"""

Returns mag at max for the model, band and magsys

Arguments:

model -- sncosmo model, e.g. SALT2

band -- sncosmo band object or string, e.g. 'bessellb'

magsys -- magnitude system, e.g. 'ab'

Keyword arguments:

t -- time relative to t0 (observer-frame), at which to evaluate

rest_frame -- default: True, overrides the redshifts

"""

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

if rest_frame:

kwargs['z'] = 0

model.set(**kwargs)

t=0, rest_frame=True):

"""

Return gradient of _get_bandmag as function of param

param, sig must be dictionaries of means and uncertainties

Best use odicts to make sure that the order of the components is correct

"""

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

out = []

if rest_frame:

if 'z' in param.keys():

param['z'] = 0

if 'z' in fixed_param.keys():

fixed_param['z'] = 0

model.set(**fixed_param)

for key,val in param.items():

model.set(**param)

h = sig[key] / 100.

model.set(**{key: val - h})

m0 = model.bandmag(band, magsys, param['t0'] + t)

model.set(**{key: val + h})

m1 = model.bandmag(band, magsys, param['t0'] + t)

out.append((m1 - m0) / (2. * h))

from itertools import combinations

SN_comb = []

ra_comb = []

dec_comb = []

for i in list(combinations(SN,2)):

SN_comb.append(i)

for i in list(combinations(ra,2)):

ra_comb.append(i)

for i in list(combinations(dec,2)):

dec_comb.append(i)

t = np.pi/180.

theta = []

for i in range(0,len(SN_comb)):

costheta = np.sin(dec_comb[i][0]*t)*np.sin(dec_comb[i][1]*t)+np.cos(dec_comb[i][0]*t)*np.cos(dec_comb[i][1]*t)*np.cos(ra_comb[i][0]*t-ra_comb[i][1]*t)

# Theta in arcsec

theta.append( np.arccos(costheta) / t * 3600 )

SN_dupl = np.array(SN_comb)[np.where(np.array(theta)<dtheta)]

theta_dupl = np.array(theta)[np.where(np.array(theta)<dtheta)]

import marshaltools as mt

from astropy.table import vstack

### No duplicates

if (nobj==1):

# Light curves

classification = data.classification

z = data.redshift

ebv_mw = data.mwebv

lc_data = data.table_sncosmo

### Duplicates

elif nobj==2:

lcs = mt.ProgramList("Cosmology")

lc1 = lcs.get_lightcurve(SN_dupl[idupl][0])

lc1_data = lc1.table_sncosmo

lc2 = lcs.get_lightcurve(SN_dupl[idupl][1])

lc2_data = lc2.table_sncosmo

classification = lc1.classification

z1 = lc1.redshift

z2 = lc2.redshift

if z1 == None:

z = z2

elif z2 == None:

z = z1

else:

z = 0.5 * (float(z1) + float(z2))

lc_data = vstack([lc1_data,lc2_data])

lc_data.meta['z'] = z

lc_data.meta['mwebv'] = lc1.mwebv

lc_data.sort('mjd')

idupl+=1

else:

lc_data = None

classification = None

if save == 1:

out = open(filename+".txt","a+")

# SN included in the analysis, good fit

if message == 0:

print("%s (%d out of %d)"%(sn,isn+1,nSN))

if save == 1:

out.write("%s "%sn)

for i in range(0,len(fit[1].parameters)):

out.write("%f "%fit[1].parameters[i])

out.write("%f %f "%(par1,par2))

lc_datag = data[np.where(data['band']=="p48g")]

lc_datar = data[np.where(data['band']=="p48r")]

if len(lc_datag)!=0 and len(lc_datar)!=0:

out.write("%f "%(lc_datag['mjd'][0] - fit[1].parameters[1]))

out.write("%f "%(lc_datar['mjd'][0] - fit[1].parameters[1]))

elif len(lc_datag)==0:

out.write("1e6 ")

out.write("%f "%(lc_datar['mjd'][0] - fit[1].parameters[1]))

elif len(lc_datar)==0:

out.write("%f "%(lc_datag['mjd'][0] - fit[1].parameters[1]))

out.write("1e6 ")

else:

out.write("1e6 ")

out.write("1e6 ")

out.write("%d "%len(lc_datag) )

out.write("%d "%len(lc_datar) )

out.write("%f %d \n"%(fit[0].chisq,fit[0].ndof))

fig = sncosmo.plot_lc(data, model=fit[1],errors=fit[0].errors,xfigsize=15,tighten_ylim=True)

fig.savefig("fits/"+sn+".pdf",bbox_inches='tight')

elif message == -1:

print("%s not considered, duplicate "%sn)

if save == 1:

out.write("#%s not considered, duplicate \n"%sn)

# Not a normal Ia

elif message == 1:

print("%s not considered, not a SN Ia norm (%s)"%(sn,par1))

if save == 1:

out.write("#%s not considered, not a SN Ia norm (%s) \n"%(sn,par1))

# Missing redshift

elif message == 2:

if par2 == 1:

print("%s not considered, missing redshift (spectra available) "%sn)

if save == 1:

out.write("#%s not considered, missing redshift (spectra available) \n"%sn)

else:

print("%s not considered, missing redshift "%sn)

if save == 1:

out.write("#%s not considered, missing redshift \n"%sn)

# Bad fit

elif message == 3:

print("%s not considered, bad fit"%sn)

if save == 1:

out.write("#%s not considered, bad fit \n"%sn)

# No data points

elif message == 4:

print("%s not considered, no data points \n"%sn)

if save == 1:

out.write("#%s not considered, no data points \n"%sn)

if save == 1:

x1cut = raw_input("Would you like a cut on x1 [y/n] ? ")

if x1cut == "y":

x1_min = np.float(input("Enter min(x1): "))

x1_max = np.float(input("Enter max(x1): "))

else:

x1_min = -11

x1_max = 11

chicut = raw_input("Would you like a cut on the reduced chisquare from SALT2 fitting [y/n] ? ")

if chicut == "y":

chisqred_max = np.float(input("Enter max(chi_sq_red): "))

else:

chisqred_max = 1e5

phasecut = raw_input("Would you like a cut on the maximum phase for the earliest g/r data point [y/n] ? ")

if phasecut == "y":

phasemax_g = np.float(input("Enter maximum phase in g band: "))

phasemax_r = np.float(input("Enter maximum phase in r band: "))

else:

phasemax_g = 1e5

phasemax_r = 1e5

ncut = raw_input("Would you like a cut on the minimum number of g/r data points [y/n] ? ")

if ncut == "y":

nmin_g = np.int(input("Enter minimum number of data points in g band: "))

nmin_r = np.int(input("Enter minimum number of data points in r band: "))

else:

nmin_g = 0

nmin_r = 0