SHINE: SHear INference Environment

A JAX-powered framework for probabilistic shear estimation in weak gravitational lensing

---

🌟 Overview

SHINE (SHear INference Environment) is a modern, high-performance framework for probabilistic shear estimation in weak gravitational lensing studies. Built on JAX, it leverages automatic differentiation and just-in-time compilation to deliver fast, scalable inference for cosmological applications.

✨ Key Features

• 🚀 JAX-powered: Automatic differentiation and JIT compilation for optimal performance
• 📊 Probabilistic Inference: Full posterior distributions for shear estimates
• 🔧 Modular Design: Flexible architecture for easy extension and customization
• 🎯 GPU Acceleration: Seamless GPU support for large-scale analyses
• 📈 Scalable: Efficient handling of large imaging surveys

📦 Installation

pip install shine-wl

For development (editable install from source):

git clone https://github.com/CosmoStat/SHINE.git
cd SHINE
pip install -e ".[dev,test]"

Euclid VIS Support

SHINE includes a full instrument backend for Euclid VIS multi-exposure shear inference (shine.euclid):

• Reads quadrant-level FITS exposures, PSF grids, background maps, and MER source catalogs
• Forward-models the scene as Sersic galaxies convolved with spatially-varying PSFs via JAX-GalSim
• Multi-tier stamp rendering (64/128/256 px) assigned per source by half-light radius
• Catalog filtering by SNR, det_quality_flag, size, and footprint visibility
• MAP and NUTS inference for shared shear (g1, g2) and per-source morphology

See notebooks/euclid_vis_map.ipynb for an interactive demo.

Quick Start

Run inference from a config file

SHINE is driven by YAML configuration files. Any parameter specified as a distribution (e.g. type: Normal) becomes a latent variable; everything else is fixed. To run the full pipeline (data generation → model building → MCMC):

python -m shine.main --config configs/test_run.yaml

Results (posterior samples in NetCDF format) are saved to the results/ directory by default. Override with --output:

python -m shine.main --config configs/test_run.yaml --output my_output/

Pedagogical example

For a step-by-step walkthrough that builds the config inline and plots diagnostics, see examples/shear_inference.py:

python examples/shear_inference.py

📖 Documentation

Full documentation is available at cosmostat.github.io/SHINE, including:

• Getting Started — installation and first run
• Configuration Reference — YAML config specification
• Validation Pipeline — bias measurement infrastructure
• API Reference — auto-generated from docstrings

🏗️ Status

⚠️ Early Development: This project is under active development. APIs may change.

🤝 Contributing

We welcome contributions! This project is in early development, and we're excited to collaborate with the community. Thanks goes to these wonderful people (emoji key):

Ezequiel Centofanti 🤔 📆

Samuel Farrens 🤔 📆

Emma Ayçoberry 🤔 📆

Francois Lanusse 🤔 📆

This project follows the all-contributors specification. Contributions of any kind welcome!

---

Born at CosmoStat, built with ❤️ for the astro community.