def test_init_with_stddev_mean_amplitude(self):
gauss = models.GaussianModel1D(amplitude=1., mean=1.5, stddev=0.5)
# check variables are correct
self.assertEqual(gauss.amplitude, 1.)
self.assertEqual(gauss.mean, 1.5)
self.assertEqual(gauss.stddev, 0.5)
# check the model works on a couple of points (evaluating astropy code here)
self.assertEqual(gauss(1.5), 1.)
self.assertAlmostEqual(gauss(1.75), 0.8825, 4)
def test_changing_stddev_recalculates_amp_from_flux(self):
""" The reason for this test is if you initialise a gaussian with flux, then change the stddev after you may
expect flux to be preserved. It is not (as amplitude is calculated from flux when set, not on the fly). Im
unsure if i should change this, raise a warning or leave it. For now im writing this test that will fail. And
raising an issue on github (issue #1)
"""
gauss = models.GaussianModel1D(mean=1.5, flux=1., stddev=0.5)
gauss.stddev = 0.6
self.assertEqual(gauss.flux, 1.)
def test_init_with_fwhm_mean_flux(self):
gauss = models.GaussianModel1D(flux=1., mean=1.5, fwhm=0.5)
# check variables are correct
self.assertEqual(gauss.flux, 1.)
self.assertAlmostEqual(gauss.amplitude, 1.879, 3)
self.assertEqual(gauss.mean, 1.5)
self.assertEqual(gauss.fwhm, 0.5)
self.assertAlmostEqual(gauss.stddev, 0.2123, 4)
# check the model works on a couple of points (evaluating astropy code here)
self.assertAlmostEqual(gauss(1.5), 1.879, 3)
self.assertAlmostEqual(gauss(1.75), 0.9394, 4)
def test_init_fails_with_stdev_mean_amplitude_flux(self):
with self.assertRaises(ValueError):
models.GaussianModel1D(amplitude=1., mean=1.5, flux=1., stddev=0.5)
def test_init_raises_Valueerror_with_stdev_fwhm_mean_amplitude(self):
with self.assertRaises(ValueError):
models.GaussianModel1D(amplitude=1.,
mean=1.5,
fwhm=0.5,
stddev=0.5)
def test_integrate(self):
gauss = models.GaussianModel1D(mean=5, flux=3, stddev=0.5)
np.testing.assert_almost_equal(
gauss.integrate(np.array([4.3, 4.7, 5.2, 6.])),
np.array([0.5805, 1.1435, 0.9655]), 4)