Welcome to discoverysamplers documentation!

discoverysamplers is a collection of bridge interfaces that connect Discovery (a JAX-based Pulsar Timing Array analysis framework) to various nested sampling and MCMC samplers.

Author	Jonas El Gammal
Source	Source code on GitHub
Documentation	Documentation on Read the Docs
License	MIT
Support	For questions, use GitHub Issues or drop me an email
Installation	Clone from GitHub

Overview

The package provides lightweight wrappers that adapt Discovery-style models (callables accepting parameter dictionaries) to the APIs expected by different sampling backends.

Key Features

Unified Interface: Consistent API across multiple sampling backends
Flexible Prior Specification: Support for multiple prior formats (dicts, tuples, callables)
JAX Integration: Optional JIT compilation for improved performance
Parameter Management: Automatic handling of fixed vs. sampled parameters
LaTeX Labels: Built-in support for publication-quality plotting

Supported Samplers

Sampler	Type	Key Features
Eryn	MCMC (Ensemble)	Parallel tempering, reversible-jump MCMC support
Nessai	Nested Sampling	Flow-based proposals, importance nested sampling
JAX-NS	Nested Sampling	Pure JAX implementation, vectorized likelihood evaluation
GPry	GP Emulation	Gaussian Process surrogate model with active learning

Quick Start

Install the package:

git clone https://github.com/jonaselgammal/discoverysamplers.git
cd discoverysamplers
pip install .

Basic usage with Nessai:

from discoverysamplers import DiscoveryNessaiBridge

# Define your model
def my_model(params):
    return -0.5 * (params['x']**2 + params['y']**2)

# Define priors
priors = {
    'x': ('uniform', -5.0, 5.0),
    'y': ('uniform', -5.0, 5.0),
}

# Create bridge and run sampler
bridge = DiscoveryNessaiBridge(my_model, priors)
results = bridge.run_sampler(nlive=1000, output='output/')

See the Quick Start Guide for more detailed examples.

Citation

If you use discoverysamplers in your research, please cite:

@software{discoverysamplers,
  author = {El Gammal, Jonas},
  title = {discoverysamplers: Tools for Bayesian inference with Discovery},
  year = {2025},
  url = {https://github.com/jonaselgammal/discoverysamplers}
}

Documentation Contents

User Guide

Examples

Example Notebooks

Acknowledgements

Acknowledgements