FAI-911: Improve API to declare selections in Python fairness (#132)

* Add data_conversion support for favorable types
* Fix formatting

Author: ruivieira

src/trustyai/metrics/fairness/group.py

@@ -6,13 +6,20 @@
 from jpype import JInt
 from org.kie.trustyai.explainability.metrics import FairnessMetrics
 
-from trustyai.model import Output, Value, PredictionProvider, Model
-from trustyai.utils.data_conversions import pandas_to_trusty
+from trustyai.model import Value, PredictionProvider, Model
+from trustyai.utils.data_conversions import (
+    pandas_to_trusty,
+    OneOutputUnionType,
+    one_output_convert,
+)
 
 ColumSelector = Union[List[int], List[str]]
 
 
 def _column_selector_to_index(columns: ColumSelector, dataframe: pd.DataFrame):
+    if len(columns) == 0:
+        raise ValueError("Must specify at least one column")
+
     if isinstance(columns[0], str):  # passing column
         columns = dataframe.columns.get_indexer(columns)
     indices = [JInt(c) for c in columns]  # Java casting
@@ -22,14 +29,15 @@ def _column_selector_to_index(columns: ColumSelector, dataframe: pd.DataFrame):
 def statistical_parity_difference(
     privileged: pd.DataFrame,
     unprivileged: pd.DataFrame,
-    favorable: List[Output],
+    favorable: OneOutputUnionType,
     outputs: Optional[List[int]] = None,
 ) -> float:
     """Calculate Statistical Parity Difference between privileged and unprivileged dataframes"""
+    favorable_prediction_object = one_output_convert(favorable)
     return FairnessMetrics.groupStatisticalParityDifference(
         pandas_to_trusty(privileged, outputs),
         pandas_to_trusty(unprivileged, outputs),
-        favorable,
+        favorable_prediction_object.outputs,
     )
 
 
@@ -39,31 +47,33 @@ def statistical_parity_difference_model(
     model: Union[PredictionProvider, Model],
     privilege_columns: ColumSelector,
     privilege_values: List[Any],
-    favorable: List[Output],
+    favorable: OneOutputUnionType,
 ) -> float:
     """Calculate Statistical Parity Difference using a samples dataframe and a model"""
+    favorable_prediction_object = one_output_convert(favorable)
     _privilege_values = [Value(v) for v in privilege_values]
     _jsamples = pandas_to_trusty(samples, no_outputs=True)
     return FairnessMetrics.groupStatisticalParityDifference(
         _jsamples,
         model,
         _column_selector_to_index(privilege_columns, samples),
         _privilege_values,
-        favorable,
+        favorable_prediction_object.outputs,
     )
 
 
 def disparate_impact_ratio(
     privileged: pd.DataFrame,
     unprivileged: pd.DataFrame,
-    favorable: List[Output],
+    favorable: OneOutputUnionType,
     outputs: Optional[List[int]] = None,
 ) -> float:
     """Calculate Disparate Impact Ration between privileged and unprivileged dataframes"""
+    favorable_prediction_object = one_output_convert(favorable)
     return FairnessMetrics.groupDisparateImpactRatio(
         pandas_to_trusty(privileged, outputs),
         pandas_to_trusty(unprivileged, outputs),
-        favorable,
+        favorable_prediction_object.outputs,
     )
 
 
@@ -73,17 +83,18 @@ def disparate_impact_ratio_model(
     model: Union[PredictionProvider, Model],
     privilege_columns: ColumSelector,
     privilege_values: List[Any],
-    favorable: List[Output],
+    favorable: OneOutputUnionType,
 ) -> float:
     """Calculate Disparate Impact Ration using a samples dataframe and a model"""
+    favorable_prediction_object = one_output_convert(favorable)
     _privilege_values = [Value(v) for v in privilege_values]
     _jsamples = pandas_to_trusty(samples, no_outputs=True)
     return FairnessMetrics.groupDisparateImpactRatio(
         _jsamples,
         model,
         _column_selector_to_index(privilege_columns, samples),
         _privilege_values,
-        favorable,
+        favorable_prediction_object.outputs,
     )
 
 
@@ -92,7 +103,7 @@ def average_odds_difference(
     test: pd.DataFrame,
     truth: pd.DataFrame,
     privilege_columns: ColumSelector,
-    privilege_values: List[Any],
+    privilege_values: OneOutputUnionType,
     positive_class: List[Any],
     outputs: Optional[List[int]] = None,
 ) -> float: