The kpoint symmetry and band plotting cleanup issue

[mfherbst]

trafficstars

Some points to cleanup left after #496 is merged

Quick n' easy

• [x] Drop enforcing pymatgen installation (also from the docs) (#534)
• [x] high_symmetry_kpath does not use atomic units (#536)
• [x] plot_bandstructure does not use atomic units consistently (#536)
• [x] plot_bandstructure interface is quirky for linear Hamiltonians (#536)

Cleanup

• [x] Use Spglib and retire direct use of spglib_jll (#876)
• [x] Merge spglib_standardize_cell and get_spglib_lattice (#568)
• [x] Drop kpoint duplication (#555)
• [x] Band structures also for supercells (#555)
• [x] Use Brillouin.jl in high_symmetry_kpath also for 1D and 2D lattices (#555)
• [ ] Integrate more with Brillouin.jl for BZ plots or band structure plots (see also #503)

Consistency

• [ ] Take care of time-reversal symmetry / absence thereof consistently for kpath (see this discussion)
• [ ] Take care of spin / magnetic moments in cell standardisation and kpath generation

[jaemolihm]

Hi, I'm planning to make a PR to Brillouin.jl for a function that gives k path for arbitrary lattice vectors (https://github.com/thchr/Brillouin.jl/issues/13#issuecomment-968625017) in a few days. After that, one can make high_symmetry_kpath in DFTK much simpler. Independently, I have a working branch that cleans up the kpath routines, e.g. deleting kpoint duplication. If nobody is working on these, I can pick it up and make a PR (after making a PR to Brillouin.jl).

[mfherbst]

@jaemolihm Yes that would be highly appreciated! Noone is working on these at the moment. Thanks for your work on this (truly annoying) aspects of solid-state modelling!