def test_evaluate(self, temperature):
bolometric_flux = 1000 * u.L_sun / (4 * np.pi * (1.5 * u.pc)**2)
b = BlackBody1D(temperature=temperature,
bolometric_flux=bolometric_flux)
assert_quantity_allclose(b(1.4 * u.micron), 4734463.93280426 * u.Jy)
assert_quantity_allclose(b(214.13747 * u.THz), 4734463.93280426 * u.Jy)
def test_fit(self):
fitter = LevMarLSQFitter()
b = BlackBody1D(3000 * u.K)
wav = np.array([0.5, 5, 10]) * u.micron
fnu = np.array([1, 10, 5]) * u.Jy
b_fit = fitter(b, wav, fnu)
assert_quantity_allclose(b_fit.temperature, 2840.7438339457754 * u.K)
assert_quantity_allclose(b_fit.bolometric_flux, 6.821837075583734e-08 * u.erg / u.cm**2 / u.s)
def func(params, return_model=False):
bb = BlackBody1D(params[3] * u.K)(wav).to(FLAM,
u.spectral_density(wav))
bb /= np.trapz(bb, wav)
sens = sens_raw * bb.value
sens /= np.nanmedian(sens)
model = params[2] * np.interp(
x * params[0] + params[1] + meanw,
wav.value,
sens,
)
if return_model:
return model
return np.nansum((data - model)**2) / (np.isfinite(model).sum())
def wavelength_calibrate(visit):
hdu = fits.open(CALIPATH + f"WFC3.IR.{visit.filter}.1st.sens.2.fits")
x = (np.arange(visit.trace.shape[0]) / visit.trace.shape[0]) - 0.5
data = np.copy(visit.trace) / np.median(visit.trace)
cent = np.average(np.arange(data.shape[0]), weights=data)
sens_raw = hdu[1].data["SENSITIVITY"]
wav = hdu[1].data["WAVELENGTH"] * u.angstrom
bb = BlackBody1D(visit.st_teff * u.K)(wav).to(FLAM,
u.spectral_density(wav))
bb /= np.trapz(bb, wav)
sens = sens_raw * bb.value
sens /= np.nanmedian(sens)
if visit.filter == "G141":
dw = 17000 - 10500
meanw = (10500 + 17000) / 2
else:
dw = 11500 - 7700
meanw = (11500 + 7700) / 2
x = x * dw
def func(params, return_model=False):
bb = BlackBody1D(params[3] * u.K)(wav).to(FLAM,
u.spectral_density(wav))
bb /= np.trapz(bb, wav)
sens = sens_raw * bb.value
sens /= np.nanmedian(sens)
model = params[2] * np.interp(
x * params[0] + params[1] + meanw,
wav.value,
sens,
)
if return_model:
return model
return np.nansum((data - model)**2) / (np.isfinite(model).sum())
r = minimize(
func,
[1, 0, 1, visit.st_teff],
method="Powell",
bounds=[
(0.1, 2),
(-1000, 1000),
(0.4, 3),
(visit.st_teff - 2000, visit.st_teff + 2000),
],
)
wavelength = (x * r.x[0] + r.x[1] + meanw) * u.angstrom
sensitivity_t = func(r.x, True)
sensitivity = np.interp(
x * r.x[0] + r.x[1] + meanw,
wav.value,
sens_raw,
)
sensitivity_raw = sensitivity.copy()
sensitivity /= np.median(sensitivity)
return (
wavelength[~visit.trace_mask],
sensitivity[~visit.trace_mask],
sensitivity_t[~visit.trace_mask],
sensitivity_raw[~visit.trace_mask],
)