node = fragment_fromstring(html)
+
+            return [node]
+
+        test_nodes = build_candidates(50)
+        candidates = score_candidates(test_nodes)
+        pscore_50 = max(c.content_score for c in candidates.values())
+
+        test_nodes = build_candidates(100)
        candidates = score_candidates(test_nodes)
-        pscore_400 = max([c.content_score for c in candidates.values()])
+        pscore_100 = max(c.content_score for c in candidates.values())

-        test_nodes = build_doc(100)
+        test_nodes = build_candidates(300)
        candidates = score_candidates(test_nodes)
-        pscore_100 = max([c.content_score for c in candidates.values()])
+        pscore_300 = max(c.content_score for c in candidates.values())

-        test_nodes = build_doc(50)
+        test_nodes = build_candidates(400)
        candidates = score_candidates(test_nodes)
-        pscore_50 = max([c.content_score for c in candidates.values()])
+        pscore_400 = max(c.content_score for c in candidates.values())

-        self.assertEqual(pscore_100, pscore_50 + 1)
-        self.assertEqual(pscore_400, pscore_50 + 3)
+        self.assertAlmostEqual(pscore_50 + 0.5, pscore_100)
+        self.assertAlmostEqual(pscore_100 + 2.0, pscore_300)
+        self.assertAlmostEqual(pscore_300, pscore_400)


 class TestLinkDensityScoring(unittest.TestCase):