relax SVN archive

Preparation of the other functions for automatically generating diffusion tensor data objects.

All the other functions (vs. the previous commit) have been prepared to return either the real value
or the MC sim data.

Simplification of the functions for automatically generating diffusion tensor data objects.

The following functions have been coalesced into their non-simulation counterparts:
self._auto_object_Dpar_unit_sim()
self._auto_object_Dx_unit_sim()
self._auto_object_Dy_unit_sim()
self._auto_object_Dz_unit_sim()

Added automatically generated MC simulation objects to the diffusion tensor data structure.

The class implementing the object is called 'DiffAutoSimArrayObject'. The following functions have
been addded to generate the objects:
self._auto_object_Diso_sim()
self._auto_object_Dpar_sim()
self._auto_object_Dper_sim()
self._auto_object_Dpar_unit_sim()
self._auto_object_Dx_sim()
self._auto_object_Dy_sim()
self._auto_object_Dz_sim()
self._auto_object_Dx_unit_sim()
self._auto_object_Dy_unit_sim()
self._auto_object_Dz_unit_sim()

These are simply copies of the non-simulation functions with the added argument 'i' for selecting
the simulation.

Redesign of the diffusion tensor data structure 'self.relax.data.diff'.

All of the Numeric data structures returned by the 'self.__getattr__()' function have been shifted
into their own instances of a new objected defined by the DiffAutoNumericObject class. This new
class when instantiated expects as an argument a function which will return a Numeric array, matrix,
etc. Using this function, the object currently implements the functions:
__add__(), implementation of addition.
__getitem__(), selecting individual elements of the array.
__iter__(), for looping over the array.
__len__(), the length of the array.
__mul__(), implementation of multiplication.
__repr__(), the 'official' string representation.
__setitm__(), throw a RelaxError if modification of the structure is attempted.

To enable the automatically generated objects, the code from 'self.__getattr__()' has been shifted
into the following 'hidden' functions of the DiffTensorElement class:
_auto_object_Dpar_unit()
_auto_object_Dx_unit()
_auto_object_Dy_unit()
_auto_object_Dz_unit()
_auto_object_rotation()
_auto_object_tensor()
_auto_object_tensor_diag()

Using the new 'self.atomic_data' data structure, the tensor PDB file is now created correctly.

The centre of mass atom, diffusion tensor representative atoms, and the atoms representing the axes
now belong to separate residues, COM, TNS, and AXS respectively. A large number of changes were
necessary to fix the breakages introduced by the change in 'self.atomic_data'.

The atom number in 'self.atomic_data' was one out.

The atom name for the tensor atoms was being set to the 'atom_id' string. This was incorrect as the
protons should have been named H1, H2, H3, etc. The 'atom_id' string was too long for the HETATM
record.

In the 'self.write_pdb_file()' function, the elements of the 'het_data' structure used for
non-standard residues has been documented in the docstring. This structure is now created prior to
the creation of the HET, HETNAM, and FORMUL records.

All the other records (HETATM, TER, CONECT, MASTER, END) are now correctly created.

The structure 'het_data' was being accessed as a Numeric array when it was just an array of arrays.

Fixed a problem with 'self.atom_add()' where 'res_num' was being accessed as 'residue_num'.

Modified the 'self.write_pdb_file()' function to properly handle the HET records.

The function 'self.create_tensor_pdb()' has been modified for the new 'self.atomic_data' structure.

The centre of mass, vector distribution, and axes have been turned into individual residues.

Readjustment of 'self.atomic_data' and the addition of epydoc fields to 'self.atom_add()'.

The record name has been removed as this can be determined by the 'self.write_pdb_file()' function.