The package currently uses pymoab

However it might be possible to use h5py

This would have the advantage of being pip installable and faster.

The data structure needed for a DAGMC compatible h5m file is shown here https://sigma.mcs.anl.gov/moab/h5m-file-format/

meshio is able to make h5m files, but these have never quite worked in the simulation code we make the h5m files for (dagmc/openmc).

It would be great to add to meshio so that the h5m files produced work in neutronics simulations.
The h5m files produced by this package work in neutronics simulations so could be used to provide a simple comparison as we can make a simple set of vertices and triangles into a dagmc/openmc compatible h5m file and compare it to the output from meshio and see what is missing