def test_data_compute_alm_sim_1d_T(self):
cosmo = self.FakeCosmology()
pol = ['T']
data = Data(self.lmax, self.n_ell_T, self.b_ell_T, pol, cosmo)
np.random.seed(10)
alm = data.compute_alm_sim(lens_power=False)
np.random.seed(10)
alm_exp = hp.synalm(data.cov_ell_nonlensed[0], new=True)
alm_sim_expec = np.zeros((1, self.nelem), dtype=complex)
alm_sim_expec[0] = alm_exp
np.testing.assert_almost_equal(alm, alm_sim_expec)
# Again for lensed.
np.random.seed(10)
alm = data.compute_alm_sim(lens_power=True)
np.random.seed(10)
alm_exp = hp.synalm(data.cov_ell_lensed[0], new=True)
alm_sim_expec = np.zeros((1, self.nelem), dtype=complex)
alm_sim_expec[0] = alm_exp
np.testing.assert_almost_equal(alm, alm_sim_expec)
def test_data_lmax(self):
pol = 'T'
cosmo = self.FakeCosmology()
data = Data(self.lmax, self.n_ell_T, self.b_ell_T, pol, cosmo)
# Test shapes.
self.assertEqual(data.b_ell.shape, (1, self.lmax + 1))
self.assertEqual(data.n_ell.shape, (1, self.lmax + 1))
self.assertEqual(data.cov_ell_lensed.shape, (1, self.lmax + 1))
self.assertEqual(data.cov_ell_nonlensed.shape, (1, self.lmax + 1))
self.assertEqual(data.icov_ell_lensed.shape, (1, self.lmax + 1))
self.assertEqual(data.icov_ell_nonlensed.shape, (1, self.lmax + 1))
lmax_new = self.lmax - 1
data.lmax = lmax_new
# Shapes should be updated
self.assertEqual(data.b_ell.shape, (1, lmax_new + 1))
self.assertEqual(data.n_ell.shape, (1, lmax_new + 1))
self.assertEqual(data.cov_ell_lensed.shape, (1, lmax_new + 1))
self.assertEqual(data.cov_ell_nonlensed.shape, (1, lmax_new + 1))
self.assertEqual(data.icov_ell_lensed.shape, (1, lmax_new + 1))
self.assertEqual(data.icov_ell_nonlensed.shape, (1, lmax_new + 1))
lmax_new = self.lmax + 1
data.lmax = lmax_new
self.assertRaises(IndexError, getattr, data, 'b_ell')
self.assertRaises(IndexError, getattr, data, 'n_ell')
self.assertRaises(IndexError, getattr, data, 'cov_ell_lensed')
self.assertRaises(IndexError, getattr, data, 'icov_ell_lensed')
self.assertRaises(IndexError, getattr, data, 'cov_ell_nonlensed')
self.assertRaises(IndexError, getattr, data, 'icov_ell_nonlensed')
def test_data_icov_diag_nonlensed_1D_E(self):
cosmo = self.FakeCosmology()
pol = ['E']
data = Data(self.lmax, self.n_ell_T, self.b_ell_T, pol, cosmo)
alm = np.random.randn(hp.Alm.getsize(self.lmax))
alm = np.reshape(alm.astype(complex), (1, self.nelem))
alm_in = alm.copy()
alm = data.icov_diag_nonlensed(alm)
alm_exp = np.zeros_like(alm_in)
alm_exp[0] = hp.almxfl(alm_in[0], data.icov_ell_nonlensed[0])
np.testing.assert_array_almost_equal(alm, alm_exp)
def test_data_init_2d(self):
pol = ['T', 'E']
cosmo = self.FakeCosmology()
data = Data(self.lmax, self.n_ell_TplusE, self.b_ell_TplusE, pol,
cosmo)
# Test if original arrays do not share memory with
# internal arrays. I want copies.
self.assertFalse(np.shares_memory(data.b_ell, self.b_ell_TplusE))
self.assertFalse(np.shares_memory(data.n_ell, self.n_ell_TplusE))
# Check initialized attributes.
self.assertEqual(data.pol, ('T', 'E'))
self.assertEqual(data.lmax, self.lmax)
self.assertEqual(data.npol, 2)
self.assertTrue(isinstance(data.cosmology, self.FakeCosmology))
np.testing.assert_almost_equal(data.b_ell, self.b_ell_TplusE)
np.testing.assert_almost_equal(data.n_ell, self.n_ell_TplusE)
# Test shapes.
self.assertEqual(data.b_ell.shape, (2, self.lmax + 1))
self.assertEqual(data.n_ell.shape, (3, self.lmax + 1))
self.assertEqual(data.cov_ell_lensed.shape, (3, self.lmax + 1))
self.assertEqual(data.cov_ell_nonlensed.shape, (3, self.lmax + 1))
self.assertEqual(data.icov_ell_lensed.shape, (3, self.lmax + 1))
self.assertEqual(data.icov_ell_nonlensed.shape, (3, self.lmax + 1))
def test_data_n_is_totcov_I(self):
pol = ['T']
cosmo = self.FakeCosmology()
n_ell_T = self.n_ell_T + 1 # To make well conditioned.
