1
|
|
|
# -*- coding: utf-8 -*- |
2
|
|
|
|
3
|
|
|
# Copyright 2014-2018 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_fingerprint. |
20
|
|
|
|
21
|
|
|
This module contains unit tests for abydos.fingerprint |
22
|
|
|
""" |
23
|
|
|
|
24
|
|
|
from __future__ import unicode_literals |
25
|
|
|
|
26
|
|
|
import unittest |
27
|
|
|
|
28
|
|
|
import abydos.phonetic as phonetic |
29
|
|
|
from abydos.fingerprint import count_fingerprint, occurrence_fingerprint, \ |
30
|
|
|
occurrence_halved_fingerprint, omission_key, phonetic_fingerprint, \ |
31
|
|
|
position_fingerprint, qgram_fingerprint, skeleton_key, str_fingerprint, \ |
32
|
|
|
synoname_toolcode |
33
|
|
|
|
34
|
|
|
from six.moves import range |
35
|
|
|
|
36
|
|
|
|
37
|
|
|
NIALL = ('Niall', 'Neal', 'Neil', 'Njall', 'Njáll', 'Nigel', 'Neel', 'Nele', |
38
|
|
|
'Nigelli', 'Nel', 'Kneale', 'Uí Néill', 'O\'Neill', 'MacNeil', |
39
|
|
|
'MacNele', 'Niall Noígíallach') |
40
|
|
|
|
41
|
|
|
|
42
|
|
|
class FingerprintTestCases(unittest.TestCase): |
43
|
|
|
"""Test fingerprint functions. |
44
|
|
|
|
45
|
|
|
abydos.fingerprint.str_fingerprint, .qgram_fingerprint, & |
46
|
|
|
.phonetic_fingerprint |
47
|
|
|
""" |
48
|
|
|
|
49
|
|
|
_testset = ('À noite, vovô Kowalsky vê o ímã cair no pé do pingüim \ |
50
|
|
|
queixoso e vovó põe açúcar no chá de tâmaras do jabuti feliz.', ) |
51
|
|
|
_anssetw = ('a acucar cair cha de do e feliz ima jabuti kowalsky no noite \ |
52
|
|
|
o pe pinguim poe queixoso tamaras ve vovo', ) |
53
|
|
|
_anssetq2 = ('abacadaialamanarasbucachcudedoeaedeieleoetevfeguhaifiminirit\ |
54
|
|
|
ixizjakokylilsmamqngnoocoeoiojokoposovowpepipoqurarnsdsksotatetiucueuiutvevowa\ |
55
|
|
|
xoyv', ) |
56
|
|
|
_anssetq1 = ('abcdefghijklmnopqrstuvwxyz', ) |
57
|
|
|
|
58
|
|
|
def test_str_fingerprint(self): |
59
|
|
|
"""Test abydos.clustering.str_fingerprint.""" |
60
|
|
|
# Base case |
61
|
|
|
self.assertEqual(str_fingerprint(''), '') |
62
|
|
|
|
63
|
|
|
for i in range(len(self._testset)): |
64
|
|
|
self.assertEqual(str_fingerprint(self._testset[i]), |
65
|
|
|
self._anssetw[i]) |
66
|
|
|
|
67
|
|
|
def test_qgram_fingerprint(self): |
68
|
|
|
"""Test abydos.clustering.qgram_fingerprint.""" |
69
|
|
|
# Base case |
70
|
|
|
self.assertEqual(qgram_fingerprint(''), '') |
71
|
|
|
|
72
|
|
|
for i in range(len(self._testset)): |
73
|
|
|
self.assertEqual(qgram_fingerprint(self._testset[i], 1), |
74
|
|
|
self._anssetq1[i]) |
75
|
|
|
self.assertEqual(qgram_fingerprint(self._testset[i], 2), |
76
|
|
|
self._anssetq2[i]) |
77
|
|
|
self.assertEqual(qgram_fingerprint(self._testset[i]), |
78
|
|
|
self._anssetq2[i]) |
79
|
|
|
|
80
|
|
|
qgram_fp_niall = ('aliallni', 'aleane', 'eiilne', 'aljallnj', |
81
|
|
|
'aljallnj', 'elgeigni', 'eeelne', 'ellene', |
82
|
|
|
'elgeiglillni', 'elne', 'aleaknlene', 'eiilinllneui', |
83
|
|
|
'eiilllneon', 'accneiilmane', 'accnellemane', |
84
|
|
|
'acalchgiiaiglalllnninooi') |
85
|
|
|
for i in range(len(NIALL)): |
86
|
|
|
self.assertEqual(qgram_fingerprint(NIALL[i]), qgram_fp_niall[i]) |
87
|
|
|
|
88
|
|
|
def test_phonetic_fingerprint(self): |
89
|
|
|
