def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ['PLP0000708.nx.hdf', 'PLP0000708.nx.hdf']
pths = [os.path.join(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
with warnings.catch_warnings():
warnings.simplefilter('ignore', RuntimeWarning)
# it should be processable
fadd = PlatypusNexus(
os.path.join(os.getcwd(), 'ADD_PLP0000708.nx.hdf'))
fadd.process()
# it should also be reduceable
reducer = PlatypusReduce(
os.path.join(self.pth, 'PLP0000711.nx.hdf'))
datasets, reduced = reducer.reduce(
os.path.join(os.getcwd(), 'ADD_PLP0000708.nx.hdf'))
assert_('y' in reduced)
# the error bars should be smaller
datasets2, reduced2 = reducer.reduce(
os.path.join(self.pth, 'PLP0000708.nx.hdf'))
assert_(np.all(reduced['y_err'] < reduced2['y_err']))
def __init__(
self,
model,
angle,
L12=2859,
footprint=60,
L2S=120,
dtheta=3.3, # angular resolution
lo_wavelength=2.8,
hi_wavelength=18,
dlambda=3.3,
rebin=2):
self.model = model
# turn off resolution smearing
self.model.dq = 0
self.bkg = model.bkg.value
self.angle = angle
# the fractional width of a square wavelength resolution
self.dlambda = dlambda / 100.
self.rebin = rebin / 100.
self.wavelength_bins = calculate_wavelength_bins(
lo_wavelength, hi_wavelength, rebin)
# nominal Q values
bin_centre = 0.5 * (self.wavelength_bins[1:] +
self.wavelength_bins[:-1])
self.q = general.q(angle, bin_centre)
# keep a tally of the direct and reflected beam
self.direct_beam = np.zeros((self.wavelength_bins.size - 1))
self.reflected_beam = np.zeros((self.wavelength_bins.size - 1))
# wavelength generator
a = PN('PLP0000711.nx.hdf')
q, i, di = a.process(normalise=False,
normalise_bins=False,
rebin_percent=0,
lo_wavelength=lo_wavelength,
hi_wavelength=hi_wavelength)
q = q.squeeze()
i = i.squeeze()
self.spectrum_dist = SpectrumDist(q, i)
# angular resolution generator, based on a trapezoidal distribution
# The slit settings are the optimised set typically used in an
# experiment
self.dtheta = dtheta / 100.
self.footprint = footprint
s1, s2 = general.slit_optimiser(footprint,
self.dtheta,
angle=angle,
L2S=L2S,
L12=L12,
verbose=False)
div, alpha, beta = general.div(s1, s2, L12=L12)
self.angular_dist = trapz(c=(alpha - beta) / 2. / alpha,
d=(alpha + beta) / 2. / alpha,
loc=-alpha,
scale=2 * alpha)
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ["PLP0000708.nx.hdf", "PLP0000708.nx.hdf"]
pths = [os.path.join(self.path, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
# it should be processable
fadd = PlatypusNexus(os.path.join(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
fadd.process()
# it should also be reduceable
reducer = ReducePlatypus(os.path.join(self.path, "PLP0000711.nx.hdf"))
reduced = reducer.reduce(os.path.join(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
assert_("ydata" in reduced)
# the error bars should be smaller
reduced2 = reducer.reduce(os.path.join(self.path, "PLP0000708.nx.hdf"))
assert_(np.all(reduced["ydata_sd"] < reduced2["ydata_sd"]))
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ["PLP0000708.nx.hdf", "PLP0000708.nx.hdf"]
pths = [pjoin(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
# it should be processable
fadd = PlatypusNexus(pjoin(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
fadd.process()
# it should also be reduceable
reducer = PlatypusReduce(pjoin(self.pth, "PLP0000711.nx.hdf"))
datasets, reduced = reducer.reduce(
pjoin(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
assert_("y" in reduced)
# the error bars should be smaller
datasets2, reduced2 = reducer.reduce(
pjoin(self.pth, "PLP0000708.nx.hdf"))
assert_(np.all(reduced["y_err"] < reduced2["y_err"]))
class TestPlatypusNexus(unittest.TestCase):
def setUp(self):
path = os.path.dirname(__file__)
self.path = path
self.f113 = PlatypusNexus(os.path.join(self.path,
'PLP0011613.nx.hdf'))
self.f641 = PlatypusNexus(os.path.join(self.path,
'PLP0011641.nx.hdf'))
return 0
def test_chod(self):
flight_length = self.f113.chod()
assert_almost_equal(flight_length[0], 7141.413818359375)
assert_almost_equal(flight_length[1], 808)
flight_length = self.f641.chod(omega=1.8, twotheta=3.6)
assert_almost_equal(flight_length[0], 7146.3567785516016)
assert_almost_equal(flight_length[1], 808)
def test_phase_angle(self):
# TODO. Add more tests where the phase_angle isn't zero.
phase_angle, master_opening = self.f641.phase_angle()
assert_almost_equal(phase_angle, 0)
assert_almost_equal(master_opening, 1.04719755)
def test_background_subtract_line(self):
