CNES/pangeo-pyinterp ↗
Created Dec 30, 2020 · View the CNES/pangeo-pyinterp repository page
Python library for optimized interpolation.
Want this for your repo?
Render a free sample of any GitHub repo in seconds.
Contributors
7
Lines of Code
1,946
From
Jun 13, 2019
To
Oct 4, 2020
About CNES/pangeo-pyinterp
Pangeo-pyinterp is a Python library for high-performance interpolation of geo-referenced data, designed specifically for geoscience applications where pure-Python implementations fall short. The core engine is written in C++ using Boost, Eigen3, and nanobind, providing optimized performance while maintaining a Python interface. It supports interpolation across 2D, 3D, and 4D fields using n-variate and bicubic interpolators, and can work with both structured regular and irregular grids as well as unstructured grids.
The library includes several tools for handling real-world geospatial data challenges. It provides methods to fill undefined or masked values in grids through FFT inpainting, local regression (LOESS), Gauss-Seidel relaxation, and multi-grid methods. For geographic indexing, it offers specialized axis classes that automatically select efficient search strategies for regular versus irregular coordinates, temporal axes for date handling, and R*Tree-based indexing for unstructured grids with transparent conversion between latitude-longitude-altitude and Earth-Centered Earth-Fixed coordinate systems.
Beyond interpolation, the library wraps Boost Geometry functionality to provide comprehensive geometric operations including area and distance calculations, centroid determination, buffering, and coordinate transformations. It supports both Cartesian and geographic coordinate spaces with multiple geodetic strategies available for precise spheroid calculations. The project also includes geohashing capabilities for encoding geographic coordinates into compact string representations suitable for database storage and indexing.