Implementation of union-find (aka disjoint-set) data structure. Currently, for performance, the structure is defined on a set ${0, \dots, n-1}$, of size $n$, which is specified at initialization.

It implements the standard operations, as well as $subset$, which returns the subset corresponding to an element. A main use case is hierarchical clustering.

The implementation is inspired by scipy's UnionFind module, and it relies on numba for performance.

Use pip install fufpy, or pip install . from the root directory of the project.

This package depends on numpy and numba, which will be installed automatically when installing via pip.

The documentation, including API, can be found here.

This software is published under the 3-clause BSD license.