krigeo

It strated as a KISS demo package featuring naive implementation of simple kriging ... ended up as a pure-python universal kriging tool with fault network infering capabilities !

Overview

This package provides tools for variable estimation via kriging. It is divided in three submodules :

• kriging provides three possibilities for kriging :
  • Simple kriging (not exposed, for learning purpose only)
  • Ordinary kriging
  • Universal kriging (or kriging with external drift)
• covariance provides standard covariance models (it allows any custom covariance model, only requirement is to implement __call__(self, lag: ArrayLike) -> ArrayLike, and it even provides an abstract Covariance base class).:
  • Gaussian
  • Cubic
  • Exponential
  • Spherical
• drifts which is a toolbox designed to deal with discontinuities, that can then be integrated in universal kriging. It allows to create drift functions associated with faults for applications in geosciences.

Usage

A demo folder is provided to show a use case of this package. It contains a global demo (global demo) that shows the whole workflow, from the import of data to the kriging, with the definition of covariance and drift models. For more details on variography and drifts, see variography and drifts, that focus on covariance and drift submodules applied to this use case.

Installation

pip install krigeo

One can also build it manually from gitlab:

pip install git+https://gitlab.com/brgm/geomodelling/spatial/krigeo

To locally build/install this package, use:

git clone https://gitlab.com/brgm/geomodelling/spatial/krigeo
cd krigeo
# regular install
python -m pip install .
# install as editable
python -m pip install -e .
# or package wheel
python -m pip wheel . -w ./dist --no-deps

Contributing

See CONTRIBUTING.md

License

krigeo is provided under a GNU Affero General Public License that can be found in the LICENSE file. By using, distributing, or contributing to this project, you agree to the terms and conditions of this license.