[contrasting] second submission

extreme_data/meteo_france_data/scm_models_data/utils_function.py

@@ -67,7 +67,6 @@ class ReturnLevelBootstrap(object):
         idxs = list(range(self.nb_bootstrap))
 
         if multiprocess is None:
-            print('multiprocessing batch')
 
             with Pool(NB_CORES) as p:
                 batchsize = math.ceil(AbstractExtractEurocodeReturnLevel.NB_BOOTSTRAP / NB_CORES)

extreme_fit/estimator/margin_estimator/abstract_margin_estimator.py

@@ -66,8 +66,8 @@ class LinearMarginEstimator(AbstractMarginEstimator):
         maxima_values = self.dataset.maxima_gev(split=split)
         coordinate_values = self.dataset.df_coordinates(split=split).values
         for maximum, coordinate in zip(maxima_values, coordinate_values):
-            assert len(
-                maximum) == 1, 'So far, only one observation for each coordinate, but code would be easy to change'
+            assert len(maximum) == 1, \
+                'So far, only one observation for each coordinate, but code would be easy to change'
             maximum = maximum[0]
             gev_params = self.function_from_fit.get_params(coordinate, is_transformed=True)
             p = gev_params.density(maximum)

extreme_fit/estimator/margin_estimator/utils.py

@@ -53,5 +53,6 @@ def _fitted_stationary_gev(fit_method, model_class, starting_year, x_gev):
         gev_param.param_name_to_confidence_interval = param_name_to_confidence_interval
     # Warning
     if not -0.5 < gev_param.shape < 0.5:
-        warnings.warn('fitted shape parameter is outside physical bounds {}'.format(gev_param.shape))
+        pass
+        # warnings.warn('fitted shape parameter is outside physical bounds {}'.format(gev_param.shape))
     return estimator, gev_param

projects/altitude_spatial_model/altitudes_fit/main_altitudes_studies.py

@@ -52,7 +52,7 @@ def main():
     elif fast:
         AbstractExtractEurocodeReturnLevel.NB_BOOTSTRAP = 10
         massif_names = ['Vanoise', 'Haute-Maurienne', 'Vercors'][:1]
-        altitudes_list = altitudes_for_groups[1:2]
+        altitudes_list = altitudes_for_groups[2:3]
     else:
         massif_names = None
         altitudes_list = altitudes_for_groups[:]
@@ -81,13 +81,13 @@ def main_loop(altitudes_list, massif_names, seasons, study_classes, model_must_p
 
 
 def plot_visualizers(massif_names, visualizer_list):
-    # plot_histogram_all_models_against_altitudes(massif_names, visualizer_list)
-    plot_histogram_all_trends_against_altitudes(massif_names, visualizer_list)
+    plot_histogram_all_models_against_altitudes(massif_names, visualizer_list)
+    plot_histogram_all_trends_against_altitudes(massif_names, visualizer_list, with_significance=True)
     # plot_shoe_plot_ratio_interval_size_against_altitude(massif_names, visualizer_list)
     for relative in [True, False]:
         plot_shoe_plot_changes_against_altitude(massif_names, visualizer_list, relative=relative)
-    plot_coherence_curves(massif_names, visualizer_list)
-    # plot_coherence_curves(['Vanoise'], visualizer_list)
+    # plot_coherence_curves(massif_names, visualizer_list)
+    plot_coherence_curves(['Vanoise'], visualizer_list)
     pass
 
 

projects/altitude_spatial_model/altitudes_fit/one_fold_analysis/altitudes_studies_visualizer_for_non_stationary_models.py

@@ -220,8 +220,8 @@ class AltitudesStudiesVisualizerForNonStationaryModels(StudyVisualizer):
 
