relax SVN archive

Created the user function 'molmol.ribbon()'.

This function applies the Molmol ribbon style and is equivalent to clicking on 'ribbon' in the
Molmol side menu.

Implemented the 'molmol.command()' user function.

The user function did previously exist, but it hadn't been implemented in the generic code. The
user function was calling the wrong function 'self.write' in 'generic_fns/molmol.py'.

Improvements to the printed out messages of the 'pdb.create_tensor_pdb()' user function.

Changed the CONECT record algorithm (for the tensor PDB) to allow more than four bonded atoms.

Multiple CONECT records are now generated for an atom if it is bonded to more than 4 other atoms.

Sorted the HETATM and CONECT records by atomic number in the tensor PDB file.

Because of the diffusion tensor symmetry, the axes where extended in both the pos and neg directions.

Improved the docstring of the user function 'pdb.create_tensor_pdb()'.

I've made it more clear that the units of the diffusion tensor are rates, not times. This is to
help users to not be confused with the fact that being parallel to the long axis of the tensor means
that the bond is moving faster!

Fixed the incorrect value of the isotropic Diso parameter.

The code for setting the diffusion tensor parameters was out of date (compared to the spheroidal and
ellipsoidal code). Diso was being explicitly set, whereas it should have been calculated on the
fly. In addition it was being calculated incorrectly with the equation Diso = 6/tm rather than
Diso = 1/(6tm)!!!

Modified and improved the comments about the uniform vector distribution.

The terms longitudinal and latitudinal have been added.

Fixed the transposed spheroidal rotation matrix.

The spheroidal rotation matrix returned from the data structure 'self.relax.data.diff[run].rotation'
was incorrect. Instead of the true rotation matrix, the transpose of the matrix was being returned.
As the spheroidal rotation matrix was transposed whereas the ellipsoidal rotation matrix was correct
(the spherical rotation matrix was the identity matrix, so no problem there), the maths for
generating the PDB geometric object representing the diffusion tensor was causing problems. Now all
three geometric objects - sphere, spheroid, and ellipsoid - should be correctly generated.