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

修改统计指标

eda/node2vec/stat/all_ctr.py

+# -*- coding: UTF-8 -*-
+import time
+import pymysql
+import datetime
+
+def con_sql(sql):
+	#从数据库的表里获取数据
+	"""
+	:type sql : str
+	:rtype : tuple
+	"""
+	db = pymysql.connect(host='10.66.157.22', port=4000, user='root', passwd='3SYz54LS9#^9sBvC', db='jerry_prod')
+	cursor = db.cursor()
+	cursor.execute(sql)
+	result = cursor.fetchall()
+	db.close()
+	return result
+
+def get_yesterday_date():
+	#自动获取昨天的日期,如"2018-08-08"
+	"""
+	:rtype : str
+	"""
+	today = datetime.date.today()
+	yesterday = today - datetime.timedelta(days=1)
+	yesterday = yesterday.strftime("%Y-%m-%d")
+	return yesterday
+
+
+
+
+OUTPUT_PATH = "/data2/models/eda/node2vec/"
+
+
+class AllStat(object):
+	def __init__(self, cid_type, uid_type, ndays=get_yesterday_date()):
+		"""
+		cid_type : diary，answer，question
+		uid_type : 8：_8结尾；6：_6结尾；6|8：_6或者_8结尾;^68：不是6或者8结尾的
+		ndays : '2018-08-30'....
+		"""
+		self.cid_type = cid_type
+		self.uid_type = uid_type
+		self.ndays = ndays
+
+	def get_uid_count(self):
+		sql = "select count(distinct(device_id)) from data_feed_click \
+			where stat_date='{0}' \
+			and (cid_type='{1}' or cid_type='diary_video') \
+			and device_id regexp '[{2}]$' \
+			and device_id not in (select distinct(device_id) from jerry_test.bl_device_list) \
+			and device_id not in (select distinct(device_id) from jerry_prod.blacklist)".format(self.ndays,\
+				self.cid_type,self.uid_type)
+		uid_count = con_sql(sql)[0][0]
+		return uid_count
+
+	def get_uid_clk_times(self):
+		sql = "select count(device_id) from data_feed_click \
+			where stat_date='{0}' \
+			and (cid_type='{1}' or cid_type='diary_video') \
+			and device_id regexp '[{2}]$' \
+			and device_id not in (select distinct(device_id) from jerry_test.bl_device_list) \
+			and device_id not in (select distinct(device_id) from jerry_prod.blacklist)".format(self.ndays,\
+				self.cid_type,self.uid_type)
+		uid_clk_times = con_sql(sql)[0][0]
+		return uid_clk_times
+
+	def get_uid_imp_times(self):
+		sql = "select count(device_id) from data_feed_exposure \
+			where stat_date='{0}' \
+			and cid_type='{1}' \
+			and device_id regexp '[{2}]$' \
+			and device_id not in (select distinct(device_id) from jerry_test.bl_device_list) \
+			and device_id not in (select distinct(device_id) from jerry_prod.blacklist)".format(self.ndays,\
+				self.cid_type,self.uid_type)
+		uid_imp_times = con_sql(sql)[0][0]
+		return uid_imp_times
+
+
+
+
+def main():
+	output = OUTPUT_PATH + "all_ctr.csv"
+	with open(output,"a+") as f:
+		tail_nu = ["1|2","3|4","5|6","7|8","0-9"]
+		resulst_ctr = []
+		for i in tail_nu:
+			class_tail = AllStat("diary",i)
+			ctr_tail = round(class_tail.get_uid_clk_times()/class_tail.get_uid_imp_times(),4)
+			resulst_ctr.append(str(ctr_tail))
+		line = get_yesterday_date()+','+','.join([i for i in resulst_ctr])+'\n'
+		f.write(line)
+
+if __name__ == '__main__':
+	main()
+		
+
+
+
+
+

eda/node2vec/stat/node2vec_ctr.py

@@ -133,7 +133,7 @@ class AllStat(object):
 
 
 def main():
-	output = OUTPUT_PATH + "ctr.csv"
+	output = OUTPUT_PATH + "nd_ctr.csv"
 	with open(output,"a+") as f:
 		print("stat" + "  " + get_yesterday_date())
 		g_class = GrayStat("diary","3|4")