|
1
|
|
|
# Copyright 2019-2020 by Christopher C. Little. |
|
2
|
|
|
# This file is part of Abydos. |
|
3
|
|
|
# |
|
4
|
|
|
# Abydos is free software: you can redistribute it and/or modify |
|
5
|
|
|
# it under the terms of the GNU General Public License as published by |
|
6
|
|
|
# the Free Software Foundation, either version 3 of the License, or |
|
7
|
|
|
# (at your option) any later version. |
|
8
|
|
|
# |
|
9
|
|
|
# Abydos is distributed in the hope that it will be useful, |
|
10
|
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11
|
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12
|
|
|
# GNU General Public License for more details. |
|
13
|
|
|
# |
|
14
|
|
|
# You should have received a copy of the GNU General Public License |
|
15
|
|
|
# along with Abydos. If not, see <http://www.gnu.org/licenses/>. |
|
16
|
|
|
|
|
17
|
|
|
"""abydos.distance._isg. |
|
18
|
|
|
|
|
19
|
1 |
|
Bouchard & Pouyez's Indice de Similitude-Guth (ISG) |
|
20
|
|
|
""" |
|
21
|
|
|
|
|
22
|
|
|
from typing import Any |
|
23
|
|
|
|
|
24
|
1 |
|
from ._distance import _Distance |
|
25
|
|
|
|
|
26
|
|
|
__all__ = ['ISG'] |
|
27
|
|
|
|
|
28
|
|
|
|
|
29
|
|
|
class ISG(_Distance): |
|
30
|
|
|
"""Indice de Similitude-Guth (ISG) similarity. |
|
31
|
1 |
|
|
|
32
|
|
|
This is an implementation of Bouchard & Pouyez's Indice de Similitude-Guth |
|
33
|
1 |
|
(ISG) :cite:`Bouchard:1980`. At its heart, ISG is Jaccard similarity, but |
|
34
|
|
|
limits on token matching are added according to part of Guth's matching |
|
35
|
|
|
criteria :cite:`Guth:1976`. |
|
36
|
1 |
|
|
|
37
|
|
|
:cite:`Bouchard:1980` is limited in its implementation details. Based on |
|
38
|
|
|
the examples given in the paper, it appears that only the first 4 of Guth's |
|
39
|
|
|
rules are considered (a letter in the first string must match a letter in |
|
40
|
|
|
the second string appearing in the same position, an adjacent position, or |
|
41
|
|
|
two positions ahead). It also appears that the distance in the paper is |
|
42
|
|
|
the greater of the distance from string 1 to string 2 and the distance |
|
43
|
|
|
from string 2 to string 1. |
|
44
|
|
|
|
|
45
|
|
|
These qualities can be specified as parameters. At initialization, specify |
|
46
|
|
|
``full_guth=True`` to apply all of Guth's rules and ``symmetric=False`` to |
|
47
|
|
|
calculate only the distance from string 1 to string 2. |
|
48
|
|
|
|
|
49
|
|
|
.. versionadded:: 0.4.1 |
|
50
|
|
|
""" |
|
51
|
|
|
|
|
52
|
|
|
def __init__( |
|
53
|
|
|
self, full_guth: bool = False, symmetric: bool = True, **kwargs: Any |
|
54
|
|
|
) -> None: |
|
55
|
|
|
"""Initialize ISG instance. |
|
56
|
|
|
|
|
57
|
|
|
Parameters |
|
58
|
|
|
---------- |
|
59
|
1 |
|
full_guth : bool |
|
60
|
|
|
Whether to apply all of Guth's matching rules |
|
61
|
|
|
symmetric : bool |
|
62
|
|
|
Whether to calculate the symmetric distance |
|
63
|
|
|
**kwargs |
|
64
|
|
|
Arbitrary keyword arguments |
|
65
|
|
|
|
|
66
|
|
|
|
|
67
|
|
|
.. versionadded:: 0.4.1 |
|
68
|
|
|
|
|
69
|
|
|
""" |
|
70
|
|
|
super(ISG, self).__init__(**kwargs) |
|
71
|
