def test_functions_callable(self):
arr = np.ones((5, 5))
mask = _sigma_clip(arr, cen_func=np.median)
mask = _sigma_clip(arr, dev_func=np.std)
# testing forced 1.0pm0.5
def _test_cen(*args, **kwargs):
return 1.0
def _test_dev(*args, **kwargs):
return 0.5
# 1.2 should not be masked with 1.0pm0.5, 2.0 and 1000 yes
arr[0, 0] = 1.2
arr[1, 1] = 2
arr[3, 2] = 1000
mask = _sigma_clip(arr, 1, cen_func=_test_cen, dev_func=_test_dev)
for i in range(5):
for j in range(5):
if (i, j) in [(1, 1), (3, 2)]:
assert_true(mask[i, j])
else:
assert_false(mask[i, j])
def test_sigmclip_types(self):
arr = np.arange(10)
# must work with numbers
_sigma_clip(arr, 1)
# must work with 2-elements array
_sigma_clip(arr, np.array([1, 2]))
_sigma_clip(arr, [1, 2])
_sigma_clip(arr, (1, 2))
def test_1D_simple(self):
with NumpyRNGContext(123):
arr = np.random.normal(5, 2, 100)
arr[32] = 1000
arr[25] = 1000
arr[12] = -1000
exp = np.zeros(100, dtype=bool)
exp[32] = True
exp[25] = True
exp[12] = True
mask = _sigma_clip(arr, 3)
assert_equal(mask, exp)
def test_invalid(self):
arr = np.ones((5, 5))
exp = np.zeros((5, 5), dtype=bool)
indx = [(1, 1), (2, 1), (2, 3)]
arr[indx[0]] = 1000
arr[indx[1]] = np.inf
arr[indx[2]] = np.nan
for i in indx:
exp[i] = True
mask = _sigma_clip(arr)
assert_equal(mask, exp)
def test_unkown_sigmaclip(self):
arr = np.arange(10)
with pytest.raises(TypeError):
_sigma_clip(arr, 'testing')
with pytest.raises(TypeError):
_sigma_clip(arr, '1,2')
# too many values must raise
with pytest.raises(ValueError):
_sigma_clip(arr, [1, 2, 3])
def test_2D_defaults(self):
# mean:4.532 median:4.847 std:282 mad_std: 2.02
arr = np.array([[6.037, 5.972, 5.841, 2.775, 0.711],
[6.539, 4.677, -1000, 5.633, 3.478],
[4.847, 7.563, 3.323, 7.952, 6.646],
[6.136, 2.690, 4.193, 1000., 4.483],
[5.673, 7.479, 3.874, 4.756, 2.021]])
# using defaults median and mad_std
expect_1 = [[0, 0, 0, 1, 1], [0, 0, 1, 0, 0], [0, 1, 0, 1, 0],
[0, 1, 0, 1, 0], [0, 1, 0, 0, 1]]
expect_1 = np.array(expect_1, dtype=bool)
expect_3 = [[0, 0, 0, 0, 0], [0, 0, 1, 0, 0], [0, 0, 0, 0, 0],
[0, 0, 0, 1, 0], [0, 0, 0, 0, 0]]
expect_3 = np.array(expect_3, dtype=bool)
mask = _sigma_clip(arr, 3)
assert_equal(mask, expect_3)
def test_functions_invalid(self):
arr = np.ones((5, 5))
with pytest.raises(KeyError, match='invalid'):
_sigma_clip(arr, cen_func='invalid')
with pytest.raises(KeyError, match='invalid'):
_sigma_clip(arr, dev_func='invalid')
with pytest.raises(KeyError, match='1'):
_sigma_clip(arr, cen_func=1)
with pytest.raises(KeyError, match='1'):
_sigma_clip(arr, dev_func=1)
with pytest.raises(TypeError):
_sigma_clip(arr, cen_func=[])
with pytest.raises(TypeError):
_sigma_clip(arr, dev_func=[])
def test_functions_names(self):
arr = np.ones((5, 5))
# all this should run
_sigma_clip(arr, cen_func='median')
_sigma_clip(arr, cen_func='mean')
_sigma_clip(arr, dev_func='std')
_sigma_clip(arr, cen_func='mean', dev_func='std')
_sigma_clip(arr, dev_func='mad_std')
_sigma_clip(arr, cen_func='median', dev_func='mad_std')