# checked each step of the background subtraction with IGOR
# so this test background correction should be correct.
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
mask = np.zeros(201, np.bool)
mask[30:70] = True
mask[130:160] = True
profile, profile_sd = plp.background_subtract_line(yvals,
yvals_sd,
mask)
verified_data = np.load(os.path.join(self.path,
'background_subtract.npy'))
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
def test_find_specular_ridge(self):
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 1000, 0, 1))
detector = np.repeat(yvals[:, np.newaxis], 1000, axis=1).T
detector_sd = np.sqrt(detector)
output = plp.find_specular_ridge(detector[np.newaxis, :], detector_sd[np.newaxis, :])
assert_(len(output) == 2)
assert_almost_equal(output[0][0], 100)
def test_background_subtract(self):
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
# now make an (N, T, Y) detector image
n_tbins = 10
detector = np.repeat(yvals, n_tbins).reshape(xvals.size, n_tbins).T
detector_sd = np.repeat(yvals_sd, n_tbins).reshape(xvals.size, n_tbins).T
detector = detector.reshape(1, n_tbins, xvals.size)
detector_sd = detector_sd.reshape(1, n_tbins, xvals.size)
mask = np.zeros((1, n_tbins, 201), np.bool)
mask[:, :, 30:70] = True
mask[:, :, 130:160] = True
det_bkg, detSD_bkg = plp.background_subtract(detector,
detector_sd,
mask)
# each of the (N, T) entries should have the same background subtracted
# entries
verified_data = np.load(os.path.join(self.path,
'background_subtract.npy'))
it = np.nditer(detector, flags=['multi_index'])
it.remove_axis(2)
while not it.finished:
profile = det_bkg[it.multi_index]
profile_sd = detSD_bkg[it.multi_index]
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
it.iternext()
def test_calculate_bins(self):
bins = plp.calculate_wavelength_bins(2., 18, 2.)
assert_almost_equal(bins[0], 1.98)
assert_almost_equal(bins[-1], 18.18)
def test_event(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
out = self.f113.process(eventmode=[0, 900, 1800], integrate=0)
assert_(np.size(out[1], axis=0) == 2)
def test_event_folder(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
out = self.f113.process(eventmode=[0, 900, 1800], integrate=0,
event_folder=self.path)
def test_multiple_acquisitions(self):
"""
TODO: add a dataset which has multiple spectra in it, and make sure it
processes.
"""
pass
def test_reduction_runs(self):
# just check it runs
self.f113.process()
# check that event mode reduction gives the same output as non-event
# mode reduction.
spectrum0 = self.f113.process(direct=True)
spectrum1 = self.f113.process(direct=True, eventmode=[], integrate=0)
assert_allclose(spectrum0[1][0], spectrum1[1][0], rtol=0.001)
# check that the wavelength resolution is roughly right, between 7 and
# 8%.
res = (self.f113.processed_spectrum['m_lambda_fwhm'][0] /
self.f113.processed_spectrum['m_lambda'][0])
assert_array_less(res, np.ones_like(res) * 0.08)
assert_array_less(np.ones_like(res) * 0.07, res)
def test_save_spectrum(self):
# test saving spectrum
self.f113.process()
# can save the spectra by supplying a filename
self.f113.write_spectrum_xml('test.xml')
self.f113.write_spectrum_dat('test.dat')
# can save by supplying file handle:
with open('test.xml', 'wb') as f:
self.f113.write_spectrum_xml(f)
# can save by supplying file handle:
with open('test.dat', 'wb') as f:
self.f113.write_spectrum_xml(f)
def test_accumulate_files(self):
fnames = ['PLP0000708.nx.hdf', 'PLP0000709.nx.hdf']
pths = [os.path.join(self.path, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
f8, f9, fadd = None, None, None
try:
f8 = h5py.File(os.path.join(self.path,
'PLP0000708.nx.hdf'), 'r')
f9 = h5py.File(os.path.join(self.path,
'PLP0000709.nx.hdf'), 'r')
fadd = h5py.File(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'), 'r')
f8d = f8['entry1/data/hmm'][0]
f9d = f9['entry1/data/hmm'][0]
faddd = fadd['entry1/data/hmm'][0]
assert_equal(faddd, f8d + f9d)
finally:
if f8 is not None:
f8.close()
if f9 is not None:
f9.close()
if fadd is not None:
fadd.close()
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ['PLP0000708.nx.hdf', 'PLP0000708.nx.hdf']
pths = [os.path.join(self.path, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
# it should be processable
fadd = PlatypusNexus(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'))
fadd.process()
# it should also be reduceable
reducer = ReducePlatypus(os.path.join(self.path,
'PLP0000711.nx.hdf'))
reduced = reducer.reduce(os.path.join(os.getcwd(), 'ADD_PLP0000708.nx.hdf'))
assert_('ydata' in reduced)
# the error bars should be smaller
reduced2 = reducer.reduce(os.path.join(self.path, 'PLP0000708.nx.hdf'))
assert_(np.all(reduced['ydata_sd'] < reduced2['ydata_sd']))
def __init__(
self,
model,
angle,
L12=2859,
footprint=60,
L2S=120,
dtheta=3.3,
lo_wavelength=2.8,
hi_wavelength=18,
dlambda=3.3,
rebin=2,
gravity=False,
force_gaussian=False,
force_uniform_wavelength=False,
):
