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

add a clickZeroUidRateDetail.py for the distribution of rate

aucCaculate.py

+
+
 from eda.ml_tools.rocCurve import get_roc_curve
 import pandas as pd
 from config import *

config.py

 
+
 DIRECTORY_PATH = '/data2/models/'
 VALIDATION_DATE = '2018-08-05'
 TEST_DATE = '2018-08-06'

delete_temp_csv.py

+
+
 import os
 import time
 from config import *

diaryCandidateSet.py

+
+
 import pymysql
 import pandas as pd
 from utils import *

diaryTraining.py

+
+
 import xlearn as xl
 from config import *
 

predictDiary.py

+
+
 from config import *
 import pandas as pd
 import pickle
@@ -40,6 +42,7 @@ def transform_ffm_format(df, device_id):
         print("ffm格式转化结束")
         predict_file_name = DIRECTORY_PATH + "result/{0}_{1}DiaryTop3000.csv".format(device_id, now)
         data.to_csv(predict_file_name, index=False,header=None)
+        print("ffm写到服务器")
         return predict_file_name
 
 
@@ -86,7 +89,7 @@ def predict_save_to_redis(user_profile, instance):
 def router(device_id):
     user_profile, not_exist = fetch_user_profile(device_id)
 
-    if not_exist:
+    if not_exist==1:
         print('Sorry, we don\'t have you.')
     else:
         predict(user_profile)
@@ -97,7 +100,7 @@ if __name__ == "__main__":
     while True:
         start = time.time()
         empty,device_id_list = get_active_users()
-        if empty:
+        if empty==1:
             time.sleep(60)
         else:
             old_device_id_list = pd.read_csv(DIRECTORY_PATH + "data_set_device_id.csv")["device_id"].values.tolist()

prepareData.py

+
+
 from utils import con_sql
 import datetime
 

processData.py

+
+
 import time
 from prepareData import fetch_data
 from utils import FFMFormatPandas

train.py

+
+
 from processData import *
 from diaryTraining import *
 from diaryCandidateSet import get_eachCityDiaryTop3000

userProfile.py

+
+
 from utils import con_sql
 from datetime import datetime
 
@@ -12,13 +14,15 @@ def get_active_users():
     device_id_df = con_sql(sql)
     if device_id_df.empty:
         print("当下这一分钟没有活跃用户，不需要预测")
-        return True,None
+        # 为了debug supervisor，修改了下面的return参数
+        return 1,[1,2]
     else:
         device_id_list = device_id_df[0].values.tolist()
         # 对device_id 进行去重
         device_id_list = list(set(device_id_list))
         print("成功获取当下一分钟内活跃用户")
-        return False,device_id_list
+        # 为了debug supervisor，修改了下面的return参数
+        return "0",device_id_list
 
 
 def fetch_user_profile(device_id):
@@ -26,11 +30,13 @@ def fetch_user_profile(device_id):
         user_profile = con_sql(sql)
         if user_profile.empty:
             print("没有获取到该用户对应的city_id")
-            return {}, user_profile.empty
+            # 为了debug supervisor，修改了下面的return参数
+            return {1:2}, 1
         else:
             user_profile = user_profile.rename(columns={0:"device_id",1:"city_id"})
             print("成功获取该用户对应的city_id")
             user_profile_dict = {}
             for i in user_profile.columns:
                 user_profile_dict[i] = user_profile.loc[0, i]
-            return user_profile_dict, user_profile.empty
+                # 为了debug supervisor，修改了下面的return参数
+            return user_profile_dict, "0"

utils.py

+
+
 # encoding = "utf-8"
 import pymysql
 import pandas as pd
@@ -12,6 +14,7 @@ def con_sql(sql):
     cursor = db.cursor()
     cursor.execute(sql)
     result = cursor.fetchall()
+    print("成功从数据库获取数据")
     df = pd.DataFrame(list(result)).dropna()
     db.close()
     return df