refactor: use database

.gitignore

@@ -1,5 +1,3 @@
-data.json
-
 # Generated by Cargo
 # will have compiled files and executables
 /target/
@@ -15,3 +13,6 @@ Cargo.lock
 # Added by cargo
 
 /target
+
+# Database
+db.sqlite

Cargo.toml

@@ -2,11 +2,10 @@
 name = "plagiarism-detector-rust"
 version = "1.0.0"
 edition = "2021"
+publish = false
 
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+[workspace]
+members = [".", "api", "service", "entity", "migration"]
 
 [dependencies]
-jieba-rs = "0.6.7"
-rocket = { version = "0.5.0-rc.3", features = ["json"] }
-serde = "1.0.157"
-serde_json = "1.0.94"
+plagiarism-detector-rust-api = { path = "api" }

README.md

@@ -1,8 +1,7 @@
 # Rust 简易论文查重
 
-![](https://img.shields.io/github/languages/code-size/SMS-COSMO/Plagiarism-Detector-Rust?color=yellow&style=flat-square)
-
 ## 部署
+
 修改配置见 `Rocket.toml`。
 
 1. 直接下载
@@ -20,40 +19,46 @@
 cargo run --release
 ```
 
-端口: `9999`
-
 **测试：**
 
 ```sh
 cargo run
 ```
 
-端口: `8000`
-
 ## API
 
-- `POST /check`
-- `POST /add`
+### 请求：
 
-`check` 为检查论文相似度
+- `POST /`
 
-`add` 为 `check` 之后将论文加入数据库 **!!暂不支持修改!!**
+| 名称    | 类型   | 内容                                                                               |
+| :------ | :----- | :--------------------------------------------------------------------------------- |
+| `id`    | String | 论文 id，不允许重复                                                                |
+| `text`  | String | 文本内容                                                                           |
+| `write` | bool   | `false`：检查论文相似度 <br> `true`：检查之后将论文加入数据库 **!!暂不支持修改!!** |
 
-**请求:**
 
 ```json
 {
     "id": "1",
-    "text": "基于 jieba 分词，tf-idf 算法求相似度   ！（*@1=-+!9）"
+    "text": "基于 jieba 分词，tf-idf 算法求相似度   ！（*@1=-+!9）",
+    "write": true
 }
 ```
 
-其中 id 应不与已有论文重复。
+### 响应：
+
+| 名称         | 类型   | 内容                                                                   |
+| :----------- | :----- | :--------------------------------------------------------------------- |
+| `code`       | i32    | 响应代码                                                               |
+| `msg`        | String | 返回信息                                                               |
+| `similarity` | []     | 表示前 5 个相似度最高的论文，每项第一个值表示相似度，第二个值为论文 id |
 
-**响应:**
 
