新增实时预测

eda/recommended_indexs/code/getActivateUidCtr.py

@@ -21,7 +21,7 @@ def get_all_click_one_rate():
 	impression_one_count = con_sql(sql)
 	impression_one_count = impression_one_count[0][0]
 	all_click_one_rate = click_one_count / impression_one_count
-	return ["all",click_one_count,impression_one_count,round(all_click_one_rate,4)]
+	return ["所有",click_one_count,impression_one_count,round(all_click_one_rate,4)]
 
 
 
@@ -35,7 +35,7 @@ def get_ios_click_one_rate():
 	impression_one_count = con_sql(sql)
 	impression_one_count = impression_one_count[0][0]
 	ios_click_one_rate = click_one_count / impression_one_count
-	return ["ios",click_one_count,impression_one_count,round(ios_click_one_rate,4)]
+	return ["苹果",click_one_count,impression_one_count,round(ios_click_one_rate,4)]
 
 
 #3 获取安卓平台的有点击用户点击率
@@ -47,7 +47,7 @@ def get_android_click_one_rate():
 	impression_one_count = con_sql(sql)
 	impression_one_count = impression_one_count[0][0]
 	android_click_one_rate = click_one_count / impression_one_count
-	return ["android",click_one_count,impression_one_count,round(android_click_one_rate,4)]
+	return ["安卓",click_one_count,impression_one_count,round(android_click_one_rate,4)]
 
 
 

eda/recommended_indexs/code/getAnswerImpRate.py

@@ -21,7 +21,7 @@ def get_all_answer_imp_rate():
 	all_imp_count = con_sql(sql)
 	all_imp_count = all_imp_count[0][0]
 	all_answer_imp_rate = answer_imp_count / all_imp_count
-	return ["all",answer_imp_count,all_imp_count,round(all_answer_imp_rate,4)]
+	return ["所有",answer_imp_count,all_imp_count,round(all_answer_imp_rate,4)]
 
 
 #2 获取ios平台的问答曝光占比
@@ -33,7 +33,7 @@ def get_ios_answer_imp_rate():
 	all_imp_count = con_sql(sql)
 	all_imp_count = all_imp_count[0][0]
 	ios_answer_imp_rate = answer_imp_count / all_imp_count
-	return ["ios",answer_imp_count,all_imp_count,round(ios_answer_imp_rate,4)]
+	return ["苹果",answer_imp_count,all_imp_count,round(ios_answer_imp_rate,4)]
 
 
 #3 获取安卓平台的问答曝光占比
@@ -45,7 +45,7 @@ def get_android_answer_imp_rate():
 	all_imp_count = con_sql(sql)
 	all_imp_count = all_imp_count[0][0]
 	android_answer_imp_rate = answer_imp_count / all_imp_count
-	return ["android",answer_imp_count,all_imp_count,round(android_answer_imp_rate,4)]
+	return ["安卓",answer_imp_count,all_imp_count,round(android_answer_imp_rate,4)]
 
 
 

eda/recommended_indexs/code/getClickAnswerUidRate.py

@@ -21,7 +21,7 @@ def get_all_click_answer_rate():
 	impression_answer_count = con_sql(sql)
 	impression_answer_count = impression_answer_count[0][0]
 	all_click_answer_rate = click_answer_count / impression_answer_count
-	return ["all",click_answer_count,impression_answer_count,round(all_click_answer_rate,4)]
+	return ["所有",click_answer_count,impression_answer_count,round(all_click_answer_rate,4)]
 
 
 
@@ -34,7 +34,7 @@ def get_ios_click_answer_rate():
 	impression_answer_count = con_sql(sql)
 	impression_answer_count = impression_answer_count[0][0]
 	ios_click_answer_rate = click_answer_count / impression_answer_count
-	return ["ios",click_answer_count,impression_answer_count,round(ios_click_answer_rate,4)]
+	return ["苹果",click_answer_count,impression_answer_count,round(ios_click_answer_rate,4)]
 
 
 
@@ -48,7 +48,7 @@ def get_android_click_answer_rate():
 	impression_answer_count = con_sql(sql)
 	impression_answer_count = impression_answer_count[0][0]
 	android_click_answer_rate = click_answer_count / impression_answer_count
-	return ["android",click_answer_count,impression_answer_count,round(android_click_answer_rate,4)]
+	return ["安卓",click_answer_count,impression_answer_count,round(android_click_answer_rate,4)]
 
