relax SVN archive

Updated the manual and changed the version number to 1.2.5.

relax version 1.2.5.

This is mainly a bugfix release. A large number of bugs have been eliminated including bug #5977,
bug #6288, bug #6383, bug #5806, bug #6384, bug #6388, bug #6389, and bug #5746. These are
described in more detail below. To fix a few of these bugs large disruptive changes were required.
This included the change of the local tm parameter name from 'tm' to 'local_tm' and the
simplification of the model-free model_statistics() function and modification of the model-free
calculate() function. Therefore regressions and new bugs may appear in this release, most likely
when the local tm parameter is encountered or during model selection.

The full list of changes is:

Features:
* The data analysis chapter has been split into four: The NOE, relaxation curve-fitting,
model-free, and reduced spectral density mapping chapters.
* A section called the 'The core design of relax' has been added to the the development
chapter of the manual.
* The 'Values, gradients, and Hessians' chapter of the manual is now complete. These are
all the equations used during optimisation.
* The Hessian kite diagram and the gradient construction diagram have been added to the
'Values, gradients, and Hessians' chapter of the manual.

Changes:
* The distribution archives now unpack to the directory 'relax-x.x.x' rather than just'
'relax', where x.x.x is the version number.
* All chapters of the manual have been edited and proofed.
* The Sparky and XEasy peak intensity reading functions have been split to allow the easy
addition of new peak intensity file types.
* Test of AIC model selection between diffusion tensors added to the test suite. (To catch
bug #5746)
* Changing of the local tm parameter name of 'tm' to 'local_tm' to distinguish it from the
global tumbling time 'tm'. (Preparation for fixing bug #6384)

Bugfixes:
* Fix of bug where self.relax.data.select_sim was treated as a list. (bug #5977)
* The value.copy() function now handles the value of None. (bug #6288)
* The global flag is now set by the relax_data.delete() function. (bug #6383)
* The model-free model_statistics() function now tests for the presence of relaxation data.
(bug #5806)
* Sorting of global from local parameters in 'generic_fns/value.py' simplified and fixed by
changing the parameter name 'tm' to 'local_tm'. (bug #6384)
* Model-free parameters now set in the OpenDX mapping test of the test-suite. (bug #6388)
* Failure of model elimination because of a 'local_tm' and 'tm' mix up. (bug #6389)
* Modification of the model-free calculate() function and elimination of the 'combined' flag
from the self.model_statistics() function. (bug #5746)

Added an entry to the changes file for relax version 1.2.5 (r2490-r2518).

Modified the release checklist to include updating the 'CHANGES' file.

Finally, a real fix to bug #5746.

The bug report is located at https://gna.org/bugs/?func=detailitem&item_id=5746 and is discussed in
the thread starting at https://mail.gna.org/public/relax-devel/2006-04/msg00018.html.

The fix follows from an idea in the post located at
https://mail.gna.org/public/relax-devel/2006-04/msg00021.html. Firstly the model-free calculate()
function was modified so that when self.param_set is equal to 'all' or 'diff' (global statistics)
then the chi-squared value calculated is summed and placed in the global data structure
self.relax.data.chi2 rather than the residue specific data structure.

This change allowed the 'combined' flag to be eliminated from the self.model_statistics() function
of 'specific_fns/model_free.py'. This flag was the source of all the problems and it's removal
allowed chunks of obscure code to be axed. The statistics only need to refer to the self.param_set
value to determine which stats should be returned.

Fix to bug #6389.

The bug is located at https://gna.org/bugs/index.php?func=detailitem&item_id=6389.

The fix was simply to change a single line so that the data structure 'tm' is referred to rather
than 'local_tm'.

Fix to bug #6388.

The bug is located at https://gna.org/bugs/?func=detailitem&item_id=6388.

The problem was simply that the model-free parameters were not set in the OpenDX mapping test of the
test-suite.

Fix to bug #6384.

The bug is located at https://gna.org/bugs/index.php?func=detailitem&item_id=6384.

The function for sorting global from local parameters in the generic_fns/value.py file has been
simplified thanks to the change of the previous revision whereby the local tm parameter is now
referred to as 'local_tm' rather than 'tm'. Hence no algorithms are required to distinguish
between the global 'tm' parameter and the local 'tm' parameter.

Changing of the local tm parameter name of 'tm' to 'local_tm'.

This is a large scale change in preparation for the fixing of bug #6384
(https://gna.org/bugs/?func=detailitem&item_id=6384). The distinction between the global tumbling
time 'tm' and the residue specific 'local tm' parameter has always been problematic. With this
change all these problems should disappear. However as the changes are large, new bugs are almost
guaranteed to appear.

Fix to bug #5806.

The bug is located at https://gna.org/bugs/?func=detailitem&item_id=5806.

To remove the bug, tests for the presence of relaxation data have been added to the
model_statistics() functions. This may cause future bugs!