Re: [Spglib-users] How to get equivalent_atoms using fortran code

[Atsushi T. <atz...@gm...>]

Hi Wang,

Without spin, spg_get_dataset, and with spin, spg_get_magnetic_dataset
is recommended to use. However for the latter, the fortran interface
has not been written yet. A possible route is using
spg_get_symmetry_with_site_tensors. The interface is

function spg_get_symmetry_with_site_tensors(rotation, translation, &
& equivalent_atoms, primitive_lattice, spin_flips, max_size, lattice, &
& position, types, tensors, tensor_rank, num_atom, with_time_reversal, &
& is_axial, symprec) bind(c)
import c_int, c_double
integer(c_int), intent(out) :: rotation(3, 3, *)
real(c_double), intent(out) :: translation(3, *)
integer(c_int), intent(out) :: equivalent_atoms(*)
real(c_double), intent(out) :: primitive_lattice(3, 3)
integer(c_int), intent(out) :: spin_flips(*)
integer(c_int), intent(in), value :: max_size
real(c_double), intent(in) :: lattice(3, 3), position(3, *)
integer(c_int), intent(in) :: types(*)
real(c_double), intent(in) :: tensors(*)
integer(c_int), intent(in), value :: tensor_rank
integer(c_int), intent(in), value :: num_atom
integer(c_int), intent(in), value :: with_time_reversal
integer(c_int), intent(in), value :: is_axial
real(c_double), intent(in), value :: symprec
integer(c_int) :: spg_get_symmetry_with_site_tensors
end function spg_get_symmetry_with_site_tensors

Unfortunately, I haven't wrote any example of it, but I hope it can be
guessed from C-api doc.
https://spglib.readthedocs.io/en/latest/api.html#spg-get-symmetry-with-site-tensors
https://spglib.readthedocs.io/en/latest/development/magnetic_symmetry_flags.html#magnetic-action-flags

If you want to specify collinear spin, the array length is the number
of atoms, tensor_rank=0, is_axial=0, and with_time_reversal =1. For
the non-collinear case, the array length is the number-of-atoms x 3,
tensor_rank=1, is_axial=1, and with_time_reversal =1. These are the
typical choices.
The spin_flips array stores if the spin is flipped or not by the found
symmety operations. Therefore the length is the number of symmetry
operations. The number of symmetry operations is unknown before
running this function, so you have to prepare large enough symmetry
operations related arrays, whose length is considered to inform to
spglib by max_size.

Togo

On Tue, Jun 13, 2023 at 8:29 PM VeiWANG via Spglib-users
<spg...@li...> wrote:
>
> Dear all,
>
> I am currently trying to use Fortran code to obtain the equivalent_atoms using spglib, with input variables of lattice vectors and atom positions. Could anyone provide guidance on how to implement this?
>
> If I understand correctly, to obtain the equivalent_atoms, I first simply call the spg_get_symmetry function, which requires input variables of max_size, lattice, position, types, num_atom, and optional parameter symprec. And then I use the spg_get_symmetry_with_collinear_spin function, which takes an additional input variable of spins and outputs equivalent_atoms directly from the function. I was wondering if anyone has experience using spin information with spglib and if so, how to incorporate it into the function. Thank you for your assistance.
>
> Best,
> Wang
> _______________________________________________
> Spglib-users mailing list
> Spg...@li...
> https://lists.sourceforge.net/lists/listinfo/spglib-users



-- 
Atsushi Togo