[refactor] remove unnecessary argument for gev_params.py

extreme_fit/distribution/gev/gev_params.py

@@ -20,10 +20,8 @@ class GevParams(AbstractExtremeParams):
     SUMMARY_NAMES = PARAM_NAMES + AbstractParams.QUANTILE_NAMES
     NB_SUMMARY_NAMES = len(SUMMARY_NAMES)
 
-    def __init__(self, loc: float, scale: float, shape: float, block_size: int = None):
+    def __init__(self, loc: float, scale: float, shape: float):
         super().__init__(loc, scale, shape)
-        self.block_size = block_size
-        self.param_name_to_confidence_interval = None
 
     @nan_if_undefined_wrapper
     def sample(self, n) -> np.ndarray:
@@ -159,8 +157,6 @@ class GevParams(AbstractExtremeParams):
         x += self.location
         return x
 
-
-
     @property
     def bound(self):
         return self.location - (self.scale / self.shape)

extreme_fit/estimator/margin_estimator/utils.py

@@ -41,16 +41,6 @@ def _fitted_stationary_gev(fit_method, model_class, starting_year, x_gev):
     estimator = fitted_linear_margin_estimator(model_class, coordinates, dataset, starting_year, fit_method)
     first_coordinate = coordinates.coordinates_values()[0]
     gev_param = estimator.function_from_fit.get_params(first_coordinate)
-    # Build confidence interval
-    fit_method_with_confidence_interval_implemented = [MarginFitMethod.is_mev_gev_fit]
-    if fit_method in fit_method_with_confidence_interval_implemented:
-        param_name_to_confidence_interval = {}
-        for i, (param_name, param_value) in enumerate(gev_param.to_dict().items()):
-            confidence_interval_half_size = estimator.result_from_model_fit.confidence_interval_half_sizes[i]
-            confidence_interval = (
-                param_value - confidence_interval_half_size, param_value + confidence_interval_half_size)
-            param_name_to_confidence_interval[param_name] = confidence_interval
-        gev_param.param_name_to_confidence_interval = param_name_to_confidence_interval
     # Warning
     if not -0.5 < gev_param.shape < 0.5:
         pass

extreme_fit/model/result_from_model_fit/abstract_result_from_model_fit.py

@@ -20,10 +20,6 @@ class AbstractResultFromModelFit(object):
         """ See Coles 2001 page 41, for an example"""
         return np.sqrt(np.diagonal(self.variance_covariance_matrix))
 
-    @property
-    def confidence_interval_half_sizes(self):
-        return [1.96 * s for s in self.standard_errors_for_mle]
-
     @staticmethod
     def get_python_dictionary(r_dictionary):
         return {name: r_dictionary.rx2(name) for name in r_dictionary.names}

projects/altitude_spatial_model/preliminary_analysis.py

@@ -182,7 +182,6 @@ class PointwiseGevStudyVisualizer(AltitudesStudies):
                     assert isinstance(param_name, int)
                     param = gev_params.return_level(return_period=param_name)
                 params.append(param)
-                # confidence_intervals.append(gev_params.param_name_to_confidence_interval[param_name])
         massif_id = self.study.all_massif_names().index(massif_name)
         plot_against_altitude(altitudes, ax, massif_id, massif_name, params, fill=False,
                               elevation_as_xaxis=elevation_as_xaxis,

projects/projection_snow_load/main_difference_between_reanalysis_and_simulations.py

-from extreme_data.meteo_france_data.adamont_data.deprecated.ensemble_simulation import EnsembleSimulation
-from extreme_data.meteo_france_data.scm_models_data.crocus.crocus import CrocusSnowLoad3Days, \
-    CrocusSweTotal
-from extreme_data.meteo_france_data.scm_models_data.visualization.study_visualizer import \
-    StudyVisualizer
-import matplotlib.pyplot as plt
-
-
-def run_plot():
-    study = CrocusSnowLoad3Days(altitude=1200)
-    study_visualizer = StudyVisualizer(study)
-    study_visualizer.visualize_max_graphs_poster('Queyras', altitude='noope', snow_abbreviation="ok", color='red')
-    plt.show()
-
-
-def density_wrt_altitude():
-    save_to_file = True
-    study_class = CrocusSweTotal
-    altitudes = [600, 900, 1200, 1500, 1800, 2100, 2400, 2700][::-1]
-
-    ensemble = EnsembleSimulation(first_winter_required_for_histo=1958)
-
-    # None means relative change, true means reanalysis values, false means simulations values
-    # mode = False
-
-    for mode in [None, True, False]:
-        for altitude in altitudes:
-
-            ax = plt.gca()
-            study = study_class(altitude=altitude)
-            # study = study_class(altitude=altitude, nb_consecutive_days=3)
-            massif_name_to_value = {}
-            df = study.observations_annual_maxima.df_maxima_gev.iloc[:2004]
-            for massif_name in study.study_massif_names:
-                s_reanalysis = df.loc[massif_name].mean()
-                if (massif_name, altitude) in ensemble.massif_name_and_altitude_to_mean_average_annual_maxima:
-                    s_simulation = ensemble.massif_name_and_altitude_to_mean_average_annual_maxima[(massif_name, altitude)]
-                    relative_change_value = 100 * (s_simulation - s_reanalysis) / s_reanalysis
-                    if mode == None:
-                        value = relative_change_value
-                    else:
-                        value = s_reanalysis if mode else s_simulation
-                    massif_name_to_value[massif_name] = value
-            print(massif_name_to_value)
-            # Plot
-            # massif_name_to_value = {m: i for i, m in enumerate(study.study_massif_names)}
-            max_values = max([abs(e) for e in massif_name_to_value.values()]) + 5
-            print(max_values)
-            variable_name = study.variable_name
-            label_relative_change = 'Relative changes in mean annual maxima\n' \
-                        'of {}\n between reanalysis and historical simulation\n' \
-                        'for 1958-2004  at {}m (%)\n' \
-                        ''.format(variable_name, study.altitude)
-            if mode is None:
-                label = label_relative_change
-            else:
-                label = 'Mean max SWE for ' + ('reanalysis' if mode else 'simulation')
-
-            vmin = -max_values if mode is None else 0.1
-            study.visualize_study(ax=ax, massif_name_to_value=massif_name_to_value,
-                                  vmin=vmin, vmax=max_values,
-                                  add_colorbar=True,
-                                  replace_blue_by_white=False,
-                                  show=False,
-                                  label=label
-                                  )
-            study_visualizer = StudyVisualizer(study, save_to_file=save_to_file)
-            study_visualizer.plot_name = 'relative_changes_in_maxima' if mode is None else label
-            study_visualizer.show_or_save_to_file()
-            ax.clear()
-            plt.close()
-
-
-if __name__ == '__main__':
-    density_wrt_altitude()
-    # test()