jpierel14 / sntd

SNTD is a Python package dedicated to measuring time delays for strongly lensed and multiply-imaged SN. The package can simulate lensed SN, fit for each SN time delay, and propagate the results all the way through to cosmological constraints.

Welcome to Supernova Time Delays (SNTD)

SNTD is a Python package dedicated to measuring time delays for (resolved) strongly lensed and multiply-imaged SN. The package can simulate lensed SN, fit for each SN time delay, and propagate the results all the way through to cosmological constraints. Documentation for this package is hosted on ReadTheDocs. This package is under active use and development, so please post GitHub issues if anything arises (bug or feature request). Please see the "notebooks" folder for jupyter notebook examples of using SNTD.

Here you can run the SNTD example notebooks directly in your browser using Google Colab:

Simulations:

Fitting and Plotting:

Constraining Cosmology:

Citation:

sntd's Issues

Bazin fits are showing double-peaked corner plots for t0, although accurate time delays

I'm guessing something systematic in the models file?

Implement chromatic microlensing fitting/simulation

Nestle likely bug

Fitting some data with a IIn template gives an error with nestle due to a negative h.

"""
File "/Users/suhaildhawan/workspaces/LSST/time_delays/core-collapse/simulate_sntd.py", line 43, in
fitCurves=sntd.fit_data(misn,snType=snType,models=template,bands=filt_ls,
File "/Users/suhaildhawan/.local/lib/python3.9/site-packages/sntd/fitting.py", line 568, in fit_data
curves = _fitparallel(args)
File "/Users/suhaildhawan/.local/lib/python3.9/site-packages/sntd/fitting.py", line 2725, in _fitparallel
res, fit = sncosmo.nest_lc(fit_table, tempMod, [x for x in args['params'] if x in tempMod.param_names],
File "/opt/anaconda3/lib/python3.9/site-packages/sncosmo/fitting.py", line 1099, in nest_lc
res = nestle.sample(loglike, prior_transform, ndim, npdim=npdim,
File "/Users/suhaildhawan/.local/lib/python3.9/site-packages/nestle.py", line 985, in sample
raise RuntimeError("Negative h encountered (h={}). Please report "
RuntimeError: Negative h encountered (h=-8.102136181477137e+23). Please report this as a likely bug.

"""

Weights sum to zero, can't be normalized

Parallel fitting issue when bounds don't encompass parameters (I think), and fitting gets zero likelihood everywhere basically.

name change for io.py?

PyCharm apparently can in some cases have an issue with a package that contains modules with names that match up to certain core python system builtin modules, such as abc.py and io.py:

https://stackoverflow.com/questions/26569828/pycharm-py-initialize-cant-initialize-sys-standard-streams

I suggest changing io.py to lightcurves.py, which I've just done in my fork. If this isn't a problem for you, then it can wait until I issue a pull request to merge my fork into your dev branch.

single tar file for batch jobs is crazy slow to read, make multiple

Microlensing code isn't always installed correctly

@srodney If you have a screenshot of the error or something, that would be helpful here.