relax SVN archive

Names of buttons are enhanced and m2-m6 and final calculation is controlled by local_tm variable.

This is the disable_models_patch file (http://gna.org/support/download.php?file_id=7720) attached to
http://gna.org/task/?6847#comment19.

This is for feedback of successful local tm calculation.

Removed the print out from _overfit_deselect_dummy().

This was totally unnecessary!

The number of elements in the atomic position arg is now checked to be 3.

The calculate() method now uses _target_fn_setup() and stores the chi2 in the cdp.

Shifted the target function initialisation into _target_fn_setup().

This will be shared by the calculate() method.

Merged revisions 10240 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3

........
r10240 | bugman | 2010-01-15 15:35:22 +0100 (Fri, 15 Jan 2010) | 5 lines

Bug fixes for the code validator script.

The 4 character spacing check was faulty.
........

Bug fixes for the code validator script.

The 4 character spacing check was faulty.

Merged revisions 10117,10157,10178-10188,10197,10204-10205 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3

........
r10117 | bugman | 2010-01-06 18:57:55 +0100 (Wed, 06 Jan 2010) | 6 lines

Modified the IO redirection methods.

The original IO streams are stored and then restore instead of sys.__stdout__ and sys.__stderr__.
This might help in the multi-processor branch where IO redirection has gone haywire.
........
r10157 | bugman | 2010-01-08 10:57:41 +0100 (Fri, 08 Jan 2010) | 6 lines

Big bug fix for the model-free grid search!!!

The lower and upper bounds were being scaled iteratively for each spin causing the values to be
ridiculously bigger and bigger. This caused a total failure of the grid search.
........
r10178 | bugman | 2010-01-12 13:20:36 +0100 (Tue, 12 Jan 2010) | 3 lines

Modified the import of euler_zyz_to_R to allow easy switching later on to a different notation.
........
r10179 | bugman | 2010-01-12 13:21:01 +0100 (Tue, 12 Jan 2010) | 3 lines

Copyright notice update.
........
r10180 | bugman | 2010-01-12 13:32:12 +0100 (Tue, 12 Jan 2010) | 3 lines

Renamed all euler_ijk_to_R() functions to euler_to_R_ijk().
........
r10181 | bugman | 2010-01-12 13:35:04 +0100 (Tue, 12 Jan 2010) | 3 lines

Ordered the euler angle conversion functions.
........
r10182 | bugman | 2010-01-12 13:40:33 +0100 (Tue, 12 Jan 2010) | 3 lines

Fixes for all the euler angle to rotation matrix unit tests.
........
r10183 | bugman | 2010-01-12 13:42:00 +0100 (Tue, 12 Jan 2010) | 3 lines

Fix for reverse_euler_zyz() due to the function renaming.
........
r10184 | bugman | 2010-01-12 14:22:09 +0100 (Tue, 12 Jan 2010) | 3 lines

Created a unit test to check all hard-coded conversion functions for all 12 static axis rotations.
........
r10185 | bugman | 2010-01-12 14:22:52 +0100 (Tue, 12 Jan 2010) | 3 lines

Fix for the euler_to_axis_angle_xzx() function.
........
r10186 | bugman | 2010-01-12 15:10:54 +0100 (Tue, 12 Jan 2010) | 5 lines

Improvements to the test_euler_cycle_1() unit test.

This includes verbose printing of data structures and checking of the rotation matrix at each cycle.
........
r10187 | bugman | 2010-01-12 15:36:01 +0100 (Tue, 12 Jan 2010) | 5 lines

Added 12 unit tests for all notations of R -> Euler -> R.

This uses a random rotation matrix and sees if the matrix is returned.
........
r10188 | bugman | 2010-01-12 15:39:31 +0100 (Tue, 12 Jan 2010) | 6 lines

Important bug fix for the R_to_euler() function.

The special angle wrapping when beta was negative is only for the ijk Euler notation when i == k!
When i != k, another type of wrapping will be required.
........
r10197 | bugman | 2010-01-13 15:15:37 +0100 (Wed, 13 Jan 2010) | 3 lines

Removed the alpha Euler angle from the isotropic code frame order model.
........
r10204 | bugman | 2010-01-13 17:33:24 +0100 (Wed, 13 Jan 2010) | 5 lines

Bug fix for the frame order isotropic cone target function.

The optimised rotation matrix for the alignment tensor orientations was being applied in reverse!!!
........
r10205 | bugman | 2010-01-13 19:21:19 +0100 (Wed, 13 Jan 2010) | 6 lines

Reverted r10204 as this was a not a bug and the revision introduced a bug!!!

The command used was:
svn merge -r10204:10203 .
........

Bug fix for the color_code_noe() method.

This was introduced by the breaking up of patch_II. The structure is now called
self.structure_r21_copy_1_copy.

A number of improvements to the final results analysis.

This corresponds to one component of the patch_II file attached to https://gna.org/task/?6847 by
Michael Bieri (https://gna.org/users/michaelbieri).

This commit is the last of those from patch_II. This patch has now been almost completely committed
to the branch in a series of individual commits (see r10206 to r10234,
https://mail.gna.org/public/relax-commits/2010-01/threads.html). The rejected changes are listed at
https://mail.gna.org/public/relax-devel/2010-01/msg00016.html.