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

change path

eda/esmm/Feature_pipline/get_tfrecord.py

-#!/usr/bin/env python
 #coding=utf-8
 
 from __future__ import absolute_import
@@ -26,8 +25,8 @@ tf.app.flags.DEFINE_integer("threads", 16, "threads num")
 #User_Fileds = set(['101','109_14','110_14','127_14','150_14','121','122','124','125','126','127','128','129'])
 #Ad_Fileds = set(['205','206','207','210','216'])
 #Context_Fileds = set(['508','509','702','853','301'])
-#Common_Fileds = {'1':'1','2':'2','3':'3','4':'4','5':'5','6':'6','7':'7','8':'8','9':'9','10':'10','11':'11','12':'12','13':'13','14':'14','15':'15','16':'16','17':'17','18':'18','19':'19','20':'20','21':'21','22':'22','23':'23','24':'24','25':'25','26':'26','27':'27','28':'28','29':'29','30':'30'}
-Common_Fileds = {'1':'1','2':'2','3':'3','4':'4','5':'5','6':'6','7':'7','8':'8','9':'9','10':'10','11':'11'}
+Common_Fileds = {'1':'1','2':'2','3':'3','4':'4','5':'5','6':'6','7':'7','8':'8','9':'9','10':'10','11':'11','12':'12','13':'13','14':'14','15':'15','16':'16','17':'17','18':'18','19':'19','20':'20','21':'21','22':'22','23':'23'}
+#Common_Fileds = {'1':'1','2':'2','3':'3','4':'4','5':'5','6':'6','7':'7','8':'8','9':'9','10':'10','11':'11'}
 UMH_Fileds = {'109_14':('u_cat','12'),'110_14':('u_shop','13'),'127_14':('u_brand','14'),'150_14':('u_int','15')}      #user multi-hot feature
 Ad_Fileds = {'206':('a_cat','16'),'207':('a_shop','17'),'210':('a_int','18'),'216':('a_brand','19')}                  #ad feature for DIN
 

eda/esmm/Model_pipline/DeepCvrMTL.py

-#!/usr/bin/env python
 #coding=utf-8
 
 #from __future__ import absolute_import
@@ -346,7 +345,7 @@ def main(_):
             print("-"*100)
             with open(FLAGS.data_dir + "/pred.txt", "w") as fo:
                 for prob in preds:
-                    fo.write("%f\t%f\n" % (prob['pctr'], prob['pcvr']))
+                    fo.write("%f\t%f\t%f\n" % (prob['pctr'], prob['pcvr'], prob['pctcvr']))
     elif FLAGS.task_type == 'export':
         print("Not Implemented, Do It Yourself!")
         #feature_spec = tf.feature_column.make_parse_example_spec(feature_columns)

eda/esmm/Model_pipline/send_mail.py

-# -*- coding: utf-8 -*-
+#coding=utf-8
 
 import smtplib
 from email.mime.text import MIMEText

eda/esmm/Model_pipline/sort_and_2sql.py

+#coding=utf-8
+
 from sqlalchemy import create_engine
 import pandas as pd
 import pymysql
@@ -17,39 +19,30 @@ def con_sql(sql):
     return result
 
 def set_join(lst):
-    return ','.join(set(lst))
+    return ','.join([str(i) for i in set(lst)])
 
 def main():
-    sql = "select device_id,city_id,cid from esmm_data2ffm_infer_native"
-    result = con_sql(sql)
-    dct = {"uid":[],"city":[],"cid_id":[]}
-    for i in result:
-        dct["uid"].append(i[0])
-        dct["city"].append(i[1])
-        dct["cid_id"].append(i[2])
-
-    df1 = pd.read_csv("/home/gaoyazhe/data/native/pred.txt",sep='\t',header=None,names=["ctr","cvr"])
-    df2 = pd.DataFrame(dct)
-    df2["ctr"],df2["cvr"] = df1["ctr"],df1["cvr"]
-    df3 = df2.groupby(by=["uid","city"]).apply(lambda x: x.sort_values(by="cvr",ascending=False)).reset_index(drop=True).groupby(by=["uid","city"]).agg({'cid_id':set_join}).reset_index(drop=False)
+
+    # native queue
+    df2 = pd.read_csv('/home/gaoyazhe/data/native.csv',usecols=[0,1,2],header=0,names=['uid','city','cid_id'],sep='\t')
+    df2['cid_id'] = df2['cid_id'].astype('object')
+
+    df1 = pd.read_csv("/home/gaoyazhe/data/native/pred.txt",sep='\t',header=None,names=["ctr","cvr","ctcvr"])
+    df2["ctr"],df2["cvr"],df2["ctcvr"] = df1["ctr"],df1["cvr"],df1["ctcvr"]
+    df3 = df2.groupby(by=["uid","city"]).apply(lambda x: x.sort_values(by="ctcvr",ascending=False)).reset_index(drop=True).groupby(by=["uid","city"]).agg({'cid_id':set_join}).reset_index(drop=False)
     ctime = int(time.time())
     df3["time"] = ctime
     df3.columns = ["device_id","city_id","native_queue","time"]
     print("native_device_count",df3.shape)
 
