#/*##########################################################################
#
# The PyMca X-Ray Fluorescence Toolkit
#
# Copyright (c) 2004-2014 European Synchrotron Radiation Facility
#
# This file is part of the PyMca X-ray Fluorescence Toolkit developed at
# the ESRF by the Software group.
#
# 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.
#
#############################################################################*/
__author__ = "V. Armando Sole - ESRF Data Analysis"
__contact__ = "sole@esrf.fr"
__license__ = "MIT"
__copyright__ = "European Synchrotron Radiation Facility, Grenoble, France"
import unittest
import sys
import os
import gc
import tempfile
[docs]class testSpecfilewrapper(unittest.TestCase):
[docs] def setUp(self):
"""
import the module
"""
try:
from PyMca5.PyMcaIO import specfilewrapper as specfile
self.specfileClass = specfile
except:
self.specfileClass = None
if self.specfileClass is not None:
text = "1.3 1 1\n"
text += "2.5 4 8\n"
text += "3.7 9 27\n"
text += "\n"
tmpFile = tempfile.mkstemp(text=False)
if sys.version < '3.0':
os.write(tmpFile[0], text)
else:
os.write(tmpFile[0], bytes(text, 'utf-8'))
os.close(tmpFile[0])
self.fname = tmpFile[1]
[docs] def tearDown(self):
"""clean up any possible files"""
# make sure the file handle is free
self._sf = None
self._scan = None
# this should free the handle
gc.collect()
if self.specfileClass is not None:
if os.path.exists(self.fname):
os.remove(self.fname)
[docs] def testSpecfilewrapperImport(self):
#"""Test successful import"""
self.assertTrue(self.specfileClass is not None,
'Unsuccessful PyMca5.PyMcaIO.specfilewrapper import')
[docs] def testSpecfilewrapperReading(self):
#"""Test specfile readout"""
self.testSpecfilewrapperImport()
self._sf = self.specfileClass.Specfile(self.fname)
# test the number of found scans
self.assertEqual(len(self._sf), 2,
'Expected to read 2 scans, read %s' %\
len(self._sf))
self.assertEqual(self._sf.scanno(), 2,
'Expected to read 2 scans, got %s' %\
self._sf.scanno())
# test scan iteration selection method
self._scan = self._sf[0]
labels = self._scan.alllabels()
expectedLabels = ['Point', 'Column 0', 'Column 1', 'Column 2']
self.assertEqual(len(labels), 4,
'Expected to read 4 labels, got %s' % len(labels))
for i in range(3):
self.assertEqual(labels[i], expectedLabels[i],
'Read "%s" instead of "%s"' %\
(labels[i], expectedLabels[i]))
# test scan number selection method
self._scan = self._sf.select('1.1')
labels = self._scan.alllabels()
expectedLabels = ['Point', 'Column 0', 'Column 1', 'Column 2']
self.assertEqual(len(labels), 4,
'Expected to read 4 labels, got %s' % len(labels))
for i in range(3):
self.assertEqual(labels[i], expectedLabels[i],
'Read "%s" instead of "%s"' %\
(labels[i], expectedLabels[i]))
# test scan number of mca
self._scan = self._sf[0]
nbmca = self._scan.nbmca()
self.assertEqual(nbmca, 0,
'Expected to read 0 mca, got %s' % nbmca)
self._scan = self._sf[1]
nbmca = self._scan.nbmca()
self.assertEqual(nbmca, 3,
'Expected to read 3 mca, got %s' % nbmca)
[docs] def testSpecfilewrapperReadingCompatibleWithUserLocale(self):
#"""Test specfile compatible with C locale"""
self.testSpecfilewrapperImport()
self._sf = self.specfileClass.Specfile(self.fname)
self._scan = self._sf[0]
datacol = self._scan.datacol(2)
data = self._scan.data()
self._sf = None
self.assertEqual(datacol[0], 1.3,
'Read %f instead of %f' %\
(datacol[0], 1.3))
self.assertEqual(datacol[1], 2.5,
'Read %f instead of %f' %\
(datacol[1], 2.5))
self.assertEqual(datacol[2], 3.7,
'Read %f instead of %f' %\
(datacol[2], 3.7))
self.assertEqual(datacol[1], data[1][1],
'Read %f instead of %f' %\
(datacol[1], data[1][1]))
gc.collect()
[docs] def testTrainingSpectrumReading(self):
from PyMca5 import PyMcaDataDir
import numpy
fname = os.path.join(PyMcaDataDir.PYMCA_DATA_DIR,
'XRFSpectrum.mca')
self._sf = self.specfileClass.Specfile(fname)
self._scan = self._sf[0]
# I find awful that starts counting at 1
# 1 is the point number
# 2 is the actual spectal data
datacol = self._scan.datacol(2)
self._scan = self._sf[1]
# The "second" scan is the readout as mca
mca = self._scan.mca(1)
self.assertTrue(numpy.alltrue(datacol == mca))
[docs]def getSuite(auto=True):
testSuite = unittest.TestSuite()
if auto:
testSuite.addTest(\
unittest.TestLoader().loadTestsFromTestCase(testSpecfilewrapper))
else:
# use a predefined order
testSuite.addTest(testSpecfilewrapper("testSpecfilewrapperImport"))
testSuite.addTest(testSpecfilewrapper("testSpecfilewrapperReading"))
testSuite.addTest(\
testSpecfilewrapper(\
"testSpecfilewrapperReadingCompatibleWithUserLocale"))
testSuite.addTest(testSpecfilewrapper("testTrainingSpectrumReading"))
return testSuite
[docs]def test(auto=False):
unittest.TextTestRunner(verbosity=2).run(getSuite(auto=auto))
if __name__ == '__main__':
test()