Provide Transformation Matrix as output of the code

[giovannipizzi]

This is being worked in branch fix_transformation_matrix

[giovannipizzi]

Missing things:

• implement triclinic
• check that tests work check the conventional matrix
• add also trasform to conventional?

[tkotani]

It is nice if seekpath can find the transformation matrix.

FYI In mathematics, a set of the three vectors which defines a lattice is called as the Z-basis. Two Z-basis, which defines the same lattice, is connected by an unimodular matrix. General theory of unimodular matrix is too difficult for me. So I wrote a simple code to find the unimodular matrix for my purpose. But it is helpful if you include a code in seekpath.

[iorin0217]

I think the "transformation_matrix" has not sufficient information about converting the input structure to the seekpath output conventional structure, and the "std_rotation_matrix" is needed. Do you agree with me?

[giovannipizzi]

Some notes from the TODO.txt file that I will remove from the repo:

• If possible, it would be nice to have the transformation matrix. NOTE: this is being worked in the branch 'fix_transformation_matrix'. a. for non-aP lattices, this should be already possible. Check if it is sufficient to combine the transf_matrix with the P matrix b. for aP lattices, we should probably check if we can get the matrix out of Niggli-reduce. This should be easy to obtain. Then, dd to the output:
  • the transformation matrix from user to standard, and from standard to primitive (and the combination)
  • a warning/flag to say if the cell has changed in the process (probably it will almost always change?)

[giovannipizzi]

Thanks @jaemolihm - I merged #97 now, this should indirectly also address this issue that I will therefore close (I also deleted my own old branch)