self.model = model
self.bkg = model.bkg.value
self.angle = angle
# dlambda refers to the FWHM of the gaussian approximation to a uniform
# distribution. The full width of the uniform distribution is
# dlambda/0.68.
self.dlambda = dlambda / 100.0
# the rebin percentage refers to the full width of the bins. You have to
# multiply this value by 0.68 to get the equivalent contribution to the
# resolution function.
self.rebin = rebin / 100.0
self.wavelength_bins = calculate_wavelength_bins(
lo_wavelength, hi_wavelength, rebin
)
bin_centre = 0.5 * (
self.wavelength_bins[1:] + self.wavelength_bins[:-1]
)
# angular deviation due to gravity
# --> no correction for gravity affecting width of angular resolution
elevations = 0
if gravity:
speeds = general.wavelength_velocity(bin_centre)
# trajectories through slits for different wavelengths
trajectories = pm.find_trajectory(L12 / 1000.0, 0, speeds)
# elevation at sample
elevations = pm.elevation(
trajectories, speeds, (L12 + L2S) / 1000.0
)
# nominal Q values
self.q = general.q(angle - elevations, bin_centre)
# keep a tally of the direct and reflected beam
self.direct_beam = np.zeros((self.wavelength_bins.size - 1))
self.reflected_beam = np.zeros((self.wavelength_bins.size - 1))
# beam monitor counts for normalisation
self.bmon_direct = 0
self.bmon_reflect = 0
self.gravity = gravity
# wavelength generator
self.force_uniform_wavelength = force_uniform_wavelength
if force_uniform_wavelength:
self.spectrum_dist = uniform(
loc=lo_wavelength - 1, scale=hi_wavelength - lo_wavelength + 1
)
else:
a = PN("PLP0000711.nx.hdf")
q, i, di = a.process(
normalise=False,
normalise_bins=False,
rebin_percent=0.5,
lo_wavelength=max(0, lo_wavelength - 1),
hi_wavelength=hi_wavelength + 1,
)
q = q.squeeze()
i = i.squeeze()
self.spectrum_dist = SpectrumDist(q, i)
self.force_gaussian = force_gaussian
# angular resolution generator, based on a trapezoidal distribution
# The slit settings are the optimised set typically used in an
# experiment. dtheta/theta refers to the FWHM of a Gaussian
# approximation to a trapezoid.
# stores the q vectors contributing towards each datapoint
self._res_kernel = {}
self._min_samples = 0
self.dtheta = dtheta / 100.0
self.footprint = footprint
self.angle = angle
self.L2S = L2S
self.L12 = L12
s1, s2 = general.slit_optimiser(
footprint,
self.dtheta,
angle=angle,
L2S=L2S,
L12=L12,
verbose=False,
)
div, alpha, beta = general.div(s1, s2, L12=L12)
self.div, self.s1, self.s2 = s1, s2, div
if force_gaussian:
self.angular_dist = norm(scale=div / 2.3548)
else:
self.angular_dist = trapz(
c=(alpha - beta) / 2.0 / alpha,
d=(alpha + beta) / 2.0 / alpha,
loc=-alpha,
scale=2 * alpha,
)
class TestPlatypusNexus(object):
@pytest.fixture(autouse=True)
def setup_method(self, tmpdir):
self.pth = os.path.dirname(os.path.abspath(__file__))
self.f113 = PlatypusNexus(os.path.join(self.pth, 'PLP0011613.nx.hdf'))
self.f641 = PlatypusNexus(os.path.join(self.pth, 'PLP0011641.nx.hdf'))
self.cwd = os.getcwd()
self.tmpdir = tmpdir.strpath
os.chdir(self.tmpdir)
return 0
def teardown_method(self):
os.chdir(self.cwd)
def test_chod(self):
flight_length = self.f113.chod()
assert_almost_equal(flight_length[0], 7141.413818359375)
assert_almost_equal(flight_length[1], 808)
flight_length = self.f641.chod(omega=1.8, twotheta=3.6)
assert_almost_equal(flight_length[0], 7146.3567785516016)
assert_almost_equal(flight_length[1], 808)
def test_phase_angle(self):
# TODO. Add more tests where the phase_angle isn't zero.
phase_angle, master_opening = self.f641.phase_angle()
assert_almost_equal(phase_angle, 0)
assert_almost_equal(master_opening, 1.04719755)
def test_background_subtract_line(self):
# checked each step of the background subtraction with IGOR
# so this test background correction should be correct.
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
mask = np.zeros(201, np.bool)
mask[30:70] = True
mask[130:160] = True
profile, profile_sd = plp.background_subtract_line(
yvals, yvals_sd, mask)
verified_data = np.load(
os.path.join(self.pth, 'background_subtract.npy'))
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
def test_find_specular_ridge(self):
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 1000, 0, 1))
detector = np.repeat(yvals[:, np.newaxis], 1000, axis=1).T
detector_sd = np.sqrt(detector)
output = plp.find_specular_ridge(detector[np.newaxis, :],
detector_sd[np.newaxis, :])
assert_(len(output) == 5)
assert_almost_equal(output[0][0], 100)
def test_background_subtract(self):
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
# now make an (N, T, Y) detector image
n_tbins = 10
detector = np.repeat(yvals, n_tbins).reshape(xvals.size, n_tbins).T
detector_sd = np.repeat(yvals_sd,
n_tbins).reshape(xvals.size, n_tbins).T
detector = detector.reshape(1, n_tbins, xvals.size)
detector_sd = detector_sd.reshape(1, n_tbins, xvals.size)
mask = np.zeros((1, n_tbins, 201), np.bool)
mask[:, :, 30:70] = True
mask[:, :, 130:160] = True
det_bkg, detSD_bkg = plp.background_subtract(detector, detector_sd,
mask)
# each of the (N, T) entries should have the same background subtracted
# entries
verified_data = np.load(
os.path.join(self.pth, 'background_subtract.npy'))
it = np.nditer(detector, flags=['multi_index'])
it.remove_axis(2)
while not it.finished:
profile = det_bkg[it.multi_index]
profile_sd = detSD_bkg[it.multi_index]
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
it.iternext()
def test_calculate_bins(self):
bins = plp.calculate_wavelength_bins(2., 18, 2.)
assert_almost_equal(bins[0], 1.98)
assert_almost_equal(bins[-1], 18.18)
def test_event(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
out = self.f113.process(eventmode=[0, 900, 1800], integrate=0)
assert_(np.size(out[1], axis=0) == 2)
def test_event_folder(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
self.f113.process(eventmode=[0, 900, 1800],
integrate=0,
event_folder=self.pth)
def test_multiple_acquisitions(self):
"""
TODO: add a dataset which has multiple spectra in it, and make sure it
processes.
