pyransame

PYthon RAndom SAmpling for MEshes

Documentation

Utilities for choosing random samples of points within cells of PyVista meshes. This package does not choose random points that define the mesh itself, rather random points on 0D vertices, 1D lines, 2D surfaces or in 3D volumes are sampled.

All linear¹ cells from vtk are supported, except for vtkConvexPointSet.

PyVista dataset accessor

On PyVista 0.48+, pip install pyransame registers a ransame accessor on every PyVista dataset. No extra import or registration step is required. PyVista discovers the accessor through the pyvista.accessors entry point and loads it lazily the first time mesh.ransame is used.

import pyvista as pv
from pyvista import examples

bunny = examples.download_bunny()
points = bunny.ransame.surface_points(500)        # numpy array, shape (500, 3)
sampled = bunny.ransame.surface_dataset(500)      # pyvista.PolyData with interpolated arrays

The accessor mirrors the top-level functions:

points and dataset infer the sampler from the cell dimensions on the mesh: vertex (0D), line (1D), surface (2D), or volume (3D). Mixed dimensions raise ValueError. Pass kind="vertex" | "line" | "surface" | "volume" to override:

sampled = mixed_mesh.ransame.dataset(500, kind="surface")

On older PyVista releases the package still installs and imports cleanly; only the mesh.ransame namespace is unavailable. Use the top-level pyransame.random_* functions in that case.

Random sampling on a 2D surface

Random sampling in a 3D volume