Re: Possible bug in relaxRe: Possible bug in relaxHi, The bug is fixed, for now. This is only for spin selections though so if you see this appear anywhere else, it would be useful to report it so I can fix unicode problems in other parts of the program. Cheers. To get the new code, you need the subversion program: $ svn co svn://svn.gna.org/svn/relax/1.3 relax-1.3 You may also need to install the repository version of minfx package as well: $ svn co svn://svn.gna.org/svn/minfx/trunk minfx Regards, Edward 2009/12/2 Olivier Serve <olivier.serve@xxxxxxxxx>:
Hi, As promised, here is a tar file containing original data, scripts and results (I also wrote the output of modsel.py in modsel-error.out and remove some results file from relax-fit, too big). To note, I am using relax 1.3.4 on a RHEL 5.3 linux. I did not know that u was prefix for Unicode, in that case, the problem lies most probably somewhere. I did not have time to test Michael solution, but it goes in the same direction of an encoding problem. Thanks a lot, Olivier Serve Le 1 déc. 2009 à 23:18, Edward d'Auvergne a écrit :Hi, Ok, I'll wait. I've noticed that u'u' is actually the 'u' character in Unicode. I'll definitely need more info to track this one down. The key might be in the results files from the mf_multimodel.py script. Cheers, Edward 2009/12/1 Olivier Serve <olivier.serve@xxxxxxxxx>:Dear Edward, Thank you for answering so quickly. I will send you all details tomorrow (It is night time for me now). But what I can say in the mean time, is that : "RelaxError: The selection string u'u' is invalid" is the full error message, quite short indeed, it took me long time to find this. The modsel.py is almost same, I just use model 0 to 5, instead of 0 to 9. But as I wrote in my previous mail, I am not at all sure that my solution is pertinent. I'll to send you a tar file tomorrow with all details. Have a nice day, Bye, Olivier Serve Olivier Serve Le 1 déc. 2009 à 22:40, Edward d'Auvergne a écrit :Hi, I have a feeling that the problem lies elsewhere. This change would probably be fatal - are the models correctly selected? Could you cut and paste the full error message? And maybe attach the modsel.py script? Without this, I'm not sure where the failure point is so I don't know what the problem could be or what the cause is. Cheers, Edward 2009/12/1 Olivier Serve <olivier.serve@xxxxxxxxx>:Hi all, I wrote possible bug in my title, because I am not sure yet it is not my mistake. Here is my problem : I ran smoothly all the following scripts : noe.py, relax-fit.py, mf_multimodel.py. But when I ran the modsel.py, I got the following error message : RelaxError: The selection string u'u' is invalid. (The name of the molecule begins by 'u' (ubiquitin)) I tracked down the error (very deep) into relax : modsel.py -> prompt/model_selection.Modsel -> generic_func/model_selection.select->specific_fns/model_free/main.duplicate_data->generic_fns/sequence.generate->generic_fns/mol_res_spin.return_spin at line 1879. The code is the following : (selection is a selection string : #<mol_name> :<res_id>[, <res_id>[, <res_id>, ...]] @<atom_id>[, <atom_id>[, <atom_id>, ...]],) if type(selection) == str: selection = [selection] select_obj = [] for i in range(len(selection)): select_obj.append(Selection(selection[i])) I think the problem here is that there is an iteration over a selection string (something like #ubi....), so when the Selection function sees 'u' alone, it does not mean anything and it raises the error. I turn the code into : select_obj=Selection(selection) And change the code accordingly a few lines after. The script seemed to have run smoothly afterward. If this is a bug, I'll be glad to open a bug report. If not, well, I'll be glad to know what I am doing wrong. I wish all of you a nice day, Best regards, Olivier Serve _______________________________________________ relax (http://nmr-relax.com) This is the relax-users mailing list relax-users@xxxxxxx To unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users_______________________________________________ relax (http://nmr-relax.com) This is the relax-users mailing list relax-users@xxxxxxxTo unsubscribe from this list, get a password reminder, or change your subscription options, visit the list information page at https://mail.gna.org/listinfo/relax-users