BayeSN-TD

About

This is the documentation for bayesn-td, an extension of BayeSN for fitting strongly-lensed Type Ia supernovae. It jointly fits all images of a lensed SN Ia — sharing the intrinsic SN parameters across images while inferring per-image time delays, magnifications, and microlensing.

Like BayeSN, bayesn-td is built on numpyro and JAX, so inference can run on GPU when JAX is installed with GPU support.

Contents

• Overview
  • What bayesn-td does
  • Relationship to BayeSN
  • The probabilistic model
  • Time conventions
  • GPU acceleration
• Installation
  • Requirements
  • GPU support
  • Installing bayesn-td
• Fitting lensed SNe within Python
  • Fitting with fit_lensed_sn
  • Preparing your data
  • MCMC configuration
  • Model options
  • Working with the output
• Running from the command line
  • Basic usage
  • YAML input file format
  • Command-line overrides
  • Example
• Output
  • chains.pkl
  • initial_chains.pkl
  • fit_summary.csv
  • sn_list.txt
• Simulating lensed light curves
  • Basic usage
  • Arguments
  • Simulating in magnitudes vs flux
  • Fixed vs sampled parameters
• Built-in models
  • Summary
  • Available models
  • Choosing a model
  • Using a custom model
• Defining filters
  • Built-in filters
  • Specifying custom filters
  • Automatic filter dropping
• API reference
  • SEDmodel
  • bayesn_td.model
• Citing bayesn-td

Support

If you find something unclear, come across any bugs, or think of any functionality you would like, please raise a GitHub issue here: https://github.com/bayesn/bayesn-td/issues.