|
1
|
|
|
# -*- coding: utf-8 -*- |
|
2
|
|
|
|
|
3
|
|
|
# Copyright 2019 by Christopher C. Little. |
|
4
|
|
|
# This file is part of Abydos. |
|
5
|
|
|
# |
|
6
|
|
|
# Abydos is free software: you can redistribute it and/or modify |
|
7
|
|
|
# it under the terms of the GNU General Public License as published by |
|
8
|
|
|
# the Free Software Foundation, either version 3 of the License, or |
|
9
|
|
|
# (at your option) any later version. |
|
10
|
|
|
# |
|
11
|
|
|
# Abydos is distributed in the hope that it will be useful, |
|
12
|
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
13
|
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
14
|
|
|
# GNU General Public License for more details. |
|
15
|
|
|
# |
|
16
|
|
|
# You should have received a copy of the GNU General Public License |
|
17
|
|
|
# along with Abydos. If not, see <http://www.gnu.org/licenses/>. |
|
18
|
|
|
|
|
19
|
1 |
|
"""abydos.distance._ssk. |
|
20
|
|
|
|
|
21
|
|
|
String subsequence kernel (SSK) similarity |
|
22
|
|
|
""" |
|
23
|
|
|
|
|
24
|
1 |
|
from __future__ import ( |
|
25
|
|
|
absolute_import, |
|
26
|
|
|
division, |
|
27
|
|
|
print_function, |
|
28
|
|
|
unicode_literals, |
|
29
|
|
|
) |
|
30
|
|
|
|
|
31
|
1 |
|
from ._token_distance import _TokenDistance |
|
32
|
1 |
|
from ..tokenizer import QSkipgrams |
|
33
|
|
|
|
|
34
|
1 |
|
__all__ = ['SSK'] |
|
35
|
|
|
|
|
36
|
|
|
|
|
37
|
1 |
|
class SSK(_TokenDistance): |
|
38
|
|
|
r"""String subsequence kernel (SSK) similarity. |
|
39
|
|
|
|
|
40
|
|
|
This is based on :cite:`Lodhi:2002`. |
|
41
|
|
|
|
|
42
|
|
|
|
|
43
|
|
|
.. versionadded:: 0.4.1 |
|
44
|
|
|
""" |
|
45
|
|
|
|
|
46
|
1 |
|
def __init__(self, tokenizer=None, ssk_lambda=0.9, **kwargs): |
|
47
|
|
|
"""Initialize SSK instance. |
|
48
|
|
|
|
|
49
|
|
|
Parameters |
|
50
|
|
|
---------- |
|
51
|
|
|
tokenizer : _Tokenizer |
|
52
|
|
|
A tokenizer instance from the :py:mod:`abydos.tokenizer` package |
|
53
|
|
|
ssk_lambda : float or Iterable |
|
54
|
|
|
A value in the range (0.0, 1.0) used for discouting gaps between |
|
55
|
|
|
characters according to the method described in :cite:`Lodhi:2002`. |
|
56
|
|
|
To supply multiple values of lambda, provide an Iterable of numeric |
|
57
|
|
|
values, such as (0.5, 0.05) or np.arange(0.05, 0.5, 0.05) |
|
58
|
|
|
**kwargs |
|
59
|
|
|
Arbitrary keyword arguments |
|
60
|
|
|
|
|
61
|
|
|
Other Parameters |
|
62
|
|
|
---------------- |
|
63
|
|
|
qval : int |
|
64
|
|
|
The length of each q-skipgram. Using this parameter and |
|
65
|
|
|
tokenizer=None will cause the instance to use the QGramskipgrams |
|
66
|
|
|
tokenizer with this q value. |
|
67
|
|
|
|
|
68
|
|
|
|
|
69
|
|
|
.. versionadded:: 0.4.1 |
|
70
|
|
|
|
|
71
|
|
|
""" |
|
72
|
1 |
|
super(SSK, self).__init__( |
|
73
|
|
|
tokenizer=tokenizer, ssk_lambda=ssk_lambda, **kwargs |
|
74
|
|
|
) |
|
75
|
|
|
|
|
76
|
1 |
|
qval = 2 if 'qval' not in self.params else self.params['qval'] |
|
77
|
1 |
|
self.params['tokenizer'] = ( |
|
78
|
|
|
tokenizer |
|
79
|
|
|
if tokenizer is not None |
|
80
|
|
|
else QSkipgrams( |
|
81
|
|
|
qval=qval, start_stop='', scaler='SSK', ssk_lambda=ssk_lambda |
