added Ozbay metric

abydos/distance/__init__.py

@@ -92,6 +92,7 @@
     - Sift4 distance (:py:class:`.Sift4` and :py:class:`.Sift4Simplest`)
     - Typo distance (:py:class:`.Typo`)
     - Synoname (:py:class:`.Synoname`)
+    - Ozbay metric (:py:class:`.Ozbay`)
 
 Most of the distance and similarity measures have ``sim`` and ``dist`` methods,
 which return a measure that is normalized to the range :math:`[0, 1]`. The
@@ -164,6 +165,7 @@
 from ._ncd_zlib import NCDzlib, dist_ncd_zlib, sim_ncd_zlib
 from ._needleman_wunsch import NeedlemanWunsch, needleman_wunsch
 from ._overlap import Overlap, dist_overlap, sim_overlap
+from ._ozbay import Ozbay
 from ._prefix import Prefix, dist_prefix, sim_prefix
 from ._ratcliff_obershelp import (
     RatcliffObershelp,
@@ -318,6 +320,7 @@
     'sim_typo',
     'Synoname',
     'synoname',
+    'Ozbay',
 ]
 
 

abydos/distance/_ozbay.py

@@ -0,0 +1,143 @@
+# -*- coding: utf-8 -*-
+
+# Copyright 2018 by Christopher C. Little.
+# This file is part of Abydos.
+#
+# Abydos is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Abydos is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Abydos. If not, see <http://www.gnu.org/licenses/>.
+
+"""abydos.distance._ozbay.
+
+Ozbay metric
+"""
+
+from __future__ import (
+    absolute_import,
+    division,
+    print_function,
+    unicode_literals,
+)
+
+from . import Jaccard, LCSstr, Levenshtein
+from ._distance import _Distance
+from ..tokenizer import QGrams
+
+__all__ = ['Ozbay']
+
+
+class Ozbay(_Distance):
+    """Ozbay metric.
+
+    The Ozbay metric :cite:`Ozbay:2015` is a string distance measure developed
+    by Hakan Ozbay, which combines Jaccard distance, Levenshtein distance, and
+    longest common substring distance.
+
+    The normalized variant should be considered experimental.
+
+    .. versionadded:: 0.4.0
+
+    """
+
+    _lev = Levenshtein()
+    _jac = Jaccard(tokenizer=QGrams(qval=1, start_stop='', scaler='set'))
+    _lcs = LCSstr()
+
+    def dist_abs(self, src, tar):
+        """Return the Ozbay metric.
+
+        Parameters
+        ----------
+        src : str
+            Source string for comparison
+        tar : str
+            Target string for comparison
+
+        Returns
+        -------
+        float
+            Ozbay metric
+
+        Examples
+        --------
+        >>> cmp = Ozbay()
+        >>> round(cmp.dist_abs('cat', 'hat'), 12)
+        0.75
+        >>> round(cmp.dist_abs('Niall', 'Neil'), 12)
+        6.0
+        >>> round(cmp.dist_abs('Colin', 'Cuilen'), 12)
+        7.714285714286
+        >>> cmp.dist_abs('ATCG', 'TAGC')
+        3.0
+
+        .. versionadded:: 0.4.0
+
+        """
+        lev_dist = self._lev.dist_abs(src, tar)
+        lev_metric = 0.0 if lev_dist == 0 else lev_dist / len(src)
+        jac_metric = self._jac.dist_abs(src, tar)
+        lcs_metric = self._lcs.sim(src, tar)
+
+        if jac_metric == 1.0:
+            ozbay_metric = lev_dist
+        elif jac_metric == 0.0:
+            ozbay_metric = lev_metric
+        else:
+            ozbay_metric = jac_metric * lev_dist
+
+        if lcs_metric > 0.0:
+            ozbay_metric /= lcs_metric
+        else:
+            ozbay_metric *= min(len(src), len(tar))
+
+        return ozbay_metric
+
+    def dist(self, src, tar):
+        """Return the normalized Ozbay distance.
+
+        Parameters
+        ----------
+        src : str
+            Source string for comparison
+        tar : str
+            Target string for comparison
+
+        Returns
+        -------
+        float
+            Normalized Ozbay distance
+
+        Examples
+        --------
+        >>> cmp = Ozbay()
+        >>> round(cmp.dist('cat', 'hat'), 12)
+        0.027777777778
+        >>> round(cmp.dist('Niall', 'Neil'), 12)
+        0.24
+        >>> round(cmp.dist('Colin', 'Cuilen'), 12)
+        0.214285714286
+        >>> cmp.dist('ATCG', 'TAGC')
+        0.140625
+
+        .. versionadded:: 0.4.0
+
+        """
+        dist = self.dist_abs(src, tar)
+        if dist:
+            return dist / (len(src) * len(tar) / self._lcs.dist(src, tar))
+        return dist
+
+
+if __name__ == '__main__':
+    import doctest
+
+    doctest.testmod()

docs/abydos.bib

@@ -888,6 +888,15 @@ @Article{Otsuka:1936
   Volume                   = {6}
 }
 
+@Misc{Ozbay:2015,
+  Title                    = {Ozbay metric},
+
+  Author                   = {Ozbay, Hakan},
+  Year                     = {2015},
+
+  Url                      = {https://github.com/hakanozbay/ozbay-metric}
+}
+
 @InProceedings{Paice:1990,
   Title                    = {Another stemmer},
   Author                   = {Paice, {Chris D.}},

