add aggregate by mean (#272)

bofire/data_models/domain/domain.py

@@ -400,7 +400,11 @@ def coerce_invalids(self, experiments: pd.DataFrame) -> pd.DataFrame:
         return experiments
 
     def aggregate_by_duplicates(
-        self, experiments: pd.DataFrame, prec: int, delimiter: str = "-"
+        self,
+        experiments: pd.DataFrame,
+        prec: int,
+        delimiter: str = "-",
+        method: Literal["mean", "median"] = "mean",
     ) -> Tuple[pd.DataFrame, list]:
         """Aggregate the dataframe by duplicate experiments
 
@@ -417,6 +421,8 @@ def aggregate_by_duplicates(
             Tuple[pd.DataFrame, list]: Dataframe holding the aggregated experiments, list of lists holding the labcodes of the duplicates
         """
         # prepare the parent frame
+        if method not in ["mean", "median"]:
+            raise ValueError(f"Unknown aggregation type provided: {method}")
 
         preprocessed = self.outputs.preprocess_experiments_any_valid_output(experiments)
         assert preprocessed is not None
@@ -437,7 +443,7 @@ def aggregate_by_duplicates(
 
         # group and aggregate
         agg: Dict[str, Any] = {
-            feat: "mean" for feat in self.get_feature_keys(ContinuousOutput)
+            feat: method for feat in self.get_feature_keys(ContinuousOutput)
         }
         agg["labcode"] = lambda x: delimiter.join(sorted(x.tolist()))
         for feat in self.get_feature_keys(Output):

tests/bofire/data_models/test_domain.py

@@ -599,7 +599,8 @@ def test_coerce_invalids():
     assert_frame_equal(experiments, expected, check_dtype=False)
 
 
-def test_aggregate_by_duplicates():
+@pytest.mark.parametrize("method", ["mean", "median"])
+def test_aggregate_by_duplicates(method):
     # dataframe with duplicates
     full = pd.DataFrame.from_dict(
         {
@@ -625,7 +626,9 @@ def test_aggregate_by_duplicates():
     domain = Domain(
         inputs=Inputs(features=[if1, if2]), outputs=Outputs(features=[of1, of2])
     )
-    aggregated, duplicated_labcodes = domain.aggregate_by_duplicates(full, prec=2)
+    aggregated, duplicated_labcodes = domain.aggregate_by_duplicates(
+        full, prec=2, method=method
+    )
     assert duplicated_labcodes == [["1", "4"]]
     assert_frame_equal(
         aggregated, expected_aggregated, check_dtype=False, check_like=True
@@ -655,10 +658,30 @@ def test_aggregate_by_duplicates():
     domain = Domain(
         inputs=Inputs(features=[if1, if2]), outputs=Outputs(features=[of1, of2])
     )
-    aggregated, duplicated_labcodes = domain.aggregate_by_duplicates(full, prec=2)
+    aggregated, duplicated_labcodes = domain.aggregate_by_duplicates(
+        full, prec=2, method=method
+    )
     assert duplicated_labcodes == []
 
 
+def test_aggregate_by_duplicates_error():
+    full = pd.DataFrame.from_dict(
+        {
+            "x1": [1.0, 2.0, 3.0, 1.0],
+            "x2": [1.0, 2.0, 3.0, 1.0],
+            "out1": [4.0, 5.0, 6.0, 3.0],
+            "out2": [-4.0, -5.0, -6.0, -3.0],
+            "valid_out1": [1, 1, 1, 1],
+            "valid_out2": [1, 1, 1, 1],
+        }
+    )
+    domain = Domain(
+        inputs=Inputs(features=[if1, if2]), outputs=Outputs(features=[of1, of2])
+    )
+    with pytest.raises(ValueError, match="Unknown aggregation type provided: 25"):
+        domain.aggregate_by_duplicates(full, prec=2, method="25")
+
+
 domain = Domain(
     inputs=Inputs(features=[if1, if2]), outputs=Outputs(features=[of1, of2, of1_, of2_])
 )