This is the analysis code used in the paper Dam, Heinesen, and Wiltshire, Mon. Not. Roy. Astron. Soc. 472 (2017) 835 [arxiv].
07/10/2022: fixes a bug in distmod.py that was affecting the
distance moduli of a single data point (thanks to Zac Lane).
This code requires NumPy, SciPy, and PyMultiNest (a Python interface of MultiNest).
The JLA dataset and covariance matrices used in this analysis are not supplied here; they can be downloaded from here and here. In addition, the covariances used in this code have been converted from FITS format to .npy format; these can be downloaded here. Please cite the respective papers if these products are used in published work.
The dataset used in the analysis computes redshifts in the CMB frame
from JLA heliocentric redshifts. Running python build.py
generates the data file jla.tsv used in this analysis
ordered as follows
zcmb, mb, x1, c, logMass, survey id, zhel, RA, DEC
Next, for fast likelihood evaluation, run python distmod.py
to produce a look-up table of luminosity distances for each
SNIa and for different cosmological parameter(s).
Running the main script snsample.py computes the evidence
for the model specified by the following command line options:
model = int(sys.argv[1]) # 1=Timescape, 2=Empty, 3=Flat
z_cut = float(sys.argv[2]) # redshift cut e.g. 0.033
zdep = int(sys.argv[3]) # redshift dependence in mean stretch and colour distributions (0 or 1)
case = int(sys.argv[4]) # redshift light curve model (1-8)
nsigma = int(sys.argv[5]) # 1, 2 or 3 sigma omega/fv prior
nlive = int(sys.argv[6]) # number of live points used in sampling
tol = float(sys.argv[7]) # stop evidence integral when next contribution less than tol