Merge branch 'master' of git.wanmeizhensuo.com:ML/ffm-baseline

mail

eda/feededa/src/main/scala/com/gmei/Recommendation_strategy_all.scala

@@ -52,8 +52,8 @@ object Recommendation_strategy_all {
       ti.tidbMapTable(dbName = "jerry_prod", tableName = "merge_queue_table")
 
 
-//      val stat_date = GmeiConfig.getMinusNDate(1)
-      val stat_date = param.date
+      val stat_date = GmeiConfig.getMinusNDate(1)
+//      val stat_date = param.date
       //println(param.date)
       val partition_date = stat_date.replace("-","")
       val decive_id_oldUser = sc.sql(
@@ -201,7 +201,7 @@ object Recommendation_strategy_all {
           .join(device_num_hit,"stat_date")
       result1.show()
 
-//      GmeiConfig.writeToJDBCTable(result1, "Recommendation_strategy_temp", SaveMode.Append)
+      GmeiConfig.writeToJDBCTable(result1, "Recommendation_strategy_temp", SaveMode.Append)
 
 
       //device_id尾号1有点击用户日记本点击数
@@ -298,7 +298,7 @@ object Recommendation_strategy_all {
         .join(device_all_cover,"stat_date")
       result2.show()
 
-//      GmeiConfig.writeToJDBCTable(result2, "strategy_other", SaveMode.Append)
+      GmeiConfig.writeToJDBCTable(result2, "strategy_other", SaveMode.Append)
 
 
 //统计新用户点击率
@@ -371,7 +371,7 @@ object Recommendation_strategy_all {
         .join(imp_count_newUser_all,"stat_date")
       result3.show()
 
-//      GmeiConfig.writeToJDBCTable(result3, "Recommendation_strategy_newUser", SaveMode.Append)
+      GmeiConfig.writeToJDBCTable(result3, "Recommendation_strategy_newUser", SaveMode.Append)
 
     }
 

tensnsorflow/device_read_time_normal.py

@@ -22,18 +22,25 @@ def normal():
     db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_test')
     sql = "select * from device_read_time"
     df = con_sql(db, sql)
-    df = df.rename(columns={0:"device_id",1:"0",2:"1",3:"10",4:"1024",5:"1080",6:"11",
-                            7:"12",8:"13",9:"2",10:"2054",11:"2214",12:"3",13:"4",14:"5",16:"6933",
-                            17:"7",18:"9",19:"922",20:"929",21:"971",22:"992"})
-    device_id = df[["device_id"]]
-    df = df.drop("device_id",axis=1)
-    minMax = MinMaxScaler()
-    result = pd.DataFrame(minMax.fit_transform(df),columns=["0","1","10","1024","1080","11",
-                            "12","13","2","2054","2214","3","4","5","6933",
-                            "7","9","922","929","971","992"])
-    result = device_id.join(result)
+    df = df.rename(columns={0:"device_id",1:"kongbai",2:"eye",3:"simi",4:"zitizhifang",5:"banyongjiu",6:"teeth",
+                            7:"kouchun",8:"ear",9:"nose",10:"banyongjiuzhuang",11:"qita",12:"lunkuo",
+                            13:"shoushen",14:"skin",16:"shenghuo",
+                            17:"breast",18:"hair",19:"kangshuai",20:"shili",21:"chanhou",22:"zhushe"})
+    # device_id = df[["device_id"]]
+    # df = df.drop("device_id",axis=1)
+    # minMax = MinMaxScaler()
+    # result = pd.DataFrame(minMax.fit_transform(df),columns=["0","1","10","1024","1080","11",
+    #                         "12","13","2","2054","2214","3","4","5","6933",
+    #                         "7","9","922","929","971","992"])
+    # result = device_id.join(result)
+    l = list(df.columns)
+    l.remove("device_id")
+    df["sum"] = df.sum(axis=1)
+    for i in l:
+        df[i] = df[i]/df["sum"]
+    df = df.drop("sum",axis=1)
     yconnect = create_engine('mysql+pymysql://root:3SYz54LS9#^9sBvC@10.66.157.22:4000/jerry_test?charset=utf8')
-    pd.io.sql.to_sql(result, "device_read_time_normal", yconnect, schema='jerry_test', if_exists='fail', index=False)
+    pd.io.sql.to_sql(df, "device_read_time_normal", yconnect, schema='jerry_test', if_exists='fail', index=False)
 
 
 if __name__ == "__main__":