Complete upstream merge

12 years ago · 7c220535df
parent be77f99be1 c8c53b304b
commit 7c220535df
3 changed files with 31 additions and 2 deletions
--- a/.gitignore
+++ b/.gitignore
@ -1,4 +1,5 @@
 *.pyc
+*.prof

 .installed.cfg
 bin
--- a/src/breadability/scoring.py
+++ b/src/breadability/scoring.py
@ -128,7 +128,7 @@ def score_candidates(nodes):

        # For every 100 characters in this paragraph, add another point. Up to
        # 3 points.
-        length_points = len(innertext) % 100 if innertext else 0
+        length_points = len(innertext) / 100 if innertext else 0
        if length_points > 3:
            content_score += 3
        else:
--- a/src/breadability/tests/test_readable.py
+++ b/src/breadability/tests/test_readable.py
@ -146,7 +146,6 @@ class TestCandidateNodes(TestCase):
            doc = fragment_fromstring(n)
            self.assertEqual(ScoredNode(doc).content_score, -5)

-
    def test_article_enables_candidate_access(self):
        """Candidates are accessible after document processing."""
        doc = Article(load_article('ars/ars.001.html'))
@ -199,6 +198,35 @@ class TestScoringNodes(TestCase):
        scores = sorted([c.content_score for c in candidates.values()])
        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.getiterator():
+                if node.tag == 'p':
+                    test_nodes.append(node)
+            return test_nodes
+
+        test_nodes = build_doc(400)
+        candidates = score_candidates(test_nodes)
+        pscore_400 = max([c.content_score for c in candidates.values()])
+
+        test_nodes = build_doc(100)
+        candidates = score_candidates(test_nodes)
+        pscore_100 = max([c.content_score for c in candidates.values()])
+
+        test_nodes = build_doc(50)
+        candidates = score_candidates(test_nodes)
+        pscore_50 = max([c.content_score for c in candidates.values()])
+
+        self.assertEqual(pscore_100, pscore_50 + 1)
+        self.assertEqual(pscore_400, pscore_50 + 3)
+

 class TestLinkDensityScoring(TestCase):
    """Link density will adjust out candidate scoresself."""