"""Test abydos.clustering.phonetic_fingerprint.""" |
90
|
|
|
# Base case |
91
|
|
|
self.assertEqual(phonetic_fingerprint(''), '') |
92
|
|
|
|
93
|
|
|
self.assertEqual(phonetic_fingerprint(' '.join(NIALL)), |
94
|
|
|
'a anl mknl njl nklk nl') |
95
|
|
|
self.assertEqual(phonetic_fingerprint(' '.join(NIALL), |
96
|
|
|
phonetic.phonet), |
97
|
|
|
'knile makneil maknele neil nel nele nial nigeli ' + |
98
|
|
|
'nigl nil noigialach oneil ui') |
99
|
|
|
self.assertEqual(phonetic_fingerprint(' '.join(NIALL), |
100
|
|
|
phonetic.soundex), |
101
|
|
|
'k540 m254 n240 n242 n400 o540 u000') |
102
|
|
|
|
103
|
|
|
|
104
|
|
|
class SPEEDCOPTestCases(unittest.TestCase): |
105
|
|
|
"""Test SPEEDCOP functions. |
106
|
|
|
|
107
|
|
|
abydos.fingerprint.skeleton_key & .omission_key |
108
|
|
|
""" |
109
|
|
|
|
110
|
|
|
def test_skeleton_key(self): |
111
|
|
|
"""Test abydos.clustering.skeleton_key.""" |
112
|
|
|
# Base case |
113
|
|
|
self.assertEqual(skeleton_key(''), '') |
114
|
|
|
|
115
|
|
|
# http://dl.acm.org/citation.cfm?id=358048 |
116
|
|
|
self.assertEqual(skeleton_key('chemogenic'), 'CHMGNEOI') |
117
|
|
|
self.assertEqual(skeleton_key('chemomagnetic'), 'CHMGNTEOAI') |
118
|
|
|
self.assertEqual(skeleton_key('chemcal'), 'CHMLEA') |
119
|
|
|
self.assertEqual(skeleton_key('chemcial'), 'CHMLEIA') |
120
|
|
|
self.assertEqual(skeleton_key('chemical'), 'CHMLEIA') |
121
|
|
|
self.assertEqual(skeleton_key('chemicial'), 'CHMLEIA') |
122
|
|
|
self.assertEqual(skeleton_key('chimical'), 'CHMLIA') |
123
|
|
|
self.assertEqual(skeleton_key('chemiluminescence'), 'CHMLNSEIU') |
124
|
|
|
self.assertEqual(skeleton_key('chemiluminescent'), 'CHMLNSTEIU') |
125
|
|
|
self.assertEqual(skeleton_key('chemicals'), 'CHMLSEIA') |
126
|
|
|
self.assertEqual(skeleton_key('chemically'), 'CHMLYEIA') |
127
|
|
|
|
128
|
|
|
def test_omission_key(self): |
129
|
|
|
"""Test abydos.clustering.omission_key.""" |
130
|
|
|
# Base case |
131
|
|
|
self.assertEqual(omission_key(''), '') |
132
|
|
|
|
133
|
|
|
# http://dl.acm.org/citation.cfm?id=358048 |
134
|
|
|
self.assertEqual(omission_key('microelectronics'), 'MCLNTSRIOE') |
135
|
|
|
self.assertEqual(omission_key('circumstantial'), 'MCLNTSRIUA') |
136
|
|
|
self.assertEqual(omission_key('luminescent'), 'MCLNTSUIE') |
137
|
|
|
self.assertEqual(omission_key('multinucleate'), 'MCLNTUIEA') |
138
|
|
|
self.assertEqual(omission_key('multinucleon'), 'MCLNTUIEO') |
139
|
|
|
self.assertEqual(omission_key('cumulene'), 'MCLNUE') |
140
|
|
|
self.assertEqual(omission_key('luminance'), 'MCLNUIAE') |
141
|
|
|
self.assertEqual(omission_key('coelomic'), 'MCLOEI') |
142
|
|
|
self.assertEqual(omission_key('molecule'), 'MCLOEU') |
143
|
|
|
self.assertEqual(omission_key('cameral'), 'MCLRAE') |
144
|
|
|
self.assertEqual(omission_key('caramel'), 'MCLRAE') |
145
|
|
|
self.assertEqual(omission_key('maceral'), 'MCLRAE') |
146
|
|
|
self.assertEqual(omission_key('lacrimal'), 'MCLRAI') |
147
|
|
|
|
148
|
|
|
|
149
|
|
|
class LightweightFingerprintsTestCases(unittest.TestCase): |
150
|
|
|
"""Test Cisłak & Grabowski lightweight fingerprint functions. |
151
|
|
|
|
152
|
|
|
