Merge branch 'offic' into 'master'

Offic

See merge request !57

eda/smart_rank/new_tag3_cold_start.py

@@ -2,7 +2,9 @@ import json
 from itertools import chain, zip_longest
 
 import redis
+
 from es_tool import es_query
+from tool import get_data_by_mysql
 
 
 def merge_lists(*lol):
@@ -35,6 +37,14 @@ keywords = [
 ]
 
 
+def get_all_city_id():
+    sql = "select distinct tag_id from api_city"
+    mysql_results = get_data_by_mysql("172.16.30.141", 3306, "work", "BJQaT9VzDcuPBqkd", "zhengxing", sql)
+    city_tag_ids = [i["tag_id"] for i in mysql_results]
+    city_tag_ids.append(-1)  # 没有城市的情况
+    return city_tag_ids
+
+
 def get_es_diary(keywords, city_tag_id=-1, version=False):
     q = {}
     sort_list = [{
@@ -289,14 +299,24 @@ if __name__ == "__main__":
     # print("tractate: " + str(get_es_tractate(keywords)[:10]) + "\n")
     # print("answer: " + str(get_es_answer(keywords)[:10]) + "\n")
 
-    diary_list = get_es_diary(keywords)
+    redis_client = redis.StrictRedis.from_url("redis://:ReDis!GmTx*0aN9@172.16.40.173:6379")
+
     tractate_list = get_es_tractate(keywords)
     answer_list = get_es_answer(keywords)
-    redis_client = redis.StrictRedis.from_url("redis://:ReDis!GmTx*0aN9@172.16.40.173:6379")
-    redis_client.rpush("doris:tag_v3:coldstart:diary", *diary_list)
     redis_client.rpush("doris:tag_v3:coldstart:tractate", *tractate_list)
     redis_client.rpush("doris:tag_v3:coldstart:answer", *answer_list)
 
-    print("diary: " + str(len(diary_list)))
-    print("tractate: " + str(len(tractate_list)))
-    print("answer: " + str(len(answer_list)))
+    all_city_tag_id = get_all_city_id()
+    for city_id in all_city_tag_id:
+        if city_id != -1:
+            diary_list = get_es_diary(keywords, city_id)
+            redis_client.rpush("doris:tag_v3:coldstart:diary:" + str(city_id), *diary_list)
+
+    diary_list = get_es_diary(keywords)
+    redis_client.rpush("doris:tag_v3:coldstart:diary:-1", *diary_list)
+
+    # print("diary: " + str(len(diary_list)))
+    # print("tractate: " + str(len(tractate_list)))
+    # print("answer: " + str(len(answer_list)))
+
+    print("done")