fix insert_value2sheet

scorecardpipeline/excel_writer.py

@@ -127,7 +127,7 @@ def move_sheet(self, worksheet, offset: int = 0, index: int = None):
 
         self.workbook.move_sheet(worksheet, offset=offset)
 
-    def insert_value2sheet(self, worksheet, insert_space, value="", style="content", auto_width=False, end_space=None):
+    def insert_value2sheet(self, worksheet, insert_space, value="", style="content", auto_width=False, end_space=None, align: dict=None, max_col_width=50):
         """
         向sheet中的某个单元格插入某种样式的内容
 
@@ -137,6 +137,8 @@ def insert_value2sheet(self, worksheet, insert_space, value="", style="content",
         :param style: 渲染的样式，参考 init_style 中初始设置的样式
         :param end_space: 如果需要合并单元格，传入需要截止的单元格位置信息，可以是 "B2" 或者 (2, 2) 任意一种形式
         :param auto_width: 是否开启自动调整列宽
+        :param align: 文本排列方式, 参考: Alignment
+        :param max_col_width: 单元格列最大宽度，默认 50
 
         :return: 返回插入元素最后一列之后、最后一行之后的位置
         """
@@ -150,6 +152,11 @@ def insert_value2sheet(self, worksheet, insert_space, value="", style="content",
         cell = worksheet[f"{start_col}{start_row}"]
         cell.style = style
 
+        if align:
+            _align = {"horizontal": "center", "vertical": "center"}
+            _align.update(align)
+            cell.alignment = Alignment(**_align)
+
         if end_space is not None:
             if isinstance(end_space, str):
                 end_col = re.findall('\D+', end_space)[0]
@@ -163,12 +170,13 @@ def insert_value2sheet(self, worksheet, insert_space, value="", style="content",
         worksheet[f"{start_col}{start_row}"] = value
 
         if auto_width:
-            # original_styles = [worksheet[f"{start_col}{i}"].fill for i in range(1, worksheet.max_row + 1)]
+            # original_styles = [worksheet[f"{start_col}{i}"].fill.copy() for i in range(1, worksheet.max_row + 1)]
             curr_width = worksheet.column_dimensions[start_col].width
-            auto_width = min(max([(self.check_contain_chinese(value)[1] * self.english_width + self.check_contain_chinese(value)[2] * self.chinese_width) * self.fontsize, 10, curr_width]), 50)
+            auto_width = min(max([(self.check_contain_chinese(value)[1] * self.english_width + self.check_contain_chinese(value)[2] * self.chinese_width) * self.fontsize, 10, curr_width]), max_col_width)
             worksheet.column_dimensions[start_col].width = auto_width
+
             # for i in range(worksheet.max_row):
-            #     worksheet[f"{start_col}{i + 1}"].fill = original_styles[i].copy()
+            #     worksheet[f"{start_col}{i + 1}"].fill = original_styles[i]
 
         if end_space is not None:
             return end_row + 1, column_index_from_string(end_col) + 1

scorecardpipeline/scorecard.py

@@ -183,7 +183,7 @@ def _transform(self, X):
 
 class BoxCoxScoreTransformer(BaseScoreTransformer):
     def __init__(self, down_lmt=300, up_lmt=1000, greater_is_better=True, cutoff=None):
-        super(BoxCoxScoreTransformer, self).__init__(down_lmt=down_lmt, up_lmt=up_lmt, greater_is_better=greater_is_better, cutoff=cutoff)
+        super().__init__(down_lmt=down_lmt, up_lmt=up_lmt, greater_is_better=greater_is_better, cutoff=cutoff)
 
     @staticmethod
     def _box_cox_optimize(x):
@@ -205,6 +205,7 @@ def fit(self, X, y=None, **fit_params):
             self.lambdas_ = np.array([self._box_cox_optimize(col) for col in X.T])
         for i, lmbda in enumerate(self.lambdas_):
             X[:, i] = stats.boxcox(X[:, i], lmbda)
+        print(X.max(), X.min(), X.mean())
         self.scaler_ = MinMaxScaler(feature_range=(self.down_lmt, self.up_lmt)).fit(X)
         return self
 
@@ -295,16 +296,16 @@ def _box_cox_inverse_tranform(x, lmbda):
 
     test_select = h2o.H2OFrame(load_pickle("/Users/lubberit/Desktop/workspace/scorecardpipeline/examples/model_report/h2o_model/test_select.pkl"))
 
-    model_path = 'model_report/h2o_model/StackedEnsemble_BestOfFamily_1_AutoML_1_20240415_162619'
+    model_path = '/Users/lubberit/Desktop/workspace/scorecardpipeline/examples/model_report/h2o_model/StackedEnsemble_BestOfFamily_1_AutoML_1_20240415_162619'
     best_model = h2o.load_model(model_path)
 
-    score_transform = StandardScoreTransformer(base_score=400, pdo=50, bad_rate=test_select["target"].mean()[0], greater_is_better=True)
-    # score_transform = BoxCoxScoreTransformer(greater_is_better=False)
+    # score_transform = StandardScoreTransformer(base_score=400, pdo=50, bad_rate=test_select["target"].mean()[0], greater_is_better=True)
+    score_transform = BoxCoxScoreTransformer(greater_is_better=False)
     y_pred = best_model.predict(test_select).as_data_frame()[["p1"]]
     score_transform.fit(y_pred)
 
     print(best_model.predict(test_select))
     score = score_transform.transform(y_pred)
     print(score)
     print(score_transform.inverse_transform(score))
-    print(score_transform.scorecard_scale())
+    # print(score_transform.scorecard_scale())