abydos.clustering.occurrence_fingerprint, .occurrence_halved_fingerprint, |
153
|
|
|
.count_fingerprint, & .position_fingerprint |
154
|
|
|
""" |
155
|
|
|
|
156
|
|
|
def test_occurrence_fingerprint(self): |
157
|
|
|
"""Test abydos.occurrence_fingerprint.""" |
158
|
|
|
# Base case |
159
|
|
|
self.assertEqual(occurrence_fingerprint(''), 0) |
160
|
|
|
|
161
|
|
|
# https://arxiv.org/pdf/1711.08475.pdf |
162
|
|
|
self.assertEqual(occurrence_fingerprint('instance'), |
163
|
|
|
0b1110111000010000) |
164
|
|
|
|
165
|
|
|
self.assertEqual(occurrence_fingerprint('inst'), |
166
|
|
|
0b0100111000000000) |
167
|
|
|
self.assertEqual(occurrence_fingerprint('instance', 15), |
168
|
|
|
0b111011100001000) |
169
|
|
|
self.assertEqual(occurrence_fingerprint('instance', 32), |
170
|
|
|
0b11101110000100000000000000000000) |
171
|
|
|
self.assertEqual(occurrence_fingerprint('instance', 64), |
172
|
|
|
0b11101110000100000000000000000000<<32) |
173
|
|
|
|
174
|
|
|
def test_occurrence_halved_fingerprint(self): |
175
|
|
|
"""Test abydos.occurrence_halved_fingerprint.""" |
176
|
|
|
# Base case |
177
|
|
|
self.assertEqual(occurrence_halved_fingerprint(''), 0) |
178
|
|
|
|
179
|
|
|
# https://arxiv.org/pdf/1711.08475.pdf |
180
|
|
|
self.assertEqual(occurrence_halved_fingerprint('instance'), |
181
|
|
|
0b0110010010111000) |
182
|
|
|
|
183
|
|
|
self.assertEqual(occurrence_halved_fingerprint('inst'), |
184
|
|
|
0b0001000010100100) |
185
|
|
|
self.assertEqual(occurrence_halved_fingerprint('instance', 15), |
186
|
|
|
0b0110010010111000) |
187
|
|
|
self.assertEqual(occurrence_halved_fingerprint('instance', 32), |
188
|
|
|
0b01100100101110000000000100000000) |
189
|
|
|
self.assertEqual(occurrence_halved_fingerprint('instance', 64), |
190
|
|
|
0b01100100101110000000000100000000<<32) |
191
|
|
|
|
192
|
|
|
|
193
|
|
|
def test_count_fingerprint(self): |
194
|
|
|
"""Test abydos.count_fingerprint.""" |
195
|
|
|
# Base case |
196
|
|
|
self.assertEqual(count_fingerprint(''), 0) |
197
|
|
|
|
198
|
|
|
# https://arxiv.org/pdf/1711.08475.pdf |
199
|
|
|
self.assertEqual(count_fingerprint('instance'), |
200
|
|
|
0b0101010001100100) |
201
|
|
|
|
202
|
|
|
self.assertEqual(count_fingerprint('inst'), |
203
|
|
|
0b0001000001010100) |
204
|
|
|
self.assertEqual(count_fingerprint('instance', 15), |
205
|
|
|
0b0101010001100100) |
206
|
|
|
self.assertEqual(count_fingerprint('instance', 32), |
207
|
|
|
0b01010100011001000000000100000000) |
208
|
|
|
self.assertEqual(count_fingerprint('instance', 64), |
209
|
|
|
0b01010100011001000000000100000000<<32) |
210
|
|
|
|
211
|
|
|
def test_position_fingerprint(self): |
212
|
|
|
"""Test abydos.position_fingerprint.""" |
213
|
|
|
# Base case |
214
|
|
|
self.assertEqual(position_fingerprint(''), |
215
|
|
|
0b1111111111111111) |
216
|
|
|
|
217
|
|
|
# https://arxiv.org/pdf/1711.08475.pdf |
218
|
|
|
self.assertEqual(position_fingerprint('instance'), |
219
|
|
|
0b1110111001110001) |
220
|
|
|
|
221
|
|
|
self.assertEqual(position_fingerprint('instance'), |
222
|
|
|
0b1110111001110001) |
223
|
|
|
self.assertEqual(position_fingerprint('instance', 15), |