 
 

eda/recommended_indexs/code/getClickDiaryUidRate.py

@@ -21,7 +21,7 @@ def get_all_click_diary_rate():
 	impression_diary_count = con_sql(sql)
 	impression_diary_count = impression_diary_count[0][0]
 	all_click_diary_rate = click_diary_count / impression_diary_count
-	return ["all",click_diary_count,impression_diary_count,round(all_click_diary_rate,4)]
+	return ["所有",click_diary_count,impression_diary_count,round(all_click_diary_rate,4)]
 
 
 
@@ -34,7 +34,7 @@ def get_ios_click_diary_rate():
 	impression_diary_count = con_sql(sql)
 	impression_diary_count = impression_diary_count[0][0]
 	ios_click_diary_rate = click_diary_count / impression_diary_count
-	return ["ios",click_diary_count,impression_diary_count,round(ios_click_diary_rate,4)]
+	return ["苹果",click_diary_count,impression_diary_count,round(ios_click_diary_rate,4)]
 
 
 #3 获取安卓平台的点击问答用户占比
@@ -46,7 +46,7 @@ def get_android_click_diary_rate():
 	impression_diary_count = con_sql(sql)
 	impression_diary_count = impression_diary_count[0][0]
 	android_click_diary_rate = click_diary_count / impression_diary_count
-	return ["android",click_diary_count,impression_diary_count,round(android_click_diary_rate,4)]
+	return ["安卓",click_diary_count,impression_diary_count,round(android_click_diary_rate,4)]
 
 
 

eda/recommended_indexs/code/getClickTimes2CountUid.py

@@ -31,12 +31,13 @@ def get_click_times_to_count_uid_df():
 
 def df2file(df,fpath):
 	with open(fpath,"w") as f:
+		tplt = "{0:^10}\t{1:^10}\n"
 		f.write("#2. Top特征\n")
 		f.write("=================================================================\n")
 		f.write("2.1用户点击次数分布(第一列：用户点击次数；第二列：独立用户数量)\n")
-		f.write("click_times"+"\t"+"count_uid"+"\n")
+		f.write(tplt.format("click_times","count_uid"))
 		for row in df.iterrows():
-			line = str(row[1][0]) + "\t" + str(row[1][1]) + "\n"
+			line = tplt.format(row[1][0],row[1][1])
 			f.write(line)
 		f.write("\n\n")
 

eda/recommended_indexs/code/getClickZeroUidRate.py

@@ -23,7 +23,7 @@ def get_all_click_zero_rate():
 	impression_zero_count = impression_zero_count[0][0]
 	click_zero_count = impression_zero_count-click_zero_count
 	all_click_zero_rate = click_zero_count / impression_zero_count
-	return ["all",click_zero_count,impression_zero_count,round(all_click_zero_rate,4)]
+	return ["所有",click_zero_count,impression_zero_count,round(all_click_zero_rate,4)]
 
 
 
@@ -37,7 +37,7 @@ def get_ios_click_zero_rate():
 	impression_zero_count = impression_zero_count[0][0]
 	click_zero_count = impression_zero_count-click_zero_count
 	ios_click_zero_rate = click_zero_count / impression_zero_count
-	return ["ios",click_zero_count,impression_zero_count,round(ios_click_zero_rate,4)]
+	return ["苹果",click_zero_count,impression_zero_count,round(ios_click_zero_rate,4)]
 
 
 #3 获取安卓平台的0点击用户占比
@@ -50,7 +50,7 @@ def get_android_click_zero_rate():
 	impression_zero_count = impression_zero_count[0][0]
 	click_zero_count = impression_zero_count-click_zero_count
 	android_click_zero_rate = click_zero_count / impression_zero_count
-	return ["android",click_zero_count,impression_zero_count,round(android_click_zero_rate,4)]
+	return ["安卓",click_zero_count,impression_zero_count,round(android_click_zero_rate,4)]
 
 
 

eda/recommended_indexs/code/getRate.py

@@ -15,78 +15,68 @@ def get_yesterday_date():
 
 def result2file(fpath):
 	with open(fpath,'w') as f:
+		tplt = "{0:\u3000<6}\t{1:\u3000<15}\t{2:\u3000<15}\t{3:\u3000<15}\n"
 		f.write("#注意：以下数据都是首页的\n")
 		f.write("#1. 比例特征\n")
 		f.write("=================================================================\n")
 		f.write("#1.1answer曝光占比(=answer被曝光数/总cid被曝光数)\n")
-		f.write("平台"+"\t"+"answer被曝光数"+"\t"+"总cid被曝光数"+"\t"+"answer被曝光数占比\n")
+		f.write(tplt.format("平台","answer被曝光数","总cid被曝光数","answer被曝光数占比"))
 		all_answer_imp_rate = get_all_answer_imp_rate()
 		ios_answer_imp_rate = get_ios_answer_imp_rate()
 		android_answer_imp_rate = get_android_answer_imp_rate()
 		lst = [all_answer_imp_rate,ios_answer_imp_rate,android_answer_imp_rate]
 		for i in lst:
-			line = ""
-			for j in i:
-				line += str(j) + '\t'
-			line = line[:-1]+'\n'
+			line = tplt.format(i[0],i[1],i[2],i[3])
 			f.write(line)
+		f.write('\n')
 		print("1.1已将answer曝光占比存入文件")
 
