bumped version to 0.2.1 and updated HISTORY

HISTORY.rst

@@ -3,6 +3,17 @@
 History
 -------
 
+0.2.1 (2014-03-14)
+++++++++++++++++++
+
+* Fixed major match grouping bug
+
+0.2.0 (2013-03-13)
+++++++++++++++++++
+
+* New utility function `find_near_matches()` for easier use
+* Additional documentation
+
 0.1.0 (2013-11-12)
 ++++++++++++++++++
 

fuzzysearch/__init__.py

@@ -13,7 +13,7 @@ Example:
 """
 __author__ = 'Tal Einat'
 __email__ = 'taleinat@gmail.com'
-__version__ = '0.2.0'
+__version__ = '0.2.1'
 
 __all__ = [
     'find_near_matches_with_ngrams',

setup.py

@@ -19,7 +19,7 @@ history = open('HISTORY.rst').read().replace('.. :changelog:', '')
 
 setup(
     name='fuzzysearch',
-    version='0.2.0',
+    version='0.2.1',
     description='fuzzysearch is useful for finding approximate subsequence matches',
     long_description=readme + '\n\n' + history,
     author='Tal Einat',

tests/test_fuzzysearch.py

@@ -102,7 +102,6 @@ class TestFuzzySearchBase(object):
     def test_substring(self):
         substring = 'PATTERN'
         text = 'aaaaaaaaaaPATTERNaaaaaaaaa'
-        idx = text.find(substring)
         expected_match = Match(start=10, end=17, dist=0)
 
         self.assertEquals(
@@ -265,6 +264,36 @@ class TestFuzzySearchBase(object):
             self.search(pattern, text, max_l_dist=2),
         )
 
+    def test_protein_search1(self):
+        # see: BioPython archives from March 14th, 2014
+        text = ''.join('''\
+            XXXXXXXXXXXXXXXXXXXGGGTTVTTSSAAAAAAAAAAAAAGGGTTLTTSSAAAAAAAAAAAA
+            AAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBGGGTTLTTSS
+        '''.split())
+        pattern = "GGGTTLTTSS"
+
+        self.assertItemsEqual(
+            [Match(start=19, end=29, dist=1),
+             Match(start=42, end=52, dist=0),
+             Match(start=99, end=109, dist=0)],
+            self.search(pattern, text, max_l_dist=2),
+        )
+
+    def test_protein_search2(self):
+        # see: BioPython archives from March 14th, 2014
+        text = ''.join('''\
+            XXXXXXXXXXXXXXXXXXXGGGTTVTTSSAAAAAAAAAAAAAGGGTTVTTSSAAAAAAAAAAA
+            AAAAAAAAAAABBBBBBBBBBBBBBBBBBBBBBBBBGGGTTLTTSS
+        '''.split())
+        pattern = "GGGTTLTTSS"
+
+        self.assertItemsEqual(
+            [Match(start=19, end=29, dist=1),
+             Match(start=42, end=52, dist=1),
+             Match(start=99, end=109, dist=0)],
+            self.search(pattern, text, max_l_dist=2),
+        )
+
 
 class TestFindNearMatchesWithNgrams(TestFuzzySearchBase, unittest.TestCase):
     def search(self, subsequence, sequence, max_l_dist):