

In [1]:

    
%matplotlib inline
import matplotlib.pyplot as plt
# import seaborn as sns
# sns.set()



In [2]:

    
import sncosmo



In [3]:

    
import analyzeSN as ans

Load the SN from an SNANA simulation

This should work with all the supernovae in the simulation, but to make this example quick, let us just load the first one and represent the collection of SN.



In [6]:

    
# location is relative or absolute path to the file
sne = ans.snanaSims.SnanaSims.fromSNANAfileroot('snana_fits', location=ans.example_data, n=1)









    



/Users/rbiswas/.local/lib/python2.7/site-packages/analyzeSN/example_data/snana_fits_HEAD.FITS

Pick out the very first SN and reformat it to look the way sncosmo expects SN to look



In [7]:

    
sn = ans.snanaSims.SnanaSims.reformat_SNANASN(sne.snList[0])

Make sure SNCosmo can understand band names



In [8]:

    
ans.snanaSims.SnanaSims.matchSNANAbandnamesinregistry()

We can look at the information in the light curve, or at the metadata



In [9]:

    
sn[:5]









    Out[9]:




<Table length=5>

MJD band FIELD TELESCOPE PHOTFLAG PHOTPROB flux fluxerr MAG MAGERR PSF_SIG1 PSF_SIG2 PSF_RATIO SKY_SIG SKY_SIG_T RDNOISE ZEROPT ZEROPT_ERR GAIN ZP ZPSYS
float64 str2 str12 str20 int32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float32 float64 str2
52881.5 i NULL ALL 0 0.0 13.44 2.846 24.679 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 27.5 ab
52881.539 r NULL ALL 0 0.0 7.74 1.701 25.278 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 27.5 ab
52881.559 z NULL ALL 0 0.0 19.72 6.539 24.263 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 27.5 ab
52886.602 i NULL ALL 0 0.0 44.52 2.909 23.379 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 27.5 ab
52900.531 i NULL ALL 0 0.0 86.57 2.914 22.657 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 27.5 ab



In [10]:

    
sn.meta









    Out[10]:





OrderedDict([('SNID', '03D1aw'),
             ('IAUC', 'UNKNOWN '),
             ('FAKE', 0),
             ('RA', 36.061607000000002),
             ('DECL', -4.5171140000000003),
             ('PIXSIZE', -9.0),
             ('NXPIX', 0),
             ('NYPIX', -9),
             ('SNTYPE', 0),
             ('NOBS', 48),
             ('PTROBS_MIN', 1),
             ('PTROBS_MAX', 48),
             ('MWEBV', 0.025),
             ('MWEBV_ERR', 0.0040000002),
             ('REDSHIFT_HELIO', -9.0),
             ('REDSHIFT_HELIO_ERR', 0.0),
             ('REDSHIFT_FINAL', 0.58170003),
             ('REDSHIFT_FINAL_ERR', 0.001),
             ('HOSTGAL_OBJID', 0),
             ('HOSTGAL_PHOTOZ', -9.0),
             ('HOSTGAL_PHOTOZ_ERR', -9.0),
             ('HOSTGAL_SPECZ', -9.0),
             ('HOSTGAL_SPECZ_ERR', -9.0),
             ('HOSTGAL_SNSEP', -9.0),
             ('HOSTGAL_LOGMASS', -9.0),
             ('HOSTGAL_LOGMASS_ERR', -9.0),
             ('HOSTGAL_MAG_g', 0.0),
             ('HOSTGAL_MAG_r', 0.0),
             ('HOSTGAL_MAG_i', 0.0),
             ('HOSTGAL_MAG_z', 0.0),
             ('HOSTGAL_SB_FLUXCAL_g', 0.0),
             ('HOSTGAL_SB_FLUXCAL_r', 0.0),
             ('HOSTGAL_SB_FLUXCAL_i', 0.0),
             ('HOSTGAL_SB_FLUXCAL_z', 0.0),
             ('PEAKMJD', 52900.422),
             ('SEARCH_TYPE', 0)])

Fitting SN to models

Instantiate a model and fit Supernovae to it. Since this is a very short demonstration for how to use the SNANA simulations rather than how to use SNCosmo correctly, we will stick to basics. A few points to note that are being ignored, even though I remember them (there are probably others that I cannot remember right away :) :

Photometric redshifts are hard even when we have the model
'salt2' can have problems with high redshift supernovae, so using salt2-extended
MW extinction MUST be added to the model, but I am ignoring for now
Model covariance should also be added for accurate results



In [11]:

    
model  = sncosmo.Model(source='salt2-extended')



In [12]:

    
model.set(z=sn.meta['REDSHIFT_FINAL'])
fit_results, fit_model = sncosmo.fit_lc(sn, model, vparam_names=['t0', 'x0', 'x1', 'c'])



In [13]:

    
print fit_model









    



<Model at 0x11456e890>
source:
  class      : SALT2Source
  name       : 'salt2-extended'
  version    : 1.0
  phases     : [-20, .., 50] days
  wavelengths: [300, .., 18000] Angstroms
parameters:
  z  = 0.58170002698898315
  t0 = 52902.658535727656
  x0 = 7.915816561859314e-06
  x1 = 0.51106105554992354
  c  = 0.26957985320119349



In [14]:

    
fig = sncosmo.plot_lc(sn, model=fit_model, band=['u', 'g', 'r', 'i', 'z', 'Y'], zp=25., zpsys='ab')



In [16]:

    
sncosmo.chisq(sn, fit_model)/ fit_results.ndof









    Out[16]:





0.88437760022231149

MJD	band	FIELD	TELESCOPE	PHOTFLAG	PHOTPROB	flux	fluxerr	MAG	MAGERR	PSF_SIG1	PSF_SIG2	PSF_RATIO	SKY_SIG	SKY_SIG_T	RDNOISE	ZEROPT	ZEROPT_ERR	GAIN	ZP	ZPSYS
float64	str2	str12	str20	int32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float32	float64	str2
52881.5	i	NULL	ALL	0	0.0	13.44	2.846	24.679	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	27.5	ab
52881.539	r	NULL	ALL	0	0.0	7.74	1.701	25.278	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	27.5	ab
52881.559	z	NULL	ALL	0	0.0	19.72	6.539	24.263	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	27.5	ab
52886.602	i	NULL	ALL	0	0.0	44.52	2.909	23.379	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	27.5	ab
52900.531	i	NULL	ALL	0	0.0	86.57	2.914	22.657	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	27.5	ab