Skip to content

M
meta_base_code
Commit c1642720 authored by litaolemo's avatar litaolemo
Browse files
Options

update

parent 55ea7c2a
Show whitespace changes
Inline Side-by-side
Showing with 305 additions and 0 deletions
new_user_analysis/new_user_behavior_analysis.py 0 → 100644
View file @ c1642720
# -*- coding:UTF-8 -*-
# @Time : 2020/11/27 10:53
# @File : new_user_behavior_analysis.py
# @email : litao@igengmei.com
# @author : litao
import hashlib
import json
from meta_base_code.utils.func_get_uesr_event import get_user_event_from_mysql
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
from elasticsearch import Elasticsearch
from elasticsearch.helpers import scan
import sys
import time
from pyspark import SparkConf
from pyspark.sql import SparkSession, DataFrame
from meta_base_code.utils.func_from_redis_get_portrait import *
import pandas as pd
# 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')
db = pymysql.connect(host='172.16.30.136', port=3306, user='doris', passwd='o5gbA27hXHHm',
db='doris_prod')
cursor = db.cursor()
cursor.execute(sql)
result = cursor.fetchall()
db.close()
return result
exists_es_dic = {}
es = Elasticsearch([
{
'host': '172.16.31.17',
'port': 9200,
}, {
'host': '172.16.31.11',
'port': 9200,
}])
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("new_user_project_protratit")
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 = []
tractate_list = []
task_days = 3
for t in range(2, 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")
tomorrow_str = (datetime.datetime.now() + datetime.timedelta(days=day_num + 1)).strftime("%Y%m%d")
today_timestamp = int(now.timestamp())
today_str = now.strftime("%Y%m%d")
today_str_format = now.strftime("%Y-%m-%d")
yesterday_str = (now + datetime.timedelta(days=-1)).strftime("%Y%m%d")
yesterday_str_format = (now + datetime.timedelta(days=-1)).strftime("%Y-%m-%d")
one_week_age_str = (now + datetime.timedelta(days=-7)).strftime("%Y%m%d")
new_urser_device_id_sql = r"""
select t2.device_id as device_id from
(select device_id from online.ml_device_day_active_status where partition_date = '{today_str}' and active_type in (1,2)
) t2
LEFT JOIN
(
select distinct device_id
from ML.ML_D_CT_DV_DEVICECLEAN_DIMEN_D
where PARTITION_DAY = '{today_str}'
AND is_abnormal_device = 'true'
)dev
on t2.device_id=dev.device_id
WHERE dev.device_id is null and t2.device_id is not null
""".format(today_str=today_str, yesterday_str_format=yesterday_str_format, today_str_format=today_str_format,
tomorrow_str=tomorrow_str)
new_urser_device_id_df = spark.sql(new_urser_device_id_sql)
new_urser_device_id_df.createOrReplaceTempView("device_id_view")
exposure_sql_lt_than_8 = """
SELECT
cl_id,
count(distinct card_id) as session_pv0
FROM
(select device_id from device_id_view) dev left join
(SELECT partition_date,
cl_id,
case when card_content_type in ('qa','answer') then 'qa'
when card_content_type in ('special_pool') then 'special' else card_content_type end as card_content_type,
CASE when transaction_type in ('fmctr','samecity_fmctr') then array('fmctr','合计')
when transaction_type in ('high_quality_fmctr') then array('high_quality_fmctr','合计')
WHEN (transaction_type like '%ctr' and transaction_type not in ('high_quality_ctr','high_quality_fmctr','fmctr','samecity_fmctr') ) THEN array('ctr预估','合计')
when transaction_type in ('high_quality_ctr') then array('high_quality_ctr','合计')
WHEN transaction_type like '%cvr' THEN array('cvr预估','合计')
WHEN transaction_type in ('-1','smr') THEN array('smr','合计')
when transaction_type in ('pgc','hotspot') then array('热点卡片')
when transaction_type in ('newdata') then array('保量卡片')
when transaction_type in ('hotspot_feed') then array('hotspot_feed','合计')
