File: similarity_test.py

package info (click to toggle)
python-wn 0.13.0-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 2,184 kB
  • sloc: python: 7,592; xml: 493; sql: 220; makefile: 12
file content (167 lines) | stat: -rw-r--r-- 6,901 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167

from math import log

import pytest

import wn
from wn import similarity as sim
from wn.taxonomy import taxonomy_depth
from wn.ic import information_content as infocont


def get_synsets(w):
    return {
        'information': w.synset('test-en-0001-n'),
        'example': w.synset('test-en-0002-n'),
        'sample': w.synset('test-en-0004-n'),
        'random sample': w.synset('test-en-0005-n'),
        'random sample2': w.synset('test-en-0008-n'),
        'datum': w.synset('test-en-0006-n'),
        'exemplify': w.synset('test-en-0003-v'),
    }


# some fake information content; computed using:
#   words = ['example', 'example', 'sample', 'random sample', 'illustrate']
#   ic = compute(words, wn.Wordnet('test-en'), distribute_weight=False)

ic = {
    'n': {'test-en-0001-n': 5.0,  # information
          'test-en-0002-n': 5.0,  # example, illustration
          'test-en-0004-n': 3.0,  # sample
          'test-en-0005-n': 2.0,  # random sample
          'test-en-0008-n': 2.0,  # random sample 2
          'test-en-0006-n': 1.0,  # datum
          None: 6.0},
    'v': {'test-en-0003-v': 2.0,  # exemplify, illustrate
          'test-en-0007-v': 1.0,  # resignate
          None: 2.0},
    'a': {None: 1.0},
    'r': {None: 1.0}
}


@pytest.mark.usefixtures('mini_db')
def test_path():
    ss = get_synsets(wn.Wordnet('test-en'))
    assert sim.path(ss['information'], ss['information']) == 1/1
    assert sim.path(ss['information'], ss['example']) == 1/2
    assert sim.path(ss['information'], ss['sample']) == 1/3
    assert sim.path(ss['information'], ss['random sample']) == 1/4
    assert sim.path(ss['random sample'], ss['datum']) == 1/5
    assert sim.path(ss['random sample2'], ss['datum']) == 0
    assert sim.path(ss['random sample2'], ss['datum'], simulate_root=True) == 1/4
    assert sim.path(
        ss['random sample'], ss['random sample2'], simulate_root=True
    ) == 1/6
    with pytest.raises(wn.Error):
        sim.path(ss['example'], ss['exemplify'])
    with pytest.raises(wn.Error):
        sim.wup(ss['example'], ss['exemplify'], simulate_root=True)


@pytest.mark.usefixtures('mini_db')
def test_wup():
    ss = get_synsets(wn.Wordnet('test-en'))
    assert sim.wup(ss['information'], ss['information']) == (2*1) / (0+0+2*1)
    assert sim.wup(ss['information'], ss['example']) == (2*1) / (0+1+2*1)
    assert sim.wup(ss['information'], ss['sample']) == (2*1) / (0+2+2*1)
    assert sim.wup(ss['information'], ss['random sample']) == (2*1) / (0+3+2*1)
    assert sim.wup(ss['random sample'], ss['datum']) == (2*1) / (3+1+2*1)
    with pytest.raises(wn.Error):
        assert sim.wup(ss['random sample2'], ss['datum'])
    assert (sim.wup(ss['random sample2'], ss['datum'], simulate_root=True)
            == (2*1) / (1+2+2*1))
    assert (sim.wup(ss['random sample'], ss['random sample2'], simulate_root=True)
            == (2*1) / (4+1+2*1))
    with pytest.raises(wn.Error):
        sim.wup(ss['example'], ss['exemplify'])
    with pytest.raises(wn.Error):
        sim.wup(ss['example'], ss['exemplify'], simulate_root=True)


@pytest.mark.usefixtures('mini_db')
def test_lch():
    w = wn.Wordnet('test-en')
    ss = get_synsets(w)
    d_n = taxonomy_depth(w, 'n')
    assert sim.lch(ss['information'], ss['information'], d_n) == -log((0+1) / (2*d_n))
    assert sim.lch(ss['information'], ss['example'], d_n) == -log((1+1) / (2*d_n))
    assert sim.lch(ss['information'], ss['sample'], d_n) == -log((2+1) / (2*d_n))
    assert sim.lch(ss['information'], ss['random sample'], d_n) == -log((3+1) / (2*d_n))
    assert sim.lch(ss['random sample'], ss['datum'], d_n) == -log((4+1) / (2*d_n))
    with pytest.raises(wn.Error):
        assert sim.lch(ss['random sample2'], ss['datum'], d_n)
    assert (sim.lch(ss['random sample2'], ss['datum'], d_n, simulate_root=True)
            == -log((3+1) / (2*d_n)))
    assert (sim.lch(ss['random sample'], ss['random sample2'], d_n, simulate_root=True)
            == -log((5+1) / (2*d_n)))
    with pytest.raises(wn.Error):
        sim.lch(ss['example'], ss['exemplify'], d_n)
    with pytest.raises(wn.Error):
        sim.lch(ss['example'], ss['exemplify'], d_n, simulate_root=True)


@pytest.mark.usefixtures('mini_db')
def test_res():
    w = wn.Wordnet('test-en')
    ss = get_synsets(w)
    assert (sim.res(ss['information'], ss['information'], ic)
            == infocont(ss['information'], ic))
    assert (sim.res(ss['information'], ss['example'], ic)
            == infocont(ss['information'], ic))
    assert (sim.res(ss['information'], ss['sample'], ic)
            == infocont(ss['information'], ic))
    assert (sim.res(ss['information'], ss['random sample'], ic)
            == infocont(ss['information'], ic))
    assert (sim.res(ss['random sample'], ss['datum'], ic)
            == infocont(ss['information'], ic))
    with pytest.raises(wn.Error):
        sim.res(ss['random sample2'], ss['datum'], ic)
    with pytest.raises(wn.Error):
        sim.res(ss['example'], ss['exemplify'], ic)


@pytest.mark.usefixtures('mini_db')
def test_jcn():
    w = wn.Wordnet('test-en')
    ss = get_synsets(w)
    info_ic = infocont(ss['information'], ic)
    assert (sim.jcn(ss['information'], ss['information'], ic)
            == float('inf'))
    assert (sim.jcn(ss['information'], ss['example'], ic)
            == float('inf'))
    assert (sim.jcn(ss['information'], ss['sample'], ic)
            == 1 / ((info_ic + infocont(ss['sample'], ic)) - 2 * info_ic))
    assert (sim.jcn(ss['information'], ss['random sample'], ic)
            == 1 / ((info_ic + infocont(ss['random sample'], ic)) - 2 * info_ic))
    assert (sim.jcn(ss['random sample'], ss['datum'], ic)
            == 1 / (
                (infocont(ss['random sample'], ic) + infocont(ss['datum'], ic))
                - 2 * info_ic))
    with pytest.raises(wn.Error):
        sim.jcn(ss['random sample2'], ss['datum'], ic)
    with pytest.raises(wn.Error):
        sim.jcn(ss['example'], ss['exemplify'], ic)


@pytest.mark.usefixtures('mini_db')
def test_lin():
    w = wn.Wordnet('test-en')
    ss = get_synsets(w)
    info_ic = infocont(ss['information'], ic)
    assert (sim.lin(ss['information'], ss['information'], ic)
            == 1.0)
    assert (sim.lin(ss['information'], ss['example'], ic)
            == 1.0)
    assert (sim.lin(ss['information'], ss['sample'], ic)
            == (2 * info_ic) / (info_ic + infocont(ss['sample'], ic)))
    assert (sim.lin(ss['information'], ss['random sample'], ic)
            == (2 * info_ic) / (info_ic + infocont(ss['random sample'], ic)))
    assert (sim.lin(ss['random sample'], ss['datum'], ic)
            == ((2 * info_ic)
                / (infocont(ss['random sample'], ic) + infocont(ss['datum'], ic))))
    with pytest.raises(wn.Error):
        sim.lin(ss['random sample2'], ss['datum'], ic)
    with pytest.raises(wn.Error):
        sim.lin(ss['example'], ss['exemplify'], ic)