|
1
|
|
|
# -*- coding: utf-8 -*- |
|
2
|
|
|
|
|
3
|
|
|
# Copyright 2018 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
|
|
|
"""abydos.tests.test_stemmer_uealite. |
|
20
|
|
|
|
|
21
|
|
|
This module contains unit tests for abydos.stemmer.uealite |
|
22
|
|
|
""" |
|
23
|
|
|
|
|
24
|
|
|
from __future__ import unicode_literals |
|
25
|
|
|
|
|
26
|
|
|
import unittest |
|
27
|
|
|
|
|
28
|
|
|
from abydos.stemmer.uealite import uealite |
|
29
|
|
|
|
|
30
|
|
|
from .. import _corpus_file |
|
31
|
|
|
|
|
32
|
|
|
|
|
33
|
|
|
class UEALiteTestCases(unittest.TestCase): |
|
34
|
|
|
"""Test UEA-lite functions. |
|
35
|
|
|
|
|
36
|
|
|
abydos.stemmer.uealite |
|
37
|
|
|
""" |
|
38
|
|
|
|
|
39
|
|
|
def test_uealite(self): |
|
40
|
|
|
"""Test abydos.stemmer.uealite.""" |
|
41
|
|
|
# base case |
|
42
|
|
|
self.assertEqual(uealite(''), '') |
|
43
|
|
|
|
|
44
|
|
|
# test cases copied from Ruby port |
|
45
|
|
|
# https://github.com/ealdent/uea-stemmer/blob/master/test/uea_stemmer_test.rb |
|
46
|
|
|
# These are corrected to match the Java version's output. |
|
47
|
|
|
# stem base words to just the base word |
|
48
|
|
|
self.assertEqual(uealite('man'), 'man') |
|
49
|
|
|
self.assertEqual(uealite('happiness'), 'happiness') |
|
50
|
|
|
# stem theses as thesis but not bases as basis |
|
51
|
|
|
self.assertEqual(uealite('theses'), 'thesis') |
|
52
|
|
|
self.assertNotEqual(uealite('bases'), 'basis') |
|
53
|
|
|
# stem preterite words ending in -ed without the -ed |
|
54
|
|
|
self.assertEqual(uealite('ordained'), 'ordain') |
|
55
|
|
|
self.assertEqual(uealite('killed'), 'kill') |
|
56
|
|
|
self.assertEqual(uealite('liked'), 'lik') |
|
57
|
|
|
self.assertEqual(uealite('helped'), 'help') |
|
58
|
|
|
self.assertEqual(uealite('scarred'), 'scarre') |
|
59
|
|
|
self.assertEqual(uealite('invited'), 'invit') |
|
60
|
|
|
self.assertEqual(uealite('exited'), 'exit') |
|
61
|
|
|
self.assertEqual(uealite('debited'), 'debit') |
|
62
|
|
|
self.assertEqual(uealite('smited'), 'smit') |
|
63
|
|
|
# stem progressive verbs and gerunds without the -ing |
|
64
|
|
|
self.assertEqual(uealite('running'), 'run') |
|
65
|
|
|
self.assertEqual(uealite('settings'), 'set') |
|
66
|
|
|
self.assertEqual(uealite('timing'), 'time') |
|
67
|
|
|
self.assertEqual(uealite('dying'), 'dy') |
|
68
|
|
|
self.assertEqual(uealite('harping'), 'harp') |
|
69
|
|
|
self.assertEqual(uealite('charring'), 'char') |
|
70
|
|
|
# not stem false progressive verbs such as 'sing' |
|
71
|
|
|
self.assertEqual(uealite('ring'), 'ring') |
|
72
|
|
|
self.assertEqual(uealite('sing'), 'se') |
|
73
|
|
|
self.assertEqual(uealite('bring'), 'br') |
|
74
|
|
|
self.assertEqual(uealite('fling'), 'fle') |
|
75
|
|
|
# stem various plural nouns and 3rd-pres verbs without the -s/-es |
|
76
|
|
|
self.assertEqual(uealite('changes'), 'change') |
|
77
|
|
|
self.assertEqual(uealite('deaths'), 'death') |
|
78
|
|
|
self.assertEqual(uealite('shadows'), 'shadow') |
|
79
|
|
|
self.assertEqual(uealite('flies'), 'fly') |
|
80
|
|
|
self.assertEqual(uealite('things'), 'thing') |
|
81
|
|
|
self.assertEqual(uealite('nothings'), 'nothing') |
|
82
|
|
|
self.assertEqual(uealite('witches'), 'witch') |
|
83
|
|
|
self.assertEqual(uealite('makes'), 'mak') |
|
84
|
|
|
self.assertEqual(uealite('smokes'), 'smok') |
