def test_combine_array(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
out = combine_arrays(data, method='mean', clipping_method='sigclip')
assert out.data.dtype == np.float64
assert out.mask is None
assert out.uncertainty is None
assert np.isclose(out.data[0], 2.2)
assert out.meta['REJMAP'][0] == 1
out = combine_arrays(data,
method='mean',
clipping_method='sigclip',
clipping_limits=(2, 2))
assert np.isclose(out.data[0], 2)
assert out.meta['REJMAP'][0] == 2
out = combine_arrays(data,
method='mean',
clipping_method='sigclip',
clipping_limits=(5, 5))
assert np.isclose(out.data[0], 11.09, atol=1e-2)
assert out.meta['REJMAP'][0] == 0
out = combine_arrays(data,
method='mean',
clipping_method='sigclip',
max_iters=0)
assert np.isclose(out.data[0], 11.09, atol=1e-2)
assert out.meta['REJMAP'][0] == 0
def test_combine_nddata(dtype):
data = [NDData(data=np.array([val], dtype=dtype)) for val in TEST_VALUES]
out = combine_arrays(data, method='mean', clipping_method='sigclip')
assert out.data.dtype == np.float64
assert out.mask is None
assert out.uncertainty is None
assert np.isclose(out.data[0], 2.2)
assert out.meta['REJMAP'][0] == 1
def test_combine_array_with_mask(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
mask = np.array([[0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0]], dtype=bool).T
out = combine_arrays(data,
mask=mask,
method='mean',
clipping_method='sigclip')
assert np.isclose(out.data[0], 2.)
assert out.meta['REJMAP'][0] == 4
def test_array_with_variance(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
var = np.ones_like(data)
out = combine_arrays(data,
variance=var,
method='mean',
clipping_method='sigclip')
assert isinstance(out.uncertainty, VarianceUncertainty)
assert np.isclose(out.data[0], 2.2)
assert np.isclose(out.uncertainty.array[0], 1 / 10) # 10 valid values
assert out.meta['REJMAP'][0] == 1
var = np.random.normal(size=data.shape)
out = combine_arrays(data,
variance=var,
method='mean',
clipping_method='sigclip')
assert np.isclose(out.data[0], 2.2)
assert np.isclose(out.uncertainty.array[0], np.mean(var[:10]) / 10)
def test_combine_nddata_with_mask(dtype):
mask_values = [0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0]
data = [
NDData(data=np.array([val], dtype=dtype),
mask=np.array([mask], dtype=bool))
for val, mask in zip(TEST_VALUES, mask_values)
]
out = combine_arrays(data, method='mean', clipping_method='sigclip')
assert np.isclose(out.data[0], 2.)
assert out.meta['REJMAP'][0] == 4
def test_nddata_with_variance(dtype):
data = [
NDData(data=np.array([val], dtype=dtype),
uncertainty=VarianceUncertainty(np.array([1], dtype=dtype)))
for val in TEST_VALUES
]
out = combine_arrays(data, method='mean', clipping_method='sigclip')
assert isinstance(out.uncertainty, VarianceUncertainty)
assert np.isclose(out.data[0], 2.2)
assert np.isclose(out.uncertainty.array[0], 1 / 10) # 10 valid values
assert out.meta['REJMAP'][0] == 1
def test_combine_sum(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
var = np.ones_like(data)
out = combine_arrays(data,
variance=var,
method='sum',
clipping_method='sigclip')
assert out.data.dtype == np.float64
assert out.mask is None
assert np.isclose(out.data[0], 22)
assert np.isclose(out.uncertainty.array[0], 10) # 10 valid values
assert out.meta['REJMAP'][0] == 1
def test_combine_varclip(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
var = np.ones_like(data)
var[-1] = 100
out = combine_arrays(data,
variance=var,
method='mean',
clipping_method='varclip')
assert isinstance(out.uncertainty, VarianceUncertainty)
assert np.isclose(out.data[0], 2.2)
assert np.isclose(out.uncertainty.array[0], 1 / 10) # 10 valid values
assert out.meta['REJMAP'][0] == 1
var[-1] = 100_000
out = combine_arrays(data,
variance=var,
method='mean',
clipping_method='varclip')
assert isinstance(out.uncertainty, VarianceUncertainty)
assert np.isclose(out.data[0], 11.09, atol=1e-2)
assert np.isclose(out.uncertainty.array[0], (100000 + 10) / 11**2)
assert out.meta['REJMAP'][0] == 0
def test_unknow_combiner():
data = np.array([TEST_VALUES]).T
with pytest.raises(ValueError, match='unknow combination method: foo'):
combine_arrays(data, method='foo')
def test_unknow_rejector():
data = np.array([TEST_VALUES]).T
with pytest.raises(ValueError, match='unknow rejection method: foo'):
combine_arrays(data, method='mean', clipping_method='foo')
def test_combine_no_clipping(dtype):
data = np.array([TEST_VALUES], dtype=dtype).T
out = combine_arrays(data, method='mean', clipping_method='none')
assert np.isclose(out.data[0], 11.09, atol=1e-2)
assert out.meta['REJMAP'][0] == 0
