fix feature_bin_stats & rule.save

itlubber · Sep 2, 2024 · 6b826b9 · 6b826b9
1 parent 12b8a05
commit 6b826b9
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 3 deletions.
diff --git a/scorecardpipeline/processing.py b/scorecardpipeline/processing.py
@@ -838,8 +838,16 @@ def feature_bin_stats(data, feature, target="target", overdue=None, dpd=None, ru
     if overdue and dpd is None:
         raise ValueError("传入 overdue 参数时必须同时传入 dpd")
 
+    drop_empty = True if not empty_separate and data[feature].isnull().sum() <= 0 else False
+
     if overdue is None:
-        return Combiner.feature_bin_stats(data, feature, target=target, rules=rules, method=method, desc=desc, combiner=combiner, ks=ks, max_n_bins=max_n_bins, min_bin_size=min_bin_size, max_bin_size=max_bin_size, greater_is_better=greater_is_better, empty_separate=empty_separate, return_cols=return_cols, return_rules=return_rules, verbose=verbose, **kwargs)
+        table, rule = Combiner.feature_bin_stats(data, feature, target=target, rules=rules, method=method, desc=desc, combiner=combiner, ks=ks, max_n_bins=max_n_bins, min_bin_size=min_bin_size, max_bin_size=max_bin_size, greater_is_better=greater_is_better, empty_separate=empty_separate, return_cols=return_cols, return_rules=True, verbose=verbose, **kwargs)
+
+        if drop_empty:
+            table = table.iloc[:-1]
+            rule = rule[:-1]
+
+        return table, rule if return_rules else table
 
     if not isinstance(overdue, list):
         overdue = [overdue]
@@ -879,6 +887,10 @@ def feature_bin_stats(data, feature, target="target", overdue=None, dpd=None, ru
 
                 table = table.merge(_table, on=[("分箱详情", c) for c in merge_columns])
 
+    if drop_empty:
+        table = table.iloc[:-1]
+        rule = rule[:-1]
+
     if return_cols is not None:
         if not isinstance(return_cols, list):
             return_cols = [return_cols]

diff --git a/scorecardpipeline/rule.py b/scorecardpipeline/rule.py
@@ -415,7 +415,7 @@ def save(report, excel_writer, sheet_name=None, merge_column=None, percent_cols=
 
         if percent_cols:
             percent_cols = [c for c in report.columns if (isinstance(c, tuple) and c[-1] in percent_cols) or (not isinstance(c, tuple) and c in percent_cols)]
-            
+
         if condition_cols:
             condition_cols = [c for c in report.columns if (isinstance(c, tuple) and c[-1] in condition_cols) or (not isinstance(c, tuple) and c in condition_cols)]
 
@@ -425,4 +425,5 @@ def save(report, excel_writer, sheet_name=None, merge_column=None, percent_cols=
         if color_cols:
             color_cols = [c for c in report.columns if (isinstance(c, tuple) and c[-1] in color_cols) or (not isinstance(c, tuple) and c in color_cols)]
 
-        dataframe2excel(report, excel_writer, sheet_name=sheet_name, merge_column=merge_column, percent_cols=percent_cols, condition_cols=condition_cols, custom_cols=custom_cols, custom_format=custom_format, color_cols=color_cols, start_col=start_col, start_row=start_row, **kwargs)
+        end_row, end_col = dataframe2excel(report, excel_writer, sheet_name=sheet_name, merge_column=merge_column, percent_cols=percent_cols, condition_cols=condition_cols, custom_cols=custom_cols, custom_format=custom_format, color_cols=color_cols, start_col=start_col, start_row=start_row, **kwargs)
+        return end_row, end_col