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
|
#!/usr/bin/env python
# coding: utf-8
#
# Project: Azimuthal integration
# https://github.com/silx-kit/pyFAI
#
# Copyright (C) 2015-2018 European Synchrotron Radiation Facility, Grenoble, France
#
# Principal author: Jérôme Kieffer (Jerome.Kieffer@ESRF.eu)
#
# 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.
"""Test suites for pixel splitting scheme validation
see sandbox/debug_split_pixel.py for visual validation
"""
__author__ = "Jérôme Kieffer"
__contact__ = "Jerome.Kieffer@ESRF.eu"
__license__ = "MIT"
__copyright__ = "European Synchrotron Radiation Facility, Grenoble, France"
__date__ = "01/02/2021"
import unittest
import platform
import numpy
import logging
from .utilstest import UtilsTest
logger = logging.getLogger(__name__)
from ..azimuthalIntegrator import AzimuthalIntegrator
from ..detectors import Detector
from ..utils import mathutil
from ..ext import splitBBox, splitPixel
from ..method_registry import IntegrationMethod
class TestSplitPixel(unittest.TestCase):
@classmethod
def setUpClass(cls):
super(TestSplitPixel, cls).setUpClass()
img = numpy.zeros((512, 512))
for i in range(1, 6):
img[i * 100, i * 100] = 1
det = Detector(1e-4, 1e-4)
det.shape = (512, 512)
ai = AzimuthalIntegrator(1, detector=det)
cls.results = {}
cls.results_ng = {}
for i, meth in enumerate(["numpy", "cython", "splitbbox", "splitpixel", "csr_no", "csr_bbox", "csr_full"]):
cls.results[meth] = ai.integrate1d_legacy(img, 10000, method=meth, unit="2th_deg")
ai.reset()
for k, v in IntegrationMethod._registry.items():
if v.dimension == 1 and v.target is None: # exclude OpenCL engines
cls.results_ng[k] = ai.integrate1d_ng(img, 10000, method=v, unit="r_mm")
@classmethod
def tearDownClass(cls):
super(TestSplitPixel, cls).tearDownClass()
cls.results = None
def test_new_gen_algoritms(self):
"This checks that the pixel splitting scheme gives consistent results"
self.assertGreater(len(self.results_ng), 0, msg="we have some results")
thres = 7
for k1, res1 in self.results_ng.items():
if k1.split == "pseudo":
# Those are half implemented algorithms ... avoid testing them!
continue
for k2, res2 in self.results_ng.items():
if k1 == k2:
continue
if k2.split == "pseudo":
continue
R = mathutil.rwp(res1, res2)
print (f"({k1.split},{k1.algo})/({k2.split},{k2.algo})\t {R}")
if k1.split == k2.split:
self.assertTrue(R < thres, f"{k1}/{k2}")
else:
self.assertTrue(R > thres, f"{k1}/{k2}")
def test_no_split(self):
"""
Validate that all non splitting algo give the same result...
"""
thres = 7
self.assertTrue(mathutil.rwp(self.results["numpy"], self.results["cython"]) < thres, "Cython/Numpy")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["cython"]) < thres, "Cython/CSR")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["numpy"]) < thres, "CSR/numpy")
self.assertTrue(mathutil.rwp(self.results["splitbbox"], self.results["numpy"]) > thres, "splitbbox/Numpy")
self.assertTrue(mathutil.rwp(self.results["splitpixel"], self.results["numpy"]) > thres, "splitpixel/Numpy")
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["numpy"]) > thres, "csr_bbox/Numpy")
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["numpy"]) > thres, "csr_full/Numpy")
self.assertTrue(mathutil.rwp(self.results["splitbbox"], self.results["cython"]) > thres, "splitbbox/cython")
self.assertTrue(mathutil.rwp(self.results["splitpixel"], self.results["cython"]) > thres, "splitpixel/cython")
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["cython"]) > thres, "csr_bbox/cython")
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["cython"]) > thres, "csr_full/cython")
self.assertTrue(mathutil.rwp(self.results["splitbbox"], self.results["csr_no"]) > thres, "splitbbox/csr_no")
self.assertTrue(mathutil.rwp(self.results["splitpixel"], self.results["csr_no"]) > thres, "splitpixel/csr_no")
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["csr_no"]) > thres, "csr_bbox/csr_no")
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["csr_no"]) > thres, "csr_full/csr_no")
def test_split_bbox(self):
"""
Validate that all bbox splitting algo give all the same result...