-		f.write("#1.2有点击用户点击率(=有点击用户点击次数/有点击用户曝光次数)\n")
-		f.write("平台"+"\t"+"有点击用户点击次数"+"\t"+"有点击用户曝光次数"+"\t"+"有点击用户点击率\n")
+		f.write("#1.2活跃用户点击率(=活跃用户点击次数/活跃用户曝光次数)\n")
+		f.write(tplt.format("平台","active用户点击次数","active用户曝光次数","active用户点击率"))
 		all_click_one_rate = get_all_click_one_rate()
 		ios_click_one_rate = get_ios_click_one_rate()
 		android_click_one_rate = get_android_click_one_rate()
 		lst = [all_click_one_rate,ios_click_one_rate,android_click_one_rate]
 		for i in lst:
-			line = ""
-			for j in i:
-				line += str(j) + '\t'
-			line = line[:-1]+'\n'
+			line = tplt.format(i[0],i[1],i[2],i[3])
 			f.write(line)
-		print("1.2已将有点击用户点击率存入文件")
+		f.write('\n')
+		print("1.2已将活跃用户点击率存入文件")
 
 		f.write("#1.3点击answer用户占比(=点击answer用户数/曝光answer用户数)\n")
-		f.write("平台"+"\t"+"点击answer用户数"+"\t"+"曝光answer用户数"+"\t"+"点击answer用户占比\n")
+		f.write(tplt.format("平台","点击answer用户数","曝光answer用户数","击answer用户占比"))
 		all_click_answer_rate = get_all_click_answer_rate()
 		ios_click_answer_rate = get_ios_click_answer_rate()
 		android_click_answer_rate = get_android_click_answer_rate()
 		lst = [all_click_answer_rate,ios_click_answer_rate,android_click_answer_rate]
 		for i in lst:
-			line = ""
-			for j in i:
-				line += str(j) + '\t'
-			line = line[:-1]+'\n'
+			line = tplt.format(i[0],i[1],i[2],i[3])
 			f.write(line)
+		f.write('\n')
 		print("1.3已将点击answer用户占比存入文件")
 
 		f.write("#1.4点击diary用户占比(=点击diary用户数/曝光diary用户数)\n")
-		f.write("平台"+"\t"+"点击diary用户数"+"\t"+"曝光diary用户数"+"\t"+"点击diary用户占比\n")
+		f.write(tplt.format("平台","点击diary用户数","曝光diary用户数","击diary用户占比"))
 		all_click_diary_rate = get_all_click_diary_rate()
 		ios_click_diary_rate = get_ios_click_diary_rate()
 		android_click_diary_rate = get_android_click_diary_rate()
 		lst = [all_click_diary_rate,ios_click_diary_rate,android_click_diary_rate]
 		for i in lst:
-			line = ""
-			for j in i:
-				line += str(j) + '\t'
-			line = line[:-1]+'\n'
+			line = tplt.format(i[0],i[1],i[2],i[3])
 			f.write(line)
+		f.write('\n')
 		print("1.4已将点击diary用户占比存入文件")
 
 		f.write("#1.5无点击用户占比(=无点击用户数/有曝光用户数)\n")
-		f.write("平台"+"\t"+"无点击用户数"+"\t"+"有曝光用户数"+"\t"+"无点击用户占比\n")
+		f.write(tplt.format("平台","no点击用户数","have曝光用户数","no点击用户占比"))
 		all_click_zero_rate = get_all_click_zero_rate()
 		ios_click_zero_rate = get_ios_click_zero_rate()
 		android_click_zero_rate = get_android_click_zero_rate()
 		lst = [all_click_zero_rate,ios_click_zero_rate,android_click_zero_rate]
 		for i in lst:
-			line = ""
-			for j in i:
-				line += str(j) + '\t'
-			line = line[:-1]+'\n'
+			line = tplt.format(i[0],i[1],i[2],i[3])
 			f.write(line)
-		f.write('\n\n')
+		f.write('\n')
 		print("1.5已将无点击用户占比存入文件")
 
