gmpas 0.5.7

Try the viewer in your browser — the real interface on a real MPAS run, with nothing to install.

import gmpas

ds = gmpas.open_mpas("diag.2019-09-01_00.00.00.nc", mesh="maritime.region.nc")

ds.mpas.plot("mslp")        # cell field, filled Voronoi polygons
ds.mpas.plot("u")           # edge field, drawn on the cell faces themselves
ds.mpas.plot_mesh()         # where the mesh actually refines, in km

mesh= may be omitted when the file carries its own mesh information, or when a mesh file with a matching cell count sits beside it.

Installation

pip install gmpas            # core: geometry, caching, remap weights
pip install "gmpas[plot]"    # + matplotlib and cartopy, for plotting

conda env create -f environment.yml && conda activate gmpas && pip install -e . --no-deps

Mesh generation requires JIGSAW, MPI and PnetCDF.

gmpas --version

From a source install, pytest -q runs the test suite.

Full detail, including the extras and what each one pulls in: docs/installation.md.

Usage

gmpas info          history.2012-02-25_12.00.00.nc
gmpas plot          history.2012-02-25_12.00.00.nc precipw -o pw.png
gmpas view          /path/to/run/
gmpas remap         history.*.nc -o out/
gmpas prep view     mesh.nc
gmpas prep hfun     hfun.py --check
gmpas prep generate hfun.py -o mesh/     # needs $JIGSAWDIR and $MKGRIDFILE

Documentation

This document is generated using LLM (Claude)