data = Data(self.lmax,
n_ell_T,
self.b_ell_T,
pol,
cosmo,
n_is_totcov=True)
cov_ell_exp = np.zeros((1, self.nell))
cov_ell_exp[0] = n_ell_T
np.testing.assert_almost_equal(data.cov_ell_nonlensed, cov_ell_exp)
# Test if inverse is also computed correctly.
np.testing.assert_almost_equal(data.icov_ell_nonlensed,
1 / cov_ell_exp)
# Again with lensing power. This should be the same.
cov_ell_exp[0] = n_ell_T
np.testing.assert_almost_equal(data.cov_ell_lensed, cov_ell_exp)
# Test if inverse is also computed correctly.
np.testing.assert_almost_equal(data.icov_ell_lensed, 1 / cov_ell_exp)
def test_data_icov_diag_nonlensed_1d_T(self):
cosmo = self.FakeCosmology()
pol = ['T']
data = Data(self.lmax, self.n_ell_T, self.b_ell_T, pol, cosmo)
alm = np.random.randn(hp.Alm.getsize(self.lmax))
alm = np.reshape(alm.astype(complex), (1, self.nelem))
alm_in = alm.copy()
alm = data.icov_diag_nonlensed(alm)
# Test if 1D input also works
alm_1d = data.icov_diag_nonlensed(alm_in[0].copy())
self.assertTrue(alm_1d.ndim == 1)
np.testing.assert_array_almost_equal(alm_1d, alm[0])
alm_exp = np.zeros_like(alm_in)
alm_exp[0] = hp.almxfl(alm_in[0], data.icov_ell_nonlensed[0])
np.testing.assert_array_almost_equal(alm, alm_exp)
def test_data_compute_alm_sim_2d(self):
cosmo = self.FakeCosmology()
pol = ['T', 'E']
data = Data(self.lmax, self.n_ell_TplusE, self.b_ell_TplusE, pol,
cosmo)
np.random.seed(10)
alm = data.compute_alm_sim(lens_power=False)
np.random.seed(10)
cl = np.zeros((4, self.nell))
cl[:2] = data.cov_ell_nonlensed[:2]
cl[3] = data.cov_ell_nonlensed[2]
alm_exp = hp.synalm(cl, new=True)
alm_sim_expec = np.zeros((2, self.nelem), dtype=complex)
alm_sim_expec[0] = alm_exp[0]
alm_sim_expec[1] = alm_exp[1]
np.testing.assert_almost_equal(alm, alm_sim_expec)
# Again for lensed.
np.random.seed(10)
alm = data.compute_alm_sim(lens_power=True)
np.random.seed(10)
cl = np.zeros((4, self.nell))
cl[:2] = data.cov_ell_lensed[:2]
cl[3] = data.cov_ell_lensed[2]
alm_exp = hp.synalm(cl, new=True)
alm_sim_expec = np.zeros((2, self.nelem), dtype=complex)
alm_sim_expec[0] = alm_exp[0]
alm_sim_expec[1] = alm_exp[1]
np.testing.assert_almost_equal(alm, alm_sim_expec)
def test_data_compute_totcov_diag_2d(self):
cosmo = self.FakeCosmology()
pol = ['T', 'E']
data = Data(self.lmax, self.n_ell_TplusE, self.b_ell_TplusE, pol,
cosmo)
expec_totcov = np.zeros((3, self.nell))
expec_totcov[0] = self.n_ell_TplusE[0] + self.c_ell[:,0] * 2 \
* self.b_ell_TplusE[0] ** 2
expec_totcov[1] = self.n_ell_TplusE[1] + self.c_ell[:,1] * 2 \
* self.b_ell_TplusE[1] ** 2
expec_totcov[2] = self.n_ell_TplusE[2] + self.c_ell[:,3] * 2 \
* self.b_ell_TplusE[0] * self.b_ell_TplusE[1]
np.testing.assert_almost_equal(data.cov_ell_nonlensed, expec_totcov)
