|
|
|
@ -6,7 +6,7 @@ from numpy import cos, pi
|
|
|
|
|
import numpy.testing as npt
|
|
|
|
|
from numpy.testing import assert_array_almost_equal
|
|
|
|
|
from wafo.padua import (padua_points, example_functions, padua_fit,
|
|
|
|
|
padua_fit2,
|
|
|
|
|
# padua_fit2,
|
|
|
|
|
padua_cubature, padua_val)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
@ -40,17 +40,19 @@ class PaduaTestCase(unittest.TestCase):
|
|
|
|
|
|
|
|
|
|
def test_padua_fit_even_degree(self):
|
|
|
|
|
points = padua_points(10)
|
|
|
|
|
C0f, _abs_error = padua_fit(points, example_functions, 6)
|
|
|
|
|
C0f, abs_error = padua_fit(points, example_functions, 6)
|
|
|
|
|
expected = np.zeros((11, 11))
|
|
|
|
|
expected[0, 0] = 1
|
|
|
|
|
assert_array_almost_equal(C0f, expected, 15)
|
|
|
|
|
assert_array_almost_equal(abs_error, 1.2168216554799264e-15)
|
|
|
|
|
|
|
|
|
|
def test_padua_fit_odd_degree(self):
|
|
|
|
|
points = padua_points(9)
|
|
|
|
|
C0f, _abs_error = padua_fit(points, example_functions, 6)
|
|
|
|
|
C0f, abs_error = padua_fit(points, example_functions, 6)
|
|
|
|
|
expected = np.zeros((10, 10))
|
|
|
|
|
expected[0, 0] = 1
|
|
|
|
|
assert_array_almost_equal(C0f, expected, 15)
|
|
|
|
|
assert_array_almost_equal(abs_error, 4.509537093983535e-17)
|
|
|
|
|
|
|
|
|
|
# TODO: padua_fit2 does not work correctly
|
|
|
|
|
# def test_padua_fit_odd_degree2(self):
|
|
|
|
@ -63,33 +65,36 @@ class PaduaTestCase(unittest.TestCase):
|
|
|
|
|
def test_padua_cubature(self):
|
|
|
|
|
domain = [0, 1, 0, 1]
|
|
|
|
|
points = padua_points(500, domain)
|
|
|
|
|
C0f, _abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
C0f, abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
val = padua_cubature(C0f, domain)
|
|
|
|
|
expected = 4.06969589491556e-01
|
|
|
|
|
assert_array_almost_equal(val, expected, 15)
|
|
|
|
|
assert_array_almost_equal(abs_error, 3.66470417665e-16)
|
|
|
|
|
|
|
|
|
|
def test_padua_val_unordered(self):
|
|
|
|
|
domain = [0, 1, 0, 1]
|
|
|
|
|
points = padua_points(20, domain)
|
|
|
|
|
C0f, _abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
X = [0, 0.5, 1]
|
|
|
|
|
|
|
|
|
|
C0f, abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
X = np.array([0, 0.5, 1])
|
|
|
|
|
# true_val = example_functions.franke(X, X)
|
|
|
|
|
val = padua_val(X, X, C0f, domain)
|
|
|
|
|
expected = [7.664205912849228e-01, 3.2621734202884815e-01,
|
|
|
|
|
3.587865112678535e-02]
|
|
|
|
|
expected = [0.76642059128493, 0.32621734202885, 0.03587865112678]
|
|
|
|
|
assert_array_almost_equal(val, expected, 14)
|
|
|
|
|
assert_array_almost_equal(abs_error, 0.003897032262116954)
|
|
|
|
|
|
|
|
|
|
def test_padua_val_grid(self):
|
|
|
|
|
domain = [0, 1, 0, 1]
|
|
|
|
|
a, b, c, d = domain
|
|
|
|
|
points = padua_points(21, domain)
|
|
|
|
|
C0f, _abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
C0f, abs_error = padua_fit(points, example_functions, 0)
|
|
|
|
|
X1 = np.linspace(a, b, 2)
|
|
|
|
|
X2 = np.linspace(c, d, 2)
|
|
|
|
|
val = padua_val(X1, X2, C0f, domain, use_meshgrid=True)
|
|
|
|
|
expected = [[7.664205912849229e-01, 1.0757071952145181e-01],
|
|
|
|
|
[2.703371615911344e-01, 3.5734971024838565e-02]]
|
|
|
|
|
|
|
|
|
|
expected = [[0.76642059128493, 0.10757071952145],
|
|
|
|
|
[0.27033716159114, 0.03573497102484]]
|
|
|
|
|
assert_array_almost_equal(val, expected, 14)
|
|
|
|
|
assert_array_almost_equal(abs_error, 0.0022486904061664046)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
|