-
Notifications
You must be signed in to change notification settings - Fork 4
/
test_nmpy.py
206 lines (160 loc) · 6.27 KB
/
test_nmpy.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
import unittest
try:
import numpy as np
from .nmpy import NumpyFeature, StreamingNumpyDecoder, PackedNumpyEncoder
except ImportError:
np = None
from .persistence import PersistenceSettings
from .data import *
from .model import BaseModel
from .lmdbstore import LmdbDatabase
from .extractor import Node
from tempfile import mkdtemp
from shutil import rmtree
class PassThrough(Node):
def __init__(self, needs=None):
super(PassThrough, self).__init__(needs=needs)
def _process(self, data):
yield data
class BaseNumpyTest(object):
def setUp(self):
if np is None:
self.skipTest('numpy is not available')
class Settings(PersistenceSettings):
id_provider = UuidProvider()
key_builder = StringDelimitedKeyBuilder()
database = InMemoryDatabase(key_builder=key_builder)
self.Settings = self._register_database(Settings)
def _check_array(self, arr, shape, dtype, orig):
self.assertTrue(isinstance(arr, np.ndarray))
self.assertTrue(np.all(arr == orig))
self.assertEqual(shape, arr.shape)
self.assertEqual(dtype, arr.dtype)
def _build_doc(self):
class Doc(BaseModel, self.Settings):
feat = NumpyFeature(PassThrough, store=True)
packed = NumpyFeature(
PassThrough,
needs=feat,
encoder=PackedNumpyEncoder,
store=True)
return Doc
def _restore(self, data):
return data
def _arrange(self, shape=None, dtype=None):
cls = self._build_doc()
arr = np.recarray(shape, dtype=dtype) \
if isinstance(dtype, list) else np.zeros(shape, dtype=dtype)
_id = cls.process(feat=arr)
doc = cls(_id)
recovered = self._restore(doc.feat)
self._check_array(recovered, shape, dtype, arr)
def _register_database(self):
raise NotImplemented()
def test_can_store_and_retrieve_packed_array(self):
cls = self._build_doc()
arr = np.zeros((10, 9))
_id = cls.process(feat=arr)
doc = cls(_id)
recovered = self._restore(doc.packed)
self.assertEqual(np.uint8, recovered.dtype)
self.assertEqual((10, 2), recovered.shape)
def test_can_store_and_retrieve_empty_array(self):
self._arrange((0,), np.uint8)
def test_can_store_and_retrieve_1d_float32_array(self):
self._arrange((33,), np.float32)
def test_can_store_and_retreive_multidimensional_uint8_array(self):
self._arrange((12, 13), np.uint8)
def test_can_store_and_retrieve_multidimensional_float32_array(self):
self._arrange((5, 10, 11), np.float32)
def test_can_store_and_retrieve_recarray(self):
self._arrange(shape=(25,), dtype=[ \
('x', np.uint8, (509,)),
('y', 'a32')])
class GreedyNumpyTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
return settings_class.clone(
database=InMemoryDatabase(key_builder=settings_class.key_builder))
class GreedyNumpyOnDiskTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
self._dir = mkdtemp()
return settings_class.clone(database=FileSystemDatabase(
path=self._dir,
key_builder=settings_class.key_builder))
def tearDown(self):
rmtree(self._dir)
class GreedyNumpyLmdbTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
self._dir = mkdtemp()
return settings_class.clone(database=LmdbDatabase(
path=self._dir,
map_size=10000000,
key_builder=settings_class.key_builder))
def tearDown(self):
rmtree(self._dir)
class StreamingNumpyTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
return settings_class.clone(
database=InMemoryDatabase(key_builder=settings_class.key_builder))
def _build_doc(self):
class Doc(BaseModel, self.Settings):
feat = NumpyFeature(
PassThrough,
store=True,
decoder=StreamingNumpyDecoder(n_examples=3))
packed = NumpyFeature(
PassThrough,
needs=feat,
encoder=PackedNumpyEncoder,
decoder=StreamingNumpyDecoder(n_examples=3),
store=True)
return Doc
def _restore(self, data):
return np.concatenate(list(data))
class StreamingNumpyOnDiskTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
self._dir = mkdtemp()
return settings_class.clone(database=FileSystemDatabase(
path=self._dir,
key_builder=settings_class.key_builder))
def tearDown(self):
rmtree(self._dir)
def _build_doc(self):
class Doc(BaseModel, self.Settings):
feat = NumpyFeature(
PassThrough,
store=True,
decoder=StreamingNumpyDecoder(n_examples=3))
packed = NumpyFeature(
PassThrough,
needs=feat,
encoder=PackedNumpyEncoder,
decoder=StreamingNumpyDecoder(n_examples=3),
store=True)
return Doc
def _restore(self, data):
return np.concatenate(list(data))
class StreamingNumpyLmdbTest(BaseNumpyTest, unittest.TestCase):
def _register_database(self, settings_class):
self._dir = mkdtemp()
return settings_class.clone(database=LmdbDatabase(
path=self._dir,
map_size=10000000,
key_builder=settings_class.key_builder))
def tearDown(self):
rmtree(self._dir)
def _build_doc(self):
class Doc(BaseModel, self.Settings):
feat = NumpyFeature(
PassThrough,
store=True,
decoder=StreamingNumpyDecoder(n_examples=3))
packed = NumpyFeature(
PassThrough,
needs=feat,
encoder=PackedNumpyEncoder,
decoder=StreamingNumpyDecoder(n_examples=3),
store=True)
return Doc
def _restore(self, data):
return np.concatenate(list(data))