 
-    sql_nearby = "select device_id,city_id,cid from esmm_data2ffm_infer_nearby"
-    result = con_sql(sql_nearby)
-    dct = {"uid":[],"city":[],"cid_id":[]}
-    for i in result:
-        dct["uid"].append(i[0])
-        dct["city"].append(i[1])
-        dct["cid_id"].append(i[2])
+    # nearby queue
+    df2 = pd.read_csv('/home/gaoyazhe/data/nearby.csv',usecols=[0,1,2],header=0,names=['uid','city','cid_id'],sep='\t')
+    df2['cid_id'] = df2['cid_id'].astype('object')
 
-    df1 = pd.read_csv("/home/gaoyazhe/data/nearby/pred.txt",sep='\t',header=None,names=["ctr","cvr"])
-    df2 = pd.DataFrame(dct)
-    df2["ctr"],df2["cvr"] = df1["ctr"],df1["cvr"]
-    df4 = df2.groupby(by=["uid","city"]).apply(lambda x: x.sort_values(by="cvr",ascending=False)).reset_index(drop=True).groupby(by=["uid","city"]).agg({'cid_id':set_join}).reset_index(drop=False)
+    df1 = pd.read_csv("/home/gaoyazhe/data/nearby/pred.txt",sep='\t',header=None,names=["ctr","cvr","ctcvr"])
+    df2["ctr"], df2["cvr"], df2["ctcvr"] = df1["ctr"], df1["cvr"], df1["ctcvr"]
+    df4 = df2.groupby(by=["uid","city"]).apply(lambda x: x.sort_values(by="ctcvr",ascending=False)).reset_index(drop=True).groupby(by=["uid","city"]).agg({'cid_id':set_join}).reset_index(drop=False)
     df4.columns = ["device_id","city_id","nearby_queue"]
     print("nearby_device_count",df4.shape)
 

eda/esmm/Model_pipline/submit.sh

@@ -15,11 +15,8 @@ rm ${DATA_PATH}/va/*
 rm ${DATA_PATH}/native/*
 rm ${DATA_PATH}/nearby/*
 
-echo "mysql to csv"
-mysql -u root -p3SYz54LS9#^9sBvC -h 10.66.157.22 -P 4000 -D jerry_test -e "select number,data from esmm_data2ffm_train" > ${DATA_PATH}/tr.csv
-mysql -u root -p3SYz54LS9#^9sBvC -h 10.66.157.22 -P 4000 -D jerry_test -e "select number,data from esmm_data2ffm_cv" > ${DATA_PATH}/va.csv
-mysql -u root -p3SYz54LS9#^9sBvC -h 10.66.157.22 -P 4000 -D jerry_test -e "select number,data from esmm_data2ffm_infer_native" > ${DATA_PATH}/native.csv
-mysql -u root -p3SYz54LS9#^9sBvC -h 10.66.157.22 -P 4000 -D jerry_test -e "select number,data from esmm_data2ffm_infer_nearby" > ${DATA_PATH}/nearby.csv
+echo "data2ffm"
+${PYTHON_PATH} ${MODEL_PATH}/Feature_pipline/data2ffm.py > ${DATA_PATH}/infer.log
 
 echo "split data"
 split -l $((`wc -l < ${DATA_PATH}/tr.csv`/15)) ${DATA_PATH}/tr.csv -d -a 4 ${DATA_PATH}/tr/tr_ --additional-suffix=.csv
@@ -50,7 +47,7 @@ currentTimeStamp=$((timeStamp*1000+`date "+%N"`/1000000))
 echo $current
 
 echo "train..."
-${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=11 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir="${DATA_PATH}" --task_type="train"
+${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=23 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir=${DATA_PATH} --task_type=train
 
 echo "train time"
 current=$(date "+%Y-%m-%d %H:%M:%S")
@@ -59,11 +56,11 @@ currentTimeStamp=$((timeStamp*1000+`date "+%N"`/1000000))
 echo $current
 
 echo "infer native..."
-${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=11 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir="${DATA_PATH}/native" --task_type="infer" > ${DATA_PATH}/infer.log
+${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=11 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir=${DATA_PATH}/native --task_type=infer > ${DATA_PATH}/infer.log
 
 
 echo "infer nearby..."
-${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=11 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir="${DATA_PATH}/nearby" --task_type="infer" > ${DATA_PATH}/infer.log
+${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/DeepCvrMTL.py --ctr_task_wgt=0.3 --learning_rate=0.0001 --deep_layers=256,128 --dropout=0.8,0.5 --optimizer=Adam --num_epochs=1 --embedding_size=16 --batch_size=1024 --field_size=11 --feature_size=354332 --l2_reg=0.005 --log_steps=100 --num_threads=36 --model_dir=${DATA_PATH}/model_ckpt/DeepCvrMTL/ --data_dir=${DATA_PATH}/nearby --task_type=infer > ${DATA_PATH}/infer.log
 
 echo "sort and 2sql"
 ${PYTHON_PATH} ${MODEL_PATH}/Model_pipline/sort_and_2sql.py

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

@@ -69,13 +69,24 @@ object EsmmData {
       if (max_stat_date_str != param.date){
         val stat_date = param.date
         println(stat_date)
+//        val imp_data = sc.sql(
+//          s"""
+//             |select distinct stat_date,device_id,city_id as ucity_id,
+//             |  cid_id,diary_service_id
+//             |from data_feed_exposure
+//             |where cid_type = 'diary'
+//             |and stat_date ='${stat_date}'
+//         """.stripMargin
+//        )
+
         val imp_data = sc.sql(
           s"""
-             |select distinct stat_date,device_id,city_id as ucity_id,
-             |  cid_id,diary_service_id
+             |select * from
+             |(select stat_date,device_id,city_id as ucity_id,cid_id,diary_service_id
              |from data_feed_exposure
              |where cid_type = 'diary'
              |and stat_date ='${stat_date}'
+             |group by stat_date,device_id,city_id,cid_id,diary_service_id having count(*) > 1) a
          """.stripMargin
         )
         //      imp_data.show()
@@ -200,7 +211,7 @@ object EsmmData {
         )
         //      union_data_scity_id.createOrReplaceTempView("union_data_scity_id")
         union_data_scity_id.show()
-        GmeiConfig.writeToJDBCTable("jdbc:mysql://10.66.157.22:4000/jerry_test?user=root&password=3SYz54LS9#^9sBvC&rewriteBatchedStatements=true",union_data_scity_id, table="esmm_train_data",SaveMode.Append)
+        GmeiConfig.writeToJDBCTable("jdbc:mysql://10.66.157.22:4000/jerry_test?user=root&password=3SYz54LS9#^9sBvC&rewriteBatchedStatements=true",union_data_scity_id, table="esmm_train_test",SaveMode.Append)
 
       } else {
         println("esmm_train_data already have param.date data")

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

@@ -76,23 +76,23 @@ object temp_analysis {
       agency_id.createOrReplaceTempView("agency_id")
 