"""
pass
def test_reduction_runs(self):
# just check it runs
self.f113.process()
# check that event mode reduction gives the same output as non-event
# mode reduction.
spectrum0 = self.f113.process(direct=True)
spectrum1 = self.f113.process(direct=True, eventmode=[], integrate=0)
assert_allclose(spectrum0[1][0], spectrum1[1][0], rtol=0.001)
# check that the wavelength resolution is roughly right, between 7 and
# 8%.
res = (self.f113.processed_spectrum['m_lambda_fwhm'][0] /
self.f113.processed_spectrum['m_lambda'][0])
assert_array_less(res, np.ones_like(res) * 0.08)
assert_array_less(np.ones_like(res) * 0.07, res)
def test_save_spectrum(self):
# test saving spectrum
self.f113.process()
# can save the spectra by supplying a filename
self.f113.write_spectrum_xml(os.path.join(self.tmpdir, 'test.xml'))
self.f113.write_spectrum_dat(os.path.join(self.tmpdir, 'test.dat'))
# can save by supplying file handle:
with open(os.path.join(self.tmpdir, 'test.xml'), 'wb') as f:
self.f113.write_spectrum_xml(f)
def test_accumulate_files(self):
fnames = ['PLP0000708.nx.hdf', 'PLP0000709.nx.hdf']
pths = [os.path.join(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
f8, f9, fadd = None, None, None
try:
f8 = h5py.File(os.path.join(self.pth, 'PLP0000708.nx.hdf'), 'r')
f9 = h5py.File(os.path.join(self.pth, 'PLP0000709.nx.hdf'), 'r')
fadd = h5py.File(
os.path.join(self.tmpdir, 'ADD_PLP0000708.nx.hdf'), 'r')
f8d = f8['entry1/data/hmm'][0]
f9d = f9['entry1/data/hmm'][0]
faddd = fadd['entry1/data/hmm'][0]
assert_equal(faddd, f8d + f9d)
finally:
if f8 is not None:
f8.close()
if f9 is not None:
f9.close()
if fadd is not None:
fadd.close()
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ['PLP0000708.nx.hdf', 'PLP0000708.nx.hdf']
pths = [os.path.join(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
# it should be processable
fadd = PlatypusNexus(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'))
fadd.process()
# it should also be reduceable
reducer = PlatypusReduce(os.path.join(self.pth, 'PLP0000711.nx.hdf'))
datasets, reduced = reducer.reduce(
os.path.join(os.getcwd(), 'ADD_PLP0000708.nx.hdf'))
assert_('y' in reduced)
# the error bars should be smaller
datasets2, reduced2 = reducer.reduce(
os.path.join(self.pth, 'PLP0000708.nx.hdf'))
assert_(np.all(reduced['y_err'] < reduced2['y_err']))
def test_manual_beam_find(self):
# you can specify a function that finds where the specular ridge is.
def manual_beam_find(detector, detector_sd):
beam_centre = np.zeros(len(detector))
beam_sd = np.zeros(len(detector))
beam_centre += 50
beam_sd += 5
return beam_centre, beam_sd, np.array([40]), np.array([60]), [[]]
# the manual beam find is only mandatory when peak_pos == -1.
# the beam_sd is much larger than the beam divergence, so a warning
# should be raised.
with pytest.warns(UserWarning):
self.f113.process(manual_beam_find=manual_beam_find, peak_pos=-1)
assert_equal(self.f113.processed_spectrum['m_beampos'][0], 50)
# manual beam finding also specifies the lower and upper pixel of the
# foreground
assert_equal(self.f113.processed_spectrum['lopx'][0], 40)
assert_equal(self.f113.processed_spectrum['hipx'][0], 60)
def test_fore_back_region(self):
# calculation of foreground and background regions is done correctly
centres = np.array([100., 90.])
sd = np.array([5., 11.5])
lopx, hipx, background_pixels = fore_back_region(centres, sd)
assert_(len(lopx) == 2)
assert_(len(hipx) == 2)
assert_(len(background_pixels) == 2)
assert_(isinstance(lopx[0], numbers.Integral))
assert_(isinstance(hipx[0], numbers.Integral))
calc_lower = np.floor(centres - sd * EXTENT_MULT)
assert_equal(lopx, calc_lower)
calc_higher = np.ceil(centres + sd * EXTENT_MULT)
assert_equal(hipx, calc_higher)
y1 = np.atleast_1d(np.round(lopx - PIXEL_OFFSET).astype('int'))
y0 = np.atleast_1d(
np.round(lopx - PIXEL_OFFSET - EXTENT_MULT * sd).astype('int'))
y2 = np.atleast_1d(np.round(hipx + PIXEL_OFFSET).astype('int'))
y3 = np.atleast_1d(
np.round(hipx + PIXEL_OFFSET + EXTENT_MULT * sd).astype('int'))
bp = np.r_[np.arange(y0[0], y1[0] + 1), np.arange(y2[0], y3[0] + 1)]
assert_equal(bp, background_pixels[0])
def test_basename_datafile(self):
# check that the right basename is returned
pth = 'a/b/c.nx.hdf'
assert_(basename_datafile(pth) == 'c')
pth = 'c.nx.hdf'
assert_(basename_datafile(pth) == 'c')
def test_floodfield_correction(self):
# check that flood field calculation works
# the values were worked out by hand on a randomly
# generated array
test_norm = np.array(
[1.12290503, 1.23743017, 0.8603352, 0.70111732, 1.07821229])
test_norm_sd = np.array(
[0.05600541, 0.05879208, 0.04902215, 0.04425413, 0.05487956])
fname = os.path.join(self.pth, 'flood.h5')
with h5py.File(fname, 'r') as f:
norm, norm_sd = create_detector_norm(f, 3.5, -3.5)
assert_almost_equal(norm, test_norm, 6)
assert_almost_equal(norm_sd, test_norm_sd, 6)
norm, norm_sd = create_detector_norm(f, -3.5, 3.5)
assert_almost_equal(norm, test_norm, 6)
assert_almost_equal(norm_sd, test_norm_sd, 6)
class TestPlatypusNexus(unittest.TestCase):
def setUp(self):
path = os.path.dirname(__file__)
self.path = path
self.f113 = PlatypusNexus(os.path.join(self.path,
'PLP0011613.nx.hdf'))
self.f641 = PlatypusNexus(os.path.join(self.path,
'PLP0011641.nx.hdf'))
self.cwd = os.getcwd()
self.tmpdir = TemporaryDirectory()
os.chdir(self.tmpdir.name)
return 0
def tearDown(self):
os.chdir(self.cwd)
self.tmpdir.cleanup()
def test_chod(self):
flight_length = self.f113.chod()
assert_almost_equal(flight_length[0], 7141.413818359375)
assert_almost_equal(flight_length[1], 808)
flight_length = self.f641.chod(omega=1.8, twotheta=3.6)
assert_almost_equal(flight_length[0], 7146.3567785516016)
assert_almost_equal(flight_length[1], 808)
def test_phase_angle(self):