"""
thres = 7
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["splitbbox"]) < thres, "csr_bbox/splitbbox")
self.assertTrue(mathutil.rwp(self.results["numpy"], self.results["splitbbox"]) > thres, "numpy/splitbbox")
self.assertTrue(mathutil.rwp(self.results["cython"], self.results["splitbbox"]) > thres, "cython/splitbbox")
self.assertTrue(mathutil.rwp(self.results["splitpixel"], self.results["splitbbox"]) > thres, "splitpixel/splitbbox")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["splitbbox"]) > thres, "csr_no/splitbbox")
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["splitbbox"]) > thres, "csr_full/splitbbox")
self.assertTrue(mathutil.rwp(self.results["numpy"], self.results["csr_bbox"]) > thres, "numpy/csr_bbox")
self.assertTrue(mathutil.rwp(self.results["cython"], self.results["csr_bbox"]) > thres, "cython/csr_bbox")
self.assertTrue(mathutil.rwp(self.results["splitpixel"], self.results["csr_bbox"]) > thres, "splitpixel/csr_bbox")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["csr_bbox"]) > thres, "csr_no/csr_bbox")
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["csr_bbox"]) > thres, "csr_full/csr_bbox")
def test_split_full(self):
"""
Validate that all full splitting algo give all the same result...
"""
thres = 7
self.assertTrue(mathutil.rwp(self.results["csr_full"], self.results["splitpixel"]) < thres, "csr_full/splitpixel")
self.assertTrue(mathutil.rwp(self.results["numpy"], self.results["splitpixel"]) > thres, "numpy/splitpixel")
self.assertTrue(mathutil.rwp(self.results["cython"], self.results["splitpixel"]) > thres, "cython/splitpixel")
self.assertTrue(mathutil.rwp(self.results["splitbbox"], self.results["splitpixel"]) > thres, "splitpixel/splitpixel")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["splitpixel"]) > thres, "csr_no/splitpixel")
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["splitpixel"]) > thres, "csr_full/splitpixel")
self.assertTrue(mathutil.rwp(self.results["numpy"], self.results["csr_full"]) > thres, "numpy/csr_full")
self.assertTrue(mathutil.rwp(self.results["cython"], self.results["csr_full"]) > thres, "cython/csr_full")
self.assertTrue(mathutil.rwp(self.results["splitbbox"], self.results["csr_full"]) > thres, "splitpixel/csr_full")
self.assertTrue(mathutil.rwp(self.results["csr_no"], self.results["csr_full"]) > thres, "csr_no/csr_full")
self.assertTrue(mathutil.rwp(self.results["csr_bbox"], self.results["csr_full"]) > thres, "csr_full/csr_full")
class TestSplitBBoxNg(unittest.TestCase):
"""Test the equivalence of the historical SplitBBox with the one propagating
the variance"""
@classmethod
def setUpClass(cls):
super(TestSplitBBoxNg, cls).setUpClass()
det = Detector.factory("Pilatus 100k")
shape = det.shape
# The randomness of the image is not correlated to bug #1021
cls.maxi = 65000
img = numpy.random.randint(0, cls.maxi, numpy.prod(shape))
if platform.machine() in ("i386", "i686", "x86_64") and (tuple.__itemsize__ == 4):
cls.epsilon = 1e-13
else:
cls.epsilon = numpy.finfo(numpy.float64).eps
ai = AzimuthalIntegrator(1, detector=det)
ai.wavelength = 1e-10
tth = ai.center_array(shape, unit="2th_rad", scale=False).ravel()
dtth = ai.delta_array(shape, unit="2th_rad").ravel()
chi = ai.chiArray(shape).ravel()
dchi = ai.deltaChi(shape).ravel()
pos = ai.corner_array(shape, unit="2th_deg", use_cython=True, scale=False)
cls.results = {}
# Legacy implementation:
cls.results["histoBBox2d_legacy"] = splitBBox.histoBBox2d(img,
tth,
dtth,
chi,
dchi,
empty=-1)
cls.results["histoBBox2d_ng"] = splitBBox.histoBBox2d_ng(img,
tth,
dtth,
chi,
dchi,
variance=img,
empty=-1)