 
-      //每日新用户
-      val device_id_newUser = sc.sql(
-        s"""
-           |select distinct(device_id) as device_id
-           |from online.ml_device_day_active_status
-           |where active_type != '4'
-           |and first_channel_source_type not in ('yqxiu1','yqxiu2','yqxiu3','yqxiu4','yqxiu5','mxyc1','mxyc2','mxyc3'
-           |        ,'wanpu','jinshan','jx','maimai','zhuoyi','huatian','suopingjingling','mocha','mizhe','meika','lamabang'
-           |        ,'js-az1','js-az2','js-az3','js-az4','js-az5','jfq-az1','jfq-az2','jfq-az3','jfq-az4','jfq-az5','toufang1'
-           |        ,'toufang2','toufang3','toufang4','toufang5','toufang6','TF-toufang1','TF-toufang2','TF-toufang3','TF-toufang4'
-           |        ,'TF-toufang5','tf-toufang1','tf-toufang2','tf-toufang3','tf-toufang4','tf-toufang5','benzhan','promotion_aso100'
-           |        ,'promotion_qianka','promotion_xiaoyu','promotion_dianru','promotion_malioaso','promotion_malioaso-shequ'
-           |        ,'promotion_shike','promotion_julang_jl03')
-           |and partition_date ='${partition_date}'
-         """.stripMargin
-      )
-      device_id_newUser.createOrReplaceTempView("device_id_new")
+//      //每日新用户
+//      val device_id_newUser = sc.sql(
+//        s"""
+//           |select distinct(device_id) as device_id
+//           |from online.ml_device_day_active_status
+//           |where active_type != '4'
+//           |and first_channel_source_type not in ('yqxiu1','yqxiu2','yqxiu3','yqxiu4','yqxiu5','mxyc1','mxyc2','mxyc3'
+//           |        ,'wanpu','jinshan','jx','maimai','zhuoyi','huatian','suopingjingling','mocha','mizhe','meika','lamabang'
+//           |        ,'js-az1','js-az2','js-az3','js-az4','js-az5','jfq-az1','jfq-az2','jfq-az3','jfq-az4','jfq-az5','toufang1'
+//           |        ,'toufang2','toufang3','toufang4','toufang5','toufang6','TF-toufang1','TF-toufang2','TF-toufang3','TF-toufang4'
+//           |        ,'TF-toufang5','tf-toufang1','tf-toufang2','tf-toufang3','tf-toufang4','tf-toufang5','benzhan','promotion_aso100'
+//           |        ,'promotion_qianka','promotion_xiaoyu','promotion_dianru','promotion_malioaso','promotion_malioaso-shequ'
+//           |        ,'promotion_shike','promotion_julang_jl03')
+//           |and partition_date ='${partition_date}'
+//         """.stripMargin
+//      )
+//      device_id_newUser.createOrReplaceTempView("device_id_new")
 
       val blacklist_id = sc.sql(
         s"""
@@ -108,16 +108,34 @@ object temp_analysis {
            |from agency_id
            |UNION ALL
            |select device_id
-           |from device_id_new
-           |UNION ALL
-           |select device_id
            |from blacklist_id
          """.stripMargin
       )
       final_id.createOrReplaceTempView("final_id")
 
+
+      val diary_clk_all = sc.sql(
+        s"""
+           |select ov.partition_date,count(ov.cl_id) as clk_num,count(distinct(ov.cl_id)),count(ov.cl_id)/count(distinct(ov.cl_id))
+           |from online.tl_hdfs_maidian_view ov left join final_id
+           |on ov.cl_id = final_id.device_id
+           |where ov.action = "page_view"
+           |and params['page_name']="diary_detail"
+           |and ov.cl_id != "NULL"
+           |and ov.partition_date >='20181201'
+           |and final_id.device_id is  null
+           |group by ov.partition_date
+           |order by ov.partition_date
+       """.stripMargin
+      )
+      diary_clk_all.show(80)
+
+
 //日记本点击
-      val referrer=List("all_case_service_comment","all_cases","diary_detail","diary_list","diary_listof_related_service",
+      val referrer=List("about_me_message_list","all_case_service_comment","all_cases","diary_detail","diary_list"
+        ,"diary_listof_related_service","answer_detail","community_home","conversation_detail","create_diary_title","diary_listof_related_service",
+        "doctor_all_cases","hospital_all_cases","my_favor","my_order","order_detail","personal_store_diary_list","received_votes",
+        "topic_detail","welfare_detail","welfare_list","welfare_special","wiki_detail","zone_detail",
       "expert_detail","free_activity_detail","home","message_home","my_diary","organization_detail","other_homepage","question_detail",
       "search_result_diary","search_result_more","welfare_detail","zone_v3")
       for( a <- referrer ){
@@ -130,7 +148,7 @@ object temp_analysis {
              |and params['page_name']="diary_detail"
              |and params['referrer']='${a}'
              |and ov.cl_id != "NULL"
-             |and ov.partition_date >='20181101'
+             |and ov.partition_date >='20181201'
              |and final_id.device_id is  null
              |group by ov.partition_date
              |order by ov.partition_date
@@ -141,6 +159,8 @@ object temp_analysis {
 
       }
 
+
+
       //5.登录人数
       val log_device_temp = sc.sql(
         s"""

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

@@ -399,3 +399,4 @@ object testt {
 
 
 }
+