Move the tests to the testfile

jcharum · mitechie · commit 1e30e33302a6 · 2012-04-21T12:48:32.000-04:00
diff --git a/src/readability_lxml/readability.py b/src/readability_lxml/readability.py
@@ -139,6 +139,13 @@ def clean_segment_number(num_segments, index, segment):
         return segment
 
 
+def clean_segment_index(num_segments, index, segment):
+    if index == (num_segments - 1) and segment.lower() == 'index':
+        return None
+    else:
+        return segment
+
+
 def clean_segment(num_segments, index, segment):
     """
     Cleans a single segment of a URL to find the base URL.  The base URL is as
@@ -150,7 +157,8 @@ def clean_segment(num_segments, index, segment):
             clean_segment_extension,
             clean_segment_ewcms,
             clean_segment_page_number,
-            clean_segment_number
+            clean_segment_number,
+            clean_segment_index
             ]
     cleaned_segment = segment
     for func in funcs:
diff --git a/src/tests/test_readability.py b/src/tests/test_readability.py
@@ -110,3 +110,17 @@ def test_numbers(self):
                 ]
         self._run_urls(specs)
 
+    def test_index(self):
+        specs = [
+                (
+                'http://foo.com/index.html',
+                'http://foo.com',
+                'index should be stripped'
+                ),
+                (
+                'http://foo.com/path/to/index.html',
+                'http://foo.com/path/to',
+                'index should be stripped'
+                )
+                ]
+        self._run_urls(specs)
