update

eda/smart_rank/dist_update_user_history_order_tags.py

@@ -67,7 +67,7 @@ def get_user_history_order_service_tag(user_id, stat_date):
             tags_dict = cur_zhengxing.fetchall()
             tags_list = [i["tag_id"] for i in tags_dict]
             # 写gmkv
-            gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=6, socket_timeout=2000)
+            gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=2, socket_timeout=2000)
             user_history_order_tags_key = "user:history_order:tags:user_id:" + str(user_id)
             tags_list_json = json.dumps(tags_list)
             gm_kv_cli.set(user_history_order_tags_key, tags_list_json)
@@ -115,7 +115,6 @@ if __name__ == '__main__':
         spark.sparkContext.setLogLevel("WARN")
 
         device_ids_lst_rdd = spark.sparkContext.parallelize(device_ids_lst)
-        gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=6, socket_timeout=2000)
         result = device_ids_lst_rdd.repartition(100).map(lambda x: get_user_history_order_service_tag(x, stat_date))
         result.collect()
 

eda/smart_rank/dist_update_user_portrait.py

@@ -110,7 +110,7 @@ def get_user_tag_score(cl_id, all_log_df, stat_date, size=10):
             finally_score_lst = finally_score[["tag_id","tag_score"]].to_dict('record')
             tag_id_list = tag_list2dict(finally_score_lst, size)
             # 写gmkv
-            gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=6, socket_timeout=2000)
+            gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=2, socket_timeout=2000)
             cl_id_portrait_key = "user:portrait_tags:cl_id:" + str(cl_id)
             tag_id_list_json = json.dumps(tag_id_list)
             gm_kv_cli.set(cl_id_portrait_key, tag_id_list_json)
@@ -157,7 +157,6 @@ if __name__ == '__main__':
         spark.sparkContext.setLogLevel("WARN")
 
         device_ids_lst_rdd = spark.sparkContext.parallelize(device_ids_lst)
-        gm_kv_cli = redis.Redis(host="172.16.40.135", port=5379, db=6, socket_timeout=2000)
         result = device_ids_lst_rdd.repartition(100).map(lambda x: get_user_tag_score(x, all_log_df, stat_date))
         result.collect()