-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtop_test.py
223 lines (208 loc) · 10.3 KB
/
top_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
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
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
import os
import tempfile
import typing
import pytest
from sklearn import datasets, model_selection
from . import (base, raw_dataset, raw_exp_ensemble_plain_advanced,
raw_exp_ensemble_risk, raw_exp_iop, raw_pipe, raw_poisoning, \
raw_dataset_test,
top)
@pytest.mark.parametrize('config', [
top.TopLevelExpIop(
base_output_directory='',
dataset_config_to_poison=raw_dataset.DatasetToPoisonRaw(
dataset_path_training='',
dataset_path_testing='',
poisoning_input=raw_poisoning.PoisoningGenerationInfoRaw(
selector=raw_poisoning.SelectorRaw(name='__poisoning.SelectorRandom'),
performer=raw_poisoning.PerformerRaw(name='__poisoning.PerformerLabelFlippingMonoDirectional'),
perc_data_points=[5.0, 7.5, 8.5],
perform_info_clazz='_poisoning.PerformInfoMonoDirectional',
perform_info_kwargs={'from_label': 0, 'to_label': 1},
selection_info_clazz='_poisoning.SelectionInfoEmpty',
# selection_info_kwargs={'from_label': 0}
selection_info_kwargs={}
)
),
dataset_exists_ok=True,
repetitions=2,
export_config=raw_exp_iop.ExportConfigExpIoPRaw(
exists_ok=True,
export_also_raw_results=False,
),
pipelines=[
raw_pipe.PipelineRaw(name='p1',
steps_to_evaluate=[0],
steps=[
raw_pipe.StepRaw('last',
'__assignments.AssignmentRoundRobinBlind',
output_col_names_pre=['assignment'],
step_func_kwargs={'N': 3})
])
]
),
top.TopLevelExpEnsemblePlainAdvanced(
# base_model_name='__sklearn.ensemble.RandomForestClassifier',
monolithic_model=base.FuncPair(func_name='__sklearn.ensemble.RandomForestClassifier'),
base_output_directory='',
dataset_config_to_poison=raw_dataset.DatasetToPoisonRaw(
dataset_path_training='',
dataset_path_testing='',
poisoning_input=raw_poisoning.PoisoningGenerationInfoRaw(
selector=raw_poisoning.SelectorRaw(name='__poisoning.SelectorRandom'),
performer=raw_poisoning.PerformerRaw(name='__poisoning.PerformerLabelFlippingMonoDirectional'),
perc_data_points=[5.0, 7.5, 8.1],
perform_info_clazz='_poisoning.PerformInfoMonoDirectional',
perform_info_kwargs={'from_label': 0, 'to_label': 1},
selection_info_clazz='_poisoning.SelectionInfoEmpty',
selection_info_kwargs={}
)
),
dataset_exists_ok=True,
export_config=raw_exp_ensemble_plain_advanced.ExportConfigExpEnsemblePlainAdvancedRaw(
exists_ok=True
),
repetitions=2,
pipelines=[
raw_pipe.PipelineRaw(name='p1',
steps=[
raw_pipe.StepRaw('last',
'__assignments.AssignmentRoundRobinBlind',
step_func_kwargs={'N': 3})
])
]
),
top.TopLevelExpEnsembleRisk(
# base_model_name='__sklearn.ensemble.RandomForestClassifier',
monolithic_model=base.FuncPair(func_name='__sklearn.ensemble.RandomForestClassifier'),
base_output_directory='',
dataset_config_to_poison=raw_dataset.DatasetToPoisonRaw(
dataset_path_training='',
dataset_path_testing='',
poisoning_input=raw_poisoning.PoisoningGenerationInfoRaw(
selector=raw_poisoning.SelectorRaw(name='__poisoning.SelectorRandom'),
performer=raw_poisoning.PerformerRaw(name='__poisoning.PerformerLabelFlippingMonoDirectional'),
perc_data_points=[5.0, 7.5],
perform_info_clazz='_poisoning.PerformInfoMonoDirectional',
perform_info_kwargs={'from_label': 0, 'to_label': 1},
selection_info_clazz='_poisoning.SelectionInfoEmpty',
selection_info_kwargs={}
)
),
dataset_exists_ok=True,
export_config=raw_exp_ensemble_risk.ExportConfigExpEnsembleRiskRaw(
exists_ok=True
),
repetitions=2,
pipelines=[
raw_pipe.PipelineRaw(name='p1',
steps_to_evaluate=[0],
risk_idx=0,
steps=[
raw_pipe.StepRaw('last',
'__assignments.AssignmentRoundRobinBlind',
output_col_names_pre=['assignment'],
step_func_kwargs={'N': 3})
])
]
