Commit 0ad276bb authored by litaolemo's avatar litaolemo

update

parent b7184195
# -*- coding:UTF-8 -*-
# @Time : 2020/9/4 17:07
# @File : search_meigou_ctr.py
# @email : litao@igengmei.com
# @author : litao
import hashlib
import json
import pymysql
import xlwt, datetime
import redis
# from pyhive import hive
from maintenance.func_send_email_with_file import send_file_email
from typing import Dict, List
from elasticsearch_7 import Elasticsearch
from elasticsearch_7.helpers import scan
import sys
import time
from pyspark import SparkConf
from pyspark.sql import SparkSession, DataFrame
# from pyspark.sql.functions import lit
# import pytispark.pytispark as pti
def con_sql(sql):
# 从数据库的表里获取数据
db = pymysql.connect(host='172.16.40.158', port=4000, user='st_user', passwd='aqpuBLYzEV7tML5RPsN1pntUzFy',
db='jerry_prod')
cursor = db.cursor()
cursor.execute(sql)
result = cursor.fetchall()
db.close()
return result
startTime = time.time()
sparkConf = SparkConf()
sparkConf.set("spark.sql.crossJoin.enabled", True)
sparkConf.set("spark.debug.maxToStringFields", "100")
sparkConf.set("spark.tispark.plan.allow_index_double_read", False)
sparkConf.set("spark.tispark.plan.allow_index_read", True)
sparkConf.set("spark.hive.mapred.supports.subdirectories", True)
sparkConf.set("spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive", True)
sparkConf.set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
sparkConf.set("mapreduce.output.fileoutputformat.compress", False)
sparkConf.set("mapreduce.map.output.compress", False)
sparkConf.set("prod.gold.jdbcuri",
"jdbc:mysql://172.16.30.136/doris_prod?user=doris&password=o5gbA27hXHHm&rewriteBatchedStatements=true")
sparkConf.set("prod.mimas.jdbcuri",
"jdbc:mysql://172.16.30.138/mimas_prod?user=mimas&password=GJL3UJe1Ck9ggL6aKnZCq4cRvM&rewriteBatchedStatements=true")
sparkConf.set("prod.gaia.jdbcuri",
"jdbc:mysql://172.16.30.143/zhengxing?user=work&password=BJQaT9VzDcuPBqkd&rewriteBatchedStatements=true")
sparkConf.set("prod.tidb.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/eagle?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.jerry.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/jerry_prod?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.158:2379")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.170:4000")
sparkConf.set("prod.tidb.database", "jerry_prod")
sparkConf.setAppName("search_answer_ctr")
spark = (SparkSession.builder.config(conf=sparkConf).config("spark.sql.extensions", "org.apache.spark.sql.TiExtensions")
.config("spark.tispark.pd.addresses", "172.16.40.170:2379").enableHiveSupport().getOrCreate())
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/brickhouse-0.7.1-SNAPSHOT.jar")
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/hive-udf-1.0-SNAPSHOT.jar")
spark.sql("CREATE TEMPORARY FUNCTION json_map AS 'brickhouse.udf.json.JsonMapUDF'")
spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJsonFormatCheck'")
spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'")
task_list = []
task_days = 60
for t in range(1, task_days):
day_num = 0 - t
now = (datetime.datetime.now() + datetime.timedelta(days=day_num))
last_30_day_str = (now + datetime.timedelta(days=-30)).strftime("%Y%m%d")
today_str = now.strftime("%Y%m%d")
yesterday_str = (now + datetime.timedelta(days=-1)).strftime("%Y%m%d")
one_week_age_str = (now + datetime.timedelta(days=-7)).strftime("%Y%m%d")
sql_search_ctr = r"""
select D.ACTIVE_TYPE,D.DEVICE_OS_TYPE,sum(T.CLICK_NUM) as CLICK_NUM,sum(C.EXPOSURE) as EXPOSURE from
(SELECT T.DEVICE_ID, --设备ID
T.CARD_ID, --卡片ID
