forked from cupslab/neural_network_cracking
-
Notifications
You must be signed in to change notification settings - Fork 0
/
markov_model_tests.py
205 lines (180 loc) · 8.06 KB
/
markov_model_tests.py
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
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
import unittest
from unittest.mock import Mock, MagicMock
import string
import tempfile
import io
import numpy as np
import pwd_guess as pg
import markov_model as mm
class MarkovModelTest(unittest.TestCase):
def test_train_one_pwd_no_smoothing(self):
config = Mock()
config.char_bag = string.ascii_lowercase + pg.PASSWORD_END
m = mm.MarkovModel(config, smoothing='none', order=2)
m.train([('pass', 1)])
self.assertAlmostEqual(m.probability_next_char('p', 'a'), 1)
self.assertAlmostEqual(m.probability_next_char('pa', 's'), 1)
self.assertAlmostEqual(
m.probability_next_char('pass', pg.PASSWORD_END), .5)
self.assertAlmostEqual(m.probability_next_char('pas', 's'), .5)
self.assertAlmostEqual(m.probability_next_char('', 'p'), 1)
self.assertAlmostEqual(m.probability_next_char('pas', 'k'), 0)
self.assertAlmostEqual(m.probability_next_char('', 'j'), 0)
def test_train_two_pwd_no_smoothing(self):
config = Mock()
config.char_bag = string.ascii_lowercase + pg.PASSWORD_END
m = mm.MarkovModel(config, smoothing='none', order=2)
m.train([('pass', 1), ('past', 1)])
self.assertAlmostEqual(m.probability_next_char('', 'p'), 1)
self.assertAlmostEqual(m.probability_next_char('p', 'a'), 1)
self.assertAlmostEqual(m.probability_next_char('pa', 's'), 1)
self.assertAlmostEqual(
m.probability_next_char('pass', pg.PASSWORD_END), 1/3.)
self.assertAlmostEqual(m.probability_next_char('pas', 's'), 1/3.)
self.assertAlmostEqual(m.probability_next_char('pas', 't'), 1/3.)
self.assertAlmostEqual(m.probability_next_char('pas', 'k'), 0)
self.assertAlmostEqual(m.probability_next_char('', 'j'), 0)
def test_train_pwd_long(self):
config = Mock()
config.char_bag = string.ascii_lowercase + pg.PASSWORD_END
m = mm.MarkovModel(config, smoothing='none', order=4)
m.train([('pa', 1)])
self.assertEqual(m.freq_dict, {
'p' : 1, 'pa' : 1, 'pa' + pg.PASSWORD_END : 1
})
def test_train_high_order_no_smoothing(self):
config = Mock()
config.char_bag = string.ascii_lowercase + pg.PASSWORD_END
m = mm.MarkovModel(config, smoothing='none', order=3)
m.train([('pass', 1), ('past', 1), ('ashen', 1)])
self.assertAlmostEqual(m.probability_next_char('', 'p'), 2./3.)
self.assertAlmostEqual(m.probability_next_char('p', 'a'), 1)
self.assertAlmostEqual(m.probability_next_char('pa', 's'), 1)
self.assertAlmostEqual(
m.probability_next_char('pass', pg.PASSWORD_END), 1)
self.assertAlmostEqual(m.probability_next_char('pas', 's'), 1./3.)
self.assertAlmostEqual(m.probability_next_char('pas', 't'), 1./3.)
self.assertAlmostEqual(m.probability_next_char('as', 'h'), 1./3.)
self.assertAlmostEqual(m.probability_next_char('as', 't'), 1./3.)
self.assertAlmostEqual(m.probability_next_char('pas', 'k'), 0)
self.assertAlmostEqual(m.probability_next_char('', 'j'), 0)
def test_save_load_model(self):
config = Mock()
config.char_bag = string.ascii_lowercase + pg.PASSWORD_END
m = mm.MarkovModel(config, smoothing='none', order=2)
m.train([('pass', 1), ('past', 1), ('ashen', 1)])
self.assertAlmostEqual(m.probability_next_char('', 'p'), 2./3.)
with tempfile.NamedTemporaryFile('w') as tempf:
m.saveModel(tempf.name)
new_model = mm.MarkovModel.fromModelFile(
tempf.name, config, smoothing='none', order=2)
self.assertAlmostEqual(new_model.probability_next_char('', 'p'), 2./3.)
def test_predict(self):
config = Mock()
config.char_bag = pg.PASSWORD_END + 'aehnpst'
m = mm.MarkovModel(config, smoothing='none', order=2)
m.train([('pass', 1), ('past', 1), ('ashen', 1)])
answer = np.zeros((len(config.char_bag), ), dtype=np.float64)
m.predict('pa', answer)
np.testing.assert_array_equal(answer, np.array([
0, 0, 0, 0, 0, 0, 1, 0
]))
class MarkovGuesserTest(unittest.TestCase):
def test_build(self):
config = pg.ModelDefaults(char_bag = pg.PASSWORD_END + 'aehnpst',
guesser_class = 'markov_model')
pg.GuesserBuilder.other_class_builders[
'markov_model'] = mm.MarkovGuesser
model = mm.MarkovModel(config, smoothing='none', order=2)
model.train([('pass', 1), ('past', 1), ('ashen', 1)])
guesser_builder = pg.GuesserBuilder(config)
guesser_builder.add_model(model)
ostream = io.StringIO()
guesser_builder.add_stream(ostream)
guesser = guesser_builder.build()
self.assertEqual(type(guesser), mm.MarkovGuesser)
np.testing.assert_array_almost_equal(
guesser.conditional_probs_many(['pa']), np.array([[[
0, 0, 0, 0, 0, 0, 1, 0
]]], dtype=np.float64))
class AdditiveSmoothingTest(unittest.TestCase):
def test_predict(self):
config = Mock()
config.char_bag = 'abc'
config.additive_smoothing_amount = 1
sm = mm.AdditiveSmoothingSmoother({
'a' : 1,
'b' : 2
}, config)
answer = np.zeros((3, ), dtype=np.float64)
sm.predict('', answer)
np.testing.assert_array_almost_equal(
answer, np.array([1/3., 1/2., 1/6.], dtype=np.float64))
class BackoffMarkovModelTest(unittest.TestCase):
def test_train_one(self):
config = Mock()
config.char_bag = (
string.ascii_lowercase + pg.PASSWORD_END)
m = mm.BackoffMarkovModel(config, order=2)
m.train([('pass', 1)])
self.assertEqual(set(m.freq_dict.items()), set([
(mm.PASSWORD_START, 1), ('p', 1), ('a', 1), ('s', 2),
(pg.PASSWORD_END, 1), (mm.PASSWORD_START + 'p', 1),
('pa', 1), ('as', 1), ('ss', 1), ('s' + pg.PASSWORD_END, 1)
]))
def test_train_two(self):
config = Mock()
config.char_bag = (
string.ascii_lowercase + pg.PASSWORD_END)
m = mm.BackoffMarkovModel(config, order=2)
m.train([('pass', 1), ('task', 1)])
self.assertEqual(set(m.freq_dict.items()), set([
(mm.PASSWORD_START, 2),
(pg.PASSWORD_END, 2),
('p', 1),
('a', 2),
('s', 3),
('k', 1),
('t', 1),
(mm.PASSWORD_START + 'p', 1),
('pa', 1),
('as', 2),
('ss', 1),
('s' + pg.PASSWORD_END, 1),
(mm.PASSWORD_START + 't', 1),
('ta', 1),
('sk', 1),
('k' + pg.PASSWORD_END, 1)
]))
def test_predict_short_context(self):
config = Mock()
config.char_bag = ('abc' + pg.PASSWORD_END)
config.backoff_smoothing_threshold = 0
config.additive_smoothing_amount = 0
m = mm.BackoffMarkovModel(config, order=2)
m.train([('abc', 1)])
answer = np.zeros((len(config.char_bag), ), dtype=np.float64)
m.predict('ab', answer)
np.testing.assert_array_almost_equal(
answer, np.array([0., 0., 0., 1.], dtype=np.float64))
answer.fill(0)
m.predict('ba', answer)
np.testing.assert_array_almost_equal(
answer, np.array([0., 0., 1., 0.], dtype=np.float64))
def test_predict_longer_context(self):
config = Mock()
config.char_bag = ('abc' + pg.PASSWORD_END)
config.backoff_smoothing_threshold = 0
config.additive_smoothing_amount = 0
m = mm.BackoffMarkovModel(config, order=3)
m.train([('abc', 1), ('aaa', 1)])
answer = np.zeros((len(config.char_bag), ), dtype=np.float64)
m.predict('ab', answer)
np.testing.assert_array_almost_equal(
answer, np.array([0., 0., 0., 1.], dtype=np.float64))
answer.fill(0)
m.predict('ba', answer)
np.testing.assert_array_almost_equal(
answer, np.array([.25, .5, .25, 0.], dtype=np.float64))
if __name__=='__main__':
unittest.main()