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
|
#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Project: Fast Azimuthal Integration
# https://github.com/pyFAI/pyFAI
#
# Copyright (C) European Synchrotron Radiation Facility, Grenoble, France
#
# Principal author: Jérôme Kieffer (Jerome.Kieffer@ESRF.eu)
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
"test suite for histogramming implementations"
__author__ = "Jérôme Kieffer"
__contact__ = "Jerome.Kieffer@ESRF.eu"
__license__ = "GPLv3+"
__copyright__ = "European Synchrotron Radiation Facility, Grenoble, France"
__date__ = "2014-09-26"
import unittest
import time
import os
import numpy
import logging
import sys
from numpy import cos
from utilstest import UtilsTest, Rwp, getLogger
logger = getLogger(__file__)
pyFAI = sys.modules["pyFAI"]
from pyFAI.histogram import histogram, histogram2d
from pyFAI.splitBBoxCSR import HistoBBox1d, HistoBBox2d
if logger.getEffectiveLevel() == logging.DEBUG:
import pylab
EPS32 = (1.0 + numpy.finfo(numpy.float32).eps)
class TestHistogram1d(unittest.TestCase):
"""basic test"""
shape = (512, 512)
npt = 500
size = shape[0] * shape[1]
maxI = 1000
epsilon = 1.0e-4
y, x = numpy.ogrid[:shape[0], :shape[1]]
tth = numpy.sqrt(x * x + y * y).astype("float32")
mod = 0.5 + 0.5 * cos(tth / 12) + 0.25 * cos(tth / 6) + 0.1 * cos(tth / 4)
data = (numpy.random.poisson(maxI, shape) * mod).astype("uint16")
data_sum = data.sum(dtype="float64")
t0 = time.time()
drange = (tth.min(), tth.max() * EPS32)
unweight_numpy, bin_edges = numpy.histogram(tth, npt, range=drange)
t1 = time.time()
weight_numpy, bin_edges = numpy.histogram(tth, npt, weights=data.astype("float64"), range=drange)
t2 = time.time()
logger.info("Timing for Numpy raw histogram: %.3f", t1 - t0)
logger.info("Timing for Numpy weighted histogram: %.3f", t2 - t1)
bins_numpy = 0.5 * (bin_edges[1:] + bin_edges[:-1])
I_numpy = weight_numpy / numpy.maximum(1.0, unweight_numpy)
t3 = time.time()
bins_cython, I_cython, weight_cython, unweight_cython = histogram(tth, data, npt, pixelSize_in_Pos=0)
t4 = time.time()
logger.info("Timing for Cython both histogram: %.3f", t4 - t3)
t3 = time.time()
integrator = HistoBBox1d(tth, delta_pos0=None, pos1=None, delta_pos1=None,
bins=npt, pos0Range=drange, allow_pos0_neg=False, unit="undefined",)
t2 = time.time()
bins_csr, I_csr, weight_csr, unweight_csr = integrator.integrate(data)
t4 = time.time()
logger.info("Timing for CSR init: %.3fs, integrate: %0.3fs, both: %.3f", (t2 - t3), (t4 - t2), (t4 - t3))
def test_count_numpy(self):
"""
Test that the pixel count and the total intensity is conserved
in numpy implementation
"""
sump = self.unweight_numpy.sum(dtype="int64")
intensity_obt = self.weight_numpy.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("Numpy: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("Numpy: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, msg="check all pixels were counted")
summed_weight_hist = self.weight_numpy.sum(dtype="float64")
self.assert_(summed_weight_hist == self.data_sum, msg="check all intensity is counted expected %s got %s" % (self.data_sum, summed_weight_hist))
self.assert_(v < self.epsilon, msg="checks delta is lower than %s, got %s" % (self.epsilon, v))
def test_count_cython(self):
"""
Test that the pixel count and the total intensity is conserved
in cython implementation
"""
sump = int(self.unweight_cython.sum(dtype="float64"))
intensity_obt = self.weight_cython.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("Cython: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("Cython: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, msg="check all pixels were counted expected %s got %s" % (self.size, sump))
summed_weight_hist = self.weight_cython.sum(dtype="float64")
self.assert_(summed_weight_hist == self.data_sum, msg="check all intensity is counted expected %s got %s" % (self.data_sum, summed_weight_hist))
self.assertTrue(v < self.epsilon, msg="checks delta is lower than %s" % self.epsilon)
def test_count_csr(self):
"""
Test that the pixel count and the total intensity is conserved
in cSR sparse matrix multiplacation implementation
"""
sump = int(self.unweight_csr.sum(dtype="float64"))
intensity_obt = self.weight_csr.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("CSR: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("CSR: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, msg="check all pixels were counted expected %s got %s" % (self.size, sump))
