Find a better way for cell transformation matrices

[RandomDefaultUser]

This is currently hard coded, but can be based on mathematical intuition.

[elcorto]

Maybe this helps:

• https://elcorto.github.io/pwtools/written/background/coord_trans.html
• https://elcorto.github.io/pwtools/generated/api/pwtools.crys.coord_trans.html#pwtools.crys.coord_trans

[RandomDefaultUser]

Thanks, I will have a look! I don't anticipate problems implementing this, I just haven't found the time doing it.

[elcorto]

Yeah, this is just dot in one direction and linalg.solve in the other, but can be tricky to get right. Once all tests pass, you are safe :)

[elcorto]

Speaking of tests:

• https://github.com/elcorto/pwtools/blob/master/pwtools/test/test_coord_trans.py
• https://github.com/elcorto/pwtools/blob/master/pwtools/test/test_crys_cell_tools_fortran.py
• https://github.com/elcorto/pwtools/blob/master/pwtools/test/test_structure.py#L131

... etc, grep for coord_trans in /path/to/pwtools/pwtools/test/*.py to find all examples.