tarp

tarp is a small python package for performing statistical coverage tests to assess the quality of posterior estimators. tarp is framework-agnostic: it works with posterior estimators and samplers written in numpy, pytorch, jax.

tarp currently implements the distance to random point (DRP) test introduced in Lemos, Coogan et al 2023, which relies on posterior samples.

Installation

Soon you'll be able to install from PyPI with pip install tarp. For now,

git clone git@github.com:Ciela-Institute/tarp.git
cd tarp
pip install .

will install the tarp package.

Contributing

Please reach out to us if you're interested in contributing!

To start, follow the installation instructions, replacing the last line with

pip install -e ".[dev]"

This creates an editable install and installs the dev dependencies for generating docs, running tests and packaging for PyPI.

Please use isort and black to format your code. Open up issues for bugs/missing features. Use pull requests for additions to the code. Write tests that can be run by pytest.