relax SVN archive

Bug fix for the generic_fns.value.set() function.

The differentiation between a spin specific parameter and a global parameter was not working for
specific analyses with no spin specific parameters. The return_data_name() function was being used
to differentiate, which makes no sense. So instead the is_spin_param() function has been introduced
to allow the specific code to answer the question explicitly. Note that this will temporarily break
many things.

Changed the 5-state xz system test to start at the true parameter values.

This should speed things up considerably in this test.

Bug fix for the 5-state xz-plane system test.

The reduced tensors were incorrect. They were the tensors for the first state, not the averaged
tensors! They have been replaced by the correct ones and the number of states set to 5.

Wrote a forth unit test for the maths_fns.n_state_model.func() optimisation target function.

Wrote a third unit test for the maths_fns.n_state_model.func() optimisation target function.

Wrote another unit test for the maths_fns.n_state_model.func() optimisation target function.

Bug fix for the unit test of the maths_fns.n_state_model.func() optimisation target function.

There was a mix up between Sxz and Syz between the reduced and full tensors!

Bug fix for the maths_fns.n_state_model.func() optimisation target function.

The probability for state N was still being incorrectly calculated!!!

Bug fix for the maths_fns.n_state_model.func() optimisation target function.

The dot product is now used for the matrix multiplication R^T.X.R.

Expansion of the docstring of the unit test of maths_fns.n_state_model.func().