-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest.py
159 lines (132 loc) · 7.35 KB
/
test.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
# Copyright 2021 Google LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from train import visualize_metric
from train.train import AccuracyInverseMetric
import numpy as np
from train.data_functions import num_black_cells, gen_data_batch, get_batch, num_black_cells_in_grid
from absl import flags
FLAGS = flags.FLAGS
FLAGS.random_board_prob = 1.0
non_train_indexies = [1,2,3]
class MetricTestCase(unittest.TestCase):
def setUp(self):
FLAGS(['test'])
self.metric_test_eval_datas = gen_data_batch(500, 4)
def metric_asserts(self, eval_datas, gen_boards, expected_min, expected_max, non_train_indexies):
metric_val = visualize_metric.visualize_metric(eval_datas, gen_boards, non_train_indexies)
self.assertGreaterEqual(metric_val, expected_min)
self.assertLessEqual(metric_val, expected_max)
def test_same_data_should_be_1_even_when_reordered(self):
metric_test_gen_datas = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 2], self.metric_test_eval_datas[:, 3],
self.metric_test_eval_datas[:, 1], self.metric_test_eval_datas[:, 4]], axis=1)
self.metric_asserts(self.metric_test_eval_datas, metric_test_gen_datas, .999, 1.01, non_train_indexies)
def test_no_credit_repeating_same_state(self):
metric_test_gen_datas = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 1], self.metric_test_eval_datas[:, 2],
self.metric_test_eval_datas[:, 1], self.metric_test_eval_datas[:, 4]], axis=1)
# (1+1+0) / 3
self.metric_asserts(self.metric_test_eval_datas, metric_test_gen_datas, .666, .8, non_train_indexies)
def test_repeating_first_state(self):
metric_test_gen_datas = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0],
self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0]], axis=1)
self.metric_asserts(self.metric_test_eval_datas, metric_test_gen_datas, .0, .07, non_train_indexies)
def test_all_zeros(self):
self.metric_asserts(np.zeros_like(self.metric_test_eval_datas), np.zeros_like(self.metric_test_eval_datas), .999, 1.01, non_train_indexies)
def test_all_ones(self):
self.metric_asserts(np.ones_like(self.metric_test_eval_datas), np.ones_like(self.metric_test_eval_datas), .999, 1.01, non_train_indexies)
def test_gen_zeros(self):
self.metric_asserts(np.zeros_like(self.metric_test_eval_datas), self.metric_test_eval_datas, .0, .01, non_train_indexies)
def test_no_credit_for_start_or_end_state(self):
metric_test_gen_datas = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0],
self.metric_test_eval_datas[:, 4], self.metric_test_eval_datas[:, 4]], axis=1)
# It should be around 0
self.metric_asserts(self.metric_test_eval_datas, metric_test_gen_datas, .0, .3, non_train_indexies)
def test_combine_metric_works(self):
metric_test_gen_datas = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 2], self.metric_test_eval_datas[:, 0],
self.metric_test_eval_datas[:, 4], self.metric_test_eval_datas[:, 4]], axis=1)
combine_val = visualize_metric.combine_metric(self.metric_test_eval_datas, self.metric_test_eval_datas, metric_test_gen_datas, non_train_indexies)
print(combine_val)
self.assertGreaterEqual(combine_val, 1)
self.assertLessEqual(combine_val, 1.1)
def test_game3_model4_steps(self):
gen_boards = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 1],
self.metric_test_eval_datas[:, 1],
self.metric_test_eval_datas[:, 2], self.metric_test_eval_datas[:, 3]], axis=1)
eval_datas = self.metric_test_eval_datas[:, :4]
self.metric_asserts(eval_datas, gen_boards, .99, 1.01, non_train_indexies)
def test_game3_model4_steps2(self):
gen_boards = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 0],
self.metric_test_eval_datas[:, 3],
self.metric_test_eval_datas[:, 2], self.metric_test_eval_datas[:, 3]], axis=1)
eval_datas = self.metric_test_eval_datas[:, :4]
# expected score: (1+0) / 2 = .5
self.metric_asserts(eval_datas, gen_boards, .5, .67, non_train_indexies)
def test_game4_model3_steps(self):
gen_boards = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 1],
self.metric_test_eval_datas[:, 2], self.metric_test_eval_datas[:, 3]], axis=1)
eval_datas = self.metric_test_eval_datas
# expected score: (1+1+0) / 2 = 1.0
self.metric_asserts(eval_datas, gen_boards, 1.0, 1.2, [1,2])
def test_game4_model3_steps_2(self):
gen_boards = np.stack(
[self.metric_test_eval_datas[:, 0], self.metric_test_eval_datas[:, 3],
self.metric_test_eval_datas[:, 4], self.metric_test_eval_datas[:, 4]], axis=1)
eval_datas = self.metric_test_eval_datas
# expected score: (0+1+0) / 2 = .5
self.metric_asserts(eval_datas, gen_boards, .54, .8, [1,2])
def test_game3_model2_steps(self):
gen_boards = np.stack(
[self.metric_test_eval_datas[:, 0], np.zeros_like(self.metric_test_eval_datas[:, 4]), self.metric_test_eval_datas[:, 4]], axis=1)
eval_datas = self.metric_test_eval_datas[:, :4]
# expected score: (0+0) / 1 = .0
self.metric_asserts(eval_datas, gen_boards, .0, .37, [1])
class AccuracyInverseMetricTestCase(unittest.TestCase):
def setUp(self):
FLAGS(['test'])
def test_board_size(self):
metric = AccuracyInverseMetric(FLAGS.board_size)
y_pred = np.array([21.3, 10.6, 70.8, 90, 35.2]) / (FLAGS.board_size **2)
y_true = np.array([21, 10, 70, 97, 35]) / (FLAGS.board_size **2)
metric.update_state(y_true, y_pred)
self.assertAlmostEqual(1-2/5, metric.result().numpy())
def test_size_2(self):
metric = AccuracyInverseMetric(2)
y_pred = np.array([1.49, .51, 1.51, 3.4, 3.4]) / 4.0
y_true = np.array([1, 0, 2, 4, 3]) / 4.0
metric.update_state(y_true, y_pred)
self.assertAlmostEqual(1-3/5, metric.result().numpy())
class DataTestCase(unittest.TestCase):
def setUp(self):
FLAGS(['test'])
def test_grid(self):
n = FLAGS.board_size
g = FLAGS.grid_size
X = np.random.rand(1, 1, n, n, 1)
y = np.zeros_like(X)
for i in range(n):
for j in range(n):
a = i // g
b = j // g
y[0, 0, i, j, 0] = np.mean(X[0, 0, a * g:(a + 1) * g, b * g:(b + 1) * g, 0])
np.testing.assert_array_almost_equal(y, num_black_cells_in_grid(X))
if __name__ == '__main__':
unittest.main()