# Test if inverse is also computed correctly.
expec_inv_totcov = np.zeros_like(expec_totcov)
# Inverse of [[a,b],[b,d]] = [[d,-b],[-b,a]] / (ad - bb).
expec_inv_totcov[0] = expec_totcov[1]
expec_inv_totcov[1] = expec_totcov[0]
expec_inv_totcov[2] = -expec_totcov[2]
det = expec_totcov[0] * expec_totcov[1] - expec_totcov[2]**2
expec_inv_totcov /= det
np.testing.assert_almost_equal(data.icov_ell_nonlensed,
expec_inv_totcov)
# With lensing power.
expec_totcov = np.zeros((3, self.nell))
expec_totcov[0] = self.n_ell_TplusE[0] + self.c_ell[:,0] \
* self.b_ell_TplusE[0] ** 2
expec_totcov[1] = self.n_ell_TplusE[1] + self.c_ell[:,1] \
* self.b_ell_TplusE[1] ** 2
expec_totcov[2] = self.n_ell_TplusE[2] + self.c_ell[:,3] \
* self.b_ell_TplusE[0] * self.b_ell_TplusE[1]
np.testing.assert_almost_equal(data.cov_ell_lensed, expec_totcov)
expec_inv_totcov = np.zeros_like(expec_totcov)
expec_inv_totcov[0] = expec_totcov[1]
expec_inv_totcov[1] = expec_totcov[0]
expec_inv_totcov[2] = -expec_totcov[2]
det = expec_totcov[0] * expec_totcov[1] - expec_totcov[2]**2
expec_inv_totcov /= det
np.testing.assert_almost_equal(data.icov_ell_lensed, expec_inv_totcov)
def test_data_n_is_totcov_pol(self):
cosmo = self.FakeCosmology()
pol = ['T', 'E']
n_ell_TplusE = self.n_ell_TplusE + 1
data = Data(self.lmax,
n_ell_TplusE,
self.b_ell_TplusE,
pol,
cosmo,
n_is_totcov=True)
expec_totcov = np.zeros((3, self.nell))
expec_totcov[0] = n_ell_TplusE[0]
expec_totcov[1] = n_ell_TplusE[1]
expec_totcov[2] = n_ell_TplusE[2]
np.testing.assert_almost_equal(data.cov_ell_nonlensed, expec_totcov)
# Test if inverse is also computed correctly.
expec_inv_totcov = np.zeros_like(expec_totcov)
# Inverse of [[a,b],[b,d]] = [[d,-b],[-b,a]] / (ad - bb).
expec_inv_totcov[0] = expec_totcov[1]
expec_inv_totcov[1] = expec_totcov[0]
expec_inv_totcov[2] = -expec_totcov[2]
det = expec_totcov[0] * expec_totcov[1] - expec_totcov[2]**2
expec_inv_totcov /= det
np.testing.assert_almost_equal(data.icov_ell_nonlensed,
expec_inv_totcov)
# With lensing power.
expec_totcov = np.zeros((3, self.nell))
expec_totcov[0] = n_ell_TplusE[0]
expec_totcov[1] = n_ell_TplusE[1]
expec_totcov[2] = n_ell_TplusE[2]
np.testing.assert_almost_equal(data.cov_ell_lensed, expec_totcov)
expec_inv_totcov = np.zeros_like(expec_totcov)
expec_inv_totcov[0] = expec_totcov[1]
expec_inv_totcov[1] = expec_totcov[0]
expec_inv_totcov[2] = -expec_totcov[2]
det = expec_totcov[0] * expec_totcov[1] - expec_totcov[2]**2
expec_inv_totcov /= det
np.testing.assert_almost_equal(data.icov_ell_lensed, expec_inv_totcov)
def test_data_compute_totcov_diag_1d_E(self):
cosmo = self.FakeCosmology()
pol = ['E']
data = Data(self.lmax, self.n_ell_T, self.b_ell_T, pol, cosmo)
cov_ell_exp = np.zeros((1, self.nell))
cov_ell_exp[0] = self.n_ell_T + self.c_ell[:, 1] * 2 * self.b_ell_T**2
np.testing.assert_almost_equal(data.cov_ell_nonlensed, cov_ell_exp)
# Test if inverse is also computed correctly.
np.testing.assert_almost_equal(data.icov_ell_nonlensed,
1 / cov_ell_exp)
# Again with lensing power.
cov_ell_exp[0] = self.n_ell_T + self.c_ell[:, 1] * self.b_ell_T**2
np.testing.assert_almost_equal(data.cov_ell_lensed, cov_ell_exp)
# Test if inverse is also computed correctly.
np.testing.assert_almost_equal(data.icov_ell_lensed, 1 / cov_ell_exp)