GRaTer-JAX: An Accelerated Package for Debris Disk Modeling

GRaTer-JAX is a Python package for modeling debris disks, disks of dust around stars that pro- vide important clues about the formation, structure, and evolution of planetary systems. GRaTer (Generator of Ring-like, axisymmetric, optically Thin dust disks for regularized fitting) is a debris disk modeling framework originally developed for generating and fitting models of optically thin, ax- isymmetric dust disks. JAX is a Python package for high-performance computing that implements just-in-time compilation and auto-differentiation, enabling significant speedups in forward modeling workflows. GRaTer-JAX delivers a more powerful, efficient, and robust debris disk modeling framework than previous tools. Its JAX-based backend provides orders-of-magnitude speedups and analytic gradients, while its intuitive and extensible API unifies forward modeling, optimization, and inference within a single workflow. Combined with key additions to its disk model, it is capable of more advanced debris disk modeling, enabling research that was previously computationally impractical such as explorations of additional morphological parameters and disk composition via scattering phase functions.