Core API#

The Visualization Toolkit (VTK), developed by Kitware, has many mesh data types that PyVista wraps. This chapter is intended to describe these different mesh types and how we wrap and implement each of those mesh types in VTK. This chapter also highlights how all PyVista types have direct access to common VTK filters (see Filters).

All PyVista meshes inherit from the DataSet type (see Data Sets). PyVista has the following mesh types:

pyvista.PointSet is used to represent a set of points. This is to provide an interface for datasets that explicitly use “point” arrays to represent geometry. The pyvista.PointSet class is an extension of vtk.vtkPointSet.
pyvista.PolyData consists of any 1D or 2D geometries to construct vertices, lines, polygons, and triangles. We generally use pyvista.PolyData to construct scattered points and closed/open surfaces (non-volumetric datasets). The pyvista.PolyData class is an extension of vtk.vtkPolyData.
A pyvista.UnstructuredGrid is the most general dataset type that can hold any 1D, 2D, or 3D cell geometries. You can think of this as a 3D extension of pyvista.PolyData that allows volumetric cells to be present. It’s fairly uncommon to explicitly make unstructured grids but they are often the result of different processing routines that might extract subsets of larger datasets. The pyvista.UnstructuredGrid class is an extension of vtk.vtkUnstructuredGrid.
A pyvista.StructuredGrid is a regular lattice of points aligned with internal coordinate axes such that the connectivity can be defined by a grid ordering. These are commonly made from numpy.meshgrid(). The cell types of structured grids must be 2D quads or 3D hexahedra. The pyvista.StructuredGrid class is an extension of vtk.vtkStructuredGrid.
A pyvista.RectilinearGrid defines meshes with implicit geometries along the axis directions that are rectangular and regular. The pyvista.RectilinearGrid class is an extension of vtk.vtkRectilinearGrid.
Image data, commonly referred to as uniform grids, and defined by the pyvista.ImageData class are meshes with implicit geometries where cell sizes are uniformly assigned along each axis and the spatial reference is built out from an origin point. The pyvista.ImageData class is an extension of vtk.vtkImageData.
pyvista.MultiBlock datasets are containers to hold several VTK datasets in one accessible and spatially referenced object. The pyvista.MultiBlock class is an extension of vtk.vtkMultiBlockDataSet.