def test_set_data_type(self):
# set 8 bit data type which will result in an incorrect mean square
# pattern if not changed.
dp_array = np.array([
[[[10, 10], [50, 30]], [[8, 20], [50, 30]]],
[[[12, 30], [0, 30]], [[10, 10], [50, 30]]],
]).astype(np.uint8)
variance_test_diffraction_pattern = ElectronDiffraction2D(dp_array)
vargen = VarianceGenerator(variance_test_diffraction_pattern)
vardps_8 = vargen.get_diffraction_variance(dqe=1)
assert isinstance(vardps_8, DiffractionVariance2D)
corr_var_dp_8 = np.array([[-0.08, -0.66857143],
[-0.92213333, -0.88666667]])
assert np.allclose(vardps_8.data[1, 1],
corr_var_dp_8,
atol=1e-6,
equal_nan=True)
vardps_16 = vargen.get_diffraction_variance(dqe=1,
set_data_type=np.uint16)
assert isinstance(vardps_16, DiffractionVariance2D)
corr_var_dp_16 = np.array([[-0.08, 0.16734694],
[0.30666667, -0.0333333]])
assert np.allclose(vardps_16.data[1, 1],
corr_var_dp_16,
atol=1e-6,
equal_nan=True)
def variance_generator():
diffraction_pattern = ElectronDiffraction2D(
np.array([
[
[
[0.66, 0.5, 0.01, 0.54, 0.83, 0.23, 0.92, 0.14],
[0.61, 0.45, 0.01, 0.42, 0.33, 0.48, 0.35, 0.19],
[0.87, 0.98, 0.58, 0.51, 0.75, 0.95, 0.22, 0.52],
[0.29, 0.13, 0.22, 0.98, 0.04, 0.12, 0.26, 0.67],
[0.59, 0.55, 0.63, 0.57, 0.89, 0.39, 0.81, 0.43],
[0.9, 0.54, 0.24, 0.92, 0.77, 0.33, 0.22, 0.7],
[0.7, 0.3, 0.76, 0.27, 0.93, 0.28, 0.73, 0.22],
[0.33, 0.7, 0.32, 0.54, 0.27, 0.87, 0.97, 0.7],
],
[
[0.68, 0.33, 0.11, 0.26, 0.88, 0.02, 0.71, 0.98],
[0.62, 0.55, 0.22, 0.43, 0.17, 0.77, 0.42, 0.52],
[0.45, 0.2, 0.21, 0.53, 0.38, 0.64, 0.32, 0.38],
[0.5, 0.23, 0.25, 0.91, 0.64, 1.0, 0.69, 0.16],
[0.11, 0.82, 0.23, 0.91, 0.46, 0.08, 0.55, 0.86],
[0.95, 0.68, 0.74, 0.44, 0.65, 0.92, 0.04, 0.62],
[0.59, 0.08, 0.2, 0.33, 0.94, 0.67, 0.31, 0.62],
[0.77, 0.01, 0.46, 0.81, 0.92, 0.72, 0.38, 0.25],
],
],
[
[
[0.07, 0.63, 0.89, 0.71, 0.65, 0.1, 0.58, 0.78],
[0.05, 0.35, 0.02, 0.87, 0.06, 0.06, 0.23, 0.52],
[0.44, 0.77, 0.93, 0.94, 0.22, 0.29, 0.22, 0.03],
[0.9, 0.79, 0.82, 0.61, 0.6, 0.96, 0.13, 0.63],
[0.57, 0.63, 0.53, 0.81, 0.38, 0.92, 0.92, 0.07],
[0.01, 0.33, 0.69, 0.36, 0.91, 0.24, 0.05, 0.85],
[0.81, 0.38, 0.74, 0.12, 0.33, 0.29, 0.25, 0.06],
[0.28, 0.61, 0.37, 0.17, 0.38, 0.52, 0.36, 0.69],
],
[
[0.77, 0.68, 0.76, 0.62, 0.52, 0.69, 0.63, 0.11],
[0.6, 0.96, 0.17, 0.58, 0.12, 0.23, 0.1, 0.2],
[0.4, 0.24, 0.25, 0.61, 0.27, 0.02, 0.03, 0.12],
[0.83, 0.66, 0.15, 0.74, 0.91, 0.19, 0.97, 0.58],
[0.5, 0.27, 0.93, 0.1, 0.78, 0.73, 0.56, 0.82],
[0.49, 0.35, 0.9, 0.48, 0.04, 0.46, 0.16, 0.1],
[1.0, 0.13, 0.56, 0.12, 0.16, 0.49, 0.63, 0.5],
[0.97, 0.71, 0.78, 0.38, 0.08, 0.79, 0.41, 0.07],
],
],
]))
return VarianceGenerator(diffraction_pattern)
def test_get_image_variance(self, variance_generator: VarianceGenerator,
dqe):
varims = variance_generator.get_image_variance(dqe)
assert isinstance(varims, ImageVariance)
mean_im = np.array([[0.51765625, 0.504375], [0.47625,
0.47125]]).reshape(2, 2)
meansq_im = np.array([[0.34353281, 0.33261875],
[0.31661875, 0.3085875]]).reshape(2, 2)
var_im = np.array([[0.28199196, 0.30749375],
[0.39593968, 0.38955456]]).reshape(2, 2)
corr_var_im = var_im - np.divide(dqe, mean_im)
assert np.allclose(varims.data[0, 0], mean_im, atol=1e-6)
assert np.allclose(varims.data[0, 1], meansq_im, atol=1e-6)
assert np.allclose(varims.data[1, 0], var_im, atol=1e-14)
assert np.allclose(varims.data[1, 1], corr_var_im, atol=1e-14)
def test_get_diffraction_variance(self,
variance_generator: VarianceGenerator,
dqe):
vardps = variance_generator.get_diffraction_variance(dqe)
assert isinstance(vardps, DiffractionVariance2D)
mean_dp = np.array([
[0.545, 0.535, 0.4425, 0.5325, 0.72, 0.26, 0.71, 0.5025],
[0.47, 0.5775, 0.105, 0.575, 0.17, 0.385, 0.275, 0.3575],
[0.54, 0.5475, 0.4925, 0.6475, 0.405, 0.475, 0.1975, 0.2625],
[0.63, 0.4525, 0.36, 0.81, 0.5475, 0.5675, 0.5125, 0.51],
[0.4425, 0.5675, 0.58, 0.5975, 0.6275, 0.53, 0.71, 0.545],
[0.5875, 0.475, 0.6425, 0.55, 0.5925, 0.4875, 0.1175, 0.5675],
[0.775, 0.2225, 0.565, 0.21, 0.59, 0.4325, 0.48, 0.35],
[0.5875, 0.5075, 0.4825, 0.475, 0.4125, 0.725, 0.53, 0.4275],
]).reshape(8, 8)
meansq_dp = np.array([
[
0.37395,
0.30455,
0.345475,
0.311925,
0.53905,
0.13485,
0.52095,
0.400125,
],
[
0.27975,
0.387275,
0.01945,
0.36365,
0.03895,
0.21995,
0.09045,
0.154225,
],
[
0.32825,
0.412725,
0.326975,
0.449175,
0.20705,
0.34915,
0.050025,
0.107525,
],
[
0.45825,
0.282375,
0.20145,
0.67705,
0.399825,
0.493025,
0.375375,
0.30195,
],
[
0.233775,
0.361175,
0.3989,
0.454775,
0.439125,
0.38445,
0.52965,
0.40045,
],
[
0.488175,
0.24635,
0.472825,
0.35,
0.461275,
0.306125,
0.019525,
0.401725,
],
[
0.62355, 0.064425, 0.3697, 0.05265, 0.47075, 0.212875, 0.2721,
0.1716
],
[
0.430275,
0.341575,
0.264825,
0.28025,
0.267525,
0.54245,
0.34575,
0.258375,
],
]).reshape(8, 8)
var_dp = np.array([
[
0.25898493,
0.06402306,
0.76437167,
0.10004629,
0.0398341,
0.99482249,
0.03342591,
0.58461424,
],
[
0.26641014,
0.16122262,
0.76417234,
0.09988658,
0.34775087,
0.48389273,
0.19603306,
0.20670937,
],
[
0.12568587,
0.37686871,
0.34803783,
0.07136149,
0.26230758,
0.54747922,
0.28248678,
0.56045351,
],
[
0.15457294,
0.37907878,
0.55439815,
0.03193111,
0.33383374,
0.53086611,
0.42914932,
0.16089965,
],
[
0.1939098,
0.12146558,
0.18579073,
0.27385725,
0.11522039,
0.36863653,
0.05068439,
0.34820301,
],
[
0.41435944,
0.09185596,
0.14539206,
0.15702479,
0.31396322,
0.28809993,
0.41421458,
0.24737526,
],
[
0.03816857,
0.30135084,
0.15811732,
0.19387755,
0.35234128,
0.13802666,
0.18098958,
0.40081633,
],
[
0.24660933,
0.32621515,
0.13753389,
0.24210526,
0.5722314,
0.03200951,
0.23086508,
0.41376834,
],
]).reshape(8, 8)
corr_var_dp = var_dp - np.divide(dqe, mean_dp)
corr_var_dp[np.isinf(corr_var_dp)] = 0
corr_var_dp[np.isnan(corr_var_dp)] = 0
assert np.allclose(vardps.data[0, 0], mean_dp, atol=1e-6)
assert np.allclose(vardps.data[0, 1], meansq_dp, atol=1e-6)
assert np.allclose(vardps.data[1, 0],
var_dp,
atol=1e-14,
equal_nan=True)
assert np.allclose(vardps.data[1, 1],
corr_var_dp,
atol=1e-14,
equal_nan=True)