relax SVN archive

Modified generate_spin_id_data_array() so that column numbering starts at 1.

Created the read_spin_data_file() generator function.

This is designed to replace quite a few inconsistent functions throughout the generic function code
base. It reads a columnar formatted file with a spin per line.

Fixed the call from the prompt to the generic fns.

Changes to the pcs.read() and rdc.read() user function interfaces.

Columns are now numbered from 1 - note this will break many things! The spin_id_col is now accepted
and the spin_id arg is explained.

The docstrings have also been modified all over these user function classes to replace
'identification string' with 'ID string'.

Removed some empty lines.

Added the spin_id_col arg to the rdc.read() user function.

This allows 2 different formats for spin identification - spin ID string or mol name, res name and
num, and spin name and num.

Renamed the 'grace.py' sample script to 'mf_grace_S2_vs_te.py'.

Added a script for creating Molmol macros to represent the model-free motions.

Fix for the system test of Art Palmer's Modelfree4 program for the new __deselect_spins() fn.

Bug fix: Added the function __deselect_spins() to remove spins with too little data.

This fixes bug #14428 (http://gna.org/bugs/?14428) reported by Pierre-Yves Savard <pierre-yves dott
savard at bcm dot ulaval dot ca>.

The new function is almost identical to that used in the model-free analysis performed by relax in
that spins with no or too little data, or those which would be over-fit are removed from the
analysis.