From 6f51667ca7ca9fc6ab43f6be239dcce05beaf05f Mon Sep 17 00:00:00 2001
From: 22earth <zhifengle@gmail.com>
Date: Sat, 15 Oct 2022 16:15:57 +0800
Subject: [PATCH] feat: filter regex

---
 Cargo.lock          |  1 +
 Cargo.toml          |  1 +
 README.md           |  7 ++++---
 src/rss_site/mod.rs | 24 +++++++++++++++++++++++-
 4 files changed, 29 insertions(+), 4 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 837f01b..4730882 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1154,6 +1154,7 @@ dependencies = [
  "gcookie",
  "log",
  "once_cell",
+ "regex",
  "reqwest",
  "rss",
  "rusqlite",
diff --git a/Cargo.toml b/Cargo.toml
index 10e4b25..72a7684 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -22,6 +22,7 @@ gcookie = "0.0.3"
 log = "0.4.0"
 env_logger = "0.9.0"
 once_cell = "1.15.0"
+regex = "1"
 
 [profile.release]
 lto = true
diff --git a/README.md b/README.md
index 7af7cce..8d350d6 100644
--- a/README.md
+++ b/README.md
@@ -16,7 +16,7 @@
 - [x] proxy 配置
   - ~~目前写死在 build_proxy_client 里面~~
   - 读取 ALL_PROXY 或者 HTTPS_PROXY 环境变量
-- [ ] 正则过滤 filter
+- [x] 正则过滤 filter
 - [ ] Windows 定时任务
   - ~~懒得写了，我是手动配置的~~
 - [x] 不同网站的并发任务
@@ -57,7 +57,7 @@ rss2pan -u "https://mikanani.me/RSS/Bangumi?bangumiId=2739&subgroupid=12"
   "mikanani.me": [
     {
       "name": "test",
-      "filter": "简体内嵌",
+      "filter": "/简体|1080p/",
       "url": "https://mikanani.me/RSS/Bangumi?bangumiId=2739&subgroupid=12"
     }
   ],
@@ -73,7 +73,8 @@ rss2pan -u "https://mikanani.me/RSS/Bangumi?bangumiId=2739&subgroupid=12"
 ```
 
 配置了 `filter` 后，标题包含该文字的会被离线。不设置 `filter` 默认离线全部
-> 正则功能还没写
+
+`/简体|\\d{3-4}[pP]/` 使用斜线包裹的正则规则。注意转义规则
 
 cid 是离线到指定的文件夹的 id 。
 获取方法: 浏览器打开 115 的文件，地址栏像 `https://115.com/?cid=2479224057885794455&offset=0&tab=&mode=wangpan`
diff --git a/src/rss_site/mod.rs b/src/rss_site/mod.rs
index 99557ca..60fc79a 100644
--- a/src/rss_site/mod.rs
+++ b/src/rss_site/mod.rs
@@ -2,6 +2,7 @@ mod dmhy;
 mod mikanani;
 mod nyaa;
 
+use regex::Regex;
 use reqwest::Method;
 use rss::{Channel, Item};
 use std::io::BufReader;
@@ -75,7 +76,17 @@ pub async fn get_magnetitem_list(config: &RssConfig) -> Vec<MagnetItem> {
             let m = site.get_magnet_item(item);
             let mut flag = true;
             if let Some(pat) = &config.filter {
-                flag = m.title.contains(pat);
+                if pat.starts_with("/") && pat.ends_with("/") {
+                    let re = Regex::new(&pat[1..pat.len() - 1]);
+                    match re {
+                        Ok(re) => {
+                            flag = re.is_match(&m.title);
+                        }
+                        Err(_) => {}
+                    }
+                } else {
+                    flag = m.title.contains(pat);
+                }
             }
             if flag {
                 item_list.push(m)
@@ -116,4 +127,15 @@ mod tests {
         let res = service.save_items(&items, true);
         assert!(res.is_ok());
     }
+    #[test]
+    fn test_re() {
+        let str_list = [
+            "[7月新番][传颂之物 二人的白皇][Utawarerumono - Futari no Hakuoro][09][1080P][MP4][GB][简中] [241.72 MB]",
+            "【幻樱字幕组】【7月新番】【传颂之物 二人白皇 Utawarerumono-Futari no Hakuoro-】【16】【BIG5_MP4】【1920X1080】 [321.13 MB]",
+            "[动漫国字幕组&澄空学园&LoliHouse] 传颂之物 二人的白皇 / Utawarerumono Futari no Hakuoro - 16 [WebRip 1080p HEVC-10bit AAC][简繁外挂字幕] [485.4 MB]"
+        ];
+        let pat = "/澄空学园|幻樱|\\d{4}[p]/";
+        let re = Regex::new(&pat[1..pat.len() - 1]).unwrap();
+        assert_eq!(str_list.map(|s| re.is_match(s)), [true, true, true]);
+    }
 }