when transaction_type in ('aistragegy') then array('新用户AI帖优先','合计')
when transaction_type in ('excestragegy') then array('新用户精华帖优先','合计')
when transaction_type in ('FIXEDSTRATEGY') then array('新氧新用户策略一','合计')
when transaction_type in ('FIXEDSTRATEGY_VIDEO') then array('新氧新用户策略二','合计')
when transaction_type like 'deeplink%' then array('deeplink策略','合计')
end AS recommend_type,
card_id,
app_session_id
from online.ml_community_precise_exposure_detail
WHERE partition_date={partition_day}
AND action in ('page_precise_exposure','home_choiceness_card_exposure') --7745版本action改为page_precise_exposure
AND is_exposure = '1' ----精准曝光
AND page_name ='home'
AND tab_name = '精选'
AND (transaction_type in ('-1','smr','hotspot','pgc','newdata','hotspot_feed','aistragegy','excestragegy','FIXEDSTRATEGY','FIXEDSTRATEGY_VIDEO')
or transaction_type like '%ctr' or transaction_type like '%cvr' or transaction_type like 'deeplink%')
AND card_content_type in ('qa','diary','user_post','answer','special_pool')
group by partition_date,
case when card_content_type in ('qa','answer') then 'qa'
when card_content_type in ('special_pool') then 'special' else card_content_type end,
cl_id,
CASE when transaction_type in ('fmctr','samecity_fmctr') then array('fmctr','合计')
when transaction_type in ('high_quality_fmctr') then array('high_quality_fmctr','合计')
WHEN (transaction_type like '%ctr' and transaction_type not in ('high_quality_ctr','high_quality_fmctr','fmctr','samecity_fmctr')) THEN array('ctr预估','合计')
when transaction_type in ('high_quality_ctr') then array('high_quality_ctr','合计')
WHEN transaction_type like '%cvr' THEN array('cvr预估','合计')
WHEN transaction_type in ('-1','smr') THEN array('smr','合计')
when transaction_type in ('pgc','hotspot') then array('热点卡片')
when transaction_type in ('newdata') then array('保量卡片')
when transaction_type in ('hotspot_feed') then array('hotspot_feed','合计')
when transaction_type in ('aistragegy') then array('新用户AI帖优先','合计')
when transaction_type in ('excestragegy') then array('新用户精华帖优先','合计')
when transaction_type in ('FIXEDSTRATEGY') then array('新氧新用户策略一','合计')
when transaction_type in ('FIXEDSTRATEGY_VIDEO') then array('新氧新用户策略二','合计')
when transaction_type like 'deeplink%' then array('deeplink策略','合计') end,
card_id,
app_session_id
)a on a.card_id = dev.device_id
LATERAL VIEW explode (a.recommend_type) v as recommend_type
group by cl_id having session_pv0 <= 8
""".format(partition_day=today_str)
exposure_sql_gte_than_16 = """
SELECT
cl_id,
count(distinct card_id) as session_pv0
FROM
(select device_id from device_id_view) dev left join
(SELECT partition_date,
cl_id,
case when card_content_type in ('qa','answer') then 'qa'
when card_content_type in ('special_pool') then 'special' else card_content_type end as card_content_type,
CASE when transaction_type in ('fmctr','samecity_fmctr') then array('fmctr','合计')
when transaction_type in ('high_quality_fmctr') then array('high_quality_fmctr','合计')
WHEN (transaction_type like '%ctr' and transaction_type not in ('high_quality_ctr','high_quality_fmctr','fmctr','samecity_fmctr') ) THEN array('ctr预估','合计')
when transaction_type in ('high_quality_ctr') then array('high_quality_ctr','合计')
WHEN transaction_type like '%cvr' THEN array('cvr预估','合计')
WHEN transaction_type in ('-1','smr') THEN array('smr','合计')
when transaction_type in ('pgc','hotspot') then array('热点卡片')
when transaction_type in ('newdata') then array('保量卡片')
when transaction_type in ('hotspot_feed') then array('hotspot_feed','合计')
when transaction_type in ('aistragegy') then array('新用户AI帖优先','合计')
when transaction_type in ('excestragegy') then array('新用户精华帖优先','合计')
when transaction_type in ('FIXEDSTRATEGY') then array('新氧新用户策略一','合计')
when transaction_type in ('FIXEDSTRATEGY_VIDEO') then array('新氧新用户策略二','合计')