# TODO. Add more tests where the phase_angle isn't zero.
phase_angle, master_opening = self.f641.phase_angle()
assert_almost_equal(phase_angle, 0)
assert_almost_equal(master_opening, 1.04719755)
def test_background_subtract_line(self):
# checked each step of the background subtraction with IGOR
# so this test background correction should be correct.
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
mask = np.zeros(201, np.bool)
mask[30:70] = True
mask[130:160] = True
profile, profile_sd = plp.background_subtract_line(yvals,
yvals_sd,
mask)
verified_data = np.load(os.path.join(self.path,
'background_subtract.npy'))
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
def test_find_specular_ridge(self):
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 1000, 0, 1))
detector = np.repeat(yvals[:, np.newaxis], 1000, axis=1).T
detector_sd = np.sqrt(detector)
output = plp.find_specular_ridge(detector[np.newaxis, :],
detector_sd[np.newaxis, :])
assert_(len(output) == 5)
assert_almost_equal(output[0][0], 100)
def test_background_subtract(self):
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
# now make an (N, T, Y) detector image
n_tbins = 10
detector = np.repeat(yvals,
n_tbins).reshape(xvals.size, n_tbins).T
detector_sd = np.repeat(yvals_sd,
n_tbins).reshape(xvals.size, n_tbins).T
detector = detector.reshape(1, n_tbins, xvals.size)
detector_sd = detector_sd.reshape(1, n_tbins, xvals.size)
mask = np.zeros((1, n_tbins, 201), np.bool)
mask[:, :, 30:70] = True
mask[:, :, 130:160] = True
det_bkg, detSD_bkg = plp.background_subtract(detector,
detector_sd,
mask)
# each of the (N, T) entries should have the same background subtracted
# entries
verified_data = np.load(os.path.join(self.path,
'background_subtract.npy'))
it = np.nditer(detector, flags=['multi_index'])
it.remove_axis(2)
while not it.finished:
profile = det_bkg[it.multi_index]
profile_sd = detSD_bkg[it.multi_index]
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
it.iternext()
def test_calculate_bins(self):
bins = plp.calculate_wavelength_bins(2., 18, 2.)
assert_almost_equal(bins[0], 1.98)
assert_almost_equal(bins[-1], 18.18)
def test_event(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
out = self.f113.process(eventmode=[0, 900, 1800], integrate=0)
assert_(np.size(out[1], axis=0) == 2)
def test_event_folder(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
self.f113.process(eventmode=[0, 900, 1800], integrate=0,
event_folder=self.path)
def test_multiple_acquisitions(self):
"""
TODO: add a dataset which has multiple spectra in it, and make sure it
processes.
"""
pass
def test_reduction_runs(self):
# just check it runs
self.f113.process()
# check that event mode reduction gives the same output as non-event
# mode reduction.
spectrum0 = self.f113.process(direct=True)
spectrum1 = self.f113.process(direct=True, eventmode=[], integrate=0)
assert_allclose(spectrum0[1][0], spectrum1[1][0], rtol=0.001)