|
85
|
|
|
self.assertEqual(uealite('does'), 'do') |
|
86
|
|
|
# stem various words with -des suffix |
|
87
|
|
|
self.assertEqual(uealite('abodes'), 'abod') |
|
88
|
|
|
self.assertEqual(uealite('escapades'), 'escapad') |
|
89
|
|
|
self.assertEqual(uealite('crusades'), 'crusad') |
|
90
|
|
|
self.assertEqual(uealite('grades'), 'grad') |
|
91
|
|
|
# stem various words with -res suffix |
|
92
|
|
|
self.assertEqual(uealite('wires'), 'wir') |
|
93
|
|
|
self.assertEqual(uealite('acres'), 'acr') |
|
94
|
|
|
self.assertEqual(uealite('fires'), 'fir') |
|
95
|
|
|
self.assertEqual(uealite('cares'), 'car') |
|
96
|
|
|
# stem acronyms when pluralized otherwise they should be left alone |
|
97
|
|
|
self.assertEqual(uealite('USA'), 'USA') |
|
98
|
|
|
self.assertEqual(uealite('FLOSS'), 'FLOSS') |
|
99
|
|
|
self.assertEqual(uealite('MREs'), 'MRE') |
|
100
|
|
|
self.assertEqual(uealite('USAED'), 'USAED') |
|
101
|
|
|
|
|
102
|
|
|
# test cases copied from Ruby port |
|
103
|
|
|
# https://github.com/ealdent/uea-stemmer/blob/master/test/uea_stemmer_test.rb |
|
104
|
|
|
# stem base words to just the base word |
|
105
|
|
|
self.assertEqual(uealite('man', var='Adams'), 'man') |
|
106
|
|
|
self.assertEqual(uealite('happiness', var='Adams'), 'happiness') |
|
107
|
|
|
# stem theses as thesis but not bases as basis |
|
108
|
|
|
self.assertEqual(uealite('theses', var='Adams'), 'thesis') |
|
109
|
|
|
self.assertNotEqual(uealite('bases', var='Adams'), 'basis') |
|
110
|
|
|
# stem preterite words ending in -ed without the -ed |
|
111
|
|
|
self.assertEqual(uealite('ordained', var='Adams'), 'ordain') |
|
112
|
|
|
self.assertEqual(uealite('killed', var='Adams'), 'kill') |
|
113
|
|
|
self.assertEqual(uealite('liked', var='Adams'), 'like') |
|
114
|
|
|
self.assertEqual(uealite('helped', var='Adams'), 'help') |
|
115
|
|
|
# self.assertEqual(uealite('scarred', var='Adams'), 'scar') |
|
116
|
|
|
self.assertEqual(uealite('invited', var='Adams'), 'invite') |
|
117
|
|
|
self.assertEqual(uealite('exited', var='Adams'), 'exit') |
|
118
|
|
|
self.assertEqual(uealite('debited', var='Adams'), 'debit') |
|
119
|
|
|
self.assertEqual(uealite('smited', var='Adams'), 'smite') |
|
120
|
|
|
# stem progressive verbs and gerunds without the -ing |
|
121
|
|
|
self.assertEqual(uealite('running', var='Adams'), 'run') |
|
122
|
|
|
self.assertEqual(uealite('settings', var='Adams'), 'set') |
|
123
|
|
|
self.assertEqual(uealite('timing', var='Adams'), 'time') |
|
124
|
|
|
self.assertEqual(uealite('dying', var='Adams'), 'die') |
|
125
|
|
|
self.assertEqual(uealite('harping', var='Adams'), 'harp') |
|
126
|
|
|
self.assertEqual(uealite('charring', var='Adams'), 'char') |
|
127
|
|
|
# not stem false progressive verbs such as 'sing' |
|
128
|
|
|
self.assertEqual(uealite('ring', var='Adams'), 'ring') |
|
129
|
|
|
self.assertEqual(uealite('sing', var='Adams'), 'sing') |
|
130
|
|
|
self.assertEqual(uealite('ring', var='Adams'), 'ring') |
|
131
|
|
|
self.assertEqual(uealite('bring', var='Adams'), 'bring') |
|
132
|
|
|
self.assertEqual(uealite('fling', var='Adams'), 'fling') |
|
133
|
|
|
# stem various plural nouns and 3rd-pres verbs without the -s/-es |
|
134
|
|
|
self.assertEqual(uealite('changes', var='Adams'), 'change') |
|
135
|
|
|
self.assertEqual(uealite('deaths', var='Adams'), 'death') |
