|
1
|
|
|
import pytest |
|
2
|
|
|
|
|
3
|
|
|
|
|
4
|
|
|
@pytest.fixture |
|
5
|
|
|
def test_suite(): |
|
6
|
|
|
"""Path of the test suite directory.""" |
|
7
|
|
|
import os |
|
8
|
|
|
return os.path.dirname(os.path.realpath(__file__)) |
|
9
|
|
|
|
|
10
|
|
|
|
|
11
|
|
|
@pytest.fixture |
|
12
|
|
|
def test_image(test_suite): |
|
13
|
|
|
import os |
|
14
|
|
|
def get_image_file_path(file_name): |
|
15
|
|
|
return os.path.join(test_suite, 'data', file_name) |
|
16
|
|
|
return get_image_file_path |
|
17
|
|
|
|
|
18
|
|
|
|
|
19
|
|
|
@pytest.fixture |
|
20
|
|
|
def disk(): |
|
21
|
|
|
from artificial_artwork.disk_operations import Disk |
|
22
|
|
|
return Disk |
|
23
|
|
|
|
|
24
|
|
|
|
|
25
|
|
|
@pytest.fixture |
|
26
|
|
|
def session(): |
|
27
|
|
|
"""Tensorflow v1 Session, with seed defined at runtime. |
|
28
|
|
|
|
|
29
|
|
|
>>> import tensorflow as tf |
|
30
|
|
|
>>> with session(2) as test: |
|
31
|
|
|
... a_C = tf.compat.v1.random_normal([1, 4, 4, 3], mean=1, stddev=4) |
|
32
|
|
|
... a_G = tf.compat.v1.random_normal([1, 4, 4, 3], mean=1, stddev=4) |
|
33
|
|
|
... J_content = compute_cost(a_C, a_G) |
|
34
|
|
|
... assert abs(J_content.eval() - 7.0738883) < 1e-5 |
|
35
|
|
|
|
|
36
|
|
|
Returns: |
|
37
|
|
|
(MySession): A tensorflow session with a set random seed |
|
38
|
|
|
""" |
|
39
|
|
|
import tensorflow as tf |
|
40
|
|
|
class MySession(): |
|
41
|
|
|
def __init__(self, seed): |
|
42
|
|
|
tf.compat.v1.reset_default_graph() |
|
43
|
|
|
self.tf_session = tf.compat.v1.Session() |
|
44
|
|
|
self.seed = seed |
|
45
|
|
|
def __enter__(self): |
|
46
|
|
|
entering_output = self.tf_session.__enter__() |
|
47
|
|
|
tf.compat.v1.set_random_seed(self.seed) |
|
48
|
|
|
return entering_output |
|
49
|
|
|
|
|
50
|
|
|
def __exit__(self, type, value, traceback): |
|
51
|
|
|
# Exception handling here |
|
52
|
|
|
self.tf_session.__exit__(type, value, traceback) |
|
53
|
|
|
return MySession |
|
54
|
|
|
|
|
55
|
|
|
|
|
56
|
|
|
# @pytest.fixture |
|
57
|
|
|
# def default_image_processing_config(): |
|
58
|
|
|
# from artificial_artwork.image import ImageProcessingConfig |
|
59
|
|
|
# return ImageProcessingConfig.from_image_dimensions() |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
@pytest.fixture |
|
63
|
|
|
def image_factory(): |
|
64
|
|
|
"""Production Image Factory. |
|
65
|
|
|
|
|
66
|
|
|
Exposes the 'from_disk(file_path, preprocess=True)'. |
|
67
|
|
|
|
|
68
|
|
|
Returns: |
|
69
|
|
|
ImageFactory: an instance of the ImageFactory class |
|
70
|
|
|
""" |
|
71
|
|
|
from artificial_artwork.image.image_factory import ImageFactory |
|
72
|
|
|
from artificial_artwork.disk_operations import Disk |
|
73
|
|
|
return ImageFactory(Disk.load_image) |
|
74
|
|
|
|
|
75
|
|
|
|
|
76
|
|
|
@pytest.fixture |
|
77
|
|
|
def termination_condition_module(): |
|
78
|
|
|
from artificial_artwork.termination_condition.termination_condition import TerminationConditionFacility, \ |
|
79
|
|
|
TerminationConditionInterface, MaxIterations, TimeLimit, Convergence |
|
80
|
|
|
|
|
81
|
|
|
# all tests require that the Facility already contains some implementations of TerminationCondition |
|
82
|
|
|
assert TerminationConditionFacility.class_registry.subclasses == { |
|
83
|
|
|
'max-iterations': MaxIterations, |
|
84
|
|
|
'time-limit': TimeLimit, |
|
85
|
|
|
'convergence': Convergence, |
|
86
|
|
|
} |
|
87
|
|
|
return type('M', (), { |
|
88
|
|
|
'facility': TerminationConditionFacility, |
|
89
|
|
|
'interface': TerminationConditionInterface, |
|
90
|
|
|
}) |
|
91
|
|
|
|
|
92
|
|
|
|
|
93
|
|
|
@pytest.fixture |
|
94
|
|
|
def termination_condition(termination_condition_module): |
|
95
|
|
|
def create_termination_condition(term_cond_type: str, *args, **kwargs) -> termination_condition_module.interface: |
|
96
|
|
|
return termination_condition_module.facility.create(term_cond_type, *args, **kwargs) |
|
97
|
|
|
return create_termination_condition |
|
98
|
|
|
|
|
99
|
|
|
|
|
100
|
|
|
@pytest.fixture |
|
101
|
|
|
def subscribe(): |
|
102
|
|
|
def _subscribe(broadcaster, listeners): |
|
103
|
|
|
broadcaster.add(*listeners) |
|
104
|
|
|
return _subscribe |
|
105
|
|
|
|
|
106
|
|
|
|
|
107
|
|
|
|
|
108
|
|
|
@pytest.fixture |
|
109
|
|
|
def broadcaster_class(): |
|
110
|
|
|
class TestSubject: |
|
111
|
|
|
def __init__(self, subject, done_callback): |
|
112
|
|
|
self.subject = subject |
|
113
|
|
|
self.done = done_callback |
|
114
|
|
|
|
|
115
|
|
|
def iterate(self): |
|
116
|
|
|
i = 0 |
|
117
|
|
|
while not self.done(): |
|
118
|
|
|
# do something in the current iteration |
|
119
|
|
|
print('Iteration with index', i) |
|
120
|
|
|
|
|
121
|
|
|
# notify when we have completed i+1 iterations |
|
122
|
|
|
self.subject.state = type('Subject', (), { |
|
123
|
|
|
'metrics': {'iterations': i + 1}, # we have completed i+1 iterations |
|
124
|
|
|
}) |
|
125
|
|
|
self.subject.notify() |
|
126
|
|
|
i += 1 |
|
127
|
|
|
return i |
|
128
|
|
|
|
|
129
|
|
|
return TestSubject |
|
130
|
|
|
|
boromir674 /
neural-style-transfer