# check that the wavelength resolution is roughly right, between 7 and
# 8%.
res = (self.f113.processed_spectrum['m_lambda_fwhm'][0] /
self.f113.processed_spectrum['m_lambda'][0])
assert_array_less(res, np.ones_like(res) * 0.08)
assert_array_less(np.ones_like(res) * 0.07, res)
def test_save_spectrum(self):
# test saving spectrum
self.f113.process()
# can save the spectra by supplying a filename
self.f113.write_spectrum_xml('test.xml')
self.f113.write_spectrum_dat('test.dat')
# can save by supplying file handle:
with open('test.xml', 'wb') as f:
self.f113.write_spectrum_xml(f)
# can save by supplying file handle:
with open('test.dat', 'wb') as f:
self.f113.write_spectrum_xml(f)
def test_accumulate_files(self):
fnames = ['PLP0000708.nx.hdf', 'PLP0000709.nx.hdf']
pths = [os.path.join(self.path, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
f8, f9, fadd = None, None, None
try:
f8 = h5py.File(os.path.join(self.path,
'PLP0000708.nx.hdf'), 'r')
f9 = h5py.File(os.path.join(self.path,
'PLP0000709.nx.hdf'), 'r')
fadd = h5py.File(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'), 'r')
f8d = f8['entry1/data/hmm'][0]
f9d = f9['entry1/data/hmm'][0]
faddd = fadd['entry1/data/hmm'][0]
assert_equal(faddd, f8d + f9d)
finally:
if f8 is not None:
f8.close()
if f9 is not None:
f9.close()
if fadd is not None:
fadd.close()
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ['PLP0000708.nx.hdf', 'PLP0000708.nx.hdf']
pths = [os.path.join(self.path, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
# it should be processable
fadd = PlatypusNexus(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'))
fadd.process()
# it should also be reduceable
reducer = PlatypusReduce(os.path.join(self.path,
'PLP0000711.nx.hdf'))
reduced = reducer.reduce(os.path.join(os.getcwd(),
'ADD_PLP0000708.nx.hdf'))
assert_('ydata' in reduced)
# the error bars should be smaller
reduced2 = reducer.reduce(os.path.join(self.path, 'PLP0000708.nx.hdf'))
assert_(np.all(reduced['ydata_sd'] < reduced2['ydata_sd']))
def test_manual_beam_find(self):
# you can specify a function that finds where the specular ridge is.
def manual_beam_find(detector, detector_sd):
beam_centre = np.zeros(len(detector))
beam_sd = np.zeros(len(detector))
beam_centre += 50
beam_sd += 5
return beam_centre, beam_sd, np.array([40]), np.array([60]), [[]]
# the manual beam find is only mandatory when peak_pos == -1.
self.f113.process(manual_beam_find=manual_beam_find,
peak_pos=-1)
assert_equal(self.f113.processed_spectrum['m_beampos'][0], 50)
# manual beam finding also specifies the lower and upper pixel of the
# foreground
assert_equal(self.f113.processed_spectrum['lopx'][0], 40)
assert_equal(self.f113.processed_spectrum['hipx'][0], 60)
def test_fore_back_region(self):
# calculation of foreground and background regions is done correctly
centres = np.array([100., 90.])
sd = np.array([5., 11.5])
lopx, hipx, background_pixels = fore_back_region(centres, sd)
assert_(len(lopx) == 2)
assert_(len(hipx) == 2)
assert_(len(background_pixels) == 2)
assert_(isinstance(lopx[0], numbers.Integral))
assert_(isinstance(hipx[0], numbers.Integral))
calc_lower = np.floor(centres - sd * EXTENT_MULT)
assert_equal(lopx, calc_lower)
calc_higher = np.ceil(centres + sd * EXTENT_MULT)
assert_equal(hipx, calc_higher)
y1 = np.atleast_1d(
np.round(lopx - PIXEL_OFFSET).astype('int'))
y0 = np.atleast_1d(
np.round(lopx - PIXEL_OFFSET - EXTENT_MULT * sd).astype('int'))
y2 = np.atleast_1d(
np.round(hipx + PIXEL_OFFSET).astype('int'))
y3 = np.atleast_1d(
np.round(hipx + PIXEL_OFFSET + EXTENT_MULT * sd).astype('int'))
bp = np.r_[np.arange(y0[0], y1[0] + 1),
np.arange(y2[0], y3[0] + 1)]
assert_equal(bp, background_pixels[0])
def test_basename_datafile(self):
# check that the right basename is returned
pth = 'a/b/c.nx.hdf'
assert_(basename_datafile(pth) == 'c')
pth = 'c.nx.hdf'
assert_(basename_datafile(pth) == 'c')
class TestPlatypusNexus(object):
@pytest.mark.usefixtures("no_data_directory")
@pytest.fixture(autouse=True)
def setup_method(self, tmpdir, data_directory):
self.pth = pjoin(data_directory, "reduce")
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
self.f113 = PlatypusNexus(pjoin(self.pth, "PLP0011613.nx.hdf"))
self.f641 = PlatypusNexus(pjoin(self.pth, "PLP0011641.nx.hdf"))
# These PNR datasets all have different flipper settings
self.f8861 = PlatypusNexus(
pjoin(self.pth, "PNR_files/PLP0008861.nx.hdf"))
self.f8862 = PlatypusNexus(
pjoin(self.pth, "PNR_files/PLP0008862.nx.hdf"))
self.f8863 = PlatypusNexus(
pjoin(self.pth, "PNR_files/PLP0008863.nx.hdf"))
self.f8864 = PlatypusNexus(
pjoin(self.pth, "PNR_files/PLP0008864.nx.hdf"))
self.cwd = os.getcwd()
self.tmpdir = tmpdir.strpath
os.chdir(self.tmpdir)
return 0
def teardown_method(self):
os.chdir(self.cwd)
def test_chod(self):
flight_length = self.f113.chod()
assert_almost_equal(flight_length[0], 7141.413818359375)
assert_almost_equal(flight_length[1], 808)
flight_length = self.f641.chod(omega=1.8, twotheta=3.6)
assert_almost_equal(flight_length[0], 7146.3567785516016)
assert_almost_equal(flight_length[1], 808)
def test_phase_angle(self):
# TODO. Add more tests where the phase_angle isn't zero.
phase_angle, master_opening = self.f641.phase_angle()
assert_almost_equal(phase_angle, 0)
assert_almost_equal(master_opening, 60)
assert self.f641.cat.t_offset is None
def test_background_subtract_line(self):
