def test_Ma_0():
"""Check expected values for Ma = 0."""
val0 = np.array([1., 1., 1., 0., 0., 0., np.nan, np.nan])
Y0 = {var_test_sup[i]: val0[i] for i in range(len(var_test_sup))}
for v in var_test_sup:
print(v, Y0[v], cf.from_Ma( v, np.atleast_1d(0.), ga))
assert np.isclose( cf.from_Ma( v, np.atleast_1d(0.), ga), Y0[v],
atol=1e-7, equal_nan=True)
if np.isfinite(Y0[v]):
assert np.isclose( cf.to_Ma( v, np.atleast_1d(Y0[v]), ga), 0.0,
atol=1e-7, equal_nan=True)
def test_Ma_inf():
"""Check expected values for Ma = i."""
val_inf = np.array([np.inf, np.inf, np.inf, np.sqrt(2.), 0.,
np.inf, np.sqrt((ga - 1.)/ 2. / ga), 0.])
Y0 = {var_test_sup[i]: val_inf[i] for i in range(len(var_test_sup))}
for v in var_test_sup:
print(v)
assert np.isclose( cf.from_Ma( v, np.atleast_1d(np.inf), ga), Y0[v], atol=1e-7)
def test_explicit_sup():
"""Compare quantity values to CUED Data Book, supersonic."""
for v in var_test_sup:
for _ in range(Nrep):
Ysup = cf.from_Ma(v, dat_sup['Ma'], ga)
err_sup = np.abs(Ysup / dat_sup[v] - 1.)
imax = np.argmax(err_sup)
assert err_sup[imax] < 0.005, \
"Ma={0} => {1}={2} with error={3}".format(
dat_sup['Ma'][imax], v, Ysup[imax], err_sup[imax])
def test_Ma_1():
"""Check inversion for values near Ma = 1."""
dMa = 0.0001
Ma_fw = np.array([-dMa, 0., dMa]) + 1.
for v in var_test_sup:
Y_fw = cf.from_Ma( v, Ma_fw, ga)
Ma_bk = cf.to_Ma( v, Y_fw, ga)
if v in ['mcpTo_APo','A_Acrit']:
Ma_bk[-1] = cf.to_Ma( v, Y_fw[-1], ga, True)
err = Ma_fw-Ma_bk
print(v, err)
assert np.all(np.abs(err[~np.isnan(Ma_bk)])<1e-5)
def test_derivative():
"""Check explicit derivative against a numerical approximation."""
for v in var_test_sup:
Ma = np.linspace(0., 3.)
# Explicit
dYdMa = cf.derivative_from_Ma(v, Ma, ga)
# Numerical approximation
Y = cf.from_Ma(v, Ma, ga)
dYdMa_numerical = np.gradient(Y, Ma)
# Check discrepancy
err = dYdMa / dYdMa_numerical - 1.
print(v)
assert np.max(err[np.isfinite(err)]) < 0.05