awg commit algorithm components

models_processing/model_predict/model_prediction_lightgbm.py

@@ -25,7 +25,7 @@ def forecast_data_distribution(pre_data,args):
 
     if len(pre_data)==0:
         send_message('lightgbm预测组件', args['farmId'], '请注意：获取NWP数据为空，预测文件无法生成！')
-        result = pd.DataFrame({col_time:[],'farm_id':[],'power_forecast':[]})
+        result = pd.DataFrame({'farm_id':[], col_time:[], 'power_forecast':[]})
 
     elif len(pre_data[pre_data[col_time].str.contains(tomorrow)])<96:
         send_message('lightgbm预测组件', args['farmId'], "日前数据记录缺失，不足96条，用DQ代替并补值！")
@@ -34,7 +34,7 @@ def forecast_data_distribution(pre_data,args):
         date_range = pd.date_range(start=start_time, end=end_time, freq='15T').strftime('%Y-%m-%d %H:%M:%S').tolist()
         df_date = pd.DataFrame({col_time:date_range})
         result = pd.merge(df_date,pre_data,how='left',on=col_time).sort_values(by=col_time).fillna(method='ffill').fillna(method='bfill')
-        result = result[['date_time', 'farm_id', 'power_forecast']]
+        result = result[['farm_id', 'date_time', 'power_forecast']]
     else:
         df = pre_data.sort_values(by=col_time).fillna(method='ffill').fillna(method='bfill')
         mongodb_connection, mongodb_database, mongodb_model_table, model_name = "mongodb://root:sdhjfREWFWEF23e@192.168.1.43:30000/", \
@@ -50,15 +50,15 @@ def forecast_data_distribution(pre_data,args):
             diff = set(model.feature_name()) - set(pre_data.columns)
             if len(diff) > 0:
                 send_message('lightgbm预测组件', args['farmId'], f'NWP特征列缺失,使用DQ代替！features:{diff}')
-                result = pre_data[['date_time', 'farm_id', 'power_forecast']]
+                result = pre_data[['farm_id', 'date_time', 'power_forecast']]
             else:
                 df['power_forecast'] = model.predict(df[model.feature_name()])
                 df.loc[df['power_forecast'] < 0, 'power_forecast'] = 0
                 print("model predict result  successfully!")
-                result = df[['date_time', 'farm_id', 'power_forecast']]
+                result = df[['farm_id', 'date_time', 'power_forecast']]
         else:
             send_message('lightgbm预测组件', args['farmId'], "日前数据记录缺失，不足96条，用DQ代替并补值！")
-            result = pre_data[['date_time', 'farm_id', 'power_forecast']]
+            result = pre_data[['farm_id', 'date_time', 'power_forecast']]
     result['power_forecast'] = round(result['power_forecast'],2)
     return result