spatial-factorization-py

Probabilistic factor models for spatially correlated data. Includes nonnegative and real-valued versions.

Installation

Many people find it convenient to install everything in a conda environment. We recommend installing the miniforge version of conda. You can then create an environment and activate it with the following commands:

conda create -n nsf python=3.11
conda activate nsf

The package can then be installed into the environment from github:

pip install git+https://github.com/willtownes/spatial-factorization-py.git#egg=spatial-factorization

Usage

import spatial_factorization
help(spatial_factorization.ModelTrainer)
help(spatial_factorization.SpatialFactorization)

Development

Building locally with hatch

First make sure that the right tools are installed:

pip install hatch
pip install keyrings.alt

Then build with hatch:

hatch build