fix uninitialized variable

Author: maxbachmann

VERSION

@@ -1 +1 @@
-1.0.0
+1.0.1

src/py_string_metric.cpp

@@ -132,6 +132,7 @@ PyObject* levenshtein(PyObject* /*self*/, PyObject* args, PyObject* keywds)
     }
   }
 
+  // todo improve error message for None
   if (!valid_str(py_s1, "s1") || !valid_str(py_s2, "s2")) {
     return NULL;
   }
@@ -141,7 +142,7 @@ PyObject* levenshtein(PyObject* /*self*/, PyObject* args, PyObject* keywds)
 
   std::size_t result = mpark::visit(LevenshteinVisitor(insert_cost, delete_cost, replace_cost, (std::size_t)max),
                                     s1_view, s2_view);
-                                  
+
   if (result == (std::size_t)-1) {
     return PyLong_FromLong(-1);
   }
@@ -249,6 +250,7 @@ PyObject* hamming(PyObject* /*self*/, PyObject* args, PyObject* keywds)
     return NULL;
   }
 
+  // todo improve error message for None
   if (!valid_str(py_s1, "s1") || !valid_str(py_s2, "s2")) {
     return NULL;
   }

src/py_string_metric.hpp

@@ -64,7 +64,12 @@ Insertion = 1, Deletion = 1, Substitution = 1:
     Myers algorithm is described in [3]_
 
 
-Insertion = 1, Deletion = 1, Substitution = 2:
+Insertion = 1, Deletion = 1, Substitution >= Insertion + Deletion:
+  when ``Substitution >= Insertion + Deletion`` set
+  ``Substitution = Insertion + Deletion``
+  since every Substitution can be performed as Insertion + Deletion
+  so in this case treat Substitution as 2
+
   - if max is 0 the similarity can be calculated using a direct comparision,
     since no difference between the strings is allowed.  The time complexity of
     this algorithm is ``O(N)``.
@@ -151,9 +156,8 @@ R"(normalized_levenshtein($module, s1, s2, weights = (1, 1, 1), processor = None
 
 Calculates a normalized levenshtein distance using custom
 costs for insertion, deletion and substitution. So far only the following
-combinations are supported:
-- weights = (1, 1, 1)
-- weights = (1, 1, 2)
+weights are supported:
+- weights = (1, 1, N) with N >= 1
 
 further combinations might be supported in the future
 

src/rapidfuzz-cpp

@@ -3,6 +3,6 @@
 """
 __author__ = "Max Bachmann"
 __license__ = "MIT"
-__version__ = "1.0.0"
+__version__ = "1.0.1"
 
 from rapidfuzz import process, fuzz, utils, levenshtein, string_metric

src/rapidfuzz/__init__.py

@@ -0,0 +1,49 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+import unittest
+import pytest
+
+from rapidfuzz import string_metric
+
+def test_empty_string():
+    """
+    when both strings are empty this is a perfect match
+    """
+    assert string_metric.levenshtein("", "") == 0
+    assert string_metric.levenshtein("", "", (1,1,0)) == 0
+    assert string_metric.levenshtein("", "", (1,1,2)) == 0
+    assert string_metric.levenshtein("", "", (1,1,5)) == 0
+    assert string_metric.levenshtein("", "", (3,7,5)) == 0
+
+def test_simple_unicode_tests():
+    """
+    some very simple tests using unicode with scorers
+    to catch relatively obvious implementation errors
+    """
+    s1 = u"ÁÄ"
+    s2 = "ABCD"
+    assert string_metric.levenshtein(s1, s2) == 4
+    assert string_metric.levenshtein(s1, s2, (1,1,0)) == 2
+    assert string_metric.levenshtein(s1, s2, (1,1,2)) == 6
+    assert string_metric.levenshtein(s1, s2, (1,1,5)) == 6
+    assert string_metric.levenshtein(s1, s2, (3,7,5)) == 24
+
+    assert string_metric.levenshtein(s1, s1) == 0
+    assert string_metric.levenshtein(s1, s1, (1,1,0)) == 0
+    assert string_metric.levenshtein(s1, s1, (1,1,2)) == 0
+    assert string_metric.levenshtein(s1, s1, (1,1,5)) == 0
+    assert string_metric.levenshtein(s1, s1, (3,7,5)) == 0
+
+def test_help():
+    """
+    test that all help texts can be printed without throwing an exception,
+    since they are implemented in C++ aswell
+    """
+    help(string_metric.levenshtein)
+    help(string_metric.normalized_levenshtein)
+    help(string_metric.hamming)
+    help(string_metric.normalized_hamming)
+
+if __name__ == '__main__':
+    unittest.main()