Identification Toolbox: improvements, bugfixes, license, preparation for getting rid of globals

Closes #1694 (closed), #1689 (closed) Improvements for minimal state space system using alternative to minreal function (I still need to provide the mex function from SLICOT, will do so soon or open an issue for this if I fail).

Please also merge into 4.6 branch.

changed milestone to %4.6

added bug identification / sensitivity analysis labels

assigned to @sebastien

added 1 commit

• 16674bb5 - Update license

Compare with previous version

added 9 commits

• d3e90a8d - 1 commit from branch Dynare:master
• 07c83748 - Fix #1694 (closed) by robust rank tolerance and histc
• e2f4099c - fix identification strength barplots for one parameter
• 84655550 - Updated code comments
• 6798b170 - construct list of fields needed from M_, options_, oo_
• 905a4543 - Better minimal state space handling and unit tests
• bc8a7a5e - Update license
• 6e211afe - Update license
• ec8f08e5 - Merge remote-tracking branch 'origin/identification_higher_order' into identification_higher_order

Compare with previous version

resolved all threads

added 7 commits

• c4f7c416 - Fix #1694 (closed) by robust rank tolerance and histc
• a62e69cf - fix identification strength barplots for one parameter
• 46c4dea5 - Updated code comments
• 1aa3dda4 - construct list of fields needed from M_, options_, oo_
• 5525a7c5 - Better minimal state space handling and unit tests
• aa0f278e - Update license
• e843ccbd - Update license

Compare with previous version

added 1 commit

• 3d51ee9e - Added unit tests for minimal state space system

Compare with previous version

enabled an automatic merge when the pipeline for 3d51ee9e succeeds

mentioned in commit a40cfedd

merged

How important is it to backport all those commits to 4.6? I'd like to be as conservative as possible.

I will for sure backports the bugfixes and the license updates.

For the other ones, I’m less sure. In particular, commit 1aa3dda4 does not apply directly, because in master I have changed the way MEX files are called (I think I could still be able to manually merge it, though).

Hi, actually it would be good to backport all commits to 4.6 as they provide many more comments that explain much better what I am doing in these functions and also fix bugs. I am not sure, I understand your comment about the "way MEX files are called" as I did not change any call to a mex file in those commits (and also not in 1aa3dda4)?! 1aa3dda4 is just a restructuring of the code without changing any computations.

I mean the changes to your files that I made in commit 7e770f69

This commit is only present on master, but was not backported to 4.6 since it is not a bugfix. So the two branches have diverged, and that implies that backporting 1aa3dda4 cannot be done automatically.

I backported them all. Hopefully I did not mess up with the commit that had to be backported manually.

Thanks for your work.

@wmutschl Unfortunately the new minimal state space system tests fail under MATLAB R2009b: see https://git.dynare.org/Dynare/dynare/-/jobs/13853

Do you think you can quickly provide a fix, or should I revert your last changes on 4.6?