Use Levenshtein not nltk

common/djangoapps/util/password_policy_validators.py

@@ -10,10 +10,11 @@ from __future__ import division
 import string
 import unicodedata
 
-from nltk.metrics.distance import edit_distance
 from django.conf import settings
 from django.core.exceptions import ValidationError
 from django.utils.translation import ugettext_lazy as _
+from Levenshtein import distance
+from six import text_type
 
 
 def validate_password_strength(value):
@@ -118,6 +119,6 @@ def validate_password_dictionary(value):
 
     if password_max_edit_distance and password_dictionary:
         for word in password_dictionary:
-            distance = edit_distance(value, word)
-            if distance <= password_max_edit_distance:
+            edit_distance = distance(text_type(value), text_type(word))
+            if edit_distance <= password_max_edit_distance:
                 raise ValidationError(_("Too similar to a restricted dictionary word."), code="dictionary_word")

common/djangoapps/util/tests/test_password_policy_validators.py

@@ -0,0 +1,31 @@
+"""Tests for util.password_policy_validators module."""
+
+import unittest
+
+from django.core.exceptions import ValidationError
+from django.test.utils import override_settings
+
+from util.password_policy_validators import validate_password_dictionary
+
+
+class PasswordPolicyValidatorsTestCase(unittest.TestCase):
+    """ Tests for password validator utility functions """
+
+    @override_settings(PASSWORD_DICTIONARY_EDIT_DISTANCE_THRESHOLD=2)
+    @override_settings(PASSWORD_DICTIONARY=['testme'])
+    def test_validate_password_dictionary(self):
+        """ Tests dictionary checks """
+        # Direct match
+        with self.assertRaises(ValidationError):
+            validate_password_dictionary('testme')
+
+        # Off by one
+        with self.assertRaises(ValidationError):
+            validate_password_dictionary('estme')
+
+        # Off by two
+        with self.assertRaises(ValidationError):
+            validate_password_dictionary('bestmet')
+
+        # Off by three (should pass)
+        validate_password_dictionary('bestem')

requirements/edx/base.txt

@@ -85,7 +85,6 @@ Markdown>=2.6,<2.7
 mongoengine==0.10.0
 MySQL-python==1.2.5
 networkx==1.7
-nltk==3.2.5
 nose-xunitmp==0.3.2
 oauthlib==1.0.3
 path.py==8.2.1
@@ -105,6 +104,7 @@ python-memcached==1.48
 django-memcached-hashring==0.1.2
 python-openid==2.2.5
 python-dateutil==2.1
+python-Levenshtein==0.12.0
 social-auth-app-django==1.2.0
 social-auth-core==1.4.0
 pytz==2016.7