when transaction_type like 'deeplink%' then array('deeplink策略','合计')
end AS recommend_type,
card_id,
app_session_id
from online.ml_community_precise_exposure_detail
WHERE partition_date= '{partition_day}'
AND action in ('page_precise_exposure','home_choiceness_card_exposure') --7745版本action改为page_precise_exposure
AND is_exposure = '1' ----精准曝光
AND page_name ='home'
AND tab_name = '精选'
AND (transaction_type in ('-1','smr','hotspot','pgc','newdata','hotspot_feed','aistragegy','excestragegy','FIXEDSTRATEGY','FIXEDSTRATEGY_VIDEO')
or transaction_type like '%ctr' or transaction_type like '%cvr' or transaction_type like 'deeplink%')
AND card_content_type in ('qa','diary','user_post','answer','special_pool')
group by partition_date,
case when card_content_type in ('qa','answer') then 'qa'
when card_content_type in ('special_pool') then 'special' else card_content_type end,
cl_id,
CASE when transaction_type in ('fmctr','samecity_fmctr') then array('fmctr','合计')
when transaction_type in ('high_quality_fmctr') then array('high_quality_fmctr','合计')
WHEN (transaction_type like '%ctr' and transaction_type not in ('high_quality_ctr','high_quality_fmctr','fmctr','samecity_fmctr')) THEN array('ctr预估','合计')
when transaction_type in ('high_quality_ctr') then array('high_quality_ctr','合计')
WHEN transaction_type like '%cvr' THEN array('cvr预估','合计')
WHEN transaction_type in ('-1','smr') THEN array('smr','合计')
when transaction_type in ('pgc','hotspot') then array('热点卡片')
when transaction_type in ('newdata') then array('保量卡片')
when transaction_type in ('hotspot_feed') then array('hotspot_feed','合计')
when transaction_type in ('aistragegy') then array('新用户AI帖优先','合计')
when transaction_type in ('excestragegy') then array('新用户精华帖优先','合计')
when transaction_type in ('FIXEDSTRATEGY') then array('新氧新用户策略一','合计')
when transaction_type in ('FIXEDSTRATEGY_VIDEO') then array('新氧新用户策略二','合计')
when transaction_type like 'deeplink%' then array('deeplink策略','合计') end,
card_id,
app_session_id
)a on a.card_id = dev.device_id
LATERAL VIEW explode (a.recommend_type) v as recommend_type
group by cl_id having session_pv0 >= 16
""".format(partition_day=today_str)
print(new_urser_device_id_sql)
exposure_sql_lt_than_8_df = spark.sql(exposure_sql_lt_than_8)
exposure_sql_gte_than_16_df = spark.sql(exposure_sql_gte_than_16)
sql_res = exposure_sql_lt_than_8_df.collect()
res_dict = {}
portrait_dict = {
"first_demands": {},
"second_demands": {},
"first_solutions": {},
"second_solutions": {},
"first_positions": {},
"second_positions": {},
"projects": {},
'anecdote_tags': {}
}
no_portrait_device_id_list = []
print("-------------------------------")
count_not_has_portratit = 0
event_dict = {}
event_dict_reverse = {}
for count_user_count, res in enumerate(sql_res):
# print(count, res)
temp_count = 0
try:
for event_cn,projects in get_user_event_from_mysql(res.device_id,today_timestamp):
project_list = projects.split(",")
for project in project_list:
if project not in event_dict:
event_dict[project] = {}
if event_dict[project].get(event_cn):
event_dict[project][event_cn] += 1
else:
event_dict[project][event_cn] = 1
if event_cn not in event_dict_reverse:
event_dict_reverse[event_cn] = {}
if event_dict_reverse[event_cn].get(project):
event_dict_reverse[event_cn][project] += 1
else:
event_dict_reverse[event_cn][project] = 1
except Exception as e:
print("error ", e)
temp_count += 1
if not temp_count:
count_not_has_portratit += 1
no_portrait_device_id_list.append(res.device_id)
# print(portrait_dict)
# print(count_user_count + 1, count_not_has_portratit)
# print("-------------------------------")
print("event_dict",today_str,event_dict_reverse)
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