 

eda/recommended_indexs/code/getTop100Answer.py

@@ -20,20 +20,16 @@ def tuple2dict(tuple_result):
 
 def result2file(result_lst,fpath):
 	with open(fpath,'w') as f:
-		header = "平台"+'\t'+"answer_id"+'\t'+"answer被点击数"+'\t'+"answer被曝光数"+'\t'+"answer被点击率"+'\t'+"answer链接"+'\n'
+		tplt = "{0:\u3000<4}\t{1:\u3000<12}\t{2:\u3000^6}\t{3:\u3000^6}\t{4:\u3000<8}\t{5:\u3000^15}\n"
 		f.write("Top 100 Answer\n")
 		f.write("=================================================================\n")
-		f.write(header)
+		f.write(tplt.format("平台","answer_id","点击数","曝光数","点击率","answer链接"))
 		for i in result_lst:
 			for j in i:
-				line = ""
-				for k in j:
-					line += str(k) + '\t'
-				line = line[:-1] + '\n'
-				f.write(line)
+				f.write(tplt.format(j[0],j[1],j[2],j[3],j[4],j[5]))
 			f.write("=================================================================\n")
 			if i != result_lst[-1]:
-				f.write(header)
+				f.write(tplt.format("平台","answer_id","点击数","曝光数","点击率","answer链接"))
 		f.write("\n\n")
 
 
@@ -58,7 +54,7 @@ def get_all_top100_answer_rate_by_ctr(all_answer_count_by_click,all_answer_count
 	for i in all_answer_count_by_click:
 		if i in all_answer_count_by_imp.keys() and all_answer_count_by_click[i]>2:
 			url = "http://m.igengmei.com/answer/" + i[i.index('|')+1:] + '/'
-			all_top100_answer_rate_by_ctr.append(("all",i,all_answer_count_by_click[i],all_answer_count_by_imp[i], round(all_answer_count_by_click[i]/all_answer_count_by_imp[i],4),url))
+			all_top100_answer_rate_by_ctr.append(("所有",i,all_answer_count_by_click[i],all_answer_count_by_imp[i], round(all_answer_count_by_click[i]/all_answer_count_by_imp[i],4),url))
 	all_top100_answer_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 	return all_top100_answer_rate_by_ctr[:100] if len(all_top100_answer_rate_by_ctr) > 100 else all_top100_answer_rate_by_ctr
 
@@ -82,7 +78,7 @@ def get_ios_top100_answer_rate_by_ctr(ios_answer_count_by_click,ios_answer_count
 	for i in ios_answer_count_by_click:
 		if i in ios_answer_count_by_imp.keys() and ios_answer_count_by_click[i]>2:
 			url = "http://m.igengmei.com/answer/" + i[i.index('|')+1:] + '/'
-			ios_top100_answer_rate_by_ctr.append(("ios",i,ios_answer_count_by_click[i],ios_answer_count_by_imp[i], round(ios_answer_count_by_click[i]/ios_answer_count_by_imp[i],4),url))
+			ios_top100_answer_rate_by_ctr.append(("苹果",i,ios_answer_count_by_click[i],ios_answer_count_by_imp[i], round(ios_answer_count_by_click[i]/ios_answer_count_by_imp[i],4),url))
 	ios_top100_answer_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 	return ios_top100_answer_rate_by_ctr[:100] if len(ios_top100_answer_rate_by_ctr) > 100 else ios_top100_answer_rate_by_ctr
 
@@ -106,7 +102,7 @@ def get_android_top100_answer_rate_by_ctr(android_answer_count_by_click,android_
     for i in android_answer_count_by_click:
         if i in android_answer_count_by_imp.keys() and android_answer_count_by_click[i]>2:
             url = "http://m.igengmei.com/answer/" + i[i.index('|')+1:] + '/'
-            android_top100_answer_rate_by_ctr.append(("android",i,android_answer_count_by_click[i],android_answer_count_by_imp[i],round(android_answer_count_by_click[i]/android_answer_count_by_imp[i],4),url))
+            android_top100_answer_rate_by_ctr.append(("安卓",i,android_answer_count_by_click[i],android_answer_count_by_imp[i],round(android_answer_count_by_click[i]/android_answer_count_by_imp[i],4),url))
     android_top100_answer_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
     return android_top100_answer_rate_by_ctr[:100] if len(android_top100_answer_rate_by_ctr) > 100 else android_top100_answer_rate_by_ctr
 

