Fix improper row assignment for Lift charts as noted by #169

Author: smlindauer

src/sasctl/pzmm/write_json_files.py

@@ -875,7 +875,7 @@ def calculate_model_statistics(
                 json_dict[1]["data"][j].update(roc_dict[j])
 
             lift_df = pd.DataFrame(conn.CASTable("Lift", caslib="Public").to_frame())
-            lift_dict = cls.apply_dataframe_to_json(json_dict[2]["data"], i, lift_df)
+            lift_dict = cls.apply_dataframe_to_json(json_dict[2]["data"], i, lift_df, 1)
             for j in range(len(lift_dict)):
                 json_dict[2]["data"][j].update(lift_dict[j])
 
@@ -1019,7 +1019,7 @@ def stat_dataset_to_dataframe(
 
     @staticmethod
     def apply_dataframe_to_json(
-        json_dict: dict, partition: int, stat_df: DataFrame
+        json_dict: dict, partition: int, stat_df: DataFrame, is_lift: bool = False
     ) -> dict:
         """
         Map the values of the ROC or Lift charts from SAS CAS to the dictionary
@@ -1033,6 +1033,9 @@ def apply_dataframe_to_json(
             Numerical representation of the data partition. Either 0, 1, or 2.
         stat_df : pandas.DataFrame
             ROC or Lift DataFrame generated from the SAS CAS percentile action set.
+        is_lift : bool
+            Specify whether to use logic for Lift or ROC row counting. Default value is
+            False.
 
         Returns
         -------
@@ -1042,7 +1045,10 @@ def apply_dataframe_to_json(
         """
         for row_num in range(len(stat_df)):
             row_dict = stat_df.iloc[row_num].replace(float("nan"), None).to_dict()
-            json_dict[row_num + partition * len(stat_df)]["dataMap"].update(row_dict)
+            if is_lift:
+                json_dict[(row_num + partition + 1) + partition * len(stat_df)]["dataMap"].update(row_dict)
+            else:
+                json_dict[row_num + (partition * len(stat_df))]["dataMap"].update(row_dict)
         return json_dict
 
     # noinspection PyCallingNonCallable, PyNestedDecorators