[Spglib-users] get_symmetry basis

[Anders C. Riis-J. <ac...@fy...>]

I am using the built-in function get_symmetry from spglib, and I seem to have some issues with this. In this context I have a question.
I will here give an example showing my problem:

If I define the unit cell for graphene:
cell = np.array([[a, 0, 0,], [-a/2, a*sin(60*pi/180), 0], [0, 0, 5*a]])
and put the carbon atoms in the (scaled) positions [0,0,0] and [0, 2/3, 1/3]
I then run spglib.get_symmetry on the system (with the default tolerance). I get a set of symmetry transformations (I understand each transformation is a combination of a matrix multiplication and a translation). I want to have the symmetry matrices (M) in cartesian coordinates and therefore make the following basis transformation:
inverse(cell).M.cell
And here I would for example expect to find the standard 60 degrees rotation matrix, however this I dot not get. I therefore realize that the original symmetry matrices might not be written in the lattice basis as I expected.

Which basis is used and what is the proper transformation to get to cartesian coordinates?

Any answer will be greatly appreciated.

Best regards
Anders Riis-Jensen