eda/recommended_indexs/code/getTop100Diary.py

@@ -20,20 +20,16 @@ def tuple2dict(tuple_result):
 
 def result2file(result_lst,fpath):
 	with open(fpath,'w') as f:
-		header = "平台"+'\t'+"diary_id"+'\t'+"diary被点击数"+'\t'+"diary被曝光数"+'\t'+"diary被点击率"+'\t'+"diary链接"+'\n'
+		tplt = "{0:\u3000<4}\t{1:\u3000<12}\t{2:\u3000^6}\t{3:\u3000^6}\t{4:\u3000<8}\t{5:\u3000^15}\n"
 		f.write("Top 100 diary\n")
 		f.write("=================================================================\n")
-		f.write(header)
+		f.write(tplt.format("平台","diary_id","点击数","曝光数","点击率","diary链接"))
 		for i in result_lst:
 			for j in i:
-				line = ""
-				for k in j:
-					line += str(k) + '\t'
-				line = line[:-1] + '\n'
-				f.write(line)
+				f.write(tplt.format(j[0],j[1],j[2],j[3],j[4],j[5]))
 			f.write("=================================================================\n")
 			if i != result_lst[-1]:
-				f.write(header)
+				f.write(tplt.format("平台","diary_id","点击数","曝光数","点击率","diary链接"))
 		f.write("\n\n")
 
 
@@ -57,7 +53,7 @@ def get_all_top100_diary_rate_by_ctr(all_diary_count_by_click,all_diary_count_by
 	for i in all_diary_count_by_click:
 		if i in all_diary_count_by_imp.keys() and all_diary_count_by_click[i] > 4:
 			url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
-			all_top100_diary_rate_by_ctr.append(("all",i,all_diary_count_by_click[i],all_diary_count_by_imp[i], round(all_diary_count_by_click[i]/all_diary_count_by_imp[i],4),url))
+			all_top100_diary_rate_by_ctr.append(("所有",i,all_diary_count_by_click[i],all_diary_count_by_imp[i], round(all_diary_count_by_click[i]/all_diary_count_by_imp[i],4),url))
 	all_top100_diary_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 	return all_top100_diary_rate_by_ctr[:100] if len(all_top100_diary_rate_by_ctr) > 100 else all_top100_diary_rate_by_ctr
 
@@ -81,7 +77,7 @@ def get_ios_top100_diary_rate_by_ctr(ios_top100_diary_count_by_click,ios_top100_
 	for i in ios_diary_count_by_click:
 		if i in ios_diary_count_by_imp.keys() and ios_diary_count_by_click[i] > 4:
 			url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
-			ios_top100_diary_rate_by_ctr.append(("ios",i,ios_diary_count_by_click[i],ios_diary_count_by_imp[i], round(ios_diary_count_by_click[i]/ios_diary_count_by_imp[i],4),url))
+			ios_top100_diary_rate_by_ctr.append(("苹果",i,ios_diary_count_by_click[i],ios_diary_count_by_imp[i], round(ios_diary_count_by_click[i]/ios_diary_count_by_imp[i],4),url))
 	ios_top100_diary_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 	return ios_top100_diary_rate_by_ctr[:100] if len(ios_top100_diary_rate_by_ctr) > 100 else ios_top100_diary_rate_by_ctr
 
@@ -105,7 +101,7 @@ def get_android_top100_diary_rate_by_ctr(android_top100_diary_count_by_click,and
     for i in android_diary_count_by_click:
         if i in android_diary_count_by_imp.keys() and android_diary_count_by_click[i] > 4:
             url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
-            android_top100_diary_rate_by_ctr.append(("android",i,android_diary_count_by_click[i],android_diary_count_by_imp[i], round(android_diary_count_by_click[i]/android_diary_count_by_imp[i],4),url))
+            android_top100_diary_rate_by_ctr.append(("安卓",i,android_diary_count_by_click[i],android_diary_count_by_imp[i], round(android_diary_count_by_click[i]/android_diary_count_by_imp[i],4),url))
     android_top100_diary_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
     return android_top100_diary_rate_by_ctr[:100] if len(android_top100_diary_rate_by_ctr) > 100 else android_top100_diary_rate_by_ctr
 

eda/recommended_indexs/code/getTop100Question.py

@@ -20,20 +20,16 @@ def tuple2dict(tuple_result):
 