# Legacy implementation:
cls.results["fullSplit2D_legacy"] = splitPixel.fullSplit2D(pos,
img,
bins=(100, 36),
empty=-1)
cls.results["fullSplit2D_ng"] = splitPixel.pseudoSplit2D_ng(pos,
img,
bins=(100, 36),
variance=img,
empty=-1)
cls.img = img
@classmethod
def tearDownClass(cls):
super(TestSplitBBoxNg, cls).tearDownClass()
cls.results = None
cls.img = None
def test_split_bbox_2d(self):
# radial position:
tth_legacy = self.results["histoBBox2d_legacy"][1]
tth_ng = self.results["histoBBox2d_ng"].radial
self.assertEqual(abs(tth_legacy - tth_ng).max(), 0, "radial position is the same")
# azimuthal position:
chi_legacy = self.results["histoBBox2d_legacy"][2]
chi_ng = self.results["histoBBox2d_ng"].azimuthal
self.assertEqual(abs(chi_legacy - chi_ng).max(), 0, "azimuthal position is the same")
# pixel count:
count_legacy = self.results["histoBBox2d_legacy"][4]
count_ng = self.results["histoBBox2d_ng"].count
if abs(count_ng).max() == 0:
print(splitBBox)
print(count_legacy)
print(count_ng)
# print("prop", self.results["histoBBox2d_ng"][4])
# print("pos1", self.results["histoBBox2d_ng"][3])
# print("pos0", self.results["histoBBox2d_ng"][2])
# print("err", self.results["histoBBox2d_ng"][1])
# print("int", self.results["histoBBox2d_ng"][0])
self.assertLess(abs(count_legacy - count_ng).max(), self.epsilon, "count is the same")
# same for normalisation ... in this case
count_ng = self.results["histoBBox2d_ng"].normalization
self.assertLess(abs(count_legacy - count_ng).max(), self.epsilon, "norm is old-count")
# Weighted signal:
weighted_legacy = self.results["histoBBox2d_legacy"][3]
signal = self.results["histoBBox2d_ng"].signal
self.assertLess(abs(signal - weighted_legacy).max(), self.maxi * self.epsilon, "Weighted is the same")
# resulting intensity validation
int_legacy = self.results["histoBBox2d_legacy"][0]
int_ng = self.results["histoBBox2d_ng"].intensity
self.assertLess(abs(int_legacy - int_ng).max(), self.epsilon, "intensity is the same")
def test_split_pixel_2d(self):
# radial position:
tth_legacy = self.results["fullSplit2D_legacy"][1]
tth_ng = self.results["fullSplit2D_ng"].radial
self.assertEqual(abs(tth_legacy - tth_ng).max(), 0, "radial position is the same")
# azimuthal position:
chi_legacy = self.results["fullSplit2D_legacy"][2]
chi_ng = self.results["fullSplit2D_ng"].azimuthal
self.assertEqual(abs(chi_legacy - chi_ng).max(), 0, "azimuthal position is the same")
# pixel count:
count_legacy = self.results["fullSplit2D_legacy"][4]
count_ng = self.results["fullSplit2D_ng"].count
self.assertLess(abs(count_legacy - count_ng).mean(), 1, "count is almost the same")
# same for normalisation ... in this case
count_ng = self.results["fullSplit2D_ng"].normalization
self.assertLess(abs(count_legacy - count_ng).mean(), 1, "norm is almost old-count")
# # Weighted signal:
# weighted_legacy = self.results["fullSplit2D_legacy"][3]
# signal = self.results["fullSplit2D_ng"][4]["signal"]
# print(weighted_legacy)
# print(signal)
# print(abs(signal / weighted_legacy).nanmean())
# self.assertEqual(abs(signal - weighted_legacy).max(), 0, "Weighted is the same")
# resulting intensity validation
# int_legacy = self.results["fullSplit2D_legacy"][0]
# int_ng = self.results["histoBBox2d_ng"][0]
# print(int_legacy)
# print(int_ng)
# self.assertEqual(abs(int_legacy - int_ng).max(), 0, "intensity is the same")
def suite():
loader = unittest.defaultTestLoader.loadTestsFromTestCase
testsuite = unittest.TestSuite()
testsuite.addTest(loader(TestSplitPixel))
testsuite.addTest(loader(TestSplitBBoxNg))
return testsuite
if __name__ == '__main__':
runner = unittest.TextTestRunner()
runner.run(suite())
UtilsTest.clean_up()
|
