seamsh 0.4.11

Ocean mesh generation

Seamsh is a Python library wrapping gmsh, gdal and scipy to simplify the generation of unstructured meshes. While primarily developed for coastal ocean simulations, it can be used in other GIS contexts.

Main Features :

• Import ESRI shapefiles to define tagged domain boundaries, interior lines and interior points.

• Define arbitrary mesh elements size fields based on distances from lines or raster files.

• Create a low-resolution valid topology from high-resolution non-conformal (i.e. intersecting) data.

Seamesh is distributed under the GPL. See the project gitlab page for the source code and bug reports. The documentation contains examples, Python API reference and installation instructions. Binary packages for 64 bits linux, windows and OSX are available on pypi.

Changes

seamsh 0.4.10

• gmsh.merge_meshes : topological reconnection instead on coordinate based

seamsh 0.4.10

• add gmsh.merge_meshes

• add transifinite curves (physical lines with prescribed number of elements)

• avoid gmsh initalization error message

seamsh 0.4.9

• add transifinite curves (physical lines with prescribed number of elements)

seamsh 0.4.8

• numpy 1.24 compatibility (np.bool -> bool)

seamsh 0.4.7

• write WKT projection as gmsh model attribute

• default msh output version set to 4.1

seamsh 0.4.6

• License files in source package

seamsh 0.4.5

• Inpoly field (requires shapely)

seamsh 0.4.4

• fix unrefine when no identical points

• ugrid conversion (requires xarray)

• fix domain physical name

seamsh 0.4.3

• add output_srs parameter to gmsh.mesh

• add gmsh.reproject function

seamsh 0.4.2

• do not crash on features without geometry in shapefiles

seamsh 0.4.1

• do not call gmsh.initialize() if already initialized

• use a temporary gmsh model for meshing

seamsh 0.4.0

• precompute and smooth mesh size field on a quadtree-like mesh

• build library directly from setup.py, no cmake anymore

• discretize curves in their own projection, not in the model projection

• handle intersection of more than 2 interior curves on the same point

• source distribution on pypi