Some renaming

No need for abbreviations here.

nannyml/performance_calculation/metrics/multiclass_classification.py

@@ -40,10 +40,10 @@
     recall_sampling_error_components,
     specificity_sampling_error,
     specificity_sampling_error_components,
-    ap_sampling_error_components,
-    ap_sampling_error,
-    bv_sampling_error_components,
-    bv_sampling_error,
+    average_precision_sampling_error_components,
+    average_precision_sampling_error,
+    business_value_sampling_error_components,
+    business_value_sampling_error,
 )
 from nannyml.thresholds import Threshold, calculate_threshold_values
 
@@ -987,7 +987,7 @@ def _fit(self, reference_data: pd.DataFrame):
             # sampling error
             binarized_y_true = list(label_binarize(reference_data[self.y_true], classes=self.classes).T)
             y_pred_proba = [reference_data[self.y_pred_proba[clazz]].T for clazz in self.classes]
-            self._sampling_error_components = ap_sampling_error_components(
+            self._sampling_error_components = average_precision_sampling_error_components(
                 y_true_reference=binarized_y_true, y_pred_proba_reference=y_pred_proba
             )
 
@@ -1033,7 +1033,7 @@ def _sampling_error(self, data: pd.DataFrame) -> float:
             )
             return np.NaN
         else:
-            return ap_sampling_error(self._sampling_error_components, data)
+            return average_precision_sampling_error(self._sampling_error_components, data)
 
 
 @MetricFactory.register(metric='business_value', use_case=ProblemType.CLASSIFICATION_MULTICLASS)
@@ -1146,7 +1146,7 @@ def _fit(self, reference_data: pd.DataFrame):
                     f"business_value_matrix has shape {self.business_value_matrix.shape} "
                     "but we have {num_classes} classes!"
                 )
-            self._sampling_error_components = bv_sampling_error_components(
+            self._sampling_error_components = business_value_sampling_error_components(
                 y_true_reference=data[self.y_true],
                 y_pred_reference=data[self.y_pred],
                 business_value_matrix=self.business_value_matrix,
@@ -1180,4 +1180,4 @@ def _sampling_error(self, data: pd.DataFrame) -> float:
             )
             return np.NaN
         else:
-            return bv_sampling_error(self._sampling_error_components, data)
+            return business_value_sampling_error(self._sampling_error_components, data)

nannyml/performance_estimation/confidence_based/metrics.py

@@ -2354,7 +2354,8 @@ def _fit(self, reference_data: pd.DataFrame):
                     "targets."
                 )
                 raise InvalidArgumentsException(
-                    "y_pred_proba class and class probabilities dictionary does not match reference data.")
+                    "y_pred_proba class and class probabilities dictionary does not match reference data."
+                )
             # sampling error
             binarized_y_true = list(label_binarize(reference_data[self.y_true], classes=self.classes).T)
             y_pred_proba = [reference_data['uncalibrated_' + self.y_pred_proba[clazz]].T for clazz in self.classes]
@@ -3385,7 +3386,7 @@ def _fit(self, reference_data: pd.DataFrame):
             # sampling error
             binarized_y_true = list(label_binarize(reference_data[self.y_true], classes=self.classes).T)
             y_pred_proba = [reference_data['uncalibrated_' + self.y_pred_proba[clazz]].T for clazz in self.classes]
-            self._sampling_error_components = mse.ap_sampling_error_components(
+            self._sampling_error_components = mse.average_precision_sampling_error_components(
                 y_true_reference=binarized_y_true, y_pred_proba_reference=y_pred_proba
             )
 
@@ -3431,7 +3432,7 @@ def _sampling_error(self, data: pd.DataFrame) -> float:
             )
             return np.NaN
         else:
-            return mse.ap_sampling_error(self._sampling_error_components, data)
+            return mse.average_precision_sampling_error(self._sampling_error_components, data)
 
     def _realized_performance(self, data: pd.DataFrame) -> float:
         try:
@@ -3525,7 +3526,7 @@ def _fit(self, reference_data: pd.DataFrame):
                     f"business_value_matrix has shape {self.business_value_matrix.shape} "
                     f"but we have {num_classes} classes!"
                 )
-            self._sampling_error_components = mse.bv_sampling_error_components(
+            self._sampling_error_components = mse.business_value_sampling_error_components(
                 y_true_reference=data[self.y_true],
                 y_pred_reference=data[self.y_pred],
                 business_value_matrix=self.business_value_matrix,
@@ -3584,7 +3585,7 @@ def _sampling_error(self, data: pd.DataFrame) -> float:
             warnings.warn(_message)
             return np.NaN
         else:
-            return mse.bv_sampling_error(self._sampling_error_components, data)
+            return mse.business_value_sampling_error(self._sampling_error_components, data)
 
     def _realized_performance(self, data: pd.DataFrame) -> float:
         try:

nannyml/sampling_error/multiclass_classification.py

@@ -413,7 +413,9 @@ def multiclass_confusion_matrix_sampling_error(sampling_error_components: Tuple,
     return standard_errors
 
 
-def ap_sampling_error_components(y_true_reference: List[np.ndarray], y_pred_proba_reference: List[pd.Series]):
+def average_precision_sampling_error_components(
+    y_true_reference: List[np.ndarray], y_pred_proba_reference: List[pd.Series]
+):
     """Calculate sampling error components for AP using reference data.
 
     The ``y_true_reference`` and ``y_pred_proba_reference`` lists represent the binarized target values and model
@@ -451,7 +453,7 @@ def _get_class_components(y_true_reference: np.ndarray, y_pred_proba_reference:
     return class_components
 
 
-def ap_sampling_error(sampling_error_components, data) -> float:
+def average_precision_sampling_error(sampling_error_components, data) -> float:
     """Calculate the AUROC sampling error for a chunk of data.
 
     Parameters
@@ -479,16 +481,12 @@ def _calculate_business_value_per_row(
 
     Intended to be used within a pandas apply function.
     """
-    cm = confusion_matrix(
-        y_true=np.array([row.y_true]),
-        y_pred=np.array([row.y_pred]),
-        labels=classes
-    )
+    cm = confusion_matrix(y_true=np.array([row.y_true]), y_pred=np.array([row.y_pred]), labels=classes)
     bv = (cm * business_value_matrix).sum()
     return bv
 
 
-def bv_sampling_error_components(
+def business_value_sampling_error_components(
     y_true_reference: pd.Series,
     y_pred_reference: pd.Series,
     business_value_matrix: np.ndarray,
@@ -514,15 +512,17 @@ def bv_sampling_error_components(
     -------
     components: tuple
     """
-    data = pd.DataFrame({
-        'y_true': y_true_reference,
-        'y_pred': y_pred_reference,
-    })
+    data = pd.DataFrame(
+        {
+            'y_true': y_true_reference,
+            'y_pred': y_pred_reference,
+        }
+    )
     bvs = data.apply(lambda x: _calculate_business_value_per_row(x, business_value_matrix, classes), axis=1)
     return (bvs.std(), normalize_business_value)
 
 
-def bv_sampling_error(sampling_error_components: Tuple, data) -> float:
+def business_value_sampling_error(sampling_error_components: Tuple, data) -> float:
     """Calculate the false positive rate sampling error for a chunk of data.
 
     Parameters