SUM(T.CLICK_NUM) AS CLICK_NUM --点击次数
FROM ML.ML_C_ET_CK_CLICK_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.PAGE_CODE = 'search_result_question_answer'
AND T.ACTION IN ('on_click_card')
GROUP BY T.DEVICE_ID,
T.CARD_ID) T
left join
(SELECT T.DEVICE_ID as DEVICE_ID, --设备ID
T.CARD_ID as CARD_ID, --卡片ID
COUNT(T.CARD_ID) AS EXPOSURE --点击次数
FROM ML.MID_ML_C_ET_PE_PRECISEEXPOSURE_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.PAGE_CODE = 'search_result_question_answer'
GROUP BY T.DEVICE_ID,
T.CARD_ID) C on T.DEVICE_ID=C.DEVICE_ID and T.CARD_ID = C.CARD_ID
LEFT JOIN
(
SELECT T.DEVICE_ID,
T.DEVICE_OS_TYPE,
T.ACTIVE_TYPE
FROM ML.ML_C_CT_DV_DEVICE_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.ACTIVE_TYPE IN ('1', '2', '4'))
D on T.DEVICE_ID = D.DEVICE_ID
LEFT JOIN
(
SELECT DISTINCT device_id
FROM ml.ml_d_ct_dv_devicespam_d --去除机构刷单设备,即作弊设备(浏览和曝光事件去除)
WHERE partition_day='{partition_day}'
UNION ALL
SELECT DISTINCT device_id
FROM dim.dim_device_user_staff --去除内网用户
)spam_pv
on spam_pv.device_id=T.DEVICE_ID
LEFT JOIN
(
SELECT partition_date,device_id
FROM
(--找出user_id当天活跃的第一个设备id
SELECT user_id,partition_date,
if(size(device_list) > 0, device_list [ 0 ], '') AS device_id
FROM online.ml_user_updates
WHERE partition_date>='{partition_day}' AND partition_date<'{end_date}'
)t1
JOIN
( --医生账号
SELECT distinct user_id
FROM online.tl_hdfs_doctor_view
WHERE partition_date = '{partition_day}'
--马甲账号/模特用户
UNION ALL
SELECT user_id
FROM ml.ml_c_ct_ui_user_dimen_d
WHERE partition_day = '{partition_day}'
AND (is_puppet = 'true' or is_classifyuser = 'true')
UNION ALL
--公司内网覆盖用户
select distinct user_id
from dim.dim_device_user_staff
UNION ALL
--登陆过医生设备
SELECT distinct t1.user_id
FROM
(
SELECT user_id, v.device_id as device_id
FROM online.ml_user_history_detail
LATERAL VIEW EXPLODE(device_history_list) v AS device_id
WHERE partition_date = '{partition_day}'
)t1
JOIN
(
SELECT device_id
FROM online.ml_device_history_detail
WHERE partition_date = '{partition_day}'
AND is_login_doctor = '1'
)t2
ON t1.device_id = t2.device_id
)t2
on t1.user_id=t2.user_id
group by partition_date,device_id
)dev
on T.DEVICE_ID=dev.device_id
WHERE (spam_pv.device_id IS NULL or spam_pv.device_id = '')
and (dev.device_id is null or dev.device_id='')
GROUP by D.DEVICE_OS_TYPE,
D.ACTIVE_TYPE
""".format(partition_day=yesterday_str, end_date=today_str)
print(sql_search_ctr)
search_ctr_df = spark.sql(sql_search_ctr)
# spam_pv_df.createOrReplaceTempView("dev_view")
search_ctr_df.show(1)
sql_res = search_ctr_df.collect()
res_dict = {
"新增": {
"ios": {"click_num": 0, "exposure": 0},
"android": {"click_num": 0, "exposure": 0}
},
"老活": {
"ios": {"click_num": 0, "exposure": 0},
"android": {"click_num": 0, "exposure": 0}
}
}
print("-------------------------------")
db = pymysql.connect(host='172.16.40.158', port=4000, user='st_user', passwd='aqpuBLYzEV7tML5RPsN1pntUzFy',
db='jerry_prod')
cursor = db.cursor()
for res in sql_res:
print(res)
if res.ACTIVE_TYPE:
if res.ACTIVE_TYPE in ('1', '2'):
res_dict["新增"][res.DEVICE_OS_TYPE]["click_num"] += res.CLICK_NUM
res_dict["新增"][res.DEVICE_OS_TYPE]["exposure"] += res.EXPOSURE
else:
res_dict["老活"][res.DEVICE_OS_TYPE]["click_num"] += res.CLICK_NUM
res_dict["老活"][res.DEVICE_OS_TYPE]["exposure"] += res.EXPOSURE
for active_type in res_dict:
for device_os_type in res_dict[active_type]:
partition_date = yesterday_str
pid = hashlib.md5((partition_date + device_os_type + active_type).encode("utf8")).hexdigest()
click_num = res_dict[active_type][device_os_type]["click_num"]
exposure = res_dict[active_type][device_os_type]["exposure"]
