These changes are part of the solution for bug #7979 (https://gna.org/bugs/?7979).
The 'self._calc_x()' methods of the DiffTensorElement class, where x is a diffusion tensor
parameter, vector, or matrix, have been shifted from the 'data' directory to the diffusion tensor
code of the 'generic_fns' directory. The methods have also been converted into module functions.
The 'self._update_object()' method calls in the 'self.__setattr__()' method have been modified so
that the type of diffusion is passed into the necessary 'calc_x()' functions.
