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

add video count

ctr-56.py

@@ -34,23 +34,13 @@ def ctr_all():
     result = cursor.fetchall()
     tail56 = pd.DataFrame(list(result))[0].values.tolist()
 
-    start = "2018-09-05"
-    sql_all = "select distinct device_id from data_feed_click " \
-              "where stat_date >= '{}' and stat_date <= '{}' and cid_type = 'diary'".format(start, date)
-    cursor.execute(sql_all)
-    result_all = cursor.fetchall()
-    df = pd.DataFrame(list(result_all))
-    all_user = df[0].values.tolist()
-    print("老用户总数：")
-    print(len(all_user))
-    day_all = set(all_user) & set(tail56)
-    print("当天老用户总数：")
-    print(len(day_all))
-    cover = len(day_all & set(device_id))
+    print("当天尾号5或6活跃用户总数：")
+    print(len(tail56))
 
+    cover = len(set(tail56)&set(device_id))
     print("当天尾号5或6活跃用户覆盖数：")
     print(cover)
-    cover_percent = format(cover / len(day_all), ".3f")
+    cover_percent = format(cover / len(tail56), ".6f")
     print("当天尾号5或6活跃用户覆盖率：")
     print(cover_percent)
     sql_click = "select count(cid) from data_feed_click " \
@@ -66,7 +56,7 @@ def ctr_all():
     exp = cursor.fetchone()[0]
     print("曝光数：" + str(exp))
     print("点击率：" + str(click / exp))
-    return len(day_all),cover,cover_percent,click,exp,format(click/exp,".6f")
+    return len(tail56),cover,cover_percent,click,exp,format(click/exp,".6f")
 
 
 def ctr():

local/test_supervisor.py

@@ -196,6 +196,7 @@ def router(device_id):
         predict(user_profile)
 
 if __name__ == "__main__":
+    sql = "delete from data_feed_click where stat_date = '2018-10-17'"
     while True:
         start = time.time()
         empty,device_id_list = get_active_users()

rank01_ctr.py

+# -*- coding: UTF-8 -*-
+import pymysql
+import datetime
+import pandas as pd
+
+
+def get_yesterday_date():
+    today = datetime.date.today()
+    yesterday = today - datetime.timedelta(days=1)
+    yesterday = yesterday.strftime("%Y-%m-%d")
+    print(yesterday)
+    return yesterday
+
+
+def get_data():
+    conn2db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='eagle')
+    cursor = conn2db.cursor()
+    sql = "select device_id from ffm_diary_queue_temp where device_id regexp '[5|6]$'"
+    cursor.execute(sql)
+    result = cursor.fetchall()
+    device = tuple(pd.DataFrame(list(result))[0].values.tolist())
+    cursor.close()
+    print(device[0:2])
+
+    return device
+
+
+def ctr_all():
+    db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_prod')
+    cursor = db.cursor()
+    sql_active = "select distinct device_id from data_feed_exposure " \
+                 "where cid_type = 'diary' and device_id regexp'[5|6]$' and stat_date = '{}';".format(date)
+    cursor.execute(sql_active)
+    result = cursor.fetchall()
+    tail56 = pd.DataFrame(list(result))[0].values.tolist()
+
+    print("当天尾号5或6活跃用户总数：")
+    print(len(tail56))
+
+    cover = len(set(tail56)&set(device_id))
+    print("当天尾号5或6活跃用户覆盖数：")
+    print(cover)
+    cover_percent = format(cover / len(tail56), ".3f")
+    print("当天尾号5或6活跃用户覆盖率：")
+    print(cover_percent)
+    sql_click = "select count(cid) from data_feed_click " \
+                "where (cid_type = 'diary' or cid_type = 'diary_video') " \
+                "and stat_date = '{}' and device_id regexp '[5|6]$';".format(date)
+    cursor.execute(sql_click)
+    click = cursor.fetchone()[0]
+    print("点击数：" + str(click))
+    sql_exp = "select count(cid) from data_feed_exposure " \
+              "where (cid_type = 'diary' or cid_type = 'diary_video') and stat_date = '{}' and " \
+              "device_id regexp '[5|6]$'".format(date)
+    cursor.execute(sql_exp)
+    exp = cursor.fetchone()[0]
+    print("曝光数：" + str(exp))
+    print("点击率：" + str(click / exp))
+    return len(tail56),cover,cover_percent,click,exp,format(click/exp,".6f")
+
+
+def ctr():
+    sql_click = "select count(cid) from data_feed_click " \
+              "where (cid_type = 'diary' or cid_type = 'diary_video') " \
+                "and stat_date = '{}' and device_id in {};".format(date,device_id)
+    db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_prod')
+    cursor = db.cursor()
+    cursor.execute(sql_click)
+    click = cursor.fetchone()[0]
+    print("实验用户点击数："+str(click))
+    sql_exp = "select count(cid) from data_feed_exposure " \
+              "where (cid_type = 'diary' or cid_type = 'diary_video') and stat_date = '{}' and " \
+              "device_id in {}".format(date,device_id)
+    cursor.execute(sql_exp)
+    exp = cursor.fetchone()[0]
+    print("实验用户曝光数："+str(exp))
+    print("实验用户点击率："+str(click/exp))
+
+    return click,exp,format(click/exp,".6f")
+
+
+def rate2file():
+    output_path = DIRECTORY_PATH + "56ctr.csv"
+    with open(output_path,'a+') as f:
+        line = date.replace('-', '')+','+str(temp_data[0])+','+str(temp_data[1])+','+str(temp_data[2])+\
+               ","+str(data[0])+","+str(data[1])+","+str(data[2])+","+str(data[3])+","+str(data[4])\
+               +","+str(data[5])+'\n'
+        f.write(line)
+
+
+if __name__ == "__main__":
+    DIRECTORY_PATH = "/data2/ffm/"
+    date = get_yesterday_date()
+    device_id = get_data()
+    temp_data = ctr()
+    data = ctr_all()
+    rate2file()