ParaView Plugin

meshio++ ships a ParaView Python plugin (tools/paraview-meshioplusplus-plugin.py) with a reader and a writer, so ParaView can open any file meshio++ reads and save to any format it writes, without converting first.

Installation

The plugin runs inside ParaView's own Python, so meshio++ must be installed for that interpreter, and its compiled core must match that interpreter's version.

1. Find the Python that your ParaView uses:

   pvpython -c "import sys; print(sys.version)"

2. Install meshio++ for that Python (a conda-forge ParaView: into the same environment; a system ParaView: into the system Python, or a virtual environment on PYTHONPATH):

   pip install "meshioplusplus[all]"

3. The wheel installs the plugin as data, at a path with no ParaView version in it:

   <prefix>/share/meshioplusplus/paraview/paraview-meshioplusplus-plugin.py

   where <prefix> is the environment's root (python -c "import sys; print(sys.prefix)"). A source checkout has the same file at tools/paraview-meshioplusplus-plugin.py.

4. Load it in ParaView, either:

   • Tools → Manage Plugins → Load New, browse to the file, and tick Auto Load to load it on every start; or

   • point PV_PLUGIN_PATH at the directory before starting ParaView, which loads every plugin in it:

     export PV_PLUGIN_PATH="$(python -c 'import sys; print(sys.prefix)')/share/meshioplusplus/paraview"

From pvpython or pvbatch, paraview.simple.LoadPlugin(path, ns=globals()) loads it and defines meshioreader and meshioWriter.

Usage

After loading, every extension meshio++ knows appears in ParaView's file dialogs. The reader's FileFormat property picks the format explicitly (automatic infers it from the file), and the file is exposed as a vtkUnstructuredGrid with its point, cell and field data. The writer saves an unstructured grid through meshioplusplus.write, choosing the format from the file name.

Limitations

• Polyhedron blocks are skipped by the reader (the legacy cell layout the plugin builds has no face stream); a message on ParaView's output names each skipped block.
• The writer groups cells by VTK type and node count, so a mesh's cell order is not preserved across a write.
• Optional format dependencies (h5py, netCDF4, ...) must be installed in ParaView's Python as well.

Testing

tests/python/test_paraview_plugin.py runs pvpython in a subprocess: it loads the plugin, reads a file through the reader, writes it back through the writer, and reads the result with meshio++. It is skipped when pvpython is absent or runs a different Python than the build under test; CI runs it in its paraview-plugin job against conda-forge ParaView.