优化封装模型

predict/dict_predict.py

@@ -170,7 +170,7 @@ class DicPredict(BasePredictor):
             for col_diffrate in sort_li_diffrate[1:]:
             for col_diffrate in sort_li_diffrate[1:]:
                 temp_pre_model_res.pop(col_diffrate[0])
                 temp_pre_model_res.pop(col_diffrate[0])
 
 
-        # 若有多个参数列且没有封装列, 就先进行封装列的提取(封装率需要大于0), 再进行参数特征的数量比较, 特征最多的选为目标列
+        # 若有多个参数列且没有封装列, 就先进行封装列的提取(封装率需要大于0.3), 再进行参数特征的数量比较, 特征最多的选为目标列
         prob_param_cols = [i for i in temp_pre_model_res if temp_pre_model_res[i] == '参数']
         prob_param_cols = [i for i in temp_pre_model_res if temp_pre_model_res[i] == '参数']
         self.info.info(f'可能的参数列有{str(prob_param_cols)}')
         self.info.info(f'可能的参数列有{str(prob_param_cols)}')
         if len(prob_param_cols) >= 2:
         if len(prob_param_cols) >= 2:
@@ -178,7 +178,7 @@ class DicPredict(BasePredictor):
                 li_encap_rate = [(col, BasePredictor.get_encap_rate(not_null_dic_data[col])) for col in
                 li_encap_rate = [(col, BasePredictor.get_encap_rate(not_null_dic_data[col])) for col in
                                    prob_param_cols]
                                    prob_param_cols]
                 sort_li_encaprate = sorted(li_encap_rate, key=lambda x: x[1], reverse=True)
                 sort_li_encaprate = sorted(li_encap_rate, key=lambda x: x[1], reverse=True)
-                if sort_li_encaprate[0][1] >= 0:
+                if sort_li_encaprate[0][1] > 0.3:
                     temp_pre_model_res[sort_li_encaprate[0][0]] = '封装'
                     temp_pre_model_res[sort_li_encaprate[0][0]] = '封装'
                     prob_param_cols.remove(sort_li_encaprate[0][0])
                     prob_param_cols.remove(sort_li_encaprate[0][0])
 
 
@@ -230,11 +230,11 @@ class DicPredict(BasePredictor):
 
 
     def predict(self, dic_data, predict_type='all'):
     def predict(self, dic_data, predict_type='all'):
         self.info.info('预测类型为: %s, 接收数据: %s' % (predict_type, str(dic_data)))
         self.info.info('预测类型为: %s, 接收数据: %s' % (predict_type, str(dic_data)))
-        dic_data = self.pre_deal(dic_data)
+        dealed_dic_data = self.pre_deal(dic_data)
 
 
         if predict_type == 'all':
         if predict_type == 'all':
             head_id_res = self.head_predict(dic_data)
             head_id_res = self.head_predict(dic_data)
-            model_id_res = self.model_predict(dic_data)
+            model_id_res = self.model_predict(dealed_dic_data)
             self.info.info('表头预测结果: ' + str(head_id_res))
             self.info.info('表头预测结果: ' + str(head_id_res))
             self.info.info('模型预测结果: ' + str(model_id_res))
             self.info.info('模型预测结果: ' + str(model_id_res))
 
 
@@ -259,7 +259,7 @@ class DicPredict(BasePredictor):
                 return res
                 return res
 
 
         elif predict_type == 'model':
         elif predict_type == 'model':
-            model_id_res = self.model_predict(dic_data)
+            model_id_res = self.model_predict(dealed_dic_data)
             if model_id_res:
             if model_id_res:
                 return model_id_res
                 return model_id_res
 
 
@@ -269,9 +269,12 @@ class DicPredict(BasePredictor):
                 return head_id_res
                 return head_id_res
 
 
     def pre_deal(self, dic_data):
     def pre_deal(self, dic_data):
+        """
+        模型预测不预测空置率大于0.8的列
+        """
         new_dic_data = {}
         new_dic_data = {}
         for k, v in dic_data.items():
         for k, v in dic_data.items():
-            # 去掉空置率大于等于0.8的列
+            # 去掉空置率大于0.8的列
             counter = 0
             counter = 0
             for item in v:
             for item in v:
                 if str(item).strip() == BasePredictor.PLACEHOLDER:
                 if str(item).strip() == BasePredictor.PLACEHOLDER:

static_config.py

@@ -4,7 +4,7 @@
 li_category = ["类别", "分类", "名称", "类别名称", "类型", "产品分类"]
 li_category = ["类别", "分类", "名称", "类别名称", "类型", "产品分类"]
 li_param = ["参数", "规格", "描述", "值", "description"]
 li_param = ["参数", "规格", "描述", "值", "description"]
 li_gn = ["型号", "参考料号", "料号", "mpn", "厂商编码", "元器件", "规格型号"]
 li_gn = ["型号", "参考料号", "料号", "mpn", "厂商编码", "元器件", "规格型号"]
-li_num = ["数量", "用量(pcs)", "用量", "pcs", "quantity", "qty", "buy qty", "buy quantity", "单板数量", "采购数量"]
+li_num = ["数量", "用量(pcs)", "用量", "pcs", "quantity", "qty", "buy qty", "buy quantity", "单板数量", "采购数量", "单机用量", "单机数量", "单板用量"]
 li_brand = ["品牌", "品牌/厂商", "参考品牌", "参考供应商", "厂商", "参考供应商", "参考厂商", "manufacturer制造商", "manufacturer", "厂牌"]
 li_brand = ["品牌", "品牌/厂商", "参考品牌", "参考供应商", "厂商", "参考供应商", "参考厂商", "manufacturer制造商", "manufacturer", "厂牌"]
 li_encap = ["封装", "封装规格", "encapsulation", "footprint封装", "packagereference"]
 li_encap = ["封装", "封装规格", "encapsulation", "footprint封装", "packagereference"]
 li_position = ["位号", "位置", "标号", "点位"]
 li_position = ["位号", "位置", "标号", "点位"]