relax SVN archive

The exact chi2 values are now checked in all the frame order system tests.

Some of these values are obviously incorrect, and will be fixed later.

The frame order system test scripts can now operate as stand alone scripts.

Fix for the quasi-random numerical integration of the torsionless frame order model.

Fix for the frame order cone PDB creation for the models with the cone_s1 order parameter.

The cone angle is no longer stored as cdp.cone_theta, so is now generated on the fly.

The theta and phi Sobol angles are now correctly unpacked.

The phi angle sampling in the iso cone frame order models has been reintroduced.

Fixes for all of the free-rotor frame order models - the sigma angle needs to be sampled!

Set the variable to cut the number of Sobol integration points down to 50 from the default of 200000.

The flag for creating the cone PDB file is now also set in all scripts.

All of the non-quadratic integration frame order target functions have been converted to quasi-random.

This completes the update from Monte Carlo to quasi-random numerical integration.

Converted all of the PCS numerical integration methods to use the quasi-random Sobol' sequence data.

This is to convert from Monte Carlo to quasi-random numerical integration.