     @property
     def add_colorbar(self):
-        return True
-        # return isinstance(self.altitude_group, (VeyHighAltitudeGroup, MidAltitudeGroup))
+        # return True
+        return isinstance(self.altitude_group, (VeyHighAltitudeGroup, MidAltitudeGroup))
 
     def plot_against_years(self, method_name, order):
         ax = plt.gca()

projects/altitude_spatial_model/altitudes_fit/one_fold_analysis/one_fold_fit.py

@@ -142,6 +142,11 @@ class OneFoldFit(object):
     @cached_property
     def sorted_estimators(self):
         estimators = list(self.model_class_to_estimator.values())
+        # print(self.massif_name)
+        # print(self.altitude_group)
+        # for estimator in estimators:
+        #     print(estimator.margin_model)
+        #     print(estimator.aic())
         sorted_estimators = sorted([estimator for estimator in estimators], key=lambda e: e.aic())
         return sorted_estimators
 
@@ -154,27 +159,13 @@ class OneFoldFit(object):
                     # and self.sensitivity_of_fit_test_last_years(e)
                     ]
         else:
-            return self._sorted_estimators_without_stationary
+            assert len(self.sorted_estimators) == len(self.models_classes)
+            return self.sorted_estimators
 
     @property
     def has_at_least_one_valid_model(self):
         return len(self.sorted_estimators_with_stationary) > 0
 
-    @cached_property
-    def _sorted_estimators_without_stationary(self):
-        return [e for e in self.sorted_estimators if not isinstance(e.margin_model, StationaryAltitudinal)]
-
-    @cached_property
-    def sorted_estimators_without_stationary(self):
-        if self.only_models_that_pass_goodness_of_fit_test:
-            return [e for e in self._sorted_estimators_without_stationary if self.goodness_of_fit_test(e)]
-        else:
-            return self._sorted_estimators_without_stationary
-
-    @property
-    def has_at_least_one_valid_non_stationary_model(self):
-        return len(self.sorted_estimators_without_stationary) > 0
-
     @property
     def model_class_to_estimator_with_finite_aic(self):
         return {type(estimator.margin_model): estimator for estimator in self.sorted_estimators}
@@ -184,10 +175,6 @@ class OneFoldFit(object):
         if self.best_estimator_minimizes_total_aic and self.best_estimator_class_for_total_aic is not None:
             return self.model_class_to_estimator[self.best_estimator_class_for_total_aic]
         else:
-            # Without stationary
-            # if self.has_at_least_one_valid_non_stationary_model:
-            #     best_estimator = self.sorted_estimators_without_stationary[0]
-            #     return best_estimator
             # With stationary
             if self.has_at_least_one_valid_model:
                 best_estimator = self.sorted_estimators_with_stationary[0]
@@ -428,7 +415,6 @@ class OneFoldFit(object):
         idxs = list(range(AbstractExtractEurocodeReturnLevel.NB_BOOTSTRAP))
 
         if multiprocess is None:
-            print('multiprocessing batch')
             start = time.time()
             with Pool(NB_CORES) as p:
                 batchsize = math.ceil(AbstractExtractEurocodeReturnLevel.NB_BOOTSTRAP / NB_CORES)

projects/altitude_spatial_model/altitudes_fit/one_fold_analysis/plot_total_aic.py

@@ -14,9 +14,9 @@ from projects.exceeding_snow_loads.utils import dpi_paper1_figure
 
 
 def plots(visualizer: AltitudesStudiesVisualizerForNonStationaryModels):
-    # visualizer.plot_shape_map()
+    visualizer.plot_shape_map()
     visualizer.plot_moments()
-    # visualizer.plot_qqplots()
+    visualizer.plot_qqplots()
     # for std in [True, False]:
     #     visualizer.studies.plot_mean_maxima_against_altitude(std=std)
 

projects/altitude_spatial_model/altitudes_fit/plots/plot_histogram_altitude_studies.py

@@ -115,7 +115,7 @@ def plot_histogram_all_trends_against_altitudes(massif_names, visualizer_list: L
     ax.tick_params(axis='both', which='major', labelsize=labelsize)
     ax.set_xticks(x)
     ax.yaxis.grid()
-    ax.set_ylim([0, 89])
+    ax.set_ylim([0, 79])
     ax.set_ylim(bottom=0)
     ax.set_xticklabels([v.altitude_group.formula_upper for v in visualizer_list])