test_nft.py 2.41 KB
Newer Older
Philipp Arras's avatar
Philipp Arras committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
# Copyright(C) 2019 Max-Planck-Society
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik.

import numpy as np
import pytest
Martin Reinecke's avatar
Martin Reinecke committed
20
from numpy.testing import assert_allclose, assert_
Philipp Arras's avatar
Philipp Arras committed
21
22
23
24
25
26
27

import nifty5 as ift

np.random.seed(40)

pmp = pytest.mark.parametrize

Martin Reinecke's avatar
Martin Reinecke committed
28

Martin Reinecke's avatar
Martin Reinecke committed
29
def _l2error(a, b):
Martin Reinecke's avatar
fixes    
Martin Reinecke committed
30
    return np.sqrt(np.sum(np.abs(a-b)**2)/np.sum(np.abs(a)**2))
Philipp Arras's avatar
Philipp Arras committed
31

Martin Reinecke's avatar
fixes    
Martin Reinecke committed
32

Martin Reinecke's avatar
Martin Reinecke committed
33
@pmp('eps', [1e-2, 1e-4, 1e-7, 1e-10, 1e-11, 1e-12, 2e-13])
Philipp Arras's avatar
Philipp Arras committed
34
35
36
@pmp('nu', [12, 128])
@pmp('nv', [4, 12, 128])
@pmp('N', [1, 10, 100])
Martin Reinecke's avatar
fixes    
Martin Reinecke committed
37
def test_gridding(nu, nv, N, eps):
38
    uvw = np.random.rand(N, 3) - 0.5
Martin Reinecke's avatar
fixes    
Martin Reinecke committed
39
    vis = (np.random.randn(N) + 1j*np.random.randn(N))
Philipp Arras's avatar
Philipp Arras committed
40
41

    # Nifty
Martin Reinecke's avatar
adjust    
Martin Reinecke committed
42
43
    GM = ift.GridderMaker(ift.RGSpace((nu, nv)), uvw=uvw,
                          channel_fact=np.array([1.]), eps=eps)
Martin Reinecke's avatar
fix    
Martin Reinecke committed
44
    vis2 = ift.from_global_data(ift.UnstructuredDomain(vis.shape), vis)
Philipp Arras's avatar
Philipp Arras committed
45

Martin Reinecke's avatar
Martin Reinecke committed
46
    Op = GM.getFull()
Martin Reinecke's avatar
fix    
Martin Reinecke committed
47
    pynu = Op(vis2).to_global_data()
Philipp Arras's avatar
Philipp Arras committed
48
49
50
51
52
    # DFT
    x, y = np.meshgrid(
        *[-ss/2 + np.arange(ss) for ss in [nu, nv]], indexing='ij')
    dft = pynu*0.
    for i in range(N):
53
        dft += (vis[i]*np.exp(2j*np.pi*(x*uvw[i, 0] + y*uvw[i, 1]))).real
Martin Reinecke's avatar
Martin Reinecke committed
54
    assert_(_l2error(dft, pynu) < eps)
Philipp Arras's avatar
Philipp Arras committed
55
56


Martin Reinecke's avatar
Martin Reinecke committed
57
@pmp('eps', [1e-2, 1e-6, 2e-13])
Philipp Arras's avatar
Philipp Arras committed
58
59
60
61
62
@pmp('nu', [12, 128])
@pmp('nv', [4, 12, 128])
@pmp('N', [1, 10, 100])
def test_build(nu, nv, N, eps):
    dom = ift.RGSpace([nu, nv])
63
64
    uvw = np.random.rand(N, 3) - 0.5
    GM = ift.GridderMaker(dom, uvw=uvw, channel_fact=np.array([1.]), eps=eps)
Martin Reinecke's avatar
Martin Reinecke committed
65
    R0 = GM.getGridder()
Philipp Arras's avatar
Philipp Arras committed
66
67
    R1 = GM.getRest()
    R = R1@R0
Martin Reinecke's avatar
Martin Reinecke committed
68
    RF = GM.getFull()
Philipp Arras's avatar
Philipp Arras committed
69
70

    # Consistency checks
71
72
73
74
75
76
    flt = np.float64
    cmplx = np.complex128
    ift.extra.consistency_check(R0, cmplx, flt, only_r_linear=True)
    ift.extra.consistency_check(R1, flt, flt)
    ift.extra.consistency_check(R, cmplx, flt, only_r_linear=True)
    ift.extra.consistency_check(RF, cmplx, flt, only_r_linear=True)