group_line_by = None
# group_line_by = ['ERA-5', 'EC-Earth']
col_wrap = None
wspace = 0.05
kwrgs = {'wspace':wspace, 'col_wrap':col_wrap}
met = ['AUC-ROC', 'AUC-PR', 'BSS', 'Rel. Curve']
kwrgs = {'wspace':0.25, 'col_wrap':3, 'threshold_bin':fc.threshold_pred}
met = ['AUC-ROC', 'AUC-PR', 'BSS', 'Rel. Curve', 'Precision', 'Accuracy']
expers = list(dict_experiments.keys())
models = list(dict_experiments[expers[0]].keys())
line_dim = 'model'
fig = dfplots.valid_figures(dict_experiments, expers=expers, models=models,
line_dim=line_dim,
group_line_by=group_line_by,
met=met, **kwrgs)
if f_format == '.png':
fig.savefig(os.path.join(filename + f_format),
bbox_inches='tight') # dpi auto 600
elif f_format == '.pdf':
fig.savefig(os.path.join(pdfs_folder,f_name+ f_format),
bbox_inches='tight')
print_sett(list_fc, stat_model_l, filename)
np.save(filename + '.npy', dict_experiments)
#%%
#fcev.plot_scatter()
'wspace': 0.16,
'hspace': .25,
'col_wrap': 2,
'skip_redundant_title': True,
'lags_relcurve': [lag_rel],
'fontbase': 14,
'figaspect': 2
}
#kwrgs = {'wspace':0.25, 'col_wrap':3, 'threshold_bin':fc.threshold_pred}
met = ['AUC-ROC', 'AUC-PR', 'Precision', 'BSS', 'Accuracy', 'Rel. Curve']
line_dim = 'exper'
group_line_by = None
fig = dfplots.valid_figures(dict_merge_all,
line_dim=line_dim,
group_line_by=group_line_by,
met=met,
**kwrgs)
f_format = '.pdf'
pathfig_valid = os.path.join(filename + f_format)
fig.savefig(pathfig_valid, bbox_inches='tight') # dpi auto 600
fc = list_of_fc[0]
df, fig = fc.plot_feature_importances(lag=lag_rel)
path_feat = filename + f'ifc{1}_logitregul_l{lag_rel}' + f_format
fig.savefig(path_feat, bbox_inches='tight')
fc.dict_sum[0].loc['Precision'].loc['Precision']
fc.dict_sum[0].loc['Accuracy'].loc['Accuracy']
# dict_experiments[new] = dict_experiments.pop(old)
f_format = '.png'
filename = os.path.join(working_folder, f_name)
group_line_by = None
#group_line_by = ['ERA-5', 'EC']
kwrgs = {'wspace': 0.08}
met = ['AUC-ROC', 'AUC-PR', 'BSS', 'Rel. Curve']
expers = list(dict_experiments.keys())
models = list(dict_experiments[expers[0]].keys())
fig = dfplots.valid_figures(dict_experiments,
expers=expers,
models=models,
line_dim='exper',
group_line_by=group_line_by,
met=met,
**kwrgs)
if f_format == '.png':
fig.savefig(os.path.join(filename + f_format),
bbox_inches='tight') # dpi auto 600
elif f_format == '.pdf':
fig.savefig(os.path.join(pdfs_folder, f_name + f_format),
bbox_inches='tight')
print_sett(experiments, stat_model_l, filename)
np.save(filename + '.npy', dict_experiments)
#%%
# =============================================================================
fc.fit_models(lead_max=35)
dict_experiments = {}
fc.perform_validation(n_boot=100, blocksize='auto',
threshold_pred=(1.5, 'times_clim'))
dict_experiments['test'] = fc.dict_sum
working_folder, filename = fc._print_sett(list_of_fc=[fc])
store=True
dict_all = dfplots.merge_valid_info([fc], store=store)
if store:
dict_merge_all = functions_pp.load_hdf5(filename)
kwrgs = {'wspace':0.25, 'col_wrap':3} #, 'threshold_bin':fc.threshold_pred}
met = ['AUC-ROC', 'AUC-PR', 'BSS', 'Rel. Curve', 'Precision', 'Accuracy']
expers = list(dict_experiments.keys())
line_dim = 'model'
fig = dfplots.valid_figures(dict_merge_all,
line_dim=line_dim,
group_line_by=None,
lines_legend=None,
met=met, **kwrgs)
# remove created output folders
shutil.rmtree(rg.path_outsub1)
shutil.rmtree(os.path.join(main_dir, 'data', 'preprocessed'))
dict_all = functions_pp.load_hdf5(filename+'.h5')
kwrgs = {'wspace':0.12, 'col_wrap':None, 'lags_relcurve':[10, 20],
'skip_redundant_title':True, 'fontbase':14}
#kwrgs = {'wspace':0.25, 'col_wrap':3, 'threshold_bin':fc.threshold_pred}
# met = ['Rel. Curve']
met = ['AUC-ROC', 'AUC-PR', 'BSS', 'Rel. Curve']
line_dim = 'exper'
fig = dfplots.valid_figures(dict_all,
line_dim=line_dim,
group_line_by=None,
met=met, **kwrgs)
f_format = '.pdf'
pathfig_valid = os.path.join(filename + f_format)
fig.savefig(pathfig_valid,
bbox_inches='tight') # dpi auto 600
#%%
# im = 0
# il = 1
# ifc = 0
# f_format = '.pdf'