Database performance optimization for large quantities

- Add two indices for avoiding possible slow queries

> This is an experimental handle for the database which has some tables over one hundred thousand lines.

latest version/database/database_initialize.py

@@ -4,7 +4,7 @@ import json
 
 # 数据库初始化文件，删掉arcaea_database.db文件后运行即可，谨慎使用
 
-ARCAEA_SERVER_VERSION = 'v2.9.1.1'
+ARCAEA_SERVER_VERSION = 'v2.9.1.3'
 
 
 def main(path='./'):
@@ -208,9 +208,9 @@ def main(path='./'):
     primary key(user_id, song_id, file_name)
     );''')
     c.execute('''create table if not exists user_download(user_id int,
-    token text,
     time int,
-    primary key(user_id, token, time)
+    token text,
+    primary key(user_id, time, token)
     );''')
     c.execute('''create table if not exists item(item_id text,
     type text,
@@ -337,6 +337,11 @@ def main(path='./'):
     primary key(course_id, item_id, type)
     );''')
 
+    c.execute(
+        '''create index best_score_1 on best_score (song_id, difficulty);''')  # 排名查询优化
+    c.execute(
+        '''create index download_token_1 on download_token (song_id, file_name);''')  # 下载token判断优化
+
     # 搭档初始化
     char = ['hikari', 'tairitsu', 'kou', 'sapphire', 'lethe', 'hikari&tairitsu(reunion)', 'Tairitsu(Axium)', 'Tairitsu(Grievous Lady)', 'stella', 'Hikari & Fisica', 'ilith', 'eto', 'luna', 'shirabe', 'Hikari(Zero)', 'Hikari(Fracture)', 'Hikari(Summer)', 'Tairitsu(Summer)', 'Tairitsu & Trin',
             'ayu', 'Eto & Luna', 'yume', 'Seine & Hikari', 'saya', 'Tairitsu & Chuni Penguin', 'Chuni Penguin', 'haruna', 'nono', 'MTA-XXX', 'MDA-21', 'kanae', 'Hikari(Fantasia)', 'Tairitsu(Sonata)', 'sia', 'DORO*C', 'Tairitsu(Tempest)', 'brillante', 'Ilith(Summer)', 'etude', 'Alice & Tenniel', 'Luna & Mia', 'areus', 'seele', 'isabelle', 'mir', 'lagrange', 'linka', 'nami', 'Saya & Elizabeth', 'lily', 'kanae(midsummer)', 'alice&tenniel(minuet)', 'tairitsu(elegy)', 'marija', 'vita', 'hikari(fatalis)']