特征工程优化

spark/featureEng.py

@@ -834,7 +834,7 @@ if __name__ == '__main__':
     train_columns = model_columns + ["label", "timestamp"]
     trainSamples = samplesWithUserFeatures.select(*train_columns)
     print("write to hdfs start...")
-    splitTimestamp = int(time.mktime(time.strptime(addDays(-1), "%Y%m%d")))
+    splitTimestamp = int(time.mktime(time.strptime(addDays(0), "%Y%m%d")))
     splitAndSaveTrainingTestSamplesByTimeStamp(trainSamples, splitTimestamp, TRAIN_FILE_PATH)
     print("write to hdfs success...")
     timestmp7 = int(round(time.time()))

train/train_service.py

@@ -4,6 +4,8 @@ import pandas as pd
 import time
 import sys
 import os
+from datetime import date, timedelta
+
 
 sys.path.append(os.path.dirname(os.path.abspath(os.path.dirname(__file__))))
 import utils.connUtils as connUtils
@@ -170,12 +172,14 @@ def predict(model_path,df):
         model_new.predict(dd, batch_size=10000)
         print("测试样本数：{},测试耗时ms:{}".format(n, int(round(time.time() * 1000)) - timestmp1))
 
-
+def addDays(n, format="%Y%m%d"):
+    return (date.today() + timedelta(days=n)).strftime(format)
 
 if __name__ == '__main__':
 
     curTime = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())
     print("train_service执行时间:{}".format(curTime))
+    splitTimestamp = int(time.mktime(time.strptime(addDays(-1), "%Y%m%d")))
 
     # redis中加载数据字典
     print("redis 中加载模型字典...")