def test_cusplet_kernel_args():
arr = np.random.randn(100)
kernel = kernel_functions.power_zero_cusp
widths = [50]
# need kernel args but we don't give them
with pytest.raises(TypeError) as e_info:
_ = shocklets.cusplet(arr, widths, kernel_func=kernel)
def test_cusplet_output():
arr = np.random.randn(100)
kernel = kernel_functions.power_zero_cusp
widths = np.linspace(10, 30, 20).astype(int)
k_args = [3.]
cc, _ = shocklets.cusplet(arr, widths, kernel_func=kernel, kernel_args=k_args)
assert cc.shape == (20, 100)
def test_make_components():
arr1 = np.linspace(1, 100, 100)
arr2 = np.linspace(100, 1, 100)
arr = np.hstack((arr1, arr2)) # make a triangle
kernel = kernel_functions.power_cusp
widths = np.linspace(10, 30, 20).astype(int)
k_args = [3.]
cc, _ = shocklets.cusplet(arr, widths, kernel_func=kernel, kernel_args=k_args)
extrema, sum_cc, gez = shocklets.classify_cusps(cc, b=0.75, geval=0.5)
windows, estimated_cusp_pts = shocklets.make_components(gez, extrema)
assert len(windows) == 1
def test_classify_cusps():
arr1 = np.linspace(1, 100, 100)
arr2 = np.linspace(100, 1, 100)
arr = np.hstack((arr1, arr2)) # make a triangle
kernel = kernel_functions.power_cusp
widths = np.linspace(10, 30, 20).astype(int)
k_args = [3.]
cc, _ = shocklets.cusplet(arr, widths, kernel_func=kernel, kernel_args=k_args)
extrema, sum_cc, gez = shocklets.classify_cusps(cc, b=0.75, geval=0.5)
assert len(gez) > 0
assert max(sum_cc) > 1
assert extrema[0] == 100