),
# one test where we also have some baselines.
top.TopLevelExpEnsembleRisk(
# base_model_name='__sklearn.ensemble.RandomForestClassifier',
monolithic_model=base.FuncPair(func_name='__sklearn.ensemble.RandomForestClassifier'),
base_output_directory='',
dataset_config_to_poison=raw_dataset.DatasetToPoisonRaw(
dataset_path_training='',
dataset_path_testing='',
poisoning_input=raw_poisoning.PoisoningGenerationInfoRaw(
selector=raw_poisoning.SelectorRaw(name='__poisoning.SelectorRandom'),
performer=raw_poisoning.PerformerRaw(name='__poisoning.PerformerLabelFlippingMonoDirectional'),
perc_data_points=[5.0, 7.5],
perform_info_clazz='_poisoning.PerformInfoMonoDirectional',
perform_info_kwargs={'from_label': 0, 'to_label': 1},
selection_info_clazz='_poisoning.SelectionInfoEmpty',
selection_info_kwargs={}
)
),
dataset_exists_ok=True,
export_config=raw_exp_ensemble_risk.ExportConfigExpEnsembleRiskRaw(
exists_ok=True
),
repetitions=2,
pipelines=[
raw_pipe.PipelineRaw(name='p1',
steps_to_evaluate=[0],
risk_idx=0,
steps=[
raw_pipe.StepRaw('last',
'__assignments.AssignmentRoundRobinBlind',
output_col_names_pre=['assignment'],
step_func_kwargs={'N': 3})
])
],
know_all_pipelines=[
raw_pipe.PipelineRaw(name='BASELINE(p1)',
steps_to_evaluate=[0],
risk_idx=0,
steps=[
raw_pipe.StepRaw('last',
'__assignments.AssignmentRoundRobinBlind',
output_col_names_pre=['assignment'],
step_func_kwargs={'N': 3})
])
]
),
])
def test_exp(config: typing.Union[top.TopLevelExpIop, top.TopLevelExpEnsembleRisk, top.TopLevelExpEnsemblePlainAdvanced]):
X, y = datasets.make_classification()
X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y)
with tempfile.NamedTemporaryFile() as training_file, tempfile.NamedTemporaryFile() as testing_file:
config.dataset_config.dataset_path_training = training_file.name
config.dataset_config.dataset_path_testing = testing_file.name
raw_dataset_test.X_y_to_csv(X_train, y_train, training_file.name)
raw_dataset_test.X_y_to_csv(X_test, y_test, testing_file.name)
with tempfile.TemporaryDirectory() as temp_dir:
config.base_output_directory = temp_dir
config.do()
# just ensure it is not empty
assert len(os.listdir(temp_dir)) > 0
# now, we do a very simple test: we re-do the same work but forbidding
# result overriding.
config.export_config.exists_ok = False
with tempfile.TemporaryDirectory() as temp_dir:
# we need to create the specific subdirectory as well.
output_dir = os.path.join(os.path.abspath(temp_dir), base.BASE_OUTPUT_DIR_OUTPUT)
os.makedirs(output_dir, exist_ok=True)
# but we pass as input the base dir, because the code itself will add base.BASE_OUTPUT_DIR_OUTPUT
# at the end.
config.base_output_directory = temp_dir
with pytest.raises(ValueError):
config.do()
@pytest.mark.parametrize('config', [
top.TopLevelDataset(
base_output_directory='',
dataset_path_training='',
dataset_path_testing='',
exists_ok=True,
poisoning_input=raw_poisoning.PoisoningGenerationInfoRaw(
selector=raw_poisoning.SelectorRaw(name='__poisoning.SelectorRandom'),
performer=raw_poisoning.PerformerRaw(name='__poisoning.PerformerLabelFlippingMonoDirectional'),
perc_data_points=[5.0, 7.5],
perform_info_clazz='_poisoning.PerformInfoMonoDirectional',
perform_info_kwargs={'from_label': 0, 'to_label': 1},
selection_info_clazz='_poisoning.SelectionInfoEmpty',
selection_info_kwargs={}
)
)
])
def test_dg(config: top.TopLevelDataset):
X, y = datasets.make_classification()
X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y)
with tempfile.NamedTemporaryFile() as training_file, tempfile.NamedTemporaryFile() as testing_file:
config.dataset_path_training = training_file.name
config.dataset_path_testing = testing_file.name
raw_dataset_test.X_y_to_csv(X_train, y_train, training_file.name)
raw_dataset_test.X_y_to_csv(X_test, y_test, testing_file.name)
with tempfile.TemporaryDirectory() as temp_dir:
config.base_output_directory = temp_dir
config.do()