forked from fedbiomed/fedbiomed
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest_torch_data_manager.py
198 lines (148 loc) · 7.21 KB
/
test_torch_data_manager.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
import unittest
import fedbiomed.common.data._torch_data_manager # noqa
import numpy as np
from unittest.mock import patch
from torch.utils.data import Dataset, Subset
from fedbiomed.common.data import TorchDataManager
from fedbiomed.common.exceptions import FedbiomedTorchDataManagerError
class TestTorchDataManager(unittest.TestCase):
class CustomDataset(Dataset):
""" Create PyTorch Dataset for test purposes """
def __init__(self):
self.X_train = np.array([[1, 2, 3],
[1, 2, 3],
[1, 2, 3],
[1, 2, 3],
[1, 2, 3],
[1, 2, 3]])
self.Y_train = np.array([1, 2, 3, 4, 5, 6])
def __len__(self):
return len(self.Y_train)
def __getitem__(self, idx):
return self.X_train[idx], self.Y_train[idx]
class CustomDatasetInvalid(Dataset):
""" Create Invalid PyTorch Dataset for test purposes """
def __init__(self):
self.X_train = [[1, 2, 3],
[1, 2, 3],
[1, 2, 3],
[1, 2, 3]]
self.Y_train = [1, 2, 3]
def __getitem__(self, idx):
return self.X_train[idx], self.Y_train[idx]
class CustomDatasetAttrError(Dataset):
""" Create PyTorch Dataset that raises Attr Error for test purposes """
def __init__(self):
self.data = None
pass
def __len__(self):
raise AttributeError
def __getitem__(self, idx):
return self.data
class CustomDatasetTypeError(Dataset):
""" Create PyTorch Dataset that raises TypeError for test purposes """
def __init__(self):
self.data = None
pass
def __len__(self):
raise TypeError
def __getitem__(self, idx):
return self.data
def setUp(self):
# Setup global TorchDataManager class
self.dataset = TestTorchDataManager.CustomDataset()
self.torch_data_manager = TorchDataManager(dataset=self.dataset,
batch_size=48,
shuffle=True)
def tearDown(self):
pass
def test_torch_data_manager_01_init_failure(self):
""" Testing build failure of Torch Data Manager """
with self.assertRaises(FedbiomedTorchDataManagerError):
TorchDataManager(dataset='wrong_type',
batch_size=48,
shuffle=True)
def test_torch_data_manager_01_dataset(self):
""" Testing dataset getter method """
result = self.torch_data_manager.dataset
self.assertEqual(result, self.dataset, 'dataset() returns un expected torch Dataset object')
def test_torch_data_manager_02_split(self):
"""Testing split method of TorchDataManager class """
# Test invalid ratio argument
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(test_ratio=12)
# Test invalid ratio argument
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(test_ratio='12')
# Test invalid ratio argument
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(test_ratio=-12)
# Test proper split
try:
self.torch_data_manager.split(0.3)
except:
self.assertTrue(False, 'Error while splitting TorchDataManager')
# Test exceptions
invalid = TestTorchDataManager.CustomDatasetInvalid()
self.torch_data_manager._dataset = invalid
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(0.3)
invalid = TestTorchDataManager.CustomDatasetTypeError()
self.torch_data_manager._dataset = invalid
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(0.3)
invalid = TestTorchDataManager.CustomDatasetAttrError()
self.torch_data_manager._dataset = invalid
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager.split(0.3)
def test_torch_data_manager_05_split_results(self):
""" Test splitting result """
# Test with split
loader_train, loader_test = self.torch_data_manager.split(0.5)
self.assertEqual(len(loader_train.dataset), len(self.dataset) / 2, 'Did not properly get loader '
'of train partition')
# If test partition is zero
loader_train, loader_test = self.torch_data_manager.split(0)
self.assertIsNone(loader_test, 'Loader is not None where it should be')
# If test partition is 1
loader_train, loader_test = self.torch_data_manager.split(1)
self.assertIsNone(loader_train, 'Loader is not None where it should be')
def test_torch_data_manager_05_subset_train(self):
""" Testing the method load train partition """
# Test with split
self.torch_data_manager.split(0.5)
subset = self.torch_data_manager.subset_train()
self.assertIsInstance(subset, Subset, 'Can not get proper subset object')
def test_torch_data_manager_05_subset_test(self):
""" Testing the method load train partition """
# Test with split
self.torch_data_manager.split(0.5)
subset = self.torch_data_manager.subset_test()
self.assertIsInstance(subset, Subset, 'Can not get proper subset object')
def test_torch_data_manager_05_load_all_samples(self):
""" Testing the method load train partition """
# Test with split
self.torch_data_manager.split(0.5)
loader = self.torch_data_manager.load_all_samples()
self.assertEqual(len(loader.dataset), len(self.dataset), 'Did not properly get loader for all samples')
@patch('fedbiomed.common.data._torch_data_manager.DataLoader')
def test_torch_data_manager_06_create_torch_data_loader(self, data_loader):
""" Test function create torch data loader """
self.torch_data_manager.split(0.5)
s = self.torch_data_manager.subset_test()
data_loader.side_effect = TypeError()
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager._create_torch_data_loader(s)
data_loader.side_effect = AttributeError()
with self.assertRaises(FedbiomedTorchDataManagerError):
self.torch_data_manager._subset_loader(s)
data_loader.side_effect = None
data_loader.return_value = 'Data'
result = self.torch_data_manager._subset_loader(s)
self.assertEqual(result, 'Data')
def test_torch_data_manager_07_to_sklearn(self):
"""Test converting TorchDataManage to SkLearnDataManager"""
result = self.torch_data_manager.to_sklearn()
self.assertIsInstance(result, fedbiomed.common.data._sklearn_data_manager.SkLearnDataManager)
if __name__ == '__main__': # pragma: no cover
unittest.main()