[COORDINATES] add test_coordinate_sensitivity

extreme_estimator/extreme_models/margin_model/parametric_margin_model.py

@@ -39,7 +39,6 @@ class ParametricMarginModel(AbstractMarginModel, ABC):
         # Enforce a starting point for the temporal trend
         if self.transformed_starting_point is not None:
             # Compute the indices to modify
-            print('transformed starting point', self.transformed_starting_point)
             ind_to_modify = df_coordinates_temp.iloc[:, 0] <= self.transformed_starting_point  # type: pd.Series
             # Assert that some coordinates are selected but not all (at least 20 data should be left for temporal trend)
             assert 0 < sum(ind_to_modify) < len(ind_to_modify) - 20

spatio_temporal_dataset/coordinates/abstract_coordinates.py

@@ -36,6 +36,7 @@ class AbstractCoordinates(object):
     # Coordinates columns
     COORDINATES_NAMES = COORDINATE_SPATIAL_NAMES + [COORDINATE_T]
     # Coordinate type
+    ALL_COORDINATES_ACCEPTED_TYPES = ['int64', 'float64']
     COORDINATE_TYPE = 'float64'
 
     def __init__(self, df: pd.DataFrame, slicer_class: type, s_split_spatial: pd.Series = None,
@@ -46,8 +47,10 @@ class AbstractCoordinates(object):
         # Sort coordinates according to a specified order
         sorted_coordinates_columns = [c for c in self.COORDINATES_NAMES if c in coordinate_columns]
         self.df_all_coordinates = df.loc[:, sorted_coordinates_columns].copy()  # type: pd.DataFrame
-        # Cast df_all_coordinates to the desired type
-        self.df_all_coordinates = self.df_all_coordinates.astype(self.COORDINATE_TYPE)
+        # Check the data type of the coordinate columns
+        accepted_dtypes = ['int']
+        assert len(self.df_all_coordinates.select_dtypes(include=self.ALL_COORDINATES_ACCEPTED_TYPES).columns) \
+               == len(coordinate_columns), 'coordinates columns dtypes should belong to {}'.format(accepted_dtypes)
 
         # Slicing attributes
         self.s_split_spatial = s_split_spatial  # type: pd.Series
@@ -128,12 +131,14 @@ class AbstractCoordinates(object):
     # Normalize
 
     def transform(self, coordinate: np.ndarray) -> np.ndarray:
-        return self.transformation.transform_array(coordinate=coordinate)
+        coordinate_float = coordinate.astype(self.COORDINATE_TYPE)
+        return self.transformation.transform_array(coordinate=coordinate_float)
 
     # Split
 
     def df_coordinates(self, split: Split = Split.all) -> pd.DataFrame:
-        df_transformed_coordinates = self.transformation.transform_df(df_coord=self.df_all_coordinates)
+        df_all_coordinate_as_float = self.df_all_coordinates.astype(self.COORDINATE_TYPE)  # type: pd.DataFrame
+        df_transformed_coordinates = self.transformation.transform_df(df_all_coordinate_as_float)
         return df_sliced(df=df_transformed_coordinates, split=split, slicer=self.slicer)
 
     def coordinates_values(self, split: Split = Split.all) -> np.ndarray:

test/test_experiment/test_weird_bug.py → test/test_experiment/test_coordinate_sensitivity.py

@@ -12,28 +12,26 @@ from utils import get_display_name_from_object_type
 
 
 class TestCoordinateSensitivity(unittest.TestCase):
+    DISPLAY = False
 
-    def test_weird(self):
-        # todo: maybe the code does not like negative coordinates
-        # todo: maybe not that the sign of the x coordinate are all negative and the other are all positive, it is easier to find the perfect spatial structure
+    def test_coordinate_normalization_sensitivity(self):
         altitudes = [3000]
         transformation_classes = [BetweenZeroAndOneNormalization, BetweenMinusOneAndOneNormalization][:]
         for transformation_class in transformation_classes:
             study_classes = [CrocusSwe]
             for study in study_iterator_global(study_classes, altitudes=altitudes, verbose=False):
-                print('\n\n')
                 study_visualizer = StudyVisualizer(study, transformation_class=transformation_class)
                 study_visualizer.temporal_non_stationarity = True
-                print(study_visualizer.coordinates)
-                # trend_test = ConditionalIndedendenceLocationTrendTest(study_visualizer.dataset)
-                # # years = [1960, 1990]
-                # # mu1s = [trend_test.get_mu1(year) for year in years]
-                # # print('Stationary')
-                # # print(trend_test.get_estimator(trend_test.stationary_margin_model_class, starting_point=None).margin_function_fitted.coef_dict)
-                # print('Non Stationary')
-                # print(trend_test.get_estimator(trend_test.non_stationary_margin_model_class, starting_point=1960).margin_function_fitted.coef_dict)
-                # # print(get_display_name_from_object_type(type(transformation_2D)), 'mu1s: ', mu1s)
-                # # self.assertTrue(0.0 not in mu1s)
+                trend_test = ConditionalIndedendenceLocationTrendTest(study_visualizer.dataset)
+                years = [1960, 1990]
+                mu1s = [trend_test.get_mu1(year) for year in years]
+                if self.DISPLAY:
+                    print('Stationary')
+                    print(trend_test.get_estimator(trend_test.stationary_margin_model_class, starting_point=None).margin_function_fitted.coef_dict)
+                    print('Non Stationary')
+                    print(trend_test.get_estimator(trend_test.non_stationary_margin_model_class, starting_point=1960).margin_function_fitted.coef_dict)
+                    print(get_display_name_from_object_type(transformation_class), 'mu1s: ', mu1s)
+                self.assertTrue(0.0 not in mu1s)
 
 
 if __name__ == '__main__':