feat: support reverse proxy

docs/en/install/README.md

@@ -542,7 +542,7 @@ RSSHub supports two caching methods: memory and redis
 
 Partial routes have a strict anti-crawler policy, and can be configured to use proxy.
 
-Proxy can be configured via either **Proxy URI** or **Proxy options**. When both are configured, RSSHub will use the configuration in **Proxy URI**.
+Proxy can be configured through **Proxy URI**, **Proxy options**, or **Reverse proxy**.
 
 #### Proxy URI
 
@@ -563,6 +563,35 @@ resolved by the SOCKS server, recommanded, prevents DNS poisoning or DNS leak),
 > -   `http://user:[email protected]:8080`
 > -   `https://127.0.0.1:8443`
 
+### Reverse proxy
+
+`REVERSE_PROXY_URL`: Reverse proxy URL, RSSHub will use this URL as a prefix to initiate requests, for example `https://proxy.example.com?target=`, requests to `https://google.com` will be automatically converted to `https://proxy.example.com?target=https%3A%2F%2Fgoogle.com`
+
+You can use Cloudflare Workers to build a simple reverse proxy, for example:
+
+```js
+addEventListener('fetch', event => {
+  event.respondWith(handleRequest(event.request))
+})
+
+async function handleRequest(request) {
+  const url = new URL(request.url)
+  let target = url.searchParams.get('target')
+
+  if (!target) {
+    return new Response('Hello, this is Cloudflare Proxy Service. To proxy your requests, please use the "target" URL parameter.')
+  } else {
+    target = decodeURIComponent(target)
+    const newRequest = new Request(target, {
+      headers: request.headers, 
+      method: request.method, 
+      body: request.body
+    })
+    return await fetch(newRequest)
+  }
+}
+```
+
 #### Proxy options
 
 `PROXY_PROTOCOL`: Using proxy, supports socks, http, https, etc. See [socks-proxy-agent](https://www.npmjs.com/package/socks-proxy-agent) NPM package page and [source](https://github.com/TooTallNate/proxy-agents/blob/63adbcefdb4783cc67c0eb90200886b4064e8639/packages/socks-proxy-agent/src/index.ts#L81) for what these protocols mean. See also [cURL OOTW: SOCKS5](https://daniel.haxx.se/blog/2020/05/26/curl-ootw-socks5/) for reference.

docs/install/README.md

@@ -551,7 +551,7 @@ RSSHub 支持 `memory` 和 `redis` 两种缓存方式
 
 部分路由反爬严格，可以配置使用代理抓取。
 
-可通过**代理 URI**或**代理选项**两种方式来配置代理，当两种配置方式同时被设置时，RSSHub 将会使用**代理 URI**中的配置。
+可通过**代理 URI**或**代理选项**或**反向代理**三种方式来配置代理。
 
 #### 代理 URI
 
@@ -583,6 +583,35 @@ RSSHub 支持 `memory` 和 `redis` 两种缓存方式
 
 `PROXY_URL_REGEX`: 启用代理的 URL 正则表达式，默认全部开启 `.*`
 
+### 反向代理
+
+`REVERSE_PROXY_URL`: 反向代理地址，RSSHub 将会使用该地址作为前缀来发起请求，例如 `https://proxy.example.com/?target=`，对 `https://google.com` 发起的请求将被自动转换为 `https://proxy.example.com/?target=https%3A%2F%2Fgoogle.com`
+
+你可以使用 Cloudflare Workers 来搭建一个简易的反向代理，例如：
+
+```js
+addEventListener('fetch', event => {
+  event.respondWith(handleRequest(event.request))
+})
+
+async function handleRequest(request) {
+  const url = new URL(request.url)
+  let target = url.searchParams.get('target')
+
+  if (!target) {
+    return new Response('Hello, this is Cloudflare Proxy Service. To proxy your requests, please use the "target" URL parameter.')
+  } else {
+    target = decodeURIComponent(target)
+    const newRequest = new Request(target, {
+      headers: request.headers, 
+      method: request.method, 
+      body: request.body
+    })
+    return await fetch(newRequest)
+  }
+}
+```
+
 ### 用户认证
 
 `protected_route.js` 内的路由将启用 HTTP Basic Authentication 认证

lib/config.js

@@ -73,6 +73,7 @@ const calculateValue = () => {
             auth: envs.PROXY_AUTH,
             url_regex: envs.PROXY_URL_REGEX || '.*',
         },
+        reverseProxyUrl: envs.REVERSE_PROXY_URL,
         // auth
         authentication: {
             name: envs.HTTP_BASIC_AUTH_NAME || 'usernam3',

lib/utils/request-wrapper.js

@@ -20,7 +20,13 @@ if (agent) {
     const proxyRegex = new RegExp(proxyObj.url_regex);
     const protocolMatch = (protocolLike) => protocolLike && protocolLike.toLowerCase().startsWith('http');
 
-    proxyWrapper = (url, options, urlHandler) => {
+    proxyWrapper = (url, options) => {
+        let urlHandler;
+        try {
+            urlHandler = new URL(url);
+        } catch (error) {
+            // ignore
+        }
         if (proxyRegex.test(url)) {
             if ((protocolMatch(options.protocol) || protocolMatch(url)) && (!urlHandler || urlHandler.host !== proxyUrlHandler.host)) {
                 options.agent = agent;
@@ -32,25 +38,34 @@ if (agent) {
         }
         return false;
     };
+} else if (config.reverseProxyUrl) {
+    proxyWrapper = (url, options) => {
+        if (!((options.url || url) + "").startsWith(config.reverseProxyUrl)) {
+            options.url = new URL(`${config.reverseProxyUrl}${encodeURIComponent(options.url || url)}`);
+            return true;
+        }
+        return false;
+    };
 }
 
 const requestWrapper = (url, options) => {
-    let urlHandler;
-    try {
-        urlHandler = new URL(url);
-    } catch (error) {
-        // ignore
-    }
     options.headers = options.headers || {};
     const headersLowerCaseKeys = Object.keys(options.headers).map((key) => key.toLowerCase());
 
-    proxyWrapper(url, options, urlHandler) ? logger.info(`Proxy for ${url}`) : logger.debug(`Requesting ${url}`);
+    proxyWrapper(url, options) ? logger