 def result2file(result_lst,fpath):
 	with open(fpath,'w') as f:
-		header = "平台"+'\t'+"question_id"+'\t'+"question被点击数"+'\t'+"question被曝光数"+'\t'+"question被点击率"+'\t'+"question链接"+'\n'
-		f.write("Top 100 question\n")
+		tplt = "{0:\u3000<4}\t{1:\u3000<12}\t{2:\u3000^6}\t{3:\u3000^6}\t{4:\u3000<8}\t{5:\u3000^15}\n"
+		f.write("Top 100 Question\n")
 		f.write("=================================================================\n")
-		f.write(header)
+		f.write(tplt.format("平台","question_id","点击数","曝光数","点击率","question链接"))
 		for i in result_lst:
 			for j in i:
-				line = ""
-				for k in j:
-					line += str(k) + '\t'
-				line = line[:-1] + '\n'
-				f.write(line)
+				f.write(tplt.format(j[0],j[1],j[2],j[3],j[4],j[5]))
 			f.write("=================================================================\n")
 			if i != result_lst[-1]:
-				f.write(header)
+				f.write(tplt.format("平台","question_id","点击数","曝光数","点击率","question链接"))
 		f.write("\n\n")
 
 
@@ -58,14 +54,14 @@ def get_all_top100_question_rate_by_ctr(all_question_count_by_click,all_question
 	if all_question_count_by_imp == {}:
 		for i in all_question_count_by_click:
 			url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-			all_top100_question_rate_by_ctr.append(("all",i,all_question_count_by_click[i],0,0,url))
+			all_top100_question_rate_by_ctr.append(("所有",i,all_question_count_by_click[i],0,0,url))
 		all_top100_question_rate_by_ctr.sort(key=lambda x:x[2],reverse=True)
 		return all_top100_question_rate_by_ctr[:100] if len(all_top100_question_rate_by_ctr) > 100 else all_top100_question_rate_by_ctr
 	else:
 		for i in all_question_count_by_click:
 			if i in all_question_count_by_imp.keys() and all_question_count_by_click[i]>2:
 				url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-				all_top100_question_rate_by_ctr.append(("all",i,all_question_count_by_click[i],all_question_count_by_imp[i], round(all_question_count_by_click[i]/all_question_count_by_imp[i],4),url))
+				all_top100_question_rate_by_ctr.append(("所有",i,all_question_count_by_click[i],all_question_count_by_imp[i], round(all_question_count_by_click[i]/all_question_count_by_imp[i],4),url))
 		all_top100_question_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 		return all_top100_question_rate_by_ctr[:100] if len(all_top100_question_rate_by_ctr) > 100 else all_top100_question_rate_by_ctr
 
@@ -89,14 +85,14 @@ def get_ios_top100_question_rate_by_ctr(ios_question_count_by_click,ios_question
 	if ios_question_count_by_imp == {}:
 		for i in ios_question_count_by_click:
 			url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-			ios_top100_question_rate_by_ctr.append(("ios",i,ios_question_count_by_click[i],0,0,url))
+			ios_top100_question_rate_by_ctr.append(("苹果",i,ios_question_count_by_click[i],0,0,url))
 		ios_top100_question_rate_by_ctr.sort(key=lambda	x:x[2],reverse=True)
 		return ios_top100_question_rate_by_ctr[:100] if len(ios_top100_question_rate_by_ctr) > 100 else ios_top100_question_rate_by_ctr
 	else:
 		for i in ios_question_count_by_click:
 			if i in ios_question_count_by_imp.keys() and ios_question_count_by_click[i]>2:
 				url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-				ios_top100_question_rate_by_ctr.append(("ios",i,ios_question_count_by_click[i],ios_question_count_by_imp[i], round(ios_question_count_by_click[i]/ios_question_count_by_imp[i],4),url))
+				ios_top100_question_rate_by_ctr.append(("苹果",i,ios_question_count_by_click[i],ios_question_count_by_imp[i], round(ios_question_count_by_click[i]/ios_question_count_by_imp[i],4),url))
 		ios_top100_question_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 		return ios_top100_question_rate_by_ctr[:100] if len(ios_top100_question_rate_by_ctr) > 100 else ios_top100_question_rate_by_ctr
 
