|
1
|
|
|
import copy |
|
2
|
|
|
import pytest |
|
3
|
|
|
import math |
|
4
|
|
|
import numpy as np |
|
5
|
|
|
import pandas as pd |
|
6
|
|
|
|
|
7
|
|
|
from hyperactive.optimizers import HillClimbingOptimizer |
|
8
|
|
|
from hyperactive.experiment import BaseExperiment, add_catch |
|
9
|
|
|
from hyperactive.search_config import SearchConfig |
|
10
|
|
|
|
|
11
|
|
|
|
|
12
|
|
|
search_config = SearchConfig( |
|
13
|
|
|
x0=list(np.arange(-10, 10, 1)), |
|
14
|
|
|
) |
|
15
|
|
|
|
|
16
|
|
|
|
|
17
|
|
|
def test_catch_1(): |
|
18
|
|
|
class Experiment(BaseExperiment): |
|
19
|
|
|
@add_catch({TypeError: np.nan}) |
|
20
|
|
|
def objective_function(self, access): |
|
21
|
|
|
a = 1 + "str" |
|
22
|
|
|
|
|
23
|
|
|
return 0 |
|
24
|
|
|
|
|
25
|
|
|
experiment = Experiment() |
|
26
|
|
|
|
|
27
|
|
|
hyper = HillClimbingOptimizer() |
|
28
|
|
|
hyper.add_search( |
|
29
|
|
|
experiment, |
|
30
|
|
|
search_config, |
|
31
|
|
|
n_iter=100, |
|
32
|
|
|
) |
|
33
|
|
|
hyper.run() |
|
34
|
|
|
|
|
35
|
|
|
|
|
36
|
|
|
def test_catch_2(): |
|
37
|
|
|
class Experiment(BaseExperiment): |
|
38
|
|
|
@add_catch({ValueError: np.nan}) |
|
39
|
|
|
def objective_function(self, access): |
|
40
|
|
|
math.sqrt(-10) |
|
41
|
|
|
|
|
42
|
|
|
return 0 |
|
43
|
|
|
|
|
44
|
|
|
experiment = Experiment() |
|
45
|
|
|
|
|
46
|
|
|
hyper = HillClimbingOptimizer() |
|
47
|
|
|
hyper.add_search( |
|
48
|
|
|
experiment, |
|
49
|
|
|
search_config, |
|
50
|
|
|
n_iter=100, |
|
51
|
|
|
) |
|
52
|
|
|
hyper.run() |
|
53
|
|
|
|
|
54
|
|
|
|
|
55
|
|
|
def test_catch_3(): |
|
56
|
|
|
class Experiment(BaseExperiment): |
|
57
|
|
|
@add_catch({ZeroDivisionError: np.nan}) |
|
58
|
|
|
def objective_function(self, access): |
|
59
|
|
|
x = 1 / 0 |
|
60
|
|
|
|
|
61
|
|
|
return 0 |
|
62
|
|
|
|
|
63
|
|
|
experiment = Experiment() |
|
64
|
|
|
|
|
65
|
|
|
hyper = HillClimbingOptimizer() |
|
66
|
|
|
hyper.add_search( |
|
67
|
|
|
experiment, |
|
68
|
|
|
search_config, |
|
69
|
|
|
n_iter=100, |
|
70
|
|
|
) |
|
71
|
|
|
hyper.run() |
|
72
|
|
|
|
|
73
|
|
|
|
|
74
|
|
|
def test_catch_all_0(): |
|
75
|
|
|
class Experiment(BaseExperiment): |
|
76
|
|
|
@add_catch( |
|
77
|
|
|
{ |
|
78
|
|
|
TypeError: np.nan, |
|
79
|
|
|
ValueError: np.nan, |
|
80
|
|
|
ZeroDivisionError: np.nan, |
|
81
|
|
|
} |
|
82
|
|
|
) |
|
83
|
|
|
def objective_function(self, access): |
|
84
|
|
|
a = 1 + "str" |
|
85
|
|
|
math.sqrt(-10) |
|
86
|
|
|
x = 1 / 0 |
|
87
|
|
|
|
|
88
|
|
|
return 0 |
|
89
|
|
|
|
|
90
|
|
|
experiment = Experiment() |
|
91
|
|
|
|
|
92
|
|
|
hyper = HillClimbingOptimizer() |
|
93
|
|
|
hyper.add_search( |
|
94
|
|
|
experiment, |
|
95
|
|
|
search_config, |
|
96
|
|
|
n_iter=100, |
|
97
|
|
|
) |
|
98
|
|
|
hyper.run() |
|
99
|
|
|
|
|
100
|
|
|
nan_ = hyper.search_data(experiment)["score"].values[0] |
|
101
|
|
|
|
|
102
|
|
|
assert math.isnan(nan_) |
|
103
|
|
|
|
|
104
|
|
|
|
|
105
|
|
|
def test_catch_all_1(): |
|
106
|
|
|
catch_return = (np.nan, {"error": True}) |
|
107
|
|
|
|
|
108
|
|
|
class Experiment(BaseExperiment): |
|
109
|
|
|
@add_catch( |
|
110
|
|
|
{ |
|
111
|
|
|
TypeError: catch_return, |
|
112
|
|
|
ValueError: catch_return, |
|
113
|
|
|
ZeroDivisionError: catch_return, |
|
114
|
|
|
} |
|
115
|
|
|
) |
|
116
|
|
|
def objective_function(self, access): |
|
117
|
|
|
a = 1 + "str" |
|
118
|
|
|
math.sqrt(-10) |
|
119
|
|
|
x = 1 / 0 |
|
120
|
|
|
|
|
121
|
|
|
return 0, {"error": False} |
|
122
|
|
|
|
|
123
|
|
|
experiment = Experiment() |
|
124
|
|
|
|
|
125
|
|
|
hyper = HillClimbingOptimizer() |
|
126
|
|
|
hyper.add_search( |
|
127
|
|
|
experiment, |
|
128
|
|
|
search_config, |
|
129
|
|
|
n_iter=100, |
|
130
|
|
|
) |
|
131
|
|
|
hyper.run() |
|
132
|
|
|
|
|
133
|
|
|
nan_ = hyper.search_data(experiment)["score"].values[0] |
|
134
|
|
|
error_ = hyper.search_data(experiment)["error"].values[0] |
|
135
|
|
|
|
|
136
|
|
|
assert math.isnan(nan_) |
|
137
|
|
|
assert error_ == True |
|
138
|
|
|
|
SimonBlanke /
Hyperactive
