ferreus-rbf 0.1.1

Fast, memory efficient radial basis function (RBF) interpolation library

ferreus_rbf (Python bindings)

Python bindings for the ferreus_rbf Rust crate, providing fast global radial basis function (RBF) interpolation from Python.

Overview

Direct RBF solvers based on dense linear algebra require O(N²) memory and O(N³) work, which becomes impractical on large datasets. The underlying Rust implementation combines:

• Domain decomposition preconditioning in a Flexible GMRES (FGMRES) solver, and
• A fast multipole method (FMM) evaluator via ferreus_bbfmm,

to reduce the cost to roughly O(N log N). This makes it possible to build global interpolants and extract isosurfaces on datasets with millions of points from Python.

Limitations

The current isosurface extraction method does not guarantee manifold or watertight meshes. Surfaces may contain trifurcations or self‑intersections and may therefore be unsuitable for downstream boolean or other topology‑ sensitive operations without additional post‑processing.

Install

pip install ferreus_rbf

Then in Python:

import ferreus_rbf

See the docs and examples/ directories in this package for more detailed usage and API documentation.

Attribution and licensing

This package was developed while the author was working at Maptek and has been approved for open‑source distribution under the terms of the MIT license.

Unless otherwise stated, the following copyright applies:

Copyright (c) 2025 Maptek Pty Ltd.
All rights reserved.

This copyright applies to all files in this repository, whether or not an individual file contains an explicit notice.

The code is released under the MIT License – see the top‑level LICENSE file for details.