@@ -120,14 +116,14 @@ def get_android_top100_question_rate_by_ctr(android_question_count_by_click,andr
     if android_question_count_by_imp == {}:
     	for i in android_question_count_by_click:
     		url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-    		android_top100_question_rate_by_ctr.append(("android",i,android_question_count_by_click[i],0,0,url))
+    		android_top100_question_rate_by_ctr.append(("安卓",i,android_question_count_by_click[i],0,0,url))
     	android_top100_question_rate_by_ctr.sort(key=lambda x:x[2],reverse=True)
     	return android_top100_question_rate_by_ctr[:100] if len(android_top100_question_rate_by_ctr) > 100 else android_top100_question_rate_by_ctr
     else:
 	    for i in android_question_count_by_click:
 	        if i in android_question_count_by_imp.keys() and android_question_count_by_click[i]>2:
 	            url = "http://m.igengmei.com/question/" + i[i.index('|')+1:] + '/'
-	            android_top100_question_rate_by_ctr.append(("android",i,android_question_count_by_click[i],android_question_count_by_imp[i],round(android_question_count_by_click[i]/android_question_count_by_imp[i],4),url))
+	            android_top100_question_rate_by_ctr.append(("安卓",i,android_question_count_by_click[i],android_question_count_by_imp[i],round(android_question_count_by_click[i]/android_question_count_by_imp[i],4),url))
 	    android_top100_question_rate_by_ctr.sort(key=lambda x:x[4],reverse=True)
 	    return android_top100_question_rate_by_ctr[:100] if len(android_top100_question_rate_by_ctr) > 100 else android_top100_question_rate_by_ctr
 

eda/recommended_indexs/code/start.sh

@@ -4,4 +4,4 @@ python getTop100Diary.py
 python getTop100Answer.py
 python getTop100Question.py
 dt=$(date -d last-day +%Y%m%d)
-cat 1rate_features_$dt.txt 2click_times_to_count_uid_$dt.txt 3top100_ctr_diary_$dt.txt 4top100_ctr_answer_$dt.txt 5top100_ctr_question_$dt.txt > result_all_$dt.txt
+cat /data2/models/eda/recommended_indexs/1rate_features_$dt.txt /data2/models/eda/recommended_indexs/2click_times_to_count_uid_$dt.txt /data2/models/eda/recommended_indexs/3top100_ctr_diary_$dt.txt /data2/models/eda/recommended_indexs/4top100_ctr_answer_$dt.txt /data2/models/eda/recommended_indexs/5top100_ctr_question_$dt.txt > /data2/models/eda/recommended_indexs/result_all_$dt.txt
\ No newline at end of file

eda/test/config.py

+DIRECTORY_PATH="/data2/models/eda/recommended_indexs/"
\ No newline at end of file