try:
search_ctr = round(click_num / exposure, 5)
except:
search_ctr = 0
instert_sql = """replace into search_answer_ctr(
partition_date,device_os_type,active_type,pid,click_num,exposure,search_ctr) VALUES('{partition_date}','{device_os_type}','{active_type}','{pid}',{click_num},{exposure},{search_ctr});""".format(
partition_date=partition_date, device_os_type=device_os_type, active_type=active_type, pid=pid, click_num=click_num,
exposure=exposure, search_ctr=search_ctr
)
print(instert_sql)
# cursor.execute("set names 'UTF8'")
res = cursor.execute(instert_sql)
db.commit()
print(res)
# cursor.executemany()
db.close()
...@@ -61,7 +61,7 @@ sparkConf.set("prod.jerry.jdbcuri", ...@@ -61,7 +61,7 @@ sparkConf.set("prod.jerry.jdbcuri",
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.158:2379") sparkConf.set("prod.tispark.pd.addresses", "172.16.40.158:2379")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.170:4000") sparkConf.set("prod.tispark.pd.addresses", "172.16.40.170:4000")
sparkConf.set("prod.tidb.database", "jerry_prod") sparkConf.set("prod.tidb.database", "jerry_prod")
sparkConf.setAppName("search_diary_ctr")
spark = (SparkSession.builder.config(conf=sparkConf).config("spark.sql.extensions", "org.apache.spark.sql.TiExtensions") spark = (SparkSession.builder.config(conf=sparkConf).config("spark.sql.extensions", "org.apache.spark.sql.TiExtensions")
.config("spark.tispark.pd.addresses", "172.16.40.170:2379").appName( .config("spark.tispark.pd.addresses", "172.16.40.170:2379").appName(
"search_diary_ctr").enableHiveSupport().getOrCreate()) "search_diary_ctr").enableHiveSupport().getOrCreate())
...@@ -73,7 +73,7 @@ spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJso ...@@ -73,7 +73,7 @@ spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJso
spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'") spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'")
task_list = [] task_list = []
task_days = 90 task_days = 60
for t in range(1, task_days): for t in range(1, task_days):
day_num = 0 - t day_num = 0 - t
now = (datetime.datetime.now() + datetime.timedelta(days=day_num)) now = (datetime.datetime.now() + datetime.timedelta(days=day_num))
......
...@@ -73,7 +73,7 @@ spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJso ...@@ -73,7 +73,7 @@ spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJso
spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'") spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'")
task_list = [] task_list = []
task_days = 90 task_days = 1
for t in range(1, task_days): for t in range(1, task_days):
day_num = 0 - t day_num = 0 - t
now = (datetime.datetime.now() + datetime.timedelta(days=day_num)) now = (datetime.datetime.now() + datetime.timedelta(days=day_num))
......
# -*- coding:UTF-8 -*-
# @Time : 2020/9/4 17:07
# @File : search_meigou_ctr.py
# @email : litao@igengmei.com
# @author : litao
import hashlib
import json
import pymysql
import xlwt, datetime
import redis
# from pyhive import hive
from maintenance.func_send_email_with_file import send_file_email
from typing import Dict, List
from elasticsearch_7 import Elasticsearch
from elasticsearch_7.helpers import scan
import sys
import time
from pyspark import SparkConf
from pyspark.sql import SparkSession, DataFrame
# from pyspark.sql.functions import lit
# import pytispark.pytispark as pti
def con_sql(sql):
# 从数据库的表里获取数据
db = pymysql.connect(host='172.16.40.158', port=4000, user='st_user', passwd='aqpuBLYzEV7tML5RPsN1pntUzFy',
db='jerry_prod')
cursor = db.cursor()
cursor.execute(sql)
