relax SVN archive

Included the Eufrak font in the model-free optimisation protocol diagrams of the relax manual.

The Eufrak font, which was converted from the font used in the LaTeX eufrak package, was missing
when the images were converted from the SVG files.

WORK IN PROGRESS! addition of methods to find paths to unit test and
system directories. Plus functions to convert module names to paths.
Also added comments.

Removing the obsolete apply branch

Bug fix for the test suite 'results.read()' user function test.

This test was failing, again due to the inability to load the program state.

Bug fix for the failure of the test suite with the loading of Sparky peak heights.

This issue was that the exponential curve fitting saved state file 'rx.save.bz2' could no longer be
loaded. This is because of the complete redesign of 'self.relax.data' which has occurred in this
branch.

Ported r2837 from the 1.2 line.

The command used was:
svn merge -r2836:2837 svn+ssh://bugman@.../svn/relax/1.2

This fixes a bug in the Sparky peak intensity code.

Ported r2837 from the 1.2 line.

The command used was:
svn merge -r2836:2837 svn+ssh://bugman@.../svn/relax/1.2

This fixes a bug in the Sparky peak intensity code.

Fixed a newly introduced bug in the reading of Sparky peak intensities.

The bug was caused by r2798 (https://mail.gna.org/public/relax-commits/2006-11/msg00096.html). The
expected proton and heteronucleus names returned by the 'self.intensit_xxxxx()' functions had been
swapped but the Sparky intensity function had not been updated.

Ported everything from r2732 to r2833 of the 1.3 line. This brings this branch into sync with 1.3.

The command used was:
svn merge -r2731:2833 svn+ssh://bugman@.../svn/relax/1.3

Ported r2831 to r2833 from the 1.3 line.

The command used was:
svn merge -r2830:2833 svn+ssh://bugman@.../svn/relax/1.3

These changes to the relax user manual include an update to the core design diagram and three flow
diagrams describing the steps of model-free analysis (the technique of Mandel et al., 1995, using
AIC model selection and model elimination, and the new model-free optimisation protocol of
'full_analysis.py'.