Start to add scoring file specific tests

src/breadability/readable.py

@@ -104,7 +104,7 @@ def check_siblings(candidate_node, candidate_list):
             append = True
 
         # Give a bonus if sibling nodes and top candidates have the example
-        # same classname
+        # same class name
         if candidate_css and sibling.get('class') == candidate_css:
             content_bonus += candidate_node.content_score * 0.2
 
@@ -148,7 +148,6 @@ def prep_article(doc):
     - extra tags
 
     """
-
     def clean_document(node):
         """Remove the style attribute on every element."""
         clean_list = ['object', 'h1']
@@ -202,9 +201,6 @@ def prep_article(doc):
     def clean_conditionally(doc, clean_el):
         """Remove the clean_el if it looks like bad content based on rules."""
 
-    def clean_objects():
-        pass
-
     doc = clean_document(doc)
     return doc
 

src/breadability/tests/test_scoring.py

@@ -0,0 +1,42 @@
+from lxml.html import fragment_fromstring
+from unittest import TestCase
+
+from breadability.scoring import check_node_attr
+
+
+class TestCheckNodeAttr(TestCase):
+    """Verify a node has a class/id in the given set.
+
+    The idea is that we have sets of known good/bad ids and classes and need
+    to verify the given node does/doesn't have those classes/ids.
+
+    """
+    def test_has_class(self):
+        """Verify that a node has a class in our set."""
+        test_set = set(['test1', 'test2'])
+        test_node = fragment_fromstring('<div/>')
+        test_node.set('class', 'test2 comment')
+
+        self.assertTrue(check_node_attr(test_node, 'class', test_set))
+
+    def test_has_id(self):
+        """Verify that a node has an id in our set."""
+        test_set = set(['test1', 'test2'])
+        test_node = fragment_fromstring('<div/>')
+        test_node.set('id', 'test2')
+
+        self.assertTrue(check_node_attr(test_node, 'id', test_set))
+
+    def test_lacks_class(self):
+        """Verify that a node does not have a class in our set."""
+        test_set = set(['test1', 'test2'])
+        test_node = fragment_fromstring('<div/>')
+        test_node.set('class', 'test4 comment')
+        self.assertFalse(check_node_attr(test_node, 'class', test_set))
+
+    def test_lacks_id(self):
+        """Verify that a node does not have an id in our set."""
+        test_set = set(['test1', 'test2'])
+        test_node = fragment_fromstring('<div/>')
+        test_node.set('id', 'test4')
+        self.assertFalse(check_node_attr(test_node, 'id', test_set))