|
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._rees_levenshtein. |
|
18
|
|
|
|
|
19
|
1 |
|
Rees-Levenshtein distance |
|
20
|
|
|
""" |
|
21
|
|
|
|
|
22
|
|
|
from typing import Any, Callable, List, cast |
|
23
|
|
|
|
|
24
|
1 |
|
from numpy import int_ as np_int |
|
25
|
|
|
from numpy import zeros as np_zeros |
|
26
|
|
|
|
|
27
|
|
|
from ._distance import _Distance |
|
28
|
|
|
|
|
29
|
|
|
__all__ = ['ReesLevenshtein'] |
|
30
|
|
|
|
|
31
|
1 |
|
|
|
32
|
1 |
|
class ReesLevenshtein(_Distance): |
|
33
|
|
|
r"""Rees-Levenshtein distance. |
|
34
|
1 |
|
|
|
35
|
|
|
Rees-Levenshtein distance :cite:`Rees:2014,Rees:2013` is the "Modified |
|
36
|
1 |
|
Damerau-Levenshtein Distance Algorithm, created by Tony Rees as part of |
|
37
|
|
|
Taxamatch. |
|
38
|
|
|
|
|
39
|
1 |
|
.. versionadded:: 0.4.0 |
|
40
|
|
|
""" |
|
41
|
|
|
|
|
42
|
|
|
def __init__( |
|
43
|
|
|
self, |
|
44
|
|
|
block_limit: int = 2, |
|
45
|
|
|
normalizer: Callable[[List[float]], float] = max, |
|
46
|
|
|
**kwargs: Any |
|
47
|
|
|
) -> None: |
|
48
|
|
|
"""Initialize ReesLevenshtein instance. |
|
49
|
1 |
|
|
|
50
|
|
|
Parameters |
|
51
|
|
|
---------- |
|
52
|
|
|
block_limit : int |
|
53
|
|
|
The block length limit |
|
54
|
|
|
normalizer : function |
|
55
|
|
|
A function that takes an list and computes a normalization term |
|
56
|
|
|
by which the edit distance is divided (max by default). Another |
|
57
|
|
|
good option is the sum function. |
|
58
|
|
|
**kwargs |
|
59
|
|
|
Arbitrary keyword arguments |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
.. versionadded:: 0.4.0 |
|
63
|
|
|
|
|
64
|
|
|
""" |
|
65
|
|
|
super(ReesLevenshtein, self).__init__(**kwargs) |
|
66
|
|
|
self._normalizer = normalizer |
|
67
|
1 |
|
self._block_limit = block_limit |
|
68
|
1 |
|
|
|
69
|
1 |
|
def dist_abs(self, src: str, tar: str) -> float: |
|
70
|
|
|
"""Return the Rees-Levenshtein distance of two strings. |
|
71
|
1 |
|
|
|
72
|
|
|
This is a straightforward port of the PL/SQL implementation at |
|
73
|
|
|
https://confluence.csiro.au/public/taxamatch/the-mdld-modified-damerau-levenshtein-distance-algorithm |
|
74
|
|
|
|
|
75
|
|
|
Parameters |
|
76
|
|
|
---------- |
|
77
|
|
|
src : str |
|
78
|
|
|
Source string for comparison |
|
79
|
|
|
tar : str |
|
80
|
|
|
Target string for comparison |
|
81
|
|
|
|
|
82
|
|
|
Returns |
|
83
|
|
|
------- |
|
84
|
|
|
float |
|
85
|
|
|
Rees-Levenshtein distance |
|
86
|
|
|
|
|
87
|
|
|
Examples |
|
88
|
|
|
-------- |
|
89
|
|
|
>>> cmp = ReesLevenshtein() |
|
90
|
|
|
>>> cmp.dist_abs('cat', 'hat') |
|
91
|
|
|
1 |
|
92
|
|
|
>>> cmp.dist_abs('Niall', 'Neil') |
|
93
|
|
|
3 |
|
94
|
|
|
>>> cmp.dist_abs('aluminum', 'Catalan') |
|
95
|
|
|
7 |
|
96
|
|
|
>>> cmp.dist_abs('ATCG', 'TAGC') |
|
97
|
|
|
2 |
|
98
|
|
|
|
|
99
|
|
|
|
|
100
|
|
|
