Wrong dimensionality of atom positions vs number of atoms in new ABINIT parser
This happens on almost all ABINIT calculations. For example: aFuxFHVCb5d_tQblAGDoCncE4o-s
If you look at the archive, the atom_positions matrix is one vector to short to what you would expect by the length of atom_labels, species, number_of_atoms, etc.
As a result the system normalisation fails, e.g. there is not classification on system and symmetry.