[krige] add exact_values support

configs/kriging.yaml

@@ -50,16 +50,15 @@ learning:
   arguments:
     method: ordinary  # ordinary or universal
     variogram_model: spherical  # linear, power, gaussian, spherical, exponential
-    variogram_parameters: [0, 1, 0.1]   # [sill, range, nugget] for gaussian, spherical, exponential
+    variogram_parameters: [0.0, 1, 0.1]   # [sill, range, nugget] for gaussian, spherical, exponential
     verbose: False
-
-
-optimisation:
-  algorithm: krige
-  hyperparameters:
-    method: [ordinary, universal]
-    variogram_model: [linear, power, gaussian, spherical, exponential]
-  optimisation_output: optimisation.csv
+    exact_values: False
+#  optimisation:
+#    algorithm: krige
+#    hyperparameters:
+#      method: [ordinary, universal]
+#      variogram_model: [linear, power, gaussian, spherical, exponential]
+#    optimisation_output: optimisation.csv
 
 # outbands: number of output bands desired.
 # Bands (1:5): ['Prediction', 'Variance', 'Lower quantile',

uncoverml/krige.py

@@ -97,6 +97,7 @@ def __init__(self,
                  nlags=6,
                  weight=False,
                  n_closest_points=10,
+                 exact_values=False,
                  verbose=False
                  ):
         if method not in krige_methods.keys():
@@ -110,6 +111,7 @@ def __init__(self,
         self.n_closest_points = n_closest_points
         self.model = None  # not trained
         self.method = method
+        self.exact_values = exact_values
 
     def fit(self, x, y, *args, **kwargs):
         """
@@ -132,6 +134,7 @@ def fit(self, x, y, *args, **kwargs):
             verbose=self.verbose,
             nlags=self.nlags,
             weight=self.weight,
+            exact_values=self.exact_values,
         )
 
     def predict(self, x, *args, **kwargs):