.. versionadded:: 0.4.0 |
|
101
|
|
|
|
|
102
|
|
|
""" |
|
103
|
|
|
v_str1_length = len(src) |
|
104
|
|
|
v_str2_length = len(tar) |
|
105
|
1 |
|
|
|
106
|
1 |
|
if tar == src: |
|
107
|
|
|
return 0 |
|
108
|
1 |
|
if not src: |
|
109
|
1 |
|
return len(tar) |
|
110
|
1 |
|
if not tar: |
|
111
|
1 |
|
return len(src) |
|
112
|
1 |
|
if v_str1_length == 1 and v_str2_length == 1: |
|
113
|
1 |
|
return 1 |
|
114
|
1 |
|
|
|
115
|
1 |
|
def _substr(string: str, start: int, length: int) -> str: |
|
116
|
|
|
if start > 0: |
|
117
|
1 |
|
start -= 1 |
|
118
|
1 |
|
else: |
|
119
|
1 |
|
start += len(string) - 1 |
|
120
|
|
|
|
|
121
|
1 |
|
end = start + length |
|
122
|
|
|
|
|
123
|
1 |
|
return string[start:end] |
|
124
|
|
|
|
|
125
|
1 |
|
v_temp_str1 = str(src) |
|
126
|
|
|
v_temp_str2 = str(tar) |
|
127
|
1 |
|
|
|
128
|
1 |
|
# first trim common leading characters |
|
129
|
|
|
while v_temp_str1[:1] == v_temp_str2[:1]: |
|
130
|
|
|
v_temp_str1 = v_temp_str1[1:] |
|
131
|
1 |
|
v_temp_str2 = v_temp_str2[1:] |
|
132
|
1 |
|
|
|
133
|
1 |
|
# then trim common trailing characters |
|
134
|
|
|
while v_temp_str1[-1:] == v_temp_str2[-1:]: |
|
135
|
|
|
v_temp_str1 = v_temp_str1[:-1] |
|
136
|
1 |
|
v_temp_str2 = v_temp_str2[:-1] |
|
137
|
1 |
|
|
|
138
|
1 |
|
v_str1_length = len(v_temp_str1) |
|
139
|
|
|
v_str2_length = len(v_temp_str2) |
|
140
|
1 |
|
|
|
141
|
1 |
|
# then calculate standard Levenshtein Distance |
|
142
|
|
|
if v_str1_length == 0 or v_str2_length == 0: |
|
143
|
|
|
return max(v_str2_length, v_str1_length) |
|
144
|
1 |
|
if v_str1_length == 1 and v_str2_length == 1: |
|
145
|
1 |
|
return 1 |
|
146
|
1 |
|
|
|
147
|
1 |
|
# create table (NB: this is transposed relative to the PL/SQL version) |
|
148
|
|
|
d_mat = np_zeros((v_str1_length + 1, v_str2_length + 1), dtype=np_int) |
|
149
|
|
|
|
|
150
|
1 |
|
# enter values in first (leftmost) column |
|
151
|
|
|
for i in range(1, v_str1_length + 1): |
|
152
|
|
|
d_mat[i, 0] = i |
|
153
|
1 |
|
# populate remaining columns |
|
154
|
1 |
|
for j in range(1, v_str2_length + 1): |
|
155
|
|
|
d_mat[0, j] = j |
|
156
|
1 |
|
|
|
157
|
1 |
|
for i in range(1, v_str1_length + 1): |
|
158
|
|
|
if v_temp_str1[i - 1] == v_temp_str2[j - 1]: |
|
159
|
1 |
|
v_this_cost = 0 |
|
160
|
1 |
|
else: |
|
161
|
1 |
|
v_this_cost = 1 |
|
162
|
|
|
|
|
163
|
1 |
|
# extension to cover multiple single, double, triple, etc. |
|
164
|
|
|
# character transpositions |
|
165
|
|
|
# that includes calculation of original Levenshtein distance |
|
166
|
|
|
# when no transposition found |
|
167
|
|
|
v_temp_block_length = int( |
|
168
|
|
|
min( |
|
169
|
1 |
|
v_str1_length / 2, v_str2_length / 2, self._block_limit |
|
170
|
|
|
) |
|
171
|
|
|
) |
|
172
|
|
|
|
|
173
|
|
|
while v_temp_block_length >= 1: |
|
174
|
|
|
if ( |
|
175
|
1 |
|
