def test_invalid_arguments(self):
"""Check that exception is thrown for invalid arguments"""
field = np.ones([5, 5])
halfwidth = -1
quantiles = [-0.1, 1.1, np.nan]
thresholds = [0, 1]
for quantile in quantiles:
with self.assertRaises(ValueError) as e:
gridpp.neighbourhood_quantile_fast(field, quantile, halfwidth,
thresholds)
def test_varying_quantile(self):
""" For now check that this runs """
values = np.array([[0, 1], [2, 3], [4, 5]])
halfwidth = 1
quantiles = np.ones(values.shape) * 0.5
thresholds = [0, 0.25, 0.5, 0.75, 1]
gridpp.neighbourhood_quantile_fast(values, quantiles, halfwidth,
thresholds)
values = np.nan * np.zeros(values.shape)
np.testing.assert_array_equal(
values,
gridpp.neighbourhood_quantile_fast(values, quantiles, halfwidth,
thresholds))
def test_varying_quantile_3d(self):
""" For now check that this runs """
np.random.seed(1000)
values = np.random.rand(100, 50, 2)
halfwidth = 1
quantiles = np.ones(values[:, :, 0].shape) * 0.5
thresholds = [0, 0.25, 0.5, 0.75, 1]
gridpp.neighbourhood_quantile_fast(values, quantiles, halfwidth,
thresholds)
values = np.nan * np.zeros(values.shape)
np.testing.assert_array_equal(
values[:, :, 0],
gridpp.neighbourhood_quantile_fast(values, quantiles, halfwidth,
thresholds))
def test_3d(self):
np.random.seed(1000)
values = np.random.rand(200, 200)
values3 = np.zeros([200, 200, 5])
for i in range(5):
values3[:, :, i] = values
halfwidths = [0, 1, 5]
quantile = 0.5
thresholds = [0, 0.25, 0.5, 0.75, 1]
for halfwidth in halfwidths:
output_2d = gridpp.neighbourhood_quantile_fast(
values, quantile, halfwidth, thresholds)
output_3d = gridpp.neighbourhood_quantile_fast(
values3, quantile, halfwidth, thresholds)
np.testing.assert_array_almost_equal(output_2d, output_3d)
def test_two_thresholds(self):
"""Checks what happens when a single threshold is provided"""
thresholds = [0, 1]
field = np.reshape(np.arange(9), [3, 3])
for halfwidth in [0, 1, 2]:
output = gridpp.neighbourhood_quantile_fast(
field, 0.9, 0, thresholds)
self.assertTrue(((output >= 0) & (output <= 1)).all())
def test_single_threshold(self):
"""Checks what happens when a single threshold is provided"""
thresholds = [0]
field = np.reshape(np.arange(9), [3, 3])
for halfwidth in [0, 1, 2]:
output = gridpp.neighbourhood_quantile_fast(
field, 0.9, halfwidth, thresholds)
np.testing.assert_array_equal(output, np.zeros([3, 3]))
def test_all_same(self):
""" Check that min and max of an neighbourhood with all identical values is correct """
field = np.zeros([10, 10])
thresholds = [0, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50, 100]
for quantile in [0, 0.001, 0.999, 1]:
with self.subTest(quantile=quantile):
output = gridpp.neighbourhood_quantile_fast(
field, quantile, 5, thresholds)
np.testing.assert_array_almost_equal(output, field)
def test_empty(self):
for quantile in np.arange(0.1, 0.9, 0.1):
for num_thresholds in [1, 2]:
thresholds = gridpp.get_neighbourhood_thresholds(
values, num_thresholds)
output = gridpp.neighbourhood_quantile_fast([[]], 0.9, 1,
thresholds)
self.assertEqual(len(output.shape), 2)
self.assertEqual(output.shape[0], 0)
self.assertEqual(output.shape[1], 0)
def test_nan_quantile(self):
field = np.ones([5, 5])
halfwidth = 1
quantile = np.nan
thresholds = [0, 1]
output = gridpp.neighbourhood_quantile_fast(field, quantile, halfwidth,
thresholds)
np.testing.assert_array_almost_equal(np.nan * np.ones(output.shape),
output)
def test_quantile(self):
thresholds = gridpp.get_neighbourhood_thresholds(values, 100)
output = np.array(
gridpp.neighbourhood_quantile_fast(values, 0.5, 1, thresholds))
self.assertEqual(output[2][2], 12) # Should be 12.5
self.assertEqual(output[2][3], 12.5) # Should be 13
output = np.array(
gridpp.neighbourhood_quantile_fast(np.full([100, 100], np.nan),
0.5, 1, thresholds))
self.assertTrue(np.isnan(np.array(output)).all())
output = np.array(
gridpp.neighbourhood_quantile_fast(np.zeros([100, 100]), 0.5, 1,
thresholds))
self.assertTrue((np.array(output) == 0).all())
output = np.array(gridpp.neighbourhood_quantile(values, 0.5, 1))
self.assertEqual(output[2][2], 12.5)
self.assertEqual(output[2][3], 13)
self.assertEqual(output[0][4], 4)
def test_empty(self):
for statistic in [
gridpp.Mean, gridpp.Min, gridpp.Max, gridpp.Median, gridpp.Std,
gridpp.Variance
]:
output = gridpp.neighbourhood([[]], 1, statistic)
self.assertEqual(len(output.shape), 2)
self.assertEqual(output.shape[0], 0)
self.assertEqual(output.shape[1], 0)
for quantile in np.arange(0.1, 0.9, 0.1):
for num_thresholds in [1, 2]:
thresholds = gridpp.get_neighbourhood_thresholds(
values, num_thresholds)
output = gridpp.neighbourhood_quantile_fast([[]], 0.9, 1,
thresholds)
self.assertEqual(len(output.shape), 2)
self.assertEqual(output.shape[0], 0)
self.assertEqual(output.shape[1], 0)
def test_missing(self):
empty = np.zeros([5, 5])
empty[0:3, 0:3] = np.nan
thresholds = [0, 1]
output = gridpp.neighbourhood_quantile_fast(empty, 0.5, 1, thresholds)
self.assertTrue(np.isnan(np.array(output)[0:2, 0:2]).all())
def test_quantile_exceed(self):
input = np.reshape(range(25), [5, 5]).astype(float)
thresholds = [1, 5, 10]
output = gridpp.neighbourhood_quantile_fast(values, 0.9, 1, thresholds)
