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
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
|
# coding: utf-8
# /*##########################################################################
#
# Copyright (c) 2016-2022 European Synchrotron Radiation Facility
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in
# all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
# THE SOFTWARE.
#
# ###########################################################################*/
__authors__ = ["V.A. Sole", "T. Vincent"]
__license__ = "MIT"
__date__ = "06/12/2022"
import os
import unittest
import tempfile
import shutil
import numpy
import h5py
import hdf5plugin
from hdf5plugin.test import suite as hdf5plugin_suite
class TestHDF5PluginRead(unittest.TestCase):
"""Test reading existing files with compressed data"""
@classmethod
def setUpClass(cls):
cls.tempdir = tempfile.mkdtemp()
@classmethod
def tearDownClass(cls):
shutil.rmtree(cls.tempdir)
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.BLOSC_ID),
"Blosc filter not available")
def testBlosc(self):
"""Test reading Blosc compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
# the blosc.h5 is in fact the example.h5 file generated
# using the example.c file from the blosc respository.
fname = os.path.join(dirname, "blosc.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
data = h5["/dset"][:]
h5.close()
expected_shape = (100, 100, 100)
self.assertTrue(data.shape[0] == 100, "Incorrect shape")
self.assertTrue(data.shape[1] == 100, "Incorrect shape")
self.assertTrue(data.shape[2] == 100, "Incorrect shape")
target = numpy.arange(numpy.prod(expected_shape),
dtype=numpy.float64)
target.shape = expected_shape
self.assertTrue(numpy.allclose(data, target), "Incorrect readout")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.LZ4_ID),
"LZ4 filter not available")
def testLZ4(self):
"""Test reading lz4 compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "lz4.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
data = h5["/entry/data"][:]
h5.close()
self.assertTrue(data.shape[0] == 50, "Incorrect shape")
self.assertTrue(data.shape[1] == 2167, "Incorrect shape")
self.assertTrue(data.shape[2] == 2070, "Incorrect shape")
self.assertTrue(data[21, 1911, 1549] == 3141, "Incorrect value")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.BSHUF_ID),
"Bitshuffle filter not available")
def testBitshuffle(self):
"""Test reading bitshuffle compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "bitshuffle.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
data = h5["/entry/data/data"][:]
h5.close()
self.assertTrue(data.shape[0] == 1, "Incorrect shape")
self.assertTrue(data.shape[1] == 2167, "Incorrect shape")
self.assertTrue(data.shape[2] == 2070, "Incorrect shape")
self.assertTrue(data[0, 1372, 613] == 922, "Incorrect value")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.FCIDECOMP_ID),
"FCIDECOMP filter not available")
def testFcidecomp(self):
"""Test reading FCIDECOMP compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "fcidecomp.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
data = h5["effective_radiance"][:]
h5.close()
expected_data = numpy.arange(1800).astype(numpy.int16).reshape(60, 30)
self.assertTrue(data.shape[0] == 60, "Incorrect shape")
self.assertTrue(data.shape[1] == 30, "Incorrect shape")
self.assertTrue(data.dtype == expected_data.dtype, "Incorrect type")
self.assertTrue(numpy.all(data == expected_data), "Incorrect values read")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.ZFP_ID),
"ZFP filter not available")
def testZfp(self):
"""Test reading ZFP compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
for fname in ["zfp_050.h5", "zfp_052.h5", "zfp_054.h5"]:
fname = os.path.join(dirname, fname)
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
original = h5["original"][()]
compressed = h5["compressed"][()]
h5.close()
self.assertTrue(original.shape == compressed.shape,
"Incorrect shape")
self.assertTrue(original.dtype == compressed.dtype,
"Incorrect dtype")
self.assertFalse(numpy.all(original == compressed),
"Values should not be identical")
self.assertTrue(numpy.allclose(original, compressed),
"Values should be close")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.SPERR_ID),
"Sperr filter not available")
def testSperrV0_1(self):
"""Test reading Sperr compressed data with v0.1 of the filter"""
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "sperr_v0.1.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
with h5py.File(fname, "r") as h5:
compressed = h5["f64_sperr"][()]
original = h5["f64_original"][()]
self.assertTrue(original.shape == compressed.shape,
"Incorrect shape")
self.assertTrue(original.dtype == compressed.dtype,
"Incorrect dtype")
self.assertTrue(numpy.allclose(original, compressed, atol=1e-3),
"Values should be close")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.SPERR_ID),
"Sperr filter not available")
def testSperrV0_2MissingValue(self):
"""Test reading Sperr compressed data with NaN values"""
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "sperr_v0.2_missing_val.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
with h5py.File(fname, "r") as h5:
compressed = h5["f64_sperr"][()]
self.assertTrue(compressed.shape == (128, 64), "Incorrect shape")
self.assertTrue(compressed.dtype == numpy.float64, "Incorrect dtype")
self.assertTrue(numpy.isnan(compressed[0, 1]))
self.assertTrue(numpy.isnan(compressed[0, 10]))
self.assertTrue(numpy.isnan(compressed[1, 36]))
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.SZ_ID),
"SZ filter not available")
def testSZ(self):
"""Test reading SZ compressed data"""
dirname = os.path.abspath(os.path.dirname(__file__))
for fname in ["sz_testfloat_8_8_128.h5"]:
