krigekit 0.2.1

Python wrapper for a Fortran kriging and sequential Gaussian simulation engine

krigekit

A Python wrapper for a high-performance Fortran kriging engine parallelised with OpenMP.

Capability	Notes
Ordinary and simple kriging	Point and block support
Co-kriging	Linear Model of Coregionalisation
Universal kriging / KED	External drift variables
Sequential Gaussian Simulation	Reproducible paths, multi-realisation
Space-time kriging	Sum-metric and product-sum ST models
Multiple Indicator Kriging / SIS	Categorical variables, three cross-variogram strategies
Spatially Varying Anisotropy	Per-block variogram
Cross-validation	Leave-one-out
Kriging weight reuse	Store and replay weights

Full documentation →

---

Installation

pip

pip install krigekit

Pre-built binary wheels are available for:

Platform	Architecture	Minimum OS
Linux	x86_64	manylinux_2_28 (RHEL 8 / Ubuntu 20.04 equivalent)
macOS	arm64 (Apple Silicon)	macOS 14 Sonoma
macOS	x86_64 (Intel)	macOS 15 Sequoia
Windows	x86_64	Windows 10 / Server 2019

Requires Python 3.10, 3.11, or 3.12.

conda / mamba

conda-forge packaging is planned. In the meantime, install via pip inside a conda environment:

conda create -n krigekit python=3.12
conda activate krigekit
pip install krigekit

---

Quick start

import numpy as np
from krigekit import Kriging

obs_coord  = np.array([[0,0],[1,0],[0,1],[1,1],[0.5,0.5]], dtype=float)
obs_value  = np.array([1.0, 2.0, 3.0, 4.0, 2.5])
grid_coord = np.mgrid[0:1.1:0.25, 0:1.1:0.25].reshape(2,-1).T

k = Kriging()
k.set_obs(ivar=1, coord=obs_coord, value=obs_value)
k.set_grid(coord=grid_coord)
k.set_vgm(ivar=1, jvar=1, vtype="sph", sill=1.0, a_major=1.0)
k.set_search()
k.solve()
df = k.get_result_df()
del k

For the full class API, co-kriging, SGSIM, space-time kriging, indicator simulation, and more, see the user guide and gallery examples.

---

Repository structure

krigekit/
├── src/
│   ├── libkriging/      Fortran kriging engine
│   ├── sparks/          Pilot-point kriging/SGSIM CLI
│   └── krigekit/        Python ctypes wrapper
├── examples/            Sphinx-Gallery example scripts
├── tests/               pytest test suite
├── test_data/           CSV/image data used by tests and examples
├── docs/                Sphinx documentation source
├── build_lib.py         Fortran compile script
├── environment.yml      conda/mamba development environment
└── pyproject.toml       pip package configuration

---

Contributing

Building from source

A Fortran compiler is required (gfortran ≥ 10 or Intel ifx/ifort).

# Clone and set up the development environment
git clone https://github.com/ougx/krigekit.git
cd krigekit

# Compile the Fortran library
python build_lib.py --compiler gfortran      # Linux / macOS
python build_lib.py --compiler gfortran      # Windows (MinGW via MSYS2)
python build_lib.py --compiler ifx           # Windows (Intel oneAPI)

# Install in editable mode with dev dependencies
pip install -e ".[dev]"

Additional build options:

python build_lib.py --opt debug        # debug build, no optimisation
python build_lib.py --no-openmp        # disable OpenMP
python build_lib.py --hcache 0         # disable factor cache

Workflow

1. Fork the repository and create a feature branch.
2. Add tests for any new behaviour.
3. Run pytest to confirm all tests pass.
4. Open a pull request.

---

License

MIT — see LICENSE.