update

utils/__init__.py

+# -*- coding:UTF-8 -*-
+# @Time  : 2020/9/9 17:15
+# @File  : __init__.py.py
+# @email : litao@igengmei.com
+# @author : litao
\ No newline at end of file

utils/func_get_pv_card_id.py

+# -*- coding:UTF-8 -*-
+# @Time  : 2020/9/9 17:16
+# @File  : func_get_pv_card_id.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 get_card_id():
+    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")
+
+    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(
+            "LR PYSPARK TEST").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 = 2
+
+    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=-3)).strftime("%Y%m%d")
+        sql = r"""
+    SELECT * FROM
+        (--精准曝光，卡片id和session_id去重
+            SELECT  partition_date,
+                    card_content_type,
+                    cl_id,
+                    recommend_type,
+                    card_id
+            FROM
+            (
+               SELECT partition_date,
+                      cl_id,
+                      case when card_content_type in ('qa','answer') then 'qa' else card_content_type end as card_content_type,
+                      CASE when transaction_type in ('fmctr') then 'fmctr'
+                           WHEN transaction_type like '%ctr' THEN 'ctr预估'
+                           WHEN transaction_type like '%cvr' THEN 'cvr预估'
+                           WHEN transaction_type in ('-1','smr') THEN 'smr'
+                           when transaction_type in ('pgc','hotspot') then '热点卡片'
+                           when transaction_type in ('newdata') then '保量卡片'
+                           when transaction_type in ('hotspot_feed') then 'hotspot_feed'
+                      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')
+                    or transaction_type like '%ctr' or transaction_type like '%cvr')
+               AND card_content_type in ('qa','diary','user_post','answer')
+               group by partition_date,
+                      case when card_content_type in ('qa','answer') then 'qa' else card_content_type end,
+                      cl_id,
+                      CASE when transaction_type in ('fmctr') then 'fmctr'
+                           WHEN transaction_type like '%ctr' THEN 'ctr预估'
+                           WHEN transaction_type like '%cvr' THEN 'cvr预估'
+                           WHEN transaction_type in ('-1','smr') THEN 'smr'
+                           when transaction_type in ('pgc','hotspot') then '热点卡片'
+                           when transaction_type in ('newdata') then '保量卡片'
+                           when transaction_type in ('hotspot_feed') then 'hotspot_feed' END,
+                      card_id,
+                      app_session_id
+            )a
+            group by partition_date,card_content_type,cl_id,recommend_type,card_id
+          )t2
+       
+            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=t2.cl_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}'
+            )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 t2.partition_date=dev.partition_date and t2.cl_id=dev.device_id
+    WHERE spam_pv.device_id IS NULL
+      and dev.device_id is null
+    """.format(partition_day=yesterday_str)
+        device_df = spark.sql(sql)
+
+        device_df.show(1, False)
+        sql_res = device_df.collect()
+        res_dict = {
+            "diary": [],
+            "user_post": [],
+            "qa": []
+        }
+        for res in sql_res:
+            print(res)
+            card_content_type = res.card_content_type
+            card_id = res.card_id
+            if card_content_type in res_dict:
+                res_dict[card_content_type].append(card_id)
+        return res_dict