|
136
|
|
|
self.assertEqual(uealite('shadows', var='Adams'), 'shadow') |
|
137
|
|
|
self.assertEqual(uealite('flies', var='Adams'), 'fly') |
|
138
|
|
|
self.assertEqual(uealite('things', var='Adams'), 'thing') |
|
139
|
|
|
self.assertEqual(uealite('nothings', var='Adams'), 'nothing') |
|
140
|
|
|
self.assertEqual(uealite('witches', var='Adams'), 'witch') |
|
141
|
|
|
self.assertEqual(uealite('makes', var='Adams'), 'make') |
|
142
|
|
|
self.assertEqual(uealite('smokes', var='Adams'), 'smoke') |
|
143
|
|
|
self.assertEqual(uealite('does', var='Adams'), 'do') |
|
144
|
|
|
# stem various words with -des suffix |
|
145
|
|
|
self.assertEqual(uealite('abodes', var='Adams'), 'abode') |
|
146
|
|
|
self.assertEqual(uealite('escapades', var='Adams'), 'escapade') |
|
147
|
|
|
self.assertEqual(uealite('crusades', var='Adams'), 'crusade') |
|
148
|
|
|
self.assertEqual(uealite('grades', var='Adams'), 'grade') |
|
149
|
|
|
# stem various words with -res suffix |
|
150
|
|
|
self.assertEqual(uealite('wires', var='Adams'), 'wire') |
|
151
|
|
|
self.assertEqual(uealite('acres', var='Adams'), 'acre') |
|
152
|
|
|
self.assertEqual(uealite('fires', var='Adams'), 'fire') |
|
153
|
|
|
self.assertEqual(uealite('cares', var='Adams'), 'care') |
|
154
|
|
|
# stem acronyms when pluralized otherwise they should be left alone |
|
155
|
|
|
self.assertEqual(uealite('USA', var='Adams'), 'USA') |
|
156
|
|
|
self.assertEqual(uealite('FLOSS', var='Adams'), 'FLOSS') |
|
157
|
|
|
self.assertEqual(uealite('MREs', var='Adams'), 'MRE') |
|
158
|
|
|
self.assertEqual(uealite('USAED', var='Adams'), 'USAED') |
|
159
|
|
|
|
|
160
|
|
|
# Perl version tests |
|
161
|
|
|
self.assertEqual(uealite('ragings'), 'rage') |
|
162
|
|
|
self.assertEqual(uealite('ragings', var='Perl'), 'rag') |
|
163
|
|
|
|
|
164
|
|
|
# complete coverage |
|
165
|
|
|
self.assertEqual(uealite('was'), 'was') |
|
166
|
|
|
self.assertEqual(uealite('during'), 'during') |
|
167
|
|
|
self.assertEqual(uealite('abcdefghijklmnopqrstuvwxyz', |
|
168
|
|
|
max_word_length=20), |
|
169
|
|
|
'abcdefghijklmnopqrstuvwxyz') |
|
170
|
|
|
self.assertEqual(uealite('10'), '10') |
|
171
|
|
|
self.assertEqual(uealite('top-ten'), 'top-ten') |
|
172
|
|
|
self.assertEqual(uealite('top-10'), 'top-10') |
|
173
|
|
|
self.assertEqual(uealite('top_ten'), 'top_ten') |
|
174
|
|
|
self.assertEqual(uealite('ABCDEFGHIJKLMs', max_acro_length=8, |
|
175
|
|
|
var='Adams'), 'ABCDEFGHIJKLMs') |
|
176
|
|
|
self.assertEqual(uealite('ABCDEFGHIJKLM', max_acro_length=8, |
|
177
|
|
|
var='Adams'), 'ABCDEFGHIJKLM') |
|
178
|
|
|
self.assertEqual(uealite('abcDefGhij'), 'abcDefGhij') |
|
179
|
|
|
self.assertEqual(uealite('Tophat'), 'Tophat') |
|
180
|
|
|
self.assertEqual(uealite(''), '') |
|
181
|
|
|
self.assertEqual(uealite(''), '') |
|
182
|
|
|
|
|
183
|
|
|
def test_uealite_wsj_set(self): |
|
184
|
|
|
"""Test abydos.stemmer.uealite (WSJ testset).""" |
|
185
|
|
|
with open(_corpus_file('uea-lite_wsj.csv')) as wsj_ts: |
|
186
|
|
|
for wsj_line in wsj_ts: |
|
187
|
|
|
(word, uea, rule) = wsj_line.strip().split(',') |
|
188
|
|
|
self.assertEqual(uealite(word, return_rule_no=True), |
|
189
|
|
|
(uea, float(rule))) |
|
190
|
|
|
|
|
191
|
|
|
|
|
192
|
|
|
if __name__ == '__main__': |
|
193
|
|
|
unittest.main() |
|
194
|
|
|
|