PyVista accessor (recommended, `pyvista >= 0.48`)

May 8, 2026 · View on GitHub

######## pyacvd ########

.. image:: https://img.shields.io/pypi/v/pyacvd.svg :target: https://pypi.org/project/pyacvd/

This module takes a surface mesh and returns a uniformly meshed surface using voronoi clustering. This approach is loosely based on research by S. Valette, and J. M. Chassery in ACVD <https://github.com/valette/ACVD>_.

Installation

Installation is straightforward using pip:

.. code::

$ pip install pyacvd

Examples

PyVista accessor (recommended, `pyvista >= 0.48`)

Installing pyacvd registers an acvd namespace on every pyvista.PolyData, so uniform remeshing slots straight into a PyVista pipeline. Once pyacvd is imported (or auto-discovered via PyVista's entry-point system) you can call mesh.acvd.<method>(...) directly:

.. code:: python

import pyacvd # registers mesh.acvd from pyvista import examples

cow = examples.download_cow().triangulate()

one-shot uniform remesh: subdivide → cluster → rebuild

remesh = cow.acvd.remesh(20000, subdivide=3) remesh.plot(color='w', show_edges=True)

The accessor never mutates the source mesh and always returns a fresh pv.PolyData. Available methods:

mesh.acvd.remesh(n_clusters, subdivide=0, fast=False, ...) — one-shot uniform remesh.
mesh.acvd.clustering(weights=None, subdivide=0) — return a configured pyacvd.Clustering for fine-grained control (plotting clusters, mixing fast/uniform clustering, etc.).
mesh.acvd.cluster_ids(n_clusters, ...) — per-point cluster id array, useful for visualization.
mesh.acvd.subdivide(n) — linearly subdivide the surface.

Classic `Clustering` API

This example remeshes a non-uniform quad mesh into a uniform triangular mesh. The classic API works on every supported PyVista version; the accessor above is a thin convenience layer on top.

.. code:: python

from pyvista import examples import pyacvd