def test_fft_axis():
for a in (0, 1):
A = op.fft(shapein16, axis=a)
yield check_almost_dense_transpose, A
def test_fft_n_complex():
for n in xrange(1, shapein16[-1]):
A = op.fft(shapein16, n=n, dtype=np.complex128)
yield check_almost_dense_transpose, A
def test_fft_n_real():
for n in xrange(1, shapein16[-1]):
A = op.fft(shapein16, n=n, dtype=np.float64)
yield check_almost_dense_transpose, A
return a
# collection of linear operators to test
shapein16 = (16, 16)
shapein3_4 = (4, 4, 4)
d = np.random.rand(*shapein16)
coef = np.random.uniform(high=10.)
mask = np.random.rand(*shapein16) > .5
s = int(np.floor(np.random.uniform(low=1, high=16)))
p = generate_permutation(16)
id16 = op.ndidentity(shapein16)
d16 = op.nddiagonal(d)
h16 = op.ndhomothetic(shapein16, coef)
m16 = op.ndmask(mask)
f1_16 = op.fft(shapein16, dtype=np.complex128)
f2_16 = op.fft2(shapein16, dtype=np.complex128)
lo_list = [id16, d16, h16, m16, f1_16, f2_16]
# collection of vectors
ones16 = np.ones(shapein16)
zeros16 = np.zeros(shapein16)
rand16 = np.random.rand(*shapein16)
v_list = [ones16, zeros16, rand16 ]
# general tests
def check_matvec(A, x):
A * x.ravel()
