Pandamesh#

This package translates geospatial vector data (points, lines, or polygons) to unstructured meshes.

import pandamesh as pm

# Get some sample data in geopandas form.
south_america = pm.data.south_america()

# Explode any multi-polygon, and project it to UTM20.
south_america = south_america.explode().reset_index().to_crs(epsg=32620)

# Set a maximum cell size of 500 km and generate a mesh.
south_america["cellsize"] = 500_000.0
mesher = pm.TriangleMesher(south_america)
vertices, faces = mesher.generate()

https://raw.githubusercontent.com/Deltares/pandamesh/main/docs/_static/pandamesh-demo.png

The package converts geospatial data, presented as geopandas GeoDataFrames, to unstructured meshes using the open source high quality mesh generators:

Christophe Geuzaine and Jean-François Remacle’s Gmsh
Jonathan Shewchuk’s Triangle

utilizing the respective Python API’s, available at:

For completeness, the source code of both projects can be found at:

https://gitlab.onelab.info/gmsh/gmsh, under api/gmsh.py
drufat/triangle

These APIs are wrapped in two lightweight classes: pandamesh.TriangleMesher and pandamesh.GmshMesher. Both are initialized with a GeoDataFrame defining the geometry features of the mesh. During initialization, geometries are checked for overlaps and intersections, as the mesh generators cannot deal with these. Generated meshes are returned as two numpy arrays: the coordinates of the vertices, and the connectivity of the mesh faces to these vertices (as is usual for many unstructured grid representations).

GeoPandas is not suited for geometries that “wrap around” the world. Consequently, this package cannot generate meshes for e.g. a sphere.

Installation#

pip install pandamesh

Other projects#

Pandamesh has been developed because none of the existing packages provide a straightforward scripting based approach to converting 2D vector geometries to 2D unstructured grids.

Examples of other packages which work with unstructured meshes are listed below.

See also this list for many other mesh generation tools.

pygmsh#

The pygmsh Python package provides useful abstractions from Gmsh’s own Python interface so you can create complex geometries more easily. It also provides tools for 3D operations (e.g. extrusions).

qgis-gsmh#

qgis-gmsh generates geometry input files for the GMSH mesh generator and converts the Gmsh mesh files to shapefiles that can be imported into QGIS.

Lambrechts, J., Comblen, R., Legat, V., Geuzaine, C., & Remacle, J. F. (2008). Multiscale mesh generation on the sphere. Ocean Dynamics, 58(5-6), 461-473.
Remacle, J. F., & Lambrechts, J. (2018). Fast and robust mesh generation on the sphere—Application to coastal domains. Computer-Aided Design, 103, 14-23. https://doi.org/10.1016/j.cad.2018.03.002

Source: ccorail/qgis-gmsh

Shingle#

Shingle provides generalised self-consistent and automated domain discretisation for multi-scale geophysical models.

Candy, A. S., & Pietrzak, J. D. (2018). Shingle 2.0: generalising self-consistent and automated domain discretisation for multi-scale geophysical models. Geoscientific Model Development, 11(1), 213-234. https://doi.org/10.5194/gmd-11-213-2018

Source: shingleproject/Shingle

Website: http://shingleproject.org/index_shingle1.0.html