1994I throws nan for specphot

[RobFirth]

As title, problem occurs when running get_specphot such as in scripts/extend_all_spectra.py

[RobFirth]

I think this is linked to #24 and has now become pretty critical to figure out. It seems like it is distorting the integrated flux of all functions that use kcorr.calc_spectrum_filter_flux which underpins lots of stuff, including zeropoints and the mangling.

I get the following output from the code below:

SN2006aj_53786.270000
Filter_area = 958.904702526
Integrated flux = 65.3955702281
SN2006aj_53787.230000
Filter_area = 583.185950131
Integrated flux = 41.0298233406
SN2006aj_53788.200000
Filter_area = 1358.72907164
Integrated flux = 131.886385781
SN2006aj_53789.170000
Filter_area = 1357.72217486
Integrated flux = 146.521717503
SN2006aj_53790.140000
Filter_area = 1361.80653493
Integrated flux = 164.34213622
SN2006aj_53791.100000
Filter_area = 1356.94789874
Integrated flux = 163.802182486
SN2006aj_53792.070000
Filter_area = 1358.08475139
Integrated flux = 170.17603978
SN2006aj_53793.040000
Filter_area = 1342.63870674
Integrated flux = 156.866851118
SN2006aj_53794.010000
Filter_area = 1358.55873129
Integrated flux = 160.605022385
SN2006aj_53794.970000
Filter_area = 494.406955809
Integrated flux = 64.3739416067
SN2006aj_53795.940000
Filter_area = 1315.20012487
Integrated flux = 137.109302356
SN2006aj_53796.910000
Filter_area = 1313.60589901
Integrated flux = 121.107583058
SN2006aj_53797.880000
Filter_area = 1313.47758368
Integrated flux = 110.196456176
SN2006aj_53798.840000
Filter_area = 1315.40072684
Integrated flux = 92.2847571523
SN2006aj_53799.810000
Filter_area = 1314.80756892
Integrated flux = 91.8528874753
SN2006aj_53801.750000
Filter_area = 1298.16443106
Integrated flux = 63.0384962734
SN2006aj_53802.720000
Filter_area = 1299.82908919
Integrated flux = 45.3544694847
SN2006aj_53803.680000
Filter_area = 1296.91402526
Integrated flux = 37.1949955268
SN2006aj_53803.680001
Filter_area = 1296.93821395
Integrated flux = 51.5059756933

Notice something weird - the filter area is changing. The filter area shouldn't change. Suspect that it might be to do with the resampling of the filter response, but the fluctuations are severe - around 1300 for most, but as low as 494 for SN2006aj_53794.970000.

snname="SN2006aj"

sn = pcc.classes.SNClass(snname)
sn.load_phot(verbose=False)
sn.get_lcfit(os.path.join(pcc.defaults._default_recon_dir_path, snname + ".dat"))

sn.plot_lc(multiplot=False)

sn.load_list(path=os.path.join(pcc.defaults._default_list_dir_path, sn.name+".list"))
sn.load_spec()

# sn.load_mangledspec()
sn.load_reconspec()
sn.load_simspec()


path_to_filter = os.path.join(pcc.defaults._default_filter_dir_path,"BessellB.dat")
BessellB = pcc.classes.FilterClass()
BessellB.read_filter_file(path_to_filter)
BessellB.calculate_AB_zp()

sim_spec_flux = []
sim_mjd_obs = []

for spec in sn.sim_spec:
    print(spec)
    flux = pcc.kcorr.calc_spectrum_filter_flux(filter_object=BessellB, spectrum_object=sn.sim_spec[spec], 
                                               correct_for_area=False)
    mjd_obs = float(spec.split("_")[-1])
    
    sim_spec_flux.append(flux)
    sim_mjd_obs.append(mjd_obs)
    

[RobFirth]

After running the above, the filter object that has been loaded in is totally messed up:

[RobFirth]

For reference, it should look like this:

[RobFirth]

Recon spec for SN2006aj before:

and after bugfix:

There are obviously 2 discrepant spectra - I think this is a separate issue. Closing.