relax SVN archive

Shifted the PCS only script into its own directory.

Created a script for the back-calculation of RDC and PCS data from the PCS only fit average structure.

Created a subdirectory for testing the frame order theory with just PCS data.

Added a script for testing the frame order analysis of the rotor model with only PCSs.

The align_id variable can now be added to the data store alignment tensor object.

Fix for the align_tensor.init user function for when the align_id arg is given.

The incorrect back-end function was being called.

The alignment ID is now set for the N-domain tensors in the frame order test data.

Fix for the frame order analysis - the parameter vector is now diagonally scaled after optimisation.

The scaling was not being performed so that the pivot position was incorrectly stored.

Shifted the multiplication by the PCS constant outside of the numerical integration.

This multiplication now only happens N times per function call, where N is the number of spins with
PCS values, rather than N times hundreds of thousands.

Shifted the rotation to the average position to outside of the numerical integration.

This now occurs once per function call, rather than hundreds of thousands of times.