def test_scf_compute_nbody_twopowertriaxial():
N = int(1e5)
Mh = 11.
ah = 50. / 8.
m = Mh / N
yfactor = 1.5
zfactor = 2.5
nsamp = 10
Norder = 10
Lorder = 10
hern = potential.HernquistPotential(amp=2 * Mh, a=ah)
hern.turn_physical_off()
hdf = df.isotropicHernquistdf(hern)
numpy.random.seed(2)
samp = [hdf.sample(n=N) for i in range(nsamp)]
positions = numpy.array(
[[samp[i].x(), samp[i].y() * yfactor, samp[i].z() * zfactor]
for i in range(nsamp)])
# This is an triaxial Hernquist profile with the same mass as the above
tptp = potential.TwoPowerTriaxialPotential(amp=2. * Mh / yfactor / zfactor,
a=ah,
alpha=1.,
beta=4.,
b=yfactor,
c=zfactor)
tptp.turn_physical_off()
cc, ss = potential.scf_compute_coeffs(tptp.dens, Norder, Lorder, a=ah)
c, s = numpy.zeros((2, nsamp, Norder, Lorder, Lorder))
for i, p in enumerate(positions):
c[i], s[i] = potential.scf_compute_coeffs_nbody(p,
Norder,
Lorder,
mass=m * numpy.ones(N),
a=ah)
# Check that the difference between the coefficients is within two standard deviations
assert (cc - (numpy.mean(c, axis=0)) <= (2. * numpy.std(c, axis=0))).all()
# Repeat test for single mass
c, s = numpy.zeros((2, nsamp, Norder, Lorder, Lorder))
for i, p in enumerate(positions):
c[i], s[i] = potential.scf_compute_coeffs_nbody(p,
Norder,
Lorder,
mass=m,
a=ah)
assert (cc - (numpy.mean(c, axis=0)) <= (2. * numpy.std(c, axis=0))).all()
return None
def compare_dens_nbody():
N = int(1e5)
Mh = 11.
ah = 50. / 8.
m = Mh / N
factor = 1.
nsamp = 10
Norder = 10
Lorder = 10
hern = potential.HernquistPotential(amp=2 * Mh, a=ah)
hern.turn_physical_off()
hdf = df.isotropicHernquistdf(hern)
samp = [hdf.sample(n=N) for i in range(nsamp)]
positions = numpy.array([[samp[i].x(), samp[i].y(), samp[i].z() * factor]
for i in range(nsamp)])
tptp = potential.TwoPowerTriaxialPotential(amp=2. * Mh,
a=ah,
alpha=1.,
beta=4.,
b=1.,
c=factor)
tptp.turn_physical_off()
cc, ss = potential.scf_compute_coeffs(tptp.dens, Norder, Lorder, a=ah)
c, s = numpy.zeros((2, nsamp, Norder, Lorder, Lorder))
for i, p in enumerate(positions):
c[i], s[i] = potential.scf_compute_coeffs_nbody(p,
m * numpy.ones(N) *
factor,
Norder,
Lorder,
a=ah)
# Check that the difference between the coefficients is within two standard deviations
assert (cc - (numpy.mean(c, axis=0)) <= (2. * numpy.std(c, axis=0))).all()