align summary_method -> aggregation_method

src/seismometer/_api.py

@@ -181,7 +181,7 @@ def fairness_audit(metric_list: Optional[list[str]] = None, fairness_threshold=1
         sg.entity_keys,
         score=sg.output,
         ref_event=sg.predict_time,
-        summary_method=sg.event_aggregation_method(sg.target),
+        aggregation_method=sg.event_aggregation_method(sg.target),
     )[[sg.target, sg.output] + sensitive_groups]
 
     display_fairness_audit(
@@ -442,7 +442,7 @@ def _plot_leadtime_enc(
         entity_keys,
         score=score,
         ref_event=target_zero,
-        summary_method="first",
+        aggregation_method="first",
     )[[target_zero, ref_time, cohort_col]]
 
     # filter by group size
@@ -508,7 +508,7 @@ def _plot_cohort_evaluation(
     subgroups: list[str],
     censor_threshold: int = 10,
     per_context_id: bool = False,
-    summary_method: str = "max",
+    aggregation_method: str = "max",
     ref_time: str = None,
 ) -> HTML:
     """
@@ -533,20 +533,22 @@ def _plot_cohort_evaluation(
     censor_threshold : int
         minimum rows to allow in a plot, by default 10
     per_context_id : bool, optional
-        if true, summarize scores for each context, by default False
-    summary_method : str, optional
-        method to summarize multiple scores into a single value before calculation of performance, by default "max"
+        if true, aggregate scores for each context, by default False
+    aggregation_method : str, optional
+        method to reduce multiple scores into a single value before calculation of performance, by default "max"
         ignored if per_context_id is False
     ref_time : str, optional
-        reference time column used for summarization when per_context_id is True and summary_method is time-based
+        reference time column used for aggregation when per_context_id is True and aggregation_method is time-based
 
     Returns
     -------
     HTML
         _description_
     """
     data = (
-        pdh.event_score(dataframe, entity_keys, score=output, ref_event=ref_time, summary_method=summary_method)
+        pdh.event_score(
+            dataframe, entity_keys, score=output, ref_event=ref_time, aggregation_method=aggregation_method
+        )
         if per_context_id
         else dataframe
     )
@@ -598,7 +600,7 @@ def _model_evaluation(
     output: str,
     thresholds: Optional[list[float]],
     per_context_id: bool = False,
-    summary_method: str = "max",
+    aggregation_method: str = "max",
     ref_time: str = None,
 ) -> HTML:
     """
@@ -620,19 +622,21 @@ def _model_evaluation(
         model thresholds
     per_context_id : bool, optional
         report only the max score for a given entity context, by default False
-    summary_method : str, optional
-        method to summarize multiple scores into a single value before calculation of performance, by default "max"
+    aggregation_method : str, optional
+        method to reduce multiple scores into a single value before calculation of performance, by default "max"
         ignored if per_context_id is False
     ref_time : str, optional
-        reference time column used for summarization when per_context_id is True and summary_method is time-based
+        reference time column used for aggregation when per_context_id is True and aggregation_method is time-based
 
     Returns
     -------
     HTML
         Plot of model evaluation metrics
     """
     data = (
-        pdh.event_score(dataframe, entity_keys, score=output, ref_event=ref_time, summary_method=summary_method)
+        pdh.event_score(
+            dataframe, entity_keys, score=output, ref_event=ref_time, aggregation_method=aggregation_method
+        )
         if per_context_id
         else dataframe
     )
@@ -912,7 +916,7 @@ def show_info(plot_help: bool = False):
 
 def _style_cohort_summaries(df: pd.DataFrame, attribute: str) -> Styler:
     """
-    Adds required styling to a cohort summary dataframe.
+    Adds required styling to a cohort dataframe.
 
     Parameters
     ----------

src/seismometer/data/pandas_helpers.py

@@ -163,7 +163,7 @@ def event_score(
     pks: list[str],
     score: str,
     ref_event: Optional[str] = None,
-    summary_method: str = "max",
+    aggregation_method: str = "max",
 ) -> pd.DataFrame:
     """
     Reduces a dataframe of all predictions to a single row of significance; such as the max or most recent value for
@@ -180,40 +180,40 @@ def event_score(
         The column name containing the score value.
     ref_event : Optional[str], optional
         The column name containing the time to consider, by default None.
-    summary_method : str, optional
+    aggregation_method : str, optional
         A string describing the method to select a value, by default 'max'.
 
     Returns
     -------
     pd.DataFrame
         The reduced dataframe with one row per combination of pks.
     """
-    logger.debug(f"Combining scores using {summary_method} for {score} on {ref_event}")
+    logger.debug(f"Combining scores using {aggregation_method} for {score} on {ref_event}")
     # groupby.agg works on columns indivdually - this wants entire row where a condition is met
     # start with first/last/max/min
 
     ref_score = _resolve_score_col(merged_frame, score)
-    if summary_method == "max":
+    if aggregation_method == "max":
         ref_col = ref_score
 
         def apply_fn(gf):
             return gf.idxmax()
 
-    elif summary_method == "min":
+    elif aggregation_method == "min":
         ref_col = ref_score
 
         def apply_fn(gf):
             return gf.idxmin()
 
     # merged frame has time columns only for events in appropriate time window,
     # implicitly reduces to positive label (need method to re-add negative samples)
-    elif summary_method == "last":
+    elif aggregation_method == "last":
 
         def apply_fn(gf):
             return gf.idxmax()
 
         ref_col = _resolve_time_col(merged_frame, ref_event)
-    elif summary_method == "first":
+    elif aggregation_method == "first":
 
         def apply_fn(gf):
             return gf.idxmin()

tests/data/test_events.py

@@ -25,16 +25,16 @@ def event_data(res):
 
 class Test_Event_Score:
     @pytest.mark.parametrize("ref_event", ["Target", "PredictTime", "Reference_5_15_Time"])
-    @pytest.mark.parametrize("summary_method", ["min", "max", "first", "last"])
+    @pytest.mark.parametrize("aggregation_method", ["min", "max", "first", "last"])
     @pytest.mark.parametrize("id_, csn", [pytest.param(1, 0, id="monotonic-increasing")])
-    def test_bad_event(self, id_, csn, summary_method, ref_event, event_data):
+    def test_bad_event(self, id_, csn, aggregation_method, ref_event, event_data):
         input_frame, expected_frame = event_data
         expected_score = expected_frame.loc[
             (expected_frame["Id"] == id_)
             & (expected_frame["CSN"] == csn)
             & (expected_frame["ref_event"] == ref_event),
-            summary_method,
+            aggregation_method,
         ]
 
-        actual = undertest.event_score(input_frame, ["Id", "CSN"], "ModelScore", ref_event, summary_method)
+        actual = undertest.event_score(input_frame, ["Id", "CSN"], "ModelScore", ref_event, aggregation_method)
         assert actual["ModelScore"].tolist() == expected_score.tolist()