224
|
|
|
0b111011100111000) |
225
|
|
|
self.assertEqual(position_fingerprint('instance', 32), |
226
|
|
|
0b11101110011100000101011111111111) |
227
|
|
|
self.assertEqual(position_fingerprint('instance', 64), |
228
|
|
|
0xee7057ffefffffff) |
229
|
|
|
|
230
|
|
|
|
231
|
|
|
class SynonameToolcodeTestCases(unittest.TestCase): |
232
|
|
|
"""Test Synoname Toolcode function. |
233
|
|
|
|
234
|
|
|
abydps.fingerprint.synoname_toolcode |
235
|
|
|
""" |
236
|
|
|
|
237
|
|
|
def test_synoname_toolcode(self): |
238
|
|
|
"""Test abydos.synoname_toolcode.""" |
239
|
|
|
# Base case |
240
|
|
|
self.assertEqual(synoname_toolcode(''), ('', '', '0000000000$$')) |
241
|
|
|
|
242
|
|
|
# from Synoname demo |
243
|
|
|
self.assertEqual(synoname_toolcode('angelico', 'fra'), |
244
|
|
|
('angelico', 'fra', '0000000308$044a$af')) |
245
|
|
|
self.assertEqual(synoname_toolcode('Aelst', 'Willem van', ''), |
246
|
|
|
('aelst', 'willem van', '0000001005$143a$awv')) |
247
|
|
|
self.assertEqual(synoname_toolcode('Afro'), |
248
|
|
|
('afro', '', '0000000004$$a')) |
249
|
|
|
self.assertEqual(synoname_toolcode('Afro', 'Basaldella'), |
250
|
|
|
('afro', 'basaldella', '0000001004$$ab')) |
251
|
|
|
self.assertEqual(synoname_toolcode('Albright', 'Ivan'), |
252
|
|
|
('albright', 'ivan', '0000000408$$ai')) |
253
|
|
|
self.assertEqual(synoname_toolcode('Antonello da Messina'), |
254
|
|
|
('antonello da messina', '', '0000000020$022b$adm')) |
255
|
|
|
self.assertEqual(synoname_toolcode('Albright', 'Ivan Le Lorraine'), |
256
|
|
|
('albright', 'ivan le lorraine', |
257
|
|
|
'0000001608$067b$ail')) |
258
|
|
|
self.assertEqual(synoname_toolcode('Bazille', 'Frederic', |
259
|
|
|
'Attributed to'), |
260
|
|
|
('bazille', 'frederic', '1000000807$$bf')) |
261
|
|
|
self.assertEqual(synoname_toolcode('Bazille', 'Frederick', |
262
|
|
|
'Attributed to'), |
263
|
|
|
('bazille', 'frederick', '1000000907$$bf')) |
264
|
|
|
self.assertEqual(synoname_toolcode('Beerstraaten', 'Jan Abrahamsz.'), |
265
|
|
|
('beerstraaten', 'jan abrahamsz.', '0200001412$$bja')) |
266
|
|
|
self.assertEqual(synoname_toolcode('Bonifacio di Pitati'), |
267
|
|
|
('bonifacio di pitati', '', '0000000019$035b$bdp')) |
268
|
|
|
self.assertEqual(synoname_toolcode('Breughel the Younger', 'Jan'), |
269
|
|
|
('breughel the younger', 'jan', |
270
|
|
|
'0020000320$134b$btyj')) |
271
|
|
|
self.assertEqual(synoname_toolcode('Brown', 'W. W.'), |
272
|
|
|
('brown', 'w. w.', '0200000505$$bw')) |
273
|
|
|
self.assertEqual(synoname_toolcode('Brueghel II (the Younger)', 'Jan'), |
274
|
|
|
('brueghel ii (the younger)', 'jan', |
275
|
|
|
'0120490325$049b134b$bityj')) |
276
|
|
|
self.assertEqual(synoname_toolcode('Brueghel II (the Younger)', |
277
|
|
|
'Pieter', 'Workshop of'), |
278
|
|
|
('brueghel ii (the younger)', 'pieter', |
279
|
|
|
'3120490625$049b134b$bityp')) |
280
|
|
|
self.assertEqual(synoname_toolcode('Bugiardini', |
281
|
|
|
'Guiliano di Piero di Simone'), |
282
|
|
|
('bugiardini', 'guiliano di piero di simone', |
283
|
|
|
'0000002710$035b035b$bgdps')) |
284
|
|
|
self.assertEqual(synoname_toolcode('Caravaggio', '', 'Follower of'), |
285
|
|
|
('caravaggio', '', '3000000010$$c')) |
286
|
|
|
self.assertEqual(synoname_toolcode('Caravaggio', |
287
|
|
|
'Michelangelo Merisi da', |
288
|
|
|
'Follower of'), |
289
|
|
|
('caravaggio', 'michelangelo merisi da', |
290
|
|
|
'3000002210$022a$cmd')) |
291
|
|
|
self.assertEqual(synoname_toolcode('Oost the Younger', 'Jacob van'), |
292
|
|
|
('oost the younger', 'jacob van', |
293
|
|
|
'0020000916$134b143a$otyjv')) |
294
|
|
|
|
295
|
|
|
# additional tests for coverage |
296
|
|
|
self.assertEqual(synoname_toolcode('Cato the Elder', '', 'Copy of'), |
297
|
|
|
('cato the elder', '', '2010000014$133b$cte')) |
298
|
|
|
self.assertEqual(synoname_toolcode('Cato, the Elder', normalize=2), |
299
|
|
|
('cato the elder', '', '0110000014$133b$cte')) |
300
|
|
|
self.assertEqual(synoname_toolcode('Cato the Elder', normalize=2), |
301
|
|
|
('cato the elder', '', '0010000014$133b$cte')) |
302
|
|
|
self.assertEqual(synoname_toolcode('Lorem ipsum dolor sit amet, '+ |
303
|
|
|
'consectetur adipiscing elit, sed '+ |
304
|
|
|
'do eiusmod tempor incididunt ut '+ |
305
|
|
|
'labore et dolore magna aliqua. '+ |
306
|
|
|
'Nulla aliquet porttitor lacus '+ |
307
|
|
|
'luctus accumsan tortor posuere. '+ |
308
|
|
|
'Egestas purus viverra accumsan '+ |
309
|
|
|
'in. Ultrices mi tempus imperdiet '+ |
310
|
|
|
'nulla malesuada pellentesque '+ |
311
|
|
|
'elit eget gravida. Proin libero '+ |
312
|
|
|
'nunc consequat interdum varius '+ |
313
|
|
|
'sit amet mattis vulputate. '+ |
314
|
|
|
'Mauris ultrices eros in cursus '+ |
315
|
|
|
'turpis massa tincidunt dui. '+ |
316
|
|
|
'Faucibus in ornare quam viverra '+ |
317
|
|
|
'orci sagittis eu volutpat odio. '+ |
318
|
|
|
'Enim blandit volutpat maecenas '+ |
319
|
|
|
'volutpat blandit aliquam etiam. '+ |
320
|
|
|
'Vel quam elementum pulvinar '+ |
321
|
|
|
'etiam. Duis ut diam quam nulla '+ |
322
|
|
|
'porttitor massa id.', |
323
|
|
|
normalize=1)[2], |
324
|
|
|
'02000060626$068d$lidsacetumnpvgflo') |
325
|
|
|
self.assertEqual(synoname_toolcode('Sainte-Vincent'), |
326
|
|
|
('sainte-vincent', '', '0100000014$110c$sv')) |
327
|
|
|
self.assertEqual(synoname_toolcode('Lorem', 'Sainte-Vincent'), |
328
|
|
|
('lorem', 'sainte-vincent', |
329
|
|
|
'0100001405$068d110b$lsvlo')) |
330
|
|
|
self.assertEqual(synoname_toolcode('Louis II', 'Jean'), |
331
|
|
|
('louis ii', 'jean', '0000490408$049b068d$lijlo')) |
332
|
|
|
self.assertEqual(synoname_toolcode('Louis', 'Jean II', normalize=2), |
333
|
|
|
('louis ii', 'jean', '0000490705$049a068d$ljilo')) |
334
|
|
|
self.assertEqual(synoname_toolcode('Louis', 'Jean II ', normalize=2), |
335
|
|
|
('louis ii', 'jean', '0000490805$049b068d$ljilo')) |
336
|
|
|
self.assertEqual(synoname_toolcode('Louis V.', 'Jean', normalize=2), |
337
|
|
|
('louis v.', 'jean', '0200000408$068d$lvjlo')) |
338
|
|
|
self.assertEqual(synoname_toolcode('Louis V.', 'Ste.-Jean Ste.', |
339
|
|
|
normalize=2), |
340
|
|
|
('louis v.', 'ste.-jean ste.', |
341
|
|
|
'0200001408$068d127b127X$lvsjlo ste')) |
342
|
|
|
|
343
|
|
|
|
344
|
|
|
if __name__ == '__main__': |
345
|
|
|
unittest.main() |
346
|
|
|
|