Reading in table saved from GrainPop object does not return the same outputs

[eblur]

Example:

from newdust import make_MRN

ENERGY = np.arange(0.1, 6.1, 0.2) # keV
THETA = np.logspace(np.log10(0.5), 3, 20) # arcsec

mrn = make_MRN(amax=0.5, na=50)
mrn.calculate_ext(ENERGY, unit="kev", theta=THETA)

for name in mrn.keys:
    mrn[name].write_extinction_table("%s_10eV.fits" % (name))

new_silpop = SingleGrainPop(mrn['sil'].size, mrn['sil'].comp, 'sil10eV', md=mrn['sil'].md, scatm_from_file="sil_10eV.fits")

print(np.sum(new_silpop.tau_ext - mrn['sil'].tau_ext))

yields a non-zero value (170.12701232539766)

Why?

[eblur]

This may have to do with the nature of python, since it uses pointers. The old population changes after loading the new population.

Loading the table model in this way worked:

new_silpop = SingleGrainPop('Powerlaw', 'Silicate', 'sil10eV', 
    md=np.copy(mrn['sil'].md), amax=0.5, na=50, 
    scatm_from_file="sil_10eV.fits")

[eblur]

I want to do something internally with the code to make this workflow less stilted. (In this case, one has to know what parameters were used to run the original table model, and then initialize the new grain population with those old parameters.)

Two options:

• Use a scattering model that interpolates over the model (Issue Write a sub-class of ScatModel that will interpolate over table information #17)
• Load size.a from the scat model table file ScatModel.pars['a'] -- ndens properties should naturally follow.