LFRic Catalyst Adaptor

ParaView Catalyst adaptor implementation for the LFRic code

This package builds a library for visualising simulation data with a simple VTK visualisation pipeline. The pipeline can be defined either in C++ or using a Python script.

Building the adaptor

To build this code, you will need to build and install ParaView with Catalyst option enabled. Once this is done, build the code using CMake as follows:

mkdir build
cd build
cmake .. -DParaView_DIR=/path/to/catalyst/install/directory/lib/cmake/paraview-5.4 -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/path/to/install/dir
make

If you want to build a debug version of the code, add -DCMAKE_BUILD_TYPE=Debug to the CMake configuration, or use the ccmake configuration tool. You can add additional compiler flags using the -DCMAKE_CXX_FLAGS= option.

On a Cray XC50 system, the following build setup should work:

cmake .. -DCMAKE_CXX_COMPILER=CC -DCMAKE_EXE_LINKER_FLAGS=-dynamic -DParaView_DIR=/path/to/catalyst/install/directory/lib/cmake/paraview-5.4 -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/path/to/install/dir

Note that dynamic linking simplifies the linking process of the Fortran application significantly.

Once CMake has finished, run

make
make install

to build and install the library.

Running the test battery

If you want to test your build, add -DBUILD_TESTING=ON to your CMake configuration and run make test or ctest after building the code. This will run a number of tests that check basic functionality.

Running a simulation with the adaptor

The Catalyst adaptor and libraries are usually dynamically linked. If the build system of your code does not hardcode shared library paths, you will need to set (possibly adapting ParaView version)

export LD_LIBRARY_PATH=/path/to/catalyst/installation/lib/paraview-5.4:$LD_LIBRARY_PATH

If you want to use the Python pipeline, set PYTHONPATH to something like

export PYTHONPATH=/path/to/catalyst/installation/lib/paraview-5.4/site-packages:/path/to/catalyst/installation/lib/paraview-5.4/site-packages/vtk:$PYTHONPATH

Python scripts

The repository includes a number of Python scripts which define visualisation pipelines or provide some post-processing functionality.

full_output.py

Simple Python pipeline for writing the model grid and data field to a VTK file.

spherical_slice.py

Simple Python pipeline for creating spherical slices of model grid with a preset radius, which are written into a VTK polydata file. Full output of the model grid and data field can also be produced by setting the corresponding flag in the pipeline script.

spherical_slice_contours.py

Same as "spherical_slice.py", but includes an additional output file with contours.

spherical_slice_rendered.py

Same as "spherical_slice.py", but includes a rendered image of the slice which is stored as a png file.

spherical_slice_rendered_coastlines.py

Same as "spherical_slice_rendered.py", but overlays coastlines on the rendered image. Requires downloading coastlines data, see source file for instructions.

meridional_slice.py

Creates and stores a meridional slice for a chosen longitude, including a transformation from Cartesian to longitude-radius coordinates.

map_project.py

This Python program expects a spherical slice (as produced by the spherical_slice.py visualisation pipeline) in VTK polydata format as input and produces a VTK polydata file with a map projection as output. The program can handle partitioned datasets, but computing map projections for multiple timesteps is not supported yet.

Running

./map_project.py input.vtp output.vtp

computes a Mollweide map projection. Use flag --list-projections to get a list of projections and their short names (projections are provide by the PROJ library). Short names can be used to set another projection with the --projname flag, e.g., --projname=gall.