# checked each step of the background subtraction with IGOR
# so this test background correction should be correct.
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
mask = np.zeros(201, bool)
mask[30:70] = True
mask[130:160] = True
profile, profile_sd = plp.background_subtract_line(
yvals, yvals_sd, mask)
verified_data = np.load(pjoin(self.pth, "background_subtract.npy"))
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
def test_find_specular_ridge(self):
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 1000, 0, 1))
detector = np.repeat(yvals[:, np.newaxis], 1000, axis=1).T
detector_sd = np.sqrt(detector)
output = plp.find_specular_ridge(detector[np.newaxis, :],
detector_sd[np.newaxis, :])
assert_(len(output) == 5)
assert_almost_equal(output[0][0], 100)
def test_pickle(self):
# can we pickle a PlatypusNexus object?
self.f8864.process()
pkl = pickle.dumps(self.f8864)
o = pickle.loads(pkl)
ops, f8864 = o.processed_spectrum, self.f8864.processed_spectrum
assert_allclose(ops["m_spec"], f8864["m_spec"])
def test_background_subtract(self):
# create some test data
xvals = np.linspace(-10, 10, 201)
yvals = np.ceil(gauss(xvals, 0, 100, 0, 1) + 2 * xvals + 30)
# add some reproducible random noise
np.random.seed(1)
yvals += np.sqrt(yvals) * np.random.randn(yvals.size)
yvals_sd = np.sqrt(yvals)
# now make an (N, T, Y) detector image
n_tbins = 10
detector = np.repeat(yvals, n_tbins).reshape(xvals.size, n_tbins).T
detector_sd = (np.repeat(yvals_sd,
n_tbins).reshape(xvals.size, n_tbins).T)
detector = detector.reshape(1, n_tbins, xvals.size)
detector_sd = detector_sd.reshape(1, n_tbins, xvals.size)
mask = np.zeros((1, n_tbins, 201), bool)
mask[:, :, 30:70] = True
mask[:, :, 130:160] = True
det_bkg, detSD_bkg = plp.background_subtract(detector, detector_sd,
mask)
# each of the (N, T) entries should have the same background subtracted
# entries
verified_data = np.load(pjoin(self.pth, "background_subtract.npy"))
it = np.nditer(detector, flags=["multi_index"])
it.remove_axis(2)
while not it.finished:
profile = det_bkg[it.multi_index]
profile_sd = detSD_bkg[it.multi_index]
assert_almost_equal(verified_data, np.c_[profile, profile_sd])
it.iternext()
def test_calculate_bins(self):
bins = plp.calculate_wavelength_bins(2.0, 18, 2.0)
assert_almost_equal(bins[0], 1.98)
assert_almost_equal(bins[-1], 18.18)
def test_event(self):
# When you use event mode processing, make sure the right amount of
# spectra are created
out = self.f641.process(eventmode=[0, 900, 1800], integrate=0)
assert_(np.size(out[1], axis=0) == 2)
def test_event_folder(self):
self.f641.process(eventmode=[0, 900, 1800],
integrate=0,
event_folder=self.pth)
def test_multiple_acquisitions(self):
"""
TODO: add a dataset which has multiple spectra in it, and make sure it
processes.
"""
pass
def test_reduction_runs(self):
# just check it runs
self.f641.process()
# check that event mode reduction gives the same output as non-event
# mode reduction.
spectrum0 = self.f641.process(direct=False)
assert "reduction_options" in self.f641.processed_spectrum
spectrum1 = self.f641.process(direct=False, eventmode=[], integrate=0)
assert_allclose(spectrum0[1][0], spectrum1[1][0], rtol=0.001)
# check that the wavelength resolution is roughly right, between 7 and
# 8%.
res = (self.f641.processed_spectrum["m_lambda_fwhm"][0] /
self.f641.processed_spectrum["m_lambda"][0])
assert_array_less(res, np.ones_like(res) * 0.08)
assert_array_less(np.ones_like(res) * 0.07, res)
def test_save_spectrum(self):
# test saving spectrum
self.f113.process()
# can save the spectra by supplying a filename
self.f113.write_spectrum_xml(pjoin(self.tmpdir, "test.xml"))
self.f113.write_spectrum_dat(pjoin(self.tmpdir, "test.dat"))
# can save by supplying file handle:
with open(pjoin(self.tmpdir, "test.xml"), "wb") as f:
self.f113.write_spectrum_xml(f)
def test_accumulate_files(self):
fnames = ["PLP0000708.nx.hdf", "PLP0000709.nx.hdf"]
pths = [pjoin(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
f8, f9, fadd = None, None, None
try:
f8 = h5py.File(pjoin(self.pth, "PLP0000708.nx.hdf"), "r")
f9 = h5py.File(pjoin(self.pth, "PLP0000709.nx.hdf"), "r")
fadd = h5py.File(pjoin(self.tmpdir, "ADD_PLP0000708.nx.hdf"), "r")
f8d = f8["entry1/data/hmm"][0]
f9d = f9["entry1/data/hmm"][0]
faddd = fadd["entry1/data/hmm"][0]
assert_equal(faddd, f8d + f9d)
finally:
if f8 is not None:
f8.close()
if f9 is not None:
f9.close()
if fadd is not None:
fadd.close()
def test_accumulate_files_reduce(self):
# test by adding a file to itself. Should have smaller stats
fnames = ["PLP0000708.nx.hdf", "PLP0000708.nx.hdf"]
pths = [pjoin(self.pth, fname) for fname in fnames]
plp.accumulate_HDF_files(pths)
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
# it should be processable
fadd = PlatypusNexus(pjoin(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
fadd.process()
# it should also be reduceable
reducer = PlatypusReduce(pjoin(self.pth, "PLP0000711.nx.hdf"))
datasets, reduced = reducer.reduce(
pjoin(os.getcwd(), "ADD_PLP0000708.nx.hdf"))
assert_("y" in reduced)
# the error bars should be smaller
datasets2, reduced2 = reducer.reduce(
pjoin(self.pth, "PLP0000708.nx.hdf"))
assert_(np.all(reduced["y_err"] < reduced2["y_err"]))
def test_manual_beam_find(self):
# you can specify a function that finds where the specular ridge is.
def manual_beam_find(detector, detector_sd, name):
beam_centre = np.zeros(len(detector))
beam_sd = np.zeros(len(detector))
beam_centre += 50
beam_sd += 5
return beam_centre, beam_sd, np.array([40]), np.array([60]), [[]]