(i >= v_temp_block_length * 2) |
|
176
|
1 |
|
and (j >= v_temp_block_length * 2) |
|
177
|
|
|
and ( |
|
178
|
|
|
_substr( |
|
179
|
|
|
v_temp_str1, |
|
180
|
|
|
i - v_temp_block_length * 2 - 1, |
|
181
|
|
|
v_temp_block_length, |
|
182
|
|
|
) |
|
183
|
|
|
== _substr( |
|
184
|
|
|
v_temp_str2, |
|
185
|
|
|
j - v_temp_block_length - 1, |
|
186
|
|
|
v_temp_block_length, |
|
187
|
|
|
) |
|
188
|
|
|
) |
|
189
|
|
|
and ( |
|
190
|
|
|
_substr( |
|
191
|
|
|
v_temp_str1, |
|
192
|
|
|
i - v_temp_block_length - 1, |
|
193
|
|
|
v_temp_block_length, |
|
194
|
|
|
) |
|
195
|
|
|
== _substr( |
|
196
|
|
|
v_temp_str2, |
|
197
|
|
|
j - v_temp_block_length * 2 - 1, |
|
198
|
|
|
v_temp_block_length, |
|
199
|
|
|
) |
|
200
|
|
|
) |
|
201
|
|
|
): |
|
202
|
|
|
# transposition found |
|
203
|
|
|
d_mat[i, j] = min( |
|
204
|
|
|
d_mat[i, j - 1] + 1, |
|
205
|
1 |
|
d_mat[i - 1, j] + 1, |
|
206
|
|
|
d_mat[ |
|
207
|
|
|
i - v_temp_block_length * 2, |
|
208
|
|
|
j - v_temp_block_length * 2, |
|
209
|
|
|
] |
|
210
|
|
|
+ v_this_cost |
|
211
|
|
|
+ v_temp_block_length |
|
212
|
|
|
- 1, |
|
213
|
|
|
) |
|
214
|
|
|
v_temp_block_length = 0 |
|
215
|
|
|
elif v_temp_block_length == 1: |
|
216
|
1 |
|
# no transposition |
|
217
|
1 |
|
d_mat[i, j] = min( |
|
218
|
|
|
d_mat[i, j - 1] + 1, |
|
219
|
1 |
|
d_mat[i - 1, j] + 1, |
|
220
|
|
|
d_mat[i - 1, j - 1] + v_this_cost, |
|
221
|
|
|
) |
|
222
|
|
|
v_temp_block_length -= 1 |
|
223
|
|
|
|
|
224
|
1 |
|
return cast(float, d_mat[v_str1_length, v_str2_length]) |
|
225
|
|
|
|
|
226
|
1 |
|
def dist(self, src: str, tar: str) -> float: |
|
227
|
|
|
"""Return the normalized Rees-Levenshtein distance of two strings. |
|
228
|
1 |
|
|
|
229
|
|
|
Parameters |
|
230
|
|
|
---------- |
|
231
|
|
|
src : str |
|
232
|
|
|
Source string for comparison |
|
233
|
|
|
tar : str |
|
234
|
|
|
Target string for comparison |
|
235
|
|
|
|
|
236
|
|
|
Returns |
|
237
|
|
|
------- |
|
238
|
|
|
float |
|
239
|
|
|
Normalized Rees-Levenshtein distance |
|
240
|
|
|
|
|
241
|
|
|
Examples |
|
242
|
|
|
-------- |
|
243
|
|
|
>>> cmp = ReesLevenshtein() |
|
244
|
|
|
>>> cmp.dist('cat', 'hat') |
|
245
|
|
|
0.3333333333333333 |
|
246
|
|
|
>>> cmp.dist('Niall', 'Neil') |
|
247
|
|
|
0.6 |
|
248
|
|
|
>>> cmp.dist('aluminum', 'Catalan') |
|
249
|
|
|
0.875 |
|
250
|
|
|
>>> cmp.dist('ATCG', 'TAGC') |
|
251
|
|
|
0.5 |
|
252
|
|
|
|
|
253
|
|
|
|
|
254
|
|
|
.. versionadded:: 0.4.0 |
|
255
|
|
|
|
|
256
|
|
|
""" |
|
257
|
|
|
if src == tar: |
|
258
|
|
|
return 0.0 |
|
259
|
1 |
|
return self.dist_abs(src, tar) / ( |
|
260
|
1 |
|
self._normalizer([len(src), len(tar)]) |
|
261
|
1 |
|
) |
|
262
|
|
|
|
|
263
|
|
|
|
|
264
|
|
|
if __name__ == '__main__': |
|
265
|
|
|
import doctest |
|
266
|
|
|
|
|
267
|
|
|
doctest.testmod() |
|
268
|
|
|
|
chrislit /
abydos