tests/distance/test_distance_ozbay.py

@@ -0,0 +1,140 @@
+# -*- coding: utf-8 -*-
+
+# Copyright 2018 by Christopher C. Little.
+# This file is part of Abydos.
+#
+# Abydos is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Abydos is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Abydos. If not, see <http://www.gnu.org/licenses/>.
+
+"""abydos.tests.distance.test_distance_ozbay
+
+This module contains unit tests for abydos.distance.Ozbay
+"""
+
+from __future__ import (
+    absolute_import,
+    division,
+    print_function,
+    unicode_literals,
+)
+
+import unittest
+
+from abydos.distance import Ozbay
+
+
+class OzbayTestCases(unittest.TestCase):
+    """Test Ozbay metric functions.
+
+    abydos.distance.Ozbay
+    """
+
+    cmp = Ozbay()
+
+    def test_ozbay_dist_abs(self):
+        """Test abydos.distance.Ozbay.dist_abs."""
+        self.assertEqual(self.cmp.dist_abs('', ''), 0.0)
+
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('piccadilly', 'bandage'), 73.63636363636363
+        )
+        self.assertAlmostEqual(self.cmp.dist_abs('abcd', 'efgh'), 16)
+
+        # Test cases from https://github.com/hakanozbay/ozbay-metric
+        self.assertEqual(self.cmp.dist_abs('ban', 'ban'), 0.0)
+        self.assertAlmostEqual(self.cmp.dist_abs('ban', 'bane'), 0.3333333333)
+        self.assertAlmostEqual(self.cmp.dist_abs('ban', 'band'), 0.3333333333)
+        self.assertEqual(self.cmp.dist_abs('ban', 'bat'), 0.75)
+        self.assertAlmostEqual(self.cmp.dist_abs('ban', 'bands'), 1.3333333333)
+        self.assertEqual(self.cmp.dist_abs('ban', 'banana'), 2.0)
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('ban', 'bandana'), 2.3333333333
+        )
+        self.assertEqual(self.cmp.dist_abs('ban', 'bandit'), 3.0)
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('ban', 'bandage'), 4.6666666666
+        )
+
+        self.assertEqual(self.cmp.dist_abs('piccadilly', 'piccadilly'), 0.0)
+        self.assertEqual(self.cmp.dist_abs('piccadilly', 'piccadilyl'), 0.25)
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('piccadilly', 'piccadlily'), 0.3333333333
+        )
+        self.assertEqual(self.cmp.dist_abs('piccadilly', 'picacdilly'), 0.4)
+        self.assertEqual(self.cmp.dist_abs('piccadilly', 'picadily'), 0.4)
+        self.assertEqual(self.cmp.dist_abs('picadily', 'piccadilly'), 0.5)
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('piccadilly', 'picacdlily'), 1.3333333333
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('ipcacdily', 'piccadilly'), 1.4814814814814814
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist_abs('piccadilly', 'ipcacdily'), 1.333333333
+        )
+        self.assertEqual(self.cmp.dist_abs('piccadilly', 'pcicadlyil'), 2.0)
+
+    def test_ozbay_dist(self):
+        """Test abydos.distance.Ozbay.dist."""
+        self.assertEqual(self.cmp.dist('', ''), 0)
+
+        self.assertAlmostEqual(
+            self.cmp.dist('piccadilly', 'bandage'), 0.9467532467532467
+        )
+        self.assertAlmostEqual(self.cmp.dist('abcd', 'efgh'), 1.0)
+
+        # Test cases from https://github.com/hakanozbay/ozbay-metric
+        self.assertEqual(self.cmp.dist('ban', 'ban'), 0.0)
+        self.assertAlmostEqual(
+            self.cmp.dist('ban', 'bane'), 0.006944444444444444
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist('ban', 'band'), 0.006944444444444444
+        )
+        self.assertEqual(self.cmp.dist('ban', 'bat'), 0.02777777777777778)
+        self.assertAlmostEqual(
+            self.cmp.dist('ban', 'bands'), 0.03555555555555556
+        )
+        self.assertEqual(self.cmp.dist('ban', 'banana'), 0.05555555555555555)
+        self.assertAlmostEqual(
+            self.cmp.dist('ban', 'bandana'), 0.0634920634920635
+        )
+        self.assertEqual(self.cmp.dist('ban', 'bandit'), 0.08333333333333333)
+        self.assertAlmostEqual(
+            self.cmp.dist('ban', 'bandage'), 0.126984126984127
+        )
+
+        self.assertEqual(self.cmp.dist('piccadilly', 'piccadilly'), 0.0)
+        self.assertEqual(
+            self.cmp.dist('piccadilly', 'piccadilyl'), 0.0004999999999999999
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist('piccadilly', 'piccadlily'), 0.0013333333333333335
+        )
+        self.assertEqual(self.cmp.dist('piccadilly', 'picacdilly'), 0.002)
+        self.assertEqual(self.cmp.dist('piccadilly', 'picadily'), 0.0025)
+        self.assertEqual(self.cmp.dist('picadily', 'piccadilly'), 0.003125)
+        self.assertAlmostEqual(
+            self.cmp.dist('piccadilly', 'picacdlily'), 0.009333333333333334
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist('ipcacdily', 'piccadilly'), 0.011522633744855966
+        )
+        self.assertAlmostEqual(
+            self.cmp.dist('piccadilly', 'ipcacdily'), 0.01037037037037037
+        )
+        self.assertEqual(self.cmp.dist('piccadilly', 'pcicadlyil'), 0.014)
+
+
+if __name__ == '__main__':
+    unittest.main()