relax SVN archive

Updated the frame order test model optimisation statistics to use only the models with samplings of 1e6.

The old Monte Carlo numerical integration numbers have been eliminated.

Updated the free rotor isotropic cone frame order test data to a higher accuracy.

Now 1e6 samplings are used, rather than 1e4.

Updated the frame order system tests for the new test data.

Redesigned the test data for the frame order models.

A base script for all models, 'generate_base.py', is now used to create all of the structural data
and to back-calculate the RDC and PCS data. Rather than loading and rotating one structure per
state, the script now rotates the spin specific vectors and atomic positions. This allows sampling
of up to 1e6 states to create very accurate debugging data.

All of the README files have been removed as these are repetitive. The back_calc.py scripts and
save files have also been deleted as these are no longer used. Two new files are created and have
been added to the repository. The file 'axis_system' has a copy of the motional axis system and the
full precision Euler or spherical angles for the system. The file 'frame_order_matrix' contains the
exact values of the frame order matrix and will be used for debugging.

The frame order system tests are now printing out all parameter values on failure.

Merged revisions 15334-15335 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3

........
r15334 | bugman | 2012-02-15 10:58:37 +0100 (Wed, 15 Feb 2012) | 5 lines

Created the SplitIO.flush() method so that a flush() call now works with the -t and -l command line args.

This simply calls the flush() methods of both streams.
........
r15335 | bugman | 2012-02-15 11:00:24 +0100 (Wed, 15 Feb 2012) | 3 lines

The float arg checks now check against all the numpy float types (float16, float32, float64, float128).
........

The print_frame_order_2nd_degree() function can now write to a file.

This can then be used for debugging in the test suite.

The float arg checks now check against all the numpy float types (float16, float32, float64, float128).

Created the SplitIO.flush() method so that a flush() call now works with the -t and -l command line args.

This simply calls the flush() methods of both streams.

Updated the frame order test model optimisation statistics tables.