summed_weight_hist = self.weight_csr.sum(dtype="float64")
self.assert_(summed_weight_hist == self.data_sum, msg="check all intensity is counted expected %s got %s" % (self.data_sum, summed_weight_hist))
self.assertTrue(v < self.epsilon, msg="checks delta is lower than %s" % self.epsilon)
def test_numpy_vs_cython_vs_csr_1d(self):
"""
Compare numpy histogram with cython simple implementation ans CSR
"""
max_delta = abs(self.bins_numpy - self.bins_cython).max()
logger.info("Bin-center position for cython/numpy, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for cython/numpy, max delta=%s" % max_delta)
max_delta = abs(self.bins_numpy - self.bins_csr).max()
logger.info("Bin-center position for csr/numpy, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for csr/numpy, max delta=%s" % max_delta)
rwp1 = Rwp((self.bins_cython, self.I_cython), (self.bins_numpy, self.I_numpy))
logger.info("Rwp Cython/Numpy = %.3f" % rwp1)
self.assert_(rwp1 < self.epsilon, "Rwp Cython/Numpy = %.3f" % rwp1)
rwp2 = Rwp((self.bins_csr, self.I_csr), (self.bins_numpy, self.I_numpy))
logger.info("Rwp CSR/Numpy = %.3f" % rwp2)
self.assert_(rwp2 < 3, "Rwp Cython/Numpy = %.3f" % rwp2)
if logger.getEffectiveLevel() == logging.DEBUG:
logger.info("Plotting results")
fig = pylab.figure()
fig.suptitle('Numpy /Cython R=%.3f, Numpy/CSR R=%.3f' % (rwp1, rwp2))
sp = fig.add_subplot(111)
sp.plot(self.bins_numpy, self.I_numpy, "-b", label='numpy')
sp.plot(self.bins_cython, self.I_cython, "-r", label="cython")
sp.plot(self.bins_csr, self.I_csr, "-g", label="CSR")
handles, labels = sp.get_legend_handles_labels()
fig.legend(handles, labels)
fig.show()
raw_input("Press enter to quit")
delta_max = abs(self.unweight_numpy - self.unweight_cython).max()
logger.info("pixel count difference numpy/cython : max delta=%s", delta_max)
self.assert_(delta_max < 1, "numpy_vs_cython_1d max delta unweight = %s" % delta_max)
delta_max = abs(self.I_cython - self.I_numpy).max()
logger.info("Intensity count difference numpy/cython : max delta=%s", delta_max)
self.assert_(delta_max < self.epsilon, "Intensity count difference numpy/cython : max delta=%s" % delta_max)
# TODO: fix this !!!
delta_max = abs(self.unweight_numpy - self.unweight_csr).max()
if delta_max > 0:
logger.warning("pixel count difference numpy/csr : max delta=%s", delta_max)
self.assert_(delta_max < 10, "numpy_vs_csr_1d max delta unweight = %s" % delta_max)
delta_max = abs(self.I_csr - self.I_numpy).max()
if delta_max > self.epsilon:
logger.warning("Intensity count difference numpy/csr : max delta=%s", delta_max)
self.assert_(delta_max < 0.55, "Intensity count difference numpy/csr : max delta=%s" % delta_max)
class TestHistogram2d(unittest.TestCase):
"""basic test for 2D histogram"""
shape = (512, 512)
size = shape[0] * shape[1]
maxI = 1000
epsilon = 1.0e-4
y, x = numpy.ogrid[:shape[0], :shape[1]]
tth = numpy.sqrt(x * x + y * y).astype("float32")
mod = 0.5 + 0.5 * cos(tth / 12) + 0.25 * cos(tth / 6) + 0.1 * cos(tth / 4)
data = (numpy.random.poisson(maxI, shape) * mod).astype("uint16")
data_sum = data.sum(dtype="float64")
npt = (400, 360)
# epsilon = 3.0e-4
chi = numpy.arctan2(y, x).astype("float32")
drange = [[tth.min(), tth.max() * EPS32], [chi.min(), chi.max() * EPS32]]
t0 = time.time()
unweight_numpy, tth_edges, chi_edges = numpy.histogram2d(tth.flatten(), chi.flatten(), npt, range=drange)
t1 = time.time()
weight_numpy, tth_edges, chi_edges = numpy.histogram2d(tth.flatten(), chi.flatten(), npt, weights=data.astype("float64").flatten(), range=drange)
t2 = time.time()
logger.info("Timing for Numpy raw histogram2d: %.3f", t1 - t0)
logger.info("Timing for Numpy weighted histogram2d: %.3f", t2 - t1)
tth_numpy = 0.5 * (tth_edges[1:] + tth_edges[:-1])
chi_numpy = 0.5 * (chi_edges[1:] + chi_edges[:-1])
I_numpy = weight_numpy / numpy.maximum(1.0, unweight_numpy)
t3 = time.time()
I_cython, tth_cython, chi_cython, weight_cython, unweight_cython = histogram2d(tth.flatten(), chi.flatten(), npt, data.flatten(), split=0)
t4 = time.time()
logger.info("Timing for Cython both histogram2d: %.3f", t4 - t3)
t3 = time.time()
integrator = HistoBBox2d(tth, None, chi, delta_pos1=None,
bins=npt, allow_pos0_neg=False, unit="undefined")
t2 = time.time()
I_csr, tth_csr, chi_csr, weight_csr, unweight_csr = integrator.integrate(data)
t4 = time.time()
logger.info("Timing for CSR init: %.3fs, integrate: %0.3fs, both: %.3f", (t2 - t3), (t4 - t2), (t4 - t3))
