relax SVN archive

Added hybrid run support for the 'pdb.create_tensor_pdb()' user function.

The function 'self.create_tensor_pdb()' in the file 'generic_fns/pdb.py' has been modified to loop
over an array of runs. This array is normally the single standard run. However when a hybrid run
is encountered, the runs of the hybrid are looped over.

The chain ID letter is now taken from the 'string.ascii_uppercase' string. The letter is extracted
using the run index.

All atom IDs have had the chain ID appended to the end to make then unique when multiple chains
exist.

The data structure 'self.atomic_data' of 'generic_fns/pdb.py' has been modified to include the
record name (ATOM, HETATM, or TER) at position 1. Hence all other elements have been shifted by
one. This is to allow multiple TER records one at the end of each chain, each having their own atom
number.

The 'het_data' array in 'self.write_pdb_file()' is no longer sorted as it already is sorted because
of the sorting of atomic data arrays. This allows the HET records to be sorted first by chain ID
then by residue number.

The HETNAM and FORMUL records are now only created if the residue has not been previously
encountered. This avoids duplicated records.

All the atomic coordinate records (ATOM, HETATM, or TER) are now created sequentially by looping
over the atomic data sorted by atom number.

Due to the failure of 'state.save()', i.e. can't pickle, 'data.py' has been split up and redesigned.

A new directory 'data' has been created to all the modules for constructing 'self.relax.data'. The
file 'data.py' has been moved into the directory as 'data/main.py' and duplicated into three more
files.

All the 'self._auto_object*()' methods for automatically creating the diffusion tensor data
structures have been shifted into the module 'data.diff_tensor_auto_objects'. These have been
converted into functions and all references to 'self' have been eliminated. The local namespace of
'self.relax.data.diff[run]' is now passed into these functions as the first argument to allow its
objects to be accessed. This change fixes the pickling problem and restores the state saving and
loading functions.

In the module 'data.diff_tensor', all that remains of the original 'data.py' file is the data
containers 'DiffTensorData' and 'DiffTensorElement' as well as the 'DiffAutoNumericObject' and
'DiffAutoSimArrayObject' automatically generated objects. The first two classes assign the
'auto_object*()' functions to specific instances of the last two classes which then call the
functions to obtain the data.

The data containers 'Element', 'SpecificData', 'Residue', and 'ResidueList' are located in the
module 'data.data_classes'.

All that remains in the original 'data.py' file ('data/main.py') is the basal data container 'Data'.

The files 'relax', 'prompt/init_data.py', and 'thread_classes.py' have been modified to import the
correct modules.

corrected Test_GetFloatClass so it works

ported r2768 (fix to bug #7641)

commands used
svn merge -r2767:2768 svn+ssh://varioustoxins@.../svn/relax/1.2

ported r2768 (fix to bug #7641)

commands used
svn merge -r2767:2768 svn+ssh://varioustoxins@.../svn/relax/1.2

fixed bug 7641 Infinite Chi2 warning

bug #7641 http://gna.org/bugs/?7641
https://mail.gna.org/public/relax-devel/2006-11/msg00078.html

ported r2634,r2636 & r2716 to the test-suite branch, corrected
test_float.py

commands used
svn merge -r2633:2634 svn+ssh://varioustoxins@.../svn/relax/1.2
svn merge -r2635:2636 svn+ssh://varioustoxins@.../svn/relax/1.2
svn merge -r2715:2716 svn+ssh://varioustoxins@.../svn/relax/1.2

these changes add the ieee-754 nan/inf code and nan catching from branch 1.3

reversed merge of ported r2634,r2636 & r2716 in r2717 to separate them
from the patches to float.py

https://mail.gna.org/public/relax-devel/2006-11/msg00053.html

Spell check of the PDB and PyMOL user functions and generic_fns code.

Added extra atoms to the 'AXS' residue of the tensor PDB representation for labelling.

Two or six extra atoms are added for the spheroid and ellipsoid axes respectively. These are set to
be nitrogen atoms to allow easy selection of just those atoms. They are also placed 3 Angstrom
outside of the geometric object so that the labels do not clash with the latitudinal and
longitudinal lines as well as the Monte Carlo simulation axes.

The code for generating the PDB chemical formula in 'self.write_pdb_file()' was modified to handle
nitrogen atoms.

The PyMOL user function 'pymol.tensor_pdb()' has been modified to use the new representation. The
final selection is also deleted at the end of the user function.