|
82
|
|
|
) |
|
83
|
|
|
) |
|
84
|
|
|
|
|
85
|
1 |
|
def sim_score(self, src, tar): |
|
86
|
|
|
"""Return the SSK similarity of two strings. |
|
87
|
|
|
|
|
88
|
|
|
Parameters |
|
89
|
|
|
---------- |
|
90
|
|
|
src : str |
|
91
|
|
|
Source string for comparison |
|
92
|
|
|
tar : str |
|
93
|
|
|
Target string for comparison |
|
94
|
|
|
|
|
95
|
|
|
Returns |
|
96
|
|
|
------- |
|
97
|
|
|
float |
|
98
|
|
|
String subsequence kernel similarity |
|
99
|
|
|
|
|
100
|
|
|
Examples |
|
101
|
|
|
-------- |
|
102
|
|
|
>>> cmp = SSK() |
|
103
|
|
|
>>> cmp.dist_abs('cat', 'hat') |
|
104
|
|
|
0.6441281138790036 |
|
105
|
|
|
>>> cmp.dist_abs('Niall', 'Neil') |
|
106
|
|
|
0.5290992177869402 |
|
107
|
|
|
>>> cmp.dist_abs('aluminum', 'Catalan') |
|
108
|
|
|
0.862398428061774 |
|
109
|
|
|
>>> cmp.dist_abs('ATCG', 'TAGC') |
|
110
|
|
|
0.38591004719395017 |
|
111
|
|
|
|
|
112
|
|
|
|
|
113
|
|
|
.. versionadded:: 0.4.1 |
|
114
|
|
|
|
|
115
|
|
|
""" |
|
116
|
1 |
|
self._tokenize(src, tar) |
|
117
|
|
|
|
|
118
|
1 |
|
src_wts = self._src_tokens |
|
119
|
1 |
|
tar_wts = self._tar_tokens |
|
120
|
|
|
|
|
121
|
1 |
|
score = sum( |
|
122
|
|
|
src_wts[token] * tar_wts[token] for token in src_wts & tar_wts |
|
123
|
|
|
) |
|
124
|
|
|
|
|
125
|
1 |
|
return score |
|
126
|
|
|
|
|
127
|
1 |
|
def sim(self, src, tar): |
|
128
|
|
|
"""Return the normalized SSK similarity of two strings. |
|
129
|
|
|
|
|
130
|
|
|
Parameters |
|
131
|
|
|
---------- |
|
132
|
|
|
src : str |
|
133
|
|
|
Source string (or QGrams/Counter objects) for comparison |
|
134
|
|
|
tar : str |
|
135
|
|
|
Target string (or QGrams/Counter objects) for comparison |
|
136
|
|
|
|
|
137
|
|
|
Returns |
|
138
|
|
|
------- |
|
139
|
|
|
float |
|
140
|
|
|
Normalized string subsequence kernel similarity |
|
141
|
|
|
|
|
142
|
|
|
Examples |
|
143
|
|
|
-------- |
|
144
|
|
|
>>> cmp = SSK() |
|
145
|
|
|
>>> cmp.sim('cat', 'hat') |
|
146
|
|
|
0.3558718861209964 |
|
147
|
|
|
>>> cmp.sim('Niall', 'Neil') |
|
148
|
|
|
0.4709007822130597 |
|
149
|
|
|
>>> cmp.sim('aluminum', 'Catalan') |
|
150
|
|
|
0.13760157193822603 |
|
151
|
|
|
>>> cmp.sim('ATCG', 'TAGC') |
|
152
|
|
|
0.6140899528060498 |
|
153
|
|
|
|
|
154
|
|
|
|
|
155
|
|
|
.. versionadded:: 0.4.1 |
|
156
|
|
|
|
|
157
|
|
|
""" |
|
158
|
1 |
|
if src == tar: |
|
159
|
1 |
|
return 1.0 |
|
160
|
|
|
|
|
161
|
1 |
|
self._tokenize(src, tar) |
|
162
|
|
|
|
|
163
|
1 |
|
src_wts = self._src_tokens |
|
164
|
1 |
|
tar_wts = self._tar_tokens |
|
165
|
|
|
|
|
166
|
1 |
|
score = sum( |
|
167
|
|
|
src_wts[token] * tar_wts[token] for token in src_wts & tar_wts |
|
168
|
|
|
) |
|
169
|
|
|
|
|
170
|
1 |
|
norm = ( |
|
171
|
|
|
sum(src_wts[token] * src_wts[token] for token in src_wts) |
|
172
|
|
|
* sum(tar_wts[token] * tar_wts[token] for token in tar_wts) |
|
173
|
|
|
) ** 0.5 |
|
174
|
|
|
|
|
175
|
1 |
|
if not score: |
|
176
|
1 |
|
return 0.0 |
|
177
|
1 |
|
return score / norm |
|
178
|
|
|
|
|
179
|
|
|
|
|
180
|
|
|
if __name__ == '__main__': |
|
181
|
|
|
import doctest |
|
182
|
|
|
|
|
183
|
|
|
doctest.testmod() |
|
184
|
|
|
|
chrislit /
abydos
