relax SVN archive

Added the 'NS 2-site' model to the relax_disp.select_model user function frontend.

This is the model of the numerical solution for the 2-site Bloch-McConnell equations. It originates
as optimization function number 1 from the fitting_main_kex.py script from Mathilde Lescanne, Paul
Schanda, and Dominique Marion (see http://thread.gmane.org/gmane.science.nmr.relax.devel/4138,
https://gna.org/task/?7712#comment2 and https://gna.org/support/download.php?file_id=18262).

This commit follows step 2 of the relaxation dispersion model addition tutorial
(http://thread.gmane.org/gmane.science.nmr.relax.devel/3907).

Added the 'NS 2-site' model to the specific_analyses.relax_disp.variables module.

This is the model of the numerical solution for the 2-site Bloch-McConnell equations. It originates
as optimization function number 1 from the fitting_main_kex.py script from Mathilde Lescanne, Paul
Schanda, and Dominique Marion (see http://thread.gmane.org/gmane.science.nmr.relax.devel/4138,
https://gna.org/task/?7712#comment2 and https://gna.org/support/download.php?file_id=18262).

This commit follows step 1 of the relaxation dispersion model addition tutorial
(http://thread.gmane.org/gmane.science.nmr.relax.devel/3907).

Fix for the Relax_disp.test_hansen_cpmg_data_auto_analysis system test for model name change.

This is for the change from the 'CR72' model to 'CR72 red' model.

Improvements for the write_results() method of the dispersion auto-analysis.

The parameter value and Grace files are now correctly created for all the recent models.

Added support for model nesting in the relaxation dispersion auto-analysis.

This involves copying the parameters from the simpler nested model rather than performing a full
grid search. This is currently used to handle all models with R20A and R20B parameters where a
simpler model with the single R20 parameter is optimised first.

Small speed up for the lib.dispersion.cr72 module for the R20A != R20B case.

Replicated calculations have been minimised.

Documentation fix for the lib.dispersion.cr72 module.

Better printouts of the R20A and R20B parameters at the end of minimisation.

Fix for the dispersion specific loop_parameters() method for the R20A and R20B parameters.

The frequency index is now correctly returned for these and the R20 parameter.

Expanded the CR72 model to handle both the R20A and R20B parameters.

The 'CR72 red' model now functions as the old CR72 model where R20A = R20B = R20. All parts of the
code have been modified. The lib.dispersion.cr72.r2eff_CR72() function has been expanded to support
the full Carver and Richards 1972 equations, dropping back to the simplified form if R20A = R20B.