 ```json
 {
+    "code": 200,
+    "msg": "加入成功",
     "similarity": [
         [
             0.97821377847197237,
@@ -79,4 +84,9 @@ cargo run
 }
 ```
 
-表示前 5 个相似度最高的论文，每项第一个值表示相似度，第二个值为论文 id。
+## 技术栈
+
+- [Rocket](https://rocket.rs/) - A web framework for Rust that makes it simple to write fast, type-safe, secure web applications with incredible usability, productivity and performance
+- [SeaORM](https://www.sea-ql.org/SeaORM/) - 🐚 SeaORM is a relational ORM to help you build web services in Rust
+- [jieba-rs](https://github.com/messense/jieba-rs) - The Jieba Chinese Word Segmentation Implemented in Rust
+

Rocket.toml

@@ -1,13 +1,6 @@
 [default]
 address = "0.0.0.0"
-# no file lock, worker has to be 1
-workers = 1
 
-[debug]
-port = 8000
-limits = { json = "50 MiB" }
-
-[release]
-port = 9999
-ip_header = false
-limits = { json = "10 MiB" }
+[default.databases.sea_orm]
+# sqlite
+url = "sqlite://db.sqlite?mode=rwc"

api/Cargo.toml

@@ -0,0 +1,23 @@
+[package]
+name = "plagiarism-detector-rust-api"
+version = "0.1.0"
+edition = "2021"
+publish = false
+
+[dependencies]
+async-stream = { version = "0.3" }
+async-trait = { version = "0.1" }
+plagiarism-detector-rust-service = { path = "../service" }
+futures = { version = "0.3" }
+futures-util = { version = "0.3" }
+rocket = { version = "0.5.0-rc.4", features = ["json"] }
+rocket_dyn_templates = { version = "0.1.0-rc.1", features = ["tera"] }
+serde_json = { version = "1" }
+entity = { path = "../entity" }
+migration = { path = "../migration" }
+tokio = "1.29.0"
+jieba-rs = "0.6"
+
+[dependencies.sea-orm-rocket]
+# path = "../../../sea-orm-rocket/lib" # remove this line in your own project and uncomment the following line
+version = "0.5.3"

api/src/data.rs

@@ -0,0 +1,9 @@
+/// Get stopword list from stopwords.txt
+pub fn stop_words() -> std::collections::HashSet<String> {
+    let mut words = std::collections::HashSet::new();
+    for item in String::from_utf8_lossy(include_bytes!("../../stopwords.txt")).split_whitespace() {
+        words.insert(item.to_string());
+    }
+
+    words
+}

api/src/lib.rs

@@ -0,0 +1,72 @@
+extern crate rocket;
+
+use rocket::fairing::{self, AdHoc};
+use rocket::serde::json::Json;
+use rocket::{Build, Rocket};
+
+use migration::MigratorTrait;
+use sea_orm_rocket::{Connection, Database};
+
+mod pool;
+use ::entity::req::*;
+use pool::Db;
+
+mod data;
+mod process;
+
+pub use entity::paper;
+pub use entity::paper::Entity as Post;
+
+#[rocket::post("/", format = "json", data = "<data>")]
+async fn check(
+    data: Json<ReqData>,
+    state: &rocket::State<SharedData>,
+    conn: Connection<'_, Db>,
+) -> Json<ResData> {
+    let req: ReqData = data.into_inner();
+    let db = conn.into_inner();
+
+    match process::similarity(&req, req.write, db, &state.jieba, &state.stop_words).await {
+        Ok(r) => Json(ResData {
+            code: 200,
+            msg: format!("{}成功", if req.write { "加入" } else { "查询" }),
+            similarity: r,
+        }),
+        Err(e) => Json(ResData {
+            code: 500,
+            msg: e.to_string(),
+            similarity: [].to_vec(),
+        }),
+    }
+}
+
+async fn run_migrations(rocket: Rocket<Build>) -> fairing::Result {
+    let conn = &Db::fetch(&rocket).unwrap().conn;
+    let _ = migration::Migrator::up(conn, None).await;
+    Ok(rocket)
+}
+
+#[tokio::main]
+async fn start() -> Result<(), rocket::Error> {
+    rocket::build()
+        .attach(Db::init())
+        .attach(AdHoc::try_on_ignite("Migrations", run_migrations))
+        .mount("/", rocket::routes![check])
+        .manage(SharedData {
+            jieba: jieba_rs::Jieba::new(),
+            stop_words: data::stop_words(),
+        })
+        .launch()
+        .await
+        .map(|_| ())
+}
+
+pub fn main() {
+    let result = start();
+
+    println!("Rocket: deorbit.");
+
+    if let Some(err) = result.err() {
+        println!("Error: {err}");
+    }
+}

api/src/pool.rs

@@ -0,0 +1,42 @@
+use plagiarism_detector_rust_service::sea_orm;
+
+use async_trait::async_trait;
+use sea_orm::ConnectOptions;
+use sea_orm_rocket::{rocket::figment::Figment, Config, Database};
+use std::time::Duration;
+
+#[derive(Database, Debug)]
+#[database("sea_orm")]
+pub struct Db(SeaOrmPool);
+
+#[derive(Debug, Clone)]
+pub struct SeaOrmPool {
+    pub conn: sea_orm::DatabaseConnection,
+}
+
+#[async_trait]
+impl sea_orm_rocket::Pool for SeaOrmPool {
+    type Error = sea_orm::DbErr;
+
+    type Connection = sea_orm::DatabaseConnection;
+
+    async fn init(figment: &Figment) -> Result<Self, Self::Error> {
+        let config = figment.extract::<Config>().unwrap();
+        let mut options: ConnectOptions = config.url.into();
+        options
+            .max_connections(config.max_connections as u32)
+            .min_connections(config.min_connections.unwrap_or_default())
+            .connect_timeout(Duration::from_secs(config.connect_timeout))
+            .sqlx_logging(false);
+        if let Some(idle_timeout) = config.idle_timeout {
+            options.idle_timeout(Duration::from_secs(idle_timeout));
+        }
+        let conn = sea_orm::Database::connect(options).await?;
+
+        Ok(SeaOrmPool { conn })
+    }
+
+    fn borrow(&self) -> &Self::Connection {
+        &self.conn
+    }
+}