On Thu, Jan 22, 2009 at 12:30 PM, Pavel Kaderavek <kada@xxxxxxxxxxxxx> wrote:
Hi, I have not heard about wiki (sorry I am very low IT educated person), but if I understood in a right way I can place those files on the web-site to enable you to download them.
Sorry, for some reason I assumed your website was a wiki. I think I'm mixing up different development efforts, so please ignore that.
So the files are: https://www.ncbr.chemi.muni.cz/~kada/maths_fns_branch.tar
This code I'll look at a little later.
https://www.ncbr.chemi.muni.cz/~kada/relax-1.2.10_cst.tar
This code I'll incorporate into the new branch I've created called cst_1.2.10. The branch was copied from: svn+ssh://bugman@xxxxxxxxxxx/svn/relax/tags/1.2.10/ and is located at: svn+ssh://bugman@xxxxxxxxxxx/svn/relax/branches/cst_1.2.10 To get this code, use the command: svn co ssh://svn.gna.org/svn/relax/branches/cst_1.2.10 I'll try to incorporate your code in one massive block (this is ok in this dead-end, yet reference branch), but I will need help. I'll get back to you and Petr when I have problems, and when I need to know the origin of new files.
-the first one is only maths_fns directory for 1.3 version, still with the abundant mf_csa.py file which was not incorporated into mf.py. The content was checked only by "eyes", but I am checking it using that validator just now.
I'll look at this code, but again I will need help to get this incorporated.
-the second one is the old "dirty" version (I cut out only the most evident abundant comments)
For the purpose of reference, it doesn't matter that it is dirty. Is this copy your functional version?
So far, I tested the code only in the way that I calculated the values of relaxation times by script in octave. Then I fit these values in relax and checked, if the result of the fitted parameters is the same as those used for the calculation. For the 1.3 line I will also try to use that test system you suggest, but now I do not think I am not at that state yet (In addition I have to admitt I do not understand what really happen there, I have just looked at it briefly).
Ah, the test should come before the code. This will help with the design that Petr is doing (well, has done but needs to adapt). But if you run a relax script and check it's results (compared to the octave results), this can be incorporated unmodified and be run automatically. That will be a perfect system test. If you or Petr could help me in setting this test up, before making the changes to the 'cst' branch, that would be much appreciated. Regards, Edward