Re: Launching the GUI from relax. -- December 01, 2009 - 11:06

Ah, right.  I found it.  The problem is that certain arguments are set
to None when they should have values.  Have a look at the code for the
prompt.minimise.Minimisation.minimise() method (the minimise() user
function).  Though the code is complex, you will see that default
values for func_tol, max_iterations, constraints, scaling, and
verbosity are set.  You will have to manually supply these defaults
with the same values as the minimise() user function defaults to, if
the GUI does not allow the user to change these properties yet.  This
should fix the problem.

Also, if you are making changes to the code, could you submit these as
patches, attaching them to the task?  If the patch is too big for me
to correctly review and comment on, that almost always results in a
rejection in a group developed software project.

Cheers,

Edward



2009/12/1 Michael Bieri <michael.bieri@xxxxxx>:
> Thank you!
>
> Now, there is another problem in minimisation:
>
> Traceback (most recent call last):
>  File "/home/michael/Programme/relax_bieri_gui/gui_bieri/relax_gui.py",
> line 2695, in exec_t2_1
>    make_tx(self.resultsdir_t11.GetValue(), t1_list, relax_times_t1_1,
> self.structure_t11.GetValue(), self.nmrfreq_value_t11.GetValue(), 1, 1,
> self.unresolved_t11.GetValue(), self, 1, global_setting, file_setting,
> sequencefile)
>  File
> "/home/michael/Programme/relax_bieri_gui/gui_bieri/res/calc_rx.py", line
> 121, in make_tx
>    minimise.minimise('simplex', scaling=False, constraints=False)
>  File
> "/home/michael/Programme/relax_bieri_gui/generic_fns/minimise.py", line
> 162, in minimise
>    minimise(min_algor=min_algor, min_options=min_options,
> func_tol=func_tol, grad_tol=grad_tol, max_iterations=max_iterations,
> constraints=constraints, scaling=scaling, verbosity=verbosity)
>  File
> "/home/michael/Programme/relax_bieri_gui/specific_fns/relax_fit.py",
> line 783, in minimise
>    results = generic_minimise(func=func, dfunc=dfunc, d2func=d2func,
> args=(), x0=param_vector, min_algor=min_algor, min_options=min_options,
> func_tol=func_tol, grad_tol=grad_tol, maxiter=max_iterations, A=A, b=b,
> full_output=True, print_flag=verbosity)
>  File "/usr/local/lib/python2.6/dist-packages/minfx/generic.py", line
> 373, in generic_minimise
>    raise NameError, "Simplex minimisation cannot be setup."
> NameError: Simplex minimisation cannot be setup.
>
>
> Hope that's also an easy one..
>
> Thank you.
> Michael
>
> Edward d'Auvergne schrieb:
> > Hi,
> >
> > The problem here is that I have updated and cleaned up the grid search
> > component of the minfx project.  relax has been modified to handle the
> > newer code.  Therefore for this to work, you need to check out a copy
> > of minfx from SVN:
> >
> > $ svn co svn+ssh://michaelbieri@xxxxxxxxxxx/svn/minfx/trunk
> >
> > preferably in a directory called minfx.  Note you should also type
> > 'svn up' in your branch to get all the updates that have been
> > happening there.
> >
> > Regards,
> >
> > Edward
> >
> >
> > 2009/12/1 Michael Bieri <michael.bieri@xxxxxx>:
> >
> > > Hi Edward
> > >
> > > I have another problem by including the GUI into relax. By executing Rx
> > > calculations, I get the following problem:
> > >
> > > Traceback (most recent call last):
> > >  File "/home/michael/Programme/relax_bieri_gui/gui_bieri/relax_gui.py",
> > > line 2695, in exec_t2_1
> > >    make_tx(self.resultsdir_t11.GetValue(), t1_list, relax_times_t1_1,
> > > self.structure_t11.GetValue(), self.nmrfreq_value_t11.GetValue(), 1, 1,
> > > self.unresolved_t11.GetValue(), self, 1, global_setting, file_setting,
> > > sequencefile)
> > >  File
> > > "/home/michael/Programme/relax_bieri_gui/gui_bieri/res/calc_rx.py", line
> > > 119, in make_tx
> > >    minimise.grid_search(inc=11)
> > >  File
> > > "/home/michael/Programme/relax_bieri_gui/generic_fns/minimise.py", line
> > > 185, in grid_search
> > >    grid_search(lower=lower, upper=upper, inc=inc,
> > > constraints=constraints, verbosity=verbosity)
> > >  File
> > > "/home/michael/Programme/relax_bieri_gui/specific_fns/relax_fit.py",
> > > line 404, in grid_search
> > >    self.minimise(min_algor='grid', lower=lower, upper=upper, inc=inc,
> > > constraints=constraints, verbosity=verbosity, sim_index=sim_index)
> > >  File
> > > "/home/michael/Programme/relax_bieri_gui/specific_fns/relax_fit.py",
> > > line 684, in minimise
> > >    results = grid(func=func, args=(), num_incs=inc, lower=lower,
> > > upper=upper, A=A, b=b, verbosity=verbosity)
> > > TypeError: grid() got an unexpected keyword argument 'num_incs'
> > >
> > >
> > > This occurs after the command: minimise.grid_search(inc=11)
> > >
> > > Do you have an idea what the problem is in this case?
> > >
> > > I think the problem is buried somewhere in specific_fns/relax_fit.py.
> > >
> > > Cheers
> > > Michael
> > >
> > > Edward d'Auvergne schrieb:
> > >
> > > > 2009/11/27 Michael Bieri <michael.bieri@xxxxxx>:
> > > >
> > > >
> > > > > Hi Edward
> > > > >
> > > > > Thank you very much for the branch. I made some modifications and 
> > > > > included
> > > > > your and Sébastien's comments.
> > > > >
> > > > > Unfortunately,  I have a problem launching the relax functions. For 
> > > > > example
> > > > > by starting the noe calculation, an error appears:
> > > > > Traceback (most recent call last):
> > > > >  File "/home/michael/Programme/relax_bieri_gui/gui_bieri/relax_gui.py", 
> > > > > line
> > > > > 2645, in exec_noe1
> > > > >   make_noe(self.res_noe1.GetValue(), self.noe_ref_1.GetValue(),
> > > > > self.noe_sat_1.GetValue(), self.noe_ref_err_1.GetValue(),
> > > > > self.noe_sat_err_1.GetValue(),
> > > > > self.nmrfreq_value_noe1.GetValue(),self.structure_noe1.GetValue(),
> > > > > self.unres_noe1.GetValue(), start_relax, self, 1, global_setting,
> > > > > file_setting, sequencefile)
> > > > >  File 
> > > > > "/home/michael/Programme/relax_bieri_gui/gui_bieri/res/calc_noe.py",
> > > > > line 66, in make_noe
> > > > >   pipe.create(pipename, 'noe')
> > > > > NameError: global name 'pipe' is not defined
> > > > >
> > > > >
> > > > >
> > > > > relaxGUI is not able to feed the commands into relax. Do  I have to 
> > > > > import
> > > > > other modules?
> > > > >
> > > > > Imported relax modules:
> > > > > from float import floatAsByteArray
> > > > > from generic_fns.mol_res_spin import generate_spin_id, spin_index_loop,
> > > > > spin_loop
> > > > > from generic_fns import pipes, spectrum
> > > > > from relax_errors import RelaxError
> > > > >
> > > > > Do you have an idea how to solve this problem?
> > > > Hi,
> > > >
> > > > This is an easy one.  You can solve this by copying how the
> > > > prompt/script interface does this - and this is how you should replace
> > > > all calls to user functions.  Your GUI replaces the user functions and
> > > > should go directly to the core relax API.  So in this case, have a
> > > > look at the prompt.pipe module and look at the create() method on line
> > > > 92 (1.3 line, r9986).  You'll see that all this method does is check
> > > > that what the user input is reasonable - i.e. both the name and method
> > > > must be strings.  Then at the end it passes everything to the
> > > > generic_fns.pipes.create() function:
> > > >
> > > > # Execute the functional code.
> > > > pipes.create(pipe_name=pipe_name, pipe_type=pipe_type)
> > > >
> > > > So all you need in your code, in this case, is to replace 'pipe' with
> > > > 'pipes'.  For every single user function call you can do the same
> > > > thing - look at what is happening in the 'prompt' package and
> > > > implement the same thing in the 'bieri_gui' package.
> > > >
> > > >
> > > >
> > > >
> > > > > I have another problem relating the OMP data. relaxGUI calculated 
> > > > > local_tm
> > > > > model, but it is not able to calculate the others (also the scripts 
> > > > > can't
> > > > > calculate). There seems to be a problem with the pdb file, I just don't 
> > > > > see
> > > > > at the moment:
> > > > >
> > > > > Internal relax PDB parser.
> > > > > Opening the file '/home/mbieri/LinuxBox/OMP/1F35_mono.pdb' for reading.
> > > > > RelaxWarning: The atom number 80 from the CONECT record cannot be found
> > > > > within the ATOM and HETATM records.
> > > > > RelaxWarning: The atom number 88 from the CONECT record cannot be found
> > > > > within the ATOM and HETATM records.
> > > > > Traceback (most recent call last):
> > > > >  File "relaxGUI.py", line 4126, in exec_model_free
> > > > >   start_model_free(self, which_model)
> > > > >  File "relaxGUI.py", line 1533, in start_model_free
> > > > >   Model()
> > > > >  File "relaxGUI.py", line 1125, in __init__
> > > > >   structure.read_pdb(PDB_FILE)
> > > > >  File "/usr/local/relax-1.3.4/prompt/structure.py", line 464, in 
> > > > > read_pdb
> > > > >   generic_fns.structure.main.read_pdb(file=file, dir=dir, 
> > > > > read_mol=read_mol,
> > > > > set_mol_name=set_mol_name, read_model=read_model,
> > > > > set_model_num=set_model_num, parser=parser)
> > > > >  File "/usr/local/relax-1.3.4/generic_fns/structure/main.py", line 247, 
> > > > > in
> > > > > read_pdb
> > > > >   cdp.structure.load_pdb(file_path, read_mol=read_mol,
> > > > > set_mol_name=set_mol_name, read_model=read_model,
> > > > > set_model_num=set_model_num, verbosity=verbosity)
> > > > >  File "/usr/local/relax-1.3.4/generic_fns/structure/internal.py", line 
> > > > > 628,
> > > > > in load_pdb
> > > > >   mol.fill_object_from_pdb(mol_records)
> > > > >  File "/usr/local/relax-1.3.4/generic_fns/structure/internal.py", line 
> > > > > 1373,
> > > > > in fill_object_from_pdb
> > > > >   self.atom_connect(index1=self.__atom_index(record[1]),
> > > > > index2=self.__atom_index(record[i+2]))
> > > > >  File "/usr/local/relax-1.3.4/generic_fns/structure/internal.py", line 
> > > > > 1332,
> > > > > in atom_connect
> > > > >   if index2 not in self.bonded[index1]:
> > > > > TypeError: list indices must be integers
> > > > >
> > > > > I tried both, the monomer (extracted in pymol) and the dimer.
> > > > This one here looks like an old bug I fixed a while ago.  Does the
> > > > same problem occur when using the 1.3 repository line rather than the
> > > > 1.3.4 release?
> > > >
> > > > As for changes to the code, these should be committed in small
> > > > packets.  For repository usage, a commit should only have one function
> > > > - and if not it should be split up.  The reason is because then it
> > > > allows for commits to be reverted, ported to other branches, etc. and
> > > > allows the code to be more easily reviewed.  When you have something,
> > > > just type:
> > > >
> > > > $ svn diff > patch
> > > >
> > > > in the base directory and attach the file to the task
> > > > (https://gna.org/task/?6847).  I can then check the code and commit
> > > > it.  Don't forget to write a commit message describing what the new
> > > > code does.  All of this is explained in greater depth in the manual.
> > > >
> > > > Cheers,
> > > >
> > > > Edward
> > > _______________________________________________
> > > relax (http://nmr-relax.com)
> > >
> > > This is the relax-devel mailing list
> > > relax-devel@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-devel
>
> _______________________________________________
> relax (http://nmr-relax.com)
>
> This is the relax-devel mailing list
> relax-devel@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-devel