eda/test/getTopFeatures.py

+# -*- coding: UTF-8 -*-
+from utils import con_sql,tuple2dict,get_yesterday_date
+from config import DIRECTORY_PATH
+
+
+class TopFeatures(object):
+	def __init__(self, ndays, platform, cid_type, top_n=-1):
+		"""
+		ndays : 1;2;3;4..
+		platform : 'all';'ios';'android'
+		cid_type : 'diary';'answer';'question'...
+		top_n : the rows of the result
+		"""
+		self.ndays = ndays
+		if platform == "ios":
+			self.platform = "='AppStore'"
+		elif platform == "android":
+			self.platform = "!='AppStore'"
+		else:
+			self.platform = " is not null"
+		self.cid_type = cid_type
+		self.top_n = top_n
+
+
+	def get_click_times(self):
+		# rtype : dict
+		if self.cid_type[-2] == 'e':
+			self.cid_type = self.cid_type.replace(' ','')
+		sql = "select cid,count(cid) from data_feed_click \
+		where from_unixtime(time,'%Y-%m-%d')=date_add(curdate(), interval -{0} day) \
+		and device_type{1}} and cid_type='{2}' \
+		group by cid order by count(cid) desc".format(self.ndays, self.platform, self.cid_type)
+		clk_times = tuple2dict(con_sql(sql))
+		return clk_times
+
+
+	def get_impression_times(self):
+		# rtype : dict
+		if self.cid_type[-2] == 'e':
+			self.cid_type = self.cid_type[:-6] + ' ' + self.cid_type[:-6:]
+		sql = "select cid,count(cid) from data_feed_exposure \
+		where from_unixtime(time,'%Y-%m-%d')=date_add(curdate(), interval -{0} day) \
+		and device_type{1}} and cid_type='{2}' \
+		group by cid order by count(cid) desc".format(self.ndays, self.platform, self.cid_type)
+		imp_times = tuple2dict(con_sql(sql))
+		return imp_times
+
+	def get_result(self, clk, imp, clk_n=2, result_types="ctr"):
+		"""
+		result_types : "clk";"imp";"ctr"
+		clk : dict
+		imp : dict
+		clk_n : 获取topN点击率时，过滤的点击数
+		rtype : list
+		"""
+		topn = []
+		#获取topN的点击
+		if imp == {} or result_types == "clk":
+			for i in clk:
+				if self.cid_type == "diary":
+					url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
+				else:
+					url = "http://m.igengmei.com/{0}/".format(self.cid_type) + i[i.index('|')+1:] + '/'
+				topn.append((self.cid_type.strip(),i,clk[i],0,0,url))
+			topn.sort(key=lambda x:x[2],reverse=True)
+			return topn[:int(self.top_n)]
+		#获取topN的曝光
+		elif clk == {} or result_types == "imp":
+			for i in imp:
+				if self.cid_type == "diary":
+					url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
+				else:
+					url = "http://m.igengmei.com/{0}/".format(self.cid_type) + i[i.index('|')+1:] + '/'
+				topn.append((self.cid_type.strip(),i,0,imp[i],0,url))
+			topn.sort(key=lambda x:x[3],reverse=True)
+			return topn[:int(self.top_n)]
+		#获取topN的ctr
+		else:
+			for i in clk:
+				if i in imp.keys() and clk[i] > clk_n:
+					if self.cid_type == "diary":
+						url = "http://m.igengmei.com/diary_book/" + i[i.index('|')+1:] + '/'
+					else:
+						url = "http://m.igengmei.com/{0}/".format(self.cid_type) + i[i.index('|')+1:] + '/'
+					topn.append((self.cid_type.strip(),i,clk[i],imp[i],round(clk[i]/imp[i],4),url))
+			topn.sort(key=lambda x:x[4],reverse=True)
+			return topn[:int(self.top_n)]
+
+	def result2file(self, result_lst, fpath):
+		"""
+		result_lst : [all,ios,android]
+		fpath : output filename
+		rtype : none
+		"""
+		with open(fpath, 'w') as f:
+			tplt = "{0:\u3000<4}\t{1:\u3000<12}\t{2:\u3000^6}\t{3:\u3000^6}\t{4:\u3000<8}\t{5:\u3000^15}\n"
+			f.write("Top {0} {1}\n".format(self.top_n,self.cid_type))
+			sep = "=================================================================\n"
+			header = tplt.format("平台","{}_id".format(self.cid_type),"点击数","曝光数","点击率","{}链接".format(self.cid_type))
+			f.write(sep)
+			f.write(header)
+			for i in result_lst:
+				for j in i:
+					f.write(tplt.format(j[0],j[1],j[2],j[3],j[4],j[5]))
+				f.write(sep)
+				if i != result_lst[-1]:
+					f.write(header)
+			f.write("\n\n")
+
+
+
+
+def main():
+	top_diary = TopFeatures(1, "all", "diary")
+	clk_diary_times_all = top_diary.get_click_times()
+	imp_diary_times_all = top_diary.get_impression_times()
+	clk_diary_ctr_all = top_diary.get_result("ctr", clk_diary_times_all, imp_diary_times_all, top_n=100)
+
+	clk_diary_times_ios = top_diary.get_click_times()
+	imp_diary_times_ios = top_diary.get_impression_times()
+	clk_diary_ctr_ios = top_diary.get_result("ctr", clk_diary_times_ios, imp_diary_times_ios, top_n=100)
+
+	clk_diary_times_android = top_diary.get_click_times()
+	imp_diary_times_android = top_diary.get_impression_times()
+	clk_diary_ctr_android = top_diary.get_result("ctr", clk_diary_times_android, imp_diary_times_android, top_n=100)
+
+	result_lst = [clk_diary_ctr_all, clk_diary_ctr_ios, clk_diary_ctr_android]
+	output_path = DIRECTORY_PATH + "5top100_ctr_diary_%s.txt"
+	top_diary.result2file(result_lst, output_path)
+
+
+
+if __name__ == '__main__':
+	main()
+		
\ No newline at end of file

eda/test/utils.py

+# -*- coding: UTF-8 -*-
+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_test')
+	cursor = db.cursor()
+	cursor.execute(sql)
+	result = cursor.fetchall()
+	db.close()
+	return result
+
+def tuple2dict(tuple_result):
+	#把sql结果从tuple(tuple,tuple)格式转换成dict格式
+	"""
+	:type tuple_result : tuple
+	:rtype : dict
+	"""
+	dict_result = {}
+	for i in range(len(tuple_result)):
+		dict_result[tuple_result[i][0]] = tuple_result[i][1]
+	return dict_result
+
+def get_yesterday_date():
+	#自动获取昨天的日期,如"20180808"
+	"""
+	:rtype : str
+	"""
+	today = datetime.date.today()
+	yesterday = today - datetime.timedelta(days=1)
+	yesterday = yesterday.strftime("%Y%m%d")
+	return yesterday