# the manual beam find is only mandatory when peak_pos == -1.
# the beam_sd is much larger than the beam divergence, so a warning
# should be raised.
with pytest.warns(UserWarning):
self.f113.process(manual_beam_find=manual_beam_find, peak_pos=-1)
assert_equal(self.f113.processed_spectrum["m_beampos"][0], 50)
# manual beam finding also specifies the lower and upper pixel of the
# foreground
assert_equal(self.f113.processed_spectrum["lopx"][0], 40)
assert_equal(self.f113.processed_spectrum["hipx"][0], 60)
def test_fore_back_region(self):
# calculation of foreground and background regions is done correctly
centres = np.array([100.0, 90.0])
sd = np.array([5.0, 11.5])
lopx, hipx, background_pixels = fore_back_region(centres, sd)
assert_(len(lopx) == 2)
assert_(len(hipx) == 2)
assert_(len(background_pixels) == 2)
assert_(isinstance(lopx[0], numbers.Integral))
assert_(isinstance(hipx[0], numbers.Integral))
calc_lower = np.floor(centres - sd * EXTENT_MULT)
assert_equal(lopx, calc_lower)
calc_higher = np.ceil(centres + sd * EXTENT_MULT)
assert_equal(hipx, calc_higher)
y1 = np.atleast_1d(np.round(lopx - PIXEL_OFFSET).astype("int"))
y0 = np.atleast_1d(
np.round(lopx - PIXEL_OFFSET - EXTENT_MULT * sd).astype("int"))
y2 = np.atleast_1d(np.round(hipx + PIXEL_OFFSET).astype("int"))
y3 = np.atleast_1d(
np.round(hipx + PIXEL_OFFSET + EXTENT_MULT * sd).astype("int"))
bp = np.r_[np.arange(y0[0], y1[0] + 1), np.arange(y2[0], y3[0] + 1)]
assert_equal(bp, background_pixels[0])
def test_basename_datafile(self):
# check that the right basename is returned
pth = "a/b/c.nx.hdf"
assert_(basename_datafile(pth) == "c")
pth = "c.nx.hdf"
assert_(basename_datafile(pth) == "c")
def test_floodfield_correction(self):
# check that flood field calculation works
# the values were worked out by hand on a randomly
# generated array
test_norm = np.array(
[1.12290503, 1.23743017, 0.8603352, 0.70111732, 1.07821229])
test_norm_sd = np.array(
[0.05600541, 0.05879208, 0.04902215, 0.04425413, 0.05487956])
fname = os.path.join(self.pth, "flood.h5")
with h5py.File(fname, "r") as f:
norm, norm_sd = create_detector_norm(f, 3.5, -3.5, axis=3)
assert_almost_equal(norm, test_norm, 6)
assert_almost_equal(norm_sd, test_norm_sd, 6)
norm, norm_sd = create_detector_norm(f, -3.5, 3.5, axis=3)
assert_almost_equal(norm, test_norm, 6)
assert_almost_equal(norm_sd, test_norm_sd, 6)
def test_PNR_metadata(self):
self.f8861.process()
self.f8862.process()
self.f8863.process()
self.f8864.process()
# Check that we can read all of the flipper statuses in files
# Flipper 1 on, flipper 2 on
assert_almost_equal(self.f8861.cat.cat["pol_flip_current"], 5.0)
assert_almost_equal(self.f8861.cat.cat["anal_flip_current"], 4.5)
# Flipper 1 on, flipper 2 off
assert_almost_equal(self.f8862.cat.cat["anal_flip_current"], 0)
assert_almost_equal(self.f8862.cat.cat["pol_flip_current"], 5.0)
# Flipper 1 off, flipper 2 on
assert_almost_equal(self.f8863.cat.cat["anal_flip_current"], 4.5)
assert_almost_equal(self.f8863.cat.cat["pol_flip_current"], 0)
# Flipper 1 off, flipper 2 off
assert_almost_equal(self.f8864.cat.cat["anal_flip_current"], 0)
assert_almost_equal(self.f8864.cat.cat["pol_flip_current"], 0)
# Check SpinChannel for each file
assert self.f8861.spin_state == SpinChannel.UP_UP
assert self.f8862.spin_state == SpinChannel.UP_DOWN
assert self.f8863.spin_state == SpinChannel.DOWN_UP
assert self.f8864.spin_state == SpinChannel.DOWN_DOWN
# test spin channel setting
# non spin analysed. mode is POL, not POLANAL
pn = PlatypusNexus(pjoin(self.pth, "PLP0016427.nx.hdf"))
assert pn.spin_state == SpinChannel.UP_UP
pn = PlatypusNexus(pjoin(self.pth, "PLP0016426.nx.hdf"))
assert pn.spin_state == SpinChannel.DOWN_DOWN
def test_PNR_magnet_read(self):
self.f8861.process()
# Check magnetic field sensors
assert_almost_equal(self.f8861.cat.cat["magnet_current_set"], 0)
assert_almost_equal(self.f8861.cat.cat["magnet_output_current"], 0.001)