result = cursor.fetchall()
db.close()
return result
startTime = time.time()
sparkConf = SparkConf()
sparkConf.set("spark.sql.crossJoin.enabled", True)
sparkConf.set("spark.debug.maxToStringFields", "100")
sparkConf.set("spark.tispark.plan.allow_index_double_read", False)
sparkConf.set("spark.tispark.plan.allow_index_read", True)
sparkConf.set("spark.hive.mapred.supports.subdirectories", True)
sparkConf.set("spark.hadoop.mapreduce.input.fileinputformat.input.dir.recursive", True)
sparkConf.set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
sparkConf.set("mapreduce.output.fileoutputformat.compress", False)
sparkConf.set("mapreduce.map.output.compress", False)
sparkConf.set("prod.gold.jdbcuri",
"jdbc:mysql://172.16.30.136/doris_prod?user=doris&password=o5gbA27hXHHm&rewriteBatchedStatements=true")
sparkConf.set("prod.mimas.jdbcuri",
"jdbc:mysql://172.16.30.138/mimas_prod?user=mimas&password=GJL3UJe1Ck9ggL6aKnZCq4cRvM&rewriteBatchedStatements=true")
sparkConf.set("prod.gaia.jdbcuri",
"jdbc:mysql://172.16.30.143/zhengxing?user=work&password=BJQaT9VzDcuPBqkd&rewriteBatchedStatements=true")
sparkConf.set("prod.tidb.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/eagle?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.jerry.jdbcuri",
"jdbc:mysql://172.16.40.158:4000/jerry_prod?user=st_user&password=aqpuBLYzEV7tML5RPsN1pntUzFy&rewriteBatchedStatements=true")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.158:2379")
sparkConf.set("prod.tispark.pd.addresses", "172.16.40.170:4000")
sparkConf.set("prod.tidb.database", "jerry_prod")
sparkConf.setAppName("search_tractate_ctr")
spark = (SparkSession.builder.config(conf=sparkConf).config("spark.sql.extensions", "org.apache.spark.sql.TiExtensions")
.config("spark.tispark.pd.addresses", "172.16.40.170:2379").enableHiveSupport().getOrCreate())
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/brickhouse-0.7.1-SNAPSHOT.jar")
spark.sql("ADD JAR hdfs:///user/hive/share/lib/udf/hive-udf-1.0-SNAPSHOT.jar")
spark.sql("CREATE TEMPORARY FUNCTION json_map AS 'brickhouse.udf.json.JsonMapUDF'")
spark.sql("CREATE TEMPORARY FUNCTION is_json AS 'com.gmei.hive.common.udf.UDFJsonFormatCheck'")
spark.sql("CREATE TEMPORARY FUNCTION arrayMerge AS 'com.gmei.hive.common.udf.UDFArryMerge'")
task_list = []
task_days = 60
for t in range(1, task_days):
day_num = 0 - t
now = (datetime.datetime.now() + datetime.timedelta(days=day_num))
last_30_day_str = (now + datetime.timedelta(days=-30)).strftime("%Y%m%d")
today_str = now.strftime("%Y%m%d")
yesterday_str = (now + datetime.timedelta(days=-1)).strftime("%Y%m%d")
one_week_age_str = (now + datetime.timedelta(days=-7)).strftime("%Y%m%d")
sql_search_ctr = r"""
select D.ACTIVE_TYPE,D.DEVICE_OS_TYPE,sum(T.CLICK_NUM) as CLICK_NUM,sum(C.EXPOSURE) as EXPOSURE from
(SELECT T.DEVICE_ID, --设备ID
T.CARD_ID, --卡片ID
SUM(T.CLICK_NUM) AS CLICK_NUM --点击次数
FROM ML.ML_C_ET_CK_CLICK_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.PAGE_CODE = 'search_result_post'
AND T.ACTION IN ('search_result_click_infomation_item','on_click_topic_card')
GROUP BY T.DEVICE_ID,
T.CARD_ID) T
left join
(SELECT T.DEVICE_ID as DEVICE_ID, --设备ID
T.CARD_ID as CARD_ID, --卡片ID
COUNT(T.CARD_ID) AS EXPOSURE --点击次数
FROM ML.MID_ML_C_ET_PE_PRECISEEXPOSURE_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.PAGE_CODE = 'search_result_post'
GROUP BY T.DEVICE_ID,
T.CARD_ID) C on T.DEVICE_ID=C.DEVICE_ID and T.CARD_ID = C.CARD_ID
LEFT JOIN
(
SELECT T.DEVICE_ID,
T.DEVICE_OS_TYPE,
T.ACTIVE_TYPE
FROM ML.ML_C_CT_DV_DEVICE_DIMEN_D T
WHERE T.PARTITION_DAY = '{partition_day}'
