a small open source server build for you to simulate the card game card drawing, card and card pool could be customised, could share your settings to community
一个小型的抽卡模拟器服务端，可自定义卡牌和卡池设置以及分享你的设置到社区
this server is under developing right now, when it could be use, you can see there is a release version of this project(because I could only code this project when I have time, so the developing may be pretty slow)
开发中... 可以使用时将会发布release版本（因为是抽空写的项目，所以可能开发速度不理想）
- root
- migrations
- src
- entity
- gql
- some other code
- tantivy
- build.rs
- sqlx-data.json
the basic layout will be like this:
migrations for sqlx-cli to migrate databases(sqlx is the database driver I choose, the orm is the sea-orm), by the way I using postgresql for database
entity is the auto-generated entitys by sea-orm-cli
gql for your graphql types(more details go to async-graphql)
tantivy is not a folder for code, it is a data path for tantivy text search engine(you could not see this because it is ignore by git, but you need this folder to make this project compiled)
Current project using PGroonga for postgresql text search so no more tantivy no more cangjie
build.rs trigger recompilation when a new migration is added
sqlx-data.json for driver settings
项目基本结构：
migrations 是sqlx命令行工具的检索路径，用于存放迁移文件的sql代码，在下使用的是sea-orm作为持久层框架（这框架使用sqlx作为数据库驱动，故而sqlx的命令行工具也可以使用），顺带一提在下使用的是postgresql
entity文件夹是用sea-orm的命令行工具自动生成（本质上的操作就是你用sqlx-cli跑迁移，然后用sea-orm-cli生成entity） gql是graphql类型定义相关的代码（去async-graphql的github页面了解更多）
tantivy是用以存放同名的文本搜索引擎的数据的文件夹（你可能看不见这个文件夹，因为不是代码故而加入了.gitignore，但你需要该文件夹才能正常编译此项目）
项目现使用PGroonga作为postgresql文本搜索分词，所以不再需要tantivy和仓颉了
build.rs是sqlx提供的一个功能，这个不是很重要（因为实际上用的是sea-orm），但还是说一下，这个是当你添加了新的迁移的时候会触发项目重新编译
sqlx-data.json是sqlx的配置文件之一（.env也可以）
as you can see at the project layout part, I use sqlx and sea-orm for postgresql and async-graphql for api, let's talk about some other dependencies(in nowadays, people works for web backend may all know the api and databases things, this why I skip those).
*tantivy, a text search engine like Apache Lucene, this is used for search our card and card pool settings by keyword which is a text. because I am a chinese person, so I may also need cang-jie for chinese characters, not using es(elastic search) because it takes too much memory, I know there is something like es written in rust, but it just not ready for production right now(it's server and types crate is not on crate.io, wtf!) ps: this situation is at 2022-1-24(if it fix this things in the future, and I still not finish the search feature upon tantivy, I may turn to it).
no more tantivy and cangjie needed
but you should install PGroonga instead*
http-server is warp(I choose it is not because of actix-web's unsafe bullshits, just for the reason that I used to using warp in my work before, and in that time, the actix-web could not be found in github, rocket.rs is not support async runtime), warp is not having docs but the docs.rs provided(it is kind of a little annoying, when you see a struct and it's functions not having any descriptions at all). but becasue I using graphql as the api, that the web server it self is not that matter now. [2022-1-29 update] because subscription need server side cache for sending the right message to the right one, I add redis for this support, don't worry, in 2022, using redis to implement thing like this is super easy
如各位在上一部分所见，数据库是postgres，使用了sqlx+sea-orm的组合（由于现在的web后端基本都是crud boy，所以这方面应该无需赘述）
直接进入正题，谈谈其他的依赖。
tantivy，文本搜索引擎，和Apache Lucene很像，在下选择该引擎作为卡池配置和卡牌设置的文本搜索功能，当然由于中文的问题需要一个额外的中文词库，这里在下选择的是cang-jie，不选择成品的es是因为java写的太吃内存了，虽然有一个基于tantivy的类似es的库，但这玩意儿作为成品来使用还太早（时间为2022-1-24，之后如果它起飞了再说，而且还得于在下研究明白tantivy之前起飞），需要git下来编译安装，crates.io上完全搜不到它的几个库，也就是说，你要作为客户端连接本地部署的这玩意儿要么就是当一个rest api来用，要么就手动复制它的代码过来通过path来在cargo中依赖（cpp直呼内行）。
服务器选择的是warp（不选actix-web不是在下有unsafe洁癖，而是在下之前在工作中用过warp，相对比较熟悉，而且那时在做选型的时候正好是unsafe节奏风暴刷的最多的时候，github上搜不到actix-web，那时的rocket.rs还不支持异步运行时，实属无奈之举），warp的文档可谓是惜字如金（有些结构体和成员函数甚至是原文放送），但是由于是用的graphql，倒是也不用太多的去关注服务器本身（只需要考虑服务器性能就行了，warp的性能还行）
不再使用tantivy和仓颉，但需要手动安装PGroonga
国内对graphql的使用好像不多，这里稍微说一下，可能理解成一种特殊的rest服务，请求永远是以特定的方式进行的，而请求的类型分为三种query，mutation和subscription，顾名思义，就是请求，突变和关注，第一种获取数据，第二种操作数据，第三种监听事件（知道个大概就行了） [2022-1-29 更新] 为subscription的正常使用，需要服务端缓存来保证发消息给指定的用户，为此在下使用了redis来缓存这些相关数据，不需要担心复杂性，都2202年了，不会还有人不会用redis也找不到好用的库吧（不会吧不会吧）
- users management(用户管理)
- normal register and login
- jsonwebtoken response when login succeed
- token from header
- search user by id
- friends//this is an optional feature may not show up in early versions
- send friend invite request
- friend lists
- custom friend list
- chat with friend
- card setting and card pool setting（卡牌和卡池设置）
- level setting like how rare this card is（等级设置，或者说稀有度设置）
- custom different way to descript how rare this will be
- query settings created by current user
- query one page at a time
- query all page at a time
- card descriptions（卡牌描述）
- text fields --like name and description（文本类，像名字和简介之类的）
- mark down support for description of the card
- logo of the card
- card pool settings
- probability
- algorithm
- depth how many times you need to reach the big price（深度，也可以说是保底）
- level setting like how rare this card is（等级设置，或者说稀有度设置）
- history for card draw（抽卡记录）
- community for sharing（社区功能，主要用于分享配置）
- shared with picture（用图片分享）
- shared using article（使用社区文章分享）
- show community articles（查看社区文章）
- articles without sharing things（正常的文章）
- recomments to articles（文章评论和回复）
- thumbs and sort for articles（点赞，并为社区文章提供相关排序）
- import and export card pool settings（卡池设置的导入导出）
- save card pool settings to json
- save card descriptions to csv
- load card pool settings from json
- load card descriptions from csv
- import card pool settings from sharing link
- card description and card pool setting sharing text search（可供搜索的卡牌描述和卡池分享）
- search keyword in community articles to find the card pool settings you want
- search keyword in card descriptions to find the card you want
- classify cards by level setting or some other thing that is not settled yet
- statistics for card draw historys（抽卡记录统计分析）