fname = os.path.join(dirname, fname)
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
h5 = h5py.File(fname, "r")
compressed = h5["testfloat"][()]
h5.close()
original_shape = (128, 8, 8)
original = numpy.array([0.23454477,
0.23452051,
0.23450762,
0.23450902,
0.23451449,
0.23453577,
0.23457345,
0.23459189,
0.23454477,
0.23452051,
0.23450762,
0.23450902,
0.23451449,
0.23453577], dtype=numpy.float32)
self.assertTrue(original_shape == compressed.shape,
"Incorrect shape")
self.assertTrue(original.dtype == compressed.dtype,
"Incorrect dtype")
self.assertTrue(numpy.all(original[0:8] == compressed[0, 0, :8]),
"Values should not be identical")
@unittest.skipUnless(h5py.h5z.filter_avail(hdf5plugin.SZ3_ID),
"SZ3 filter not available")
def testSZ3(self):
"""Test reading and witing SZ3 compressed data"""
# the file contains original data and data compressed using hdf5plugin under linux
# the floating point data have been checked against data compressed using h5repack
# for absolute mode and value 1E-3
# h5repack -v -f UD=32024,1,9,0,1062232653,3539053052,0,0,0,0,0,0
# for relative move and value 1E-4
# h5repack -v -f UD=32024,1,9,1,0,0,1058682594,3944497965,0,0,0,0
dirname = os.path.abspath(os.path.dirname(__file__))
fname = os.path.join(dirname, "sz3.h5")
self.assertTrue(os.path.exists(fname),
"Cannot find %s file" % fname)
for dname in ["testfloat_8_8_128", "testdouble_8_8_128"]:
h5 = h5py.File(fname, "r")
original = h5[dname][()]
# absolute 1E-3
value = 1E-3
compressed_name = dname + "_absolute_sz3"
compressed = h5[compressed_name][()]
compressed_back = h5[compressed_name + "_back"][()]
self.assertTrue(original.shape == compressed.shape, "Incorrect shape")
self.assertFalse(numpy.all(original == compressed),
"Values should not be identical")
self.assertTrue(numpy.allclose(compressed, compressed_back),
"Compressed read back values should be identical to compressed data")
# this also tests the algorithm and not just the plugin
self.assertTrue(numpy.allclose(original, compressed, atol=value),
"Values should be within tolerance")
# create a compressed file
output_file = os.path.join(self.tempdir, compressed_name + ".h5")
with h5py.File(output_file, "w", driver="core", backing_store=False) as h5o:
h5o.create_dataset("data", data=original, dtype=original.dtype, chunks=original.shape,
compression=hdf5plugin.SZ3(absolute=value))
output_data = h5o["/data"][()]
self.assertFalse(numpy.all(original == output_data),
"Values should not be identical")
self.assertTrue(numpy.allclose(original, output_data, atol=value),
"Values should be within tolerance")
# relative 1E-4
value = 1E-4
compressed_name = dname + "_relative_sz3"
compressed = h5[compressed_name][()]
compressed_back = h5[compressed_name + "_back"][()]
self.assertTrue(original.shape == compressed.shape, "Incorrect shape")
self.assertFalse(numpy.all(original == compressed),
"Values should not be identical")
# under windows the results are not identical to linux
# therefore we cannot check for equality of decompressed values
self.assertTrue(numpy.allclose(compressed, compressed_back, rtol=0.1 * value),
"Relative read back values should be very close to compressed data")
# create a compressed file
output_file = os.path.join(self.tempdir, compressed_name + ".h5")
with h5py.File(output_file, "w", driver="core", backing_store=False) as h5o:
h5o.create_dataset("data", data=original, dtype=original.dtype, chunks=original.shape,
compression=hdf5plugin.SZ3(relative=value))
output_data = h5o["/data"][()]
self.assertFalse(numpy.all(original == output_data),
"Values should not be identical")
# see what relative and absolute differences are acceptable for this mode
difference = original - compressed_back
idx = numpy.argmax(abs(difference))
difference.shape = -1
rtol = abs(difference[idx] / original.flatten()[idx])
# TODO: Check why one needs to have such large tolerance
rtol = rtol * 5
self.assertTrue(numpy.allclose(original, output_data, rtol=rtol),
"Newly compressed data should match original compression quality")
self.assertTrue(numpy.allclose(compressed_back, output_data, rtol=1.5 * rtol),
"Compressed data should be close")
# L2 Norm
value = 0.33
compressed_name = dname + "_norm2_sz3"
compressed = h5[compressed_name][()]
compressed_back = h5[compressed_name + "_back"][()]
self.assertTrue(original.shape == compressed.shape, "Incorrect shape")
self.assertFalse(numpy.all(original == compressed),
"Values should not be identical")
# Absolute error from L2 norm param from:
# https://github.com/szcompressor/SZ3/blob/v3.1.8/include/SZ3/utils/Statistic.hpp#L44
abs_error = numpy.sqrt(3.0 / compressed.ndim) * value
self.assertTrue(numpy.allclose(compressed, compressed_back, atol=abs_error),
"Compressed read back values should be identical to compressed data")
# create a compressed file
output_file = os.path.join(self.tempdir, compressed_name + ".h5")
with h5py.File(output_file, "w", driver="core", backing_store=False) as h5o:
h5o.create_dataset("data", data=original, dtype=original.dtype, chunks=original.shape,
compression=hdf5plugin.SZ3(norm2=value))
output_data = h5o["/data"][()]
self.assertFalse(numpy.all(original == output_data),
"Values should not be identical")
self.assertTrue(numpy.all(compressed == output_data),
"Compressed data should be identical")
self.assertTrue(numpy.allclose(compressed_back, output_data, atol=abs_error),
"Newly L2 norm read back values should be identical to compressed data")
h5.close()
def suite():
testSuite = unittest.TestSuite()
testSuite.addTest(unittest.TestLoader().loadTestsFromTestCase(
TestHDF5PluginRead))
testSuite.addTest(hdf5plugin_suite())
return testSuite
if __name__ == "__main__":
import sys
result = unittest.TextTestRunner(verbosity=2).run(suite())
if result.wasSuccessful():
sys.exit(0)
else:
sys.exit(1)
|