def test_count_numpy(self):
"""
Test that the pixel count and the total intensity is conserved
in numpy implementation
"""
sump = self.unweight_numpy.sum(dtype="int64")
intensity_obt = self.weight_numpy.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("Numpy: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("Numpy: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, "Numpy: Total number of points: %s (%s expected), delta = %s" % (sump, self.size, delta))
self.assert_(v < self.epsilon, "Numpy: Total Intensity: %s (%s expected), variation = %s" % (intensity_obt, self.data_sum, v))
def test_count_cython(self):
"""
Test that the pixel count and the total intensity is conserved
in cython implementation
"""
sump = int(self.unweight_cython.sum(dtype="int64"))
intensity_obt = self.weight_cython.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("Cython: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("Cython: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, msg="check all pixels were counted")
self.assert_(v < self.epsilon, msg="checks delta is lower than %s" % self.epsilon)
def test_count_csr(self):
"""
Test that the pixel count and the total intensity is conserved
in csr implementation
"""
sump = int(self.unweight_csr.sum(dtype="int64"))
intensity_obt = self.weight_csr.sum(dtype="float64")
delta = abs(sump - self.size)
logger.info("CSR: Total number of points: %s (%s expected), delta = %s", sump, self.size, delta)
v = abs(intensity_obt - self.data_sum) / self.data_sum
logger.info("CSR: Total Intensity: %s (%s expected), variation = %s", intensity_obt, self.data_sum, v)
self.assert_(delta == 0, msg="check all pixels were counted")
self.assert_(v < self.epsilon, msg="checks delta is lower than %s" % self.epsilon)
def test_numpy_vs_cython_vs_csr_2d(self):
"""
Compare numpy histogram with cython simple implementation
"""
max_delta = abs(self.tth_numpy - self.tth_cython).max()
logger.info("Bin-center position for cython/numpy tth, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for cython/numpy tth, max delta=%s" % max_delta)
max_delta = abs(self.chi_numpy - self.chi_cython).max()
logger.info("Bin-center position for cython/numpy chi, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for cython/numpy chi, max delta=%s" % max_delta)
delta_max = abs(self.unweight_numpy - self.unweight_cython).max()
logger.info("pixel count difference numpy/cython : max delta=%s", delta_max)
self.assert_(delta_max == 0, "pixel count difference numpy/cython : max delta=%s" % delta_max)
delta_max = abs(self.I_cython - self.I_numpy).max()
logger.info("Intensity count difference numpy/cython : max delta=%s", delta_max)
self.assert_(delta_max < self.epsilon * self.maxI, "Intensity count difference numpy/cython : max delta=%s" % delta_max)
max_delta = abs(self.tth_numpy - self.tth_csr).max()
logger.info("Bin-center position for csr/numpy tth, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for csr/numpy tth, max delta=%s" % max_delta)
max_delta = abs(self.chi_numpy - self.chi_csr).max()
logger.info("Bin-center position for csr/numpy chi, max delta=%s", max_delta)
self.assert_(max_delta < self.epsilon, "Bin-center position for csr/numpy chi, max delta=%s" % max_delta)
delta_max = abs(self.unweight_numpy - self.unweight_csr.T).max()
if delta_max > 0:
logger.warning("pixel count difference numpy/csr : max delta=%s", delta_max)
self.assert_(delta_max < 2, "pixel count difference numpy/csr : max delta=%s" % delta_max)
delta_max = abs(self.I_csr.T - self.I_numpy).max()
if delta_max > self.epsilon:
logger.warning("Intensity count difference numpy/csr : max delta=%s", delta_max)
self.assert_(delta_max < 28, "Intensity count difference numpy/csr : max delta=%s" % delta_max)
def test_suite_all_Histogram():
testSuite = unittest.TestSuite()
testSuite.addTest(TestHistogram1d("test_count_numpy"))
testSuite.addTest(TestHistogram1d("test_count_cython"))
testSuite.addTest(TestHistogram1d("test_count_csr"))
testSuite.addTest(TestHistogram1d("test_numpy_vs_cython_vs_csr_1d"))
testSuite.addTest(TestHistogram2d("test_count_numpy"))
testSuite.addTest(TestHistogram2d("test_count_cython"))
testSuite.addTest(TestHistogram2d("test_count_csr"))
testSuite.addTest(TestHistogram2d("test_numpy_vs_cython_vs_csr_2d"))
return testSuite
if __name__ == '__main__':
mysuite = test_suite_all_Histogram()
runner = unittest.TextTestRunner()
runner.run(mysuite)
|