|
|
self._full_guth = full_guth |
|
72
|
|
|
self._symmetric = symmetric |
|
73
|
|
|
|
|
74
|
|
|
def _isg_i(self, src: str, tar: str) -> float: |
|
75
|
1 |
|
"""Return an individual ISG similarity (not symmetric) for src to tar. |
|
76
|
1 |
|
|
|
77
|
1 |
|
Parameters |
|
78
|
|
|
---------- |
|
79
|
1 |
|
src : str |
|
80
|
|
|
Source string for comparison |
|
81
|
|
|
tar : str |
|
82
|
|
|
Target string for comparison |
|
83
|
|
|
|
|
84
|
|
|
Returns |
|
85
|
|
|
------- |
|
86
|
|
|
float |
|
87
|
|
|
The ISG similarity |
|
88
|
|
|
|
|
89
|
|
|
|
|
90
|
|
|
.. versionadded:: 0.4.1 |
|
91
|
|
|
|
|
92
|
|
|
""" |
|
93
|
|
|
|
|
94
|
|
|
def _char_at(name: str, pos: int) -> str: |
|
95
|
|
|
if pos >= len(name): |
|
96
|
|
|
return '' |
|
97
|
|
|
return name[pos] |
|
98
|
|
|
|
|
99
|
1 |
|
matches = 0 |
|
100
|
1 |
|
for pos in range(len(src)): |
|
101
|
1 |
|
s = _char_at(src, pos) |
|
102
|
1 |
|
if s and s in set(tar[max(0, pos - 1) : pos + 3]): |
|
103
|
|
|
matches += 1 |
|
104
|
1 |
|
continue |
|
105
|
1 |
|
|
|
106
|
1 |
|
if self._full_guth: |
|
107
|
1 |
|
t = _char_at(tar, pos) |
|
108
|
1 |
|
if t and t in set(src[max(0, pos - 1) : pos + 3]): |
|
109
|
1 |
|
matches += 1 |
|
110
|
1 |
|
continue |
|
111
|
|
|
|
|
112
|
1 |
|
s = _char_at(src, pos + 1) |
|
113
|
1 |
|
t = _char_at(tar, pos + 1) |
|
114
|
1 |
|
if s and t and s == t: |
|
115
|
1 |
|
matches += 1 |
|
116
|
1 |
|
continue |
|
117
|
1 |
|
|
|
118
|
|
|
s = _char_at(src, pos + 2) |
|
119
|
1 |
|
t = _char_at(tar, pos + 2) |
|
120
|
1 |
|
if s and t and s == t: |
|
121
|
1 |
|
matches += 1 |
|
122
|
1 |
|
continue |
|
123
|
1 |
|
|
|
124
|
|
|
return matches / (len(src) + len(tar) - matches) |
|
125
|
1 |
|
|
|
126
|
1 |
|
def sim(self, src: str, tar: str) -> float: |
|
127
|
1 |
|
"""Return the Indice de Similitude-Guth (ISG) similarity of two words. |
|
128
|
1 |
|
|
|
129
|
1 |
|
Parameters |
|
130
|
|
|
---------- |
|
131
|
1 |
|
src : str |
|
132
|
|
|
Source string for comparison |
|
133
|
1 |
|
tar : str |
|
134
|
|
|
Target string for comparison |
|
135
|
|
|
|
|
136
|
|
|
Returns |
|
137
|
|
|
------- |
|
138
|
|
|
float |
|
139
|
|
|
The ISG similarity |
|
140
|
|
|
|
|
141
|
|
|
Examples |
|
142
|
|
|
-------- |
|
143
|
|
|
>>> cmp = ISG() |
|
144
|
|
|
>>> cmp.sim('cat', 'hat') |
|
145
|
|
|
0.5 |
|
146
|
|
|
>>> cmp.sim('Niall', 'Neil') |
|
147
|
|
|
0.5 |
|
148
|
|
|
>>> cmp.sim('aluminum', 'Catalan') |
|
149
|
|
|
0.15384615384615385 |
|
150
|
|
|
>>> cmp.sim('ATCG', 'TAGC') |
|
151
|
|
|
1.0 |
|
152
|
|
|
|
|
153
|
|
|
|
|
154
|
|
|
.. versionadded:: 0.4.1 |
|
155
|
|
|
|
|
156
|
|
|
""" |
|
157
|
|
|
if src == tar: |
|
158
|
|
|
return 1.0 |
|
159
|
|
|
if len(src) > len(tar): |
|
160
|
|
|
src, tar = tar, src |
|
161
|
|
|
elif self._symmetric and len(src) == len(tar): |
|
162
|
|
|
return max(self._isg_i(src, tar), self._isg_i(tar, src)) |
|
163
|
|
|
return self._isg_i(src, tar) |
|
164
|
1 |
|
|
|
165
|
1 |
|
|
|
166
|
1 |
|
if __name__ == '__main__': |
|
167
|
1 |
|
import doctest |
|
168
|
1 |
|
|
|
169
|
|
|
doctest.testmod() |
|
170
|
|
|
|
chrislit /
abydos