AND T.ACTIVE_TYPE IN ('1', '2', '4'))
D on T.DEVICE_ID = D.DEVICE_ID
LEFT JOIN
(
SELECT DISTINCT device_id
FROM ml.ml_d_ct_dv_devicespam_d --去除机构刷单设备,即作弊设备(浏览和曝光事件去除)
WHERE partition_day='{partition_day}'
UNION ALL
SELECT DISTINCT device_id
FROM dim.dim_device_user_staff --去除内网用户
)spam_pv
on spam_pv.device_id=T.DEVICE_ID
LEFT JOIN
(
SELECT partition_date,device_id
FROM
(--找出user_id当天活跃的第一个设备id
SELECT user_id,partition_date,
if(size(device_list) > 0, device_list [ 0 ], '') AS device_id
FROM online.ml_user_updates
WHERE partition_date>='{partition_day}' AND partition_date<'{end_date}'
)t1
JOIN
( --医生账号
SELECT distinct user_id
FROM online.tl_hdfs_doctor_view
WHERE partition_date = '{partition_day}'
--马甲账号/模特用户
UNION ALL
SELECT user_id
FROM ml.ml_c_ct_ui_user_dimen_d
WHERE partition_day = '{partition_day}'
AND (is_puppet = 'true' or is_classifyuser = 'true')
UNION ALL
--公司内网覆盖用户
select distinct user_id
from dim.dim_device_user_staff
UNION ALL
--登陆过医生设备
SELECT distinct t1.user_id
FROM
(
SELECT user_id, v.device_id as device_id
FROM online.ml_user_history_detail
LATERAL VIEW EXPLODE(device_history_list) v AS device_id
WHERE partition_date = '{partition_day}'
)t1
JOIN
(
SELECT device_id
FROM online.ml_device_history_detail
WHERE partition_date = '{partition_day}'
AND is_login_doctor = '1'
)t2
ON t1.device_id = t2.device_id
)t2
on t1.user_id=t2.user_id
group by partition_date,device_id
)dev
on T.DEVICE_ID=dev.device_id
WHERE (spam_pv.device_id IS NULL or spam_pv.device_id = '')
and (dev.device_id is null or dev.device_id='')
GROUP by D.DEVICE_OS_TYPE,
D.ACTIVE_TYPE
""".format(partition_day=yesterday_str, end_date=today_str)
print(sql_search_ctr)
search_ctr_df = spark.sql(sql_search_ctr)
# spam_pv_df.createOrReplaceTempView("dev_view")
search_ctr_df.show(1)
sql_res = search_ctr_df.collect()
res_dict = {
"新增": {
"ios": {"click_num": 0, "exposure": 0},
"android": {"click_num": 0, "exposure": 0}
},
"老活": {
"ios": {"click_num": 0, "exposure": 0},
"android": {"click_num": 0, "exposure": 0}
}
}
print("-------------------------------")
db = pymysql.connect(host='172.16.40.158', port=4000, user='st_user', passwd='aqpuBLYzEV7tML5RPsN1pntUzFy',
db='jerry_prod')
cursor = db.cursor()
for res in sql_res:
print(res)
if res.ACTIVE_TYPE:
if res.ACTIVE_TYPE in ('1', '2'):
res_dict["新增"][res.DEVICE_OS_TYPE]["click_num"] += res.CLICK_NUM
res_dict["新增"][res.DEVICE_OS_TYPE]["exposure"] += res.EXPOSURE
else:
res_dict["老活"][res.DEVICE_OS_TYPE]["click_num"] += res.CLICK_NUM
res_dict["老活"][res.DEVICE_OS_TYPE]["exposure"] += res.EXPOSURE
for active_type in res_dict:
for device_os_type in res_dict[active_type]:
partition_date = yesterday_str
pid = hashlib.md5((partition_date + device_os_type + active_type).encode("utf8")).hexdigest()
click_num = res_dict[active_type][device_os_type]["click_num"]
exposure = res_dict[active_type][device_os_type]["exposure"]
try:
search_ctr = round(click_num / exposure, 5)
except:
search_ctr = 0
instert_sql = """replace into search_tractate_ctr(
partition_date,device_os_type,active_type,pid,click_num,exposure,search_ctr) VALUES('{partition_date}','{device_os_type}','{active_type}','{pid}',{click_num},{exposure},{search_ctr});""".format(
partition_date=partition_date, device_os_type=device_os_type, active_type=active_type, pid=pid, click_num=click_num,
exposure=exposure, search_ctr=search_ctr
)
print(instert_sql)
# cursor.execute("set names 'UTF8'")
res = cursor.execute(instert_sql)
db.commit()
print(res)
# cursor.executemany()
db.close()
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment