Re: Is there a way to clear the Monte Carlo simulation data key? -- August 28, 2014 - 11:35

Hi,

Maybe you should consider adding this to the monte_carlo.setup user
function?  Is the problem that you cannot run Monte Carlo simulations
twice with a different number of simulations?  This should really be
handled by monte_carlo.setup.

Regards,

Edward



On 28 August 2014 11:25, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> wrote:
> Allright, I found a fix.
>
>
> # Make Carlo Simulations number
> mc_number_list = range(5, 100, 20)
>
> sim_attr_list = ['chi2_sim', 'f_count_sim', 'g_count_sim',
> 'h_count_sim', 'i0_sim', 'iter_sim', 'peak_intensity_sim',
> 'r2eff_sim', 'select_sim', 'warning_sim']
>
> # Loop over the Monte Carlo simulations:
> for number in mc_number_list:
>     # First delete old simulations.
>     for cur_spin, mol_name, resi, resn, spin_id in
> spin_loop(full_info=True, return_id=True, skip_desel=True):
>         # Loop over the simulated attributes.
>         for sim_attr in sim_attr_list:
>             if hasattr(cur_spin, sim_attr):
>                 delattr(cur_spin, sim_attr)
>
>     self.interpreter.monte_carlo.setup(number=number)
>     self.interpreter.monte_carlo.create_data()
>     self.interpreter.monte_carlo.initial_values()
>     self.interpreter.minimise.execute(min_algor=min_algor,
> constraints=constraints)
>     self.interpreter.eliminate()
>     self.interpreter.monte_carlo.error_analysis()
>
>     est_key = 'mc_%s'%number
>     est_keys.append(est_key)
>
>     # Collect data.
>     for cur_spin, mol_name, resi, resn, spin_id in
> spin_loop(full_info=True, return_id=True, skip_desel=True):
>         # Add key for estimate.
>         my_dic[spin_id][est_key] = {}
>
>         for exp_type, frq, offset, point, ei, mi, oi, di in
> loop_exp_frq_offset_point(return_indices=True):
>             # Generate the param_key.
>             param_key = return_param_key_from_data(exp_type=exp_type,
> frq=frq, offset=offset, point=point)
>
>             # Add key to dic.
>             my_dic[spin_id][est_key][param_key] = {}
>
>             # Get the value.
>             r2eff = getattr(cur_spin, 'r2eff')[param_key]
>             r2eff_err = getattr(cur_spin, 'r2eff_err')[param_key]
>             i0 = getattr(cur_spin, 'i0')[param_key]
>             i0_err = getattr(cur_spin, 'i0_err')[param_key]
>
>             # Save to dic.
>             my_dic[spin_id][est_key][param_key]['r2eff'] = r2eff
>             my_dic[spin_id][est_key][param_key]['r2eff_err'] = r2eff_err
>             my_dic[spin_id][est_key][param_key]['i0'] = i0
>             my_dic[spin_id][est_key][param_key]['i0_err'] = i0_err
>
> 2014-08-28 11:20 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:
> > Hi,
> >
> > Could you describe a situation that covers this?  What do you mean by
> > the Monte Carlo simulation data key?  In the data pipe and spin
> > containers, the Monte Carlo simulation optimisation results are
> > usually stored in the "*_sim" data structures as lists, and the errors
> > from the simulations in the "*_err" data structures which are simple
> > floats.  But this is dependent on the analysis and data type - some
> > model parameters can be single values, lists of values, or
> > dictionaries of values, and this is preserved in the Monte Carlo
> > simulation structures as well.
> >
> > Regards,
> >
> > Edward
> >
> >
> > On 28 August 2014 11:14, Troels Emtekær Linnet <tlinnet@xxxxxxxxxxxxx> 
> > wrote:
> > > Dear Edward.
> > >
> > > Is there a way to clear the Monte Carlo simulation data key?
> > >
> > > I try to run some data with increasing number of Monte Carlo simulations.
> > >
> > > Thank you.
> > > Best
> > > Troels
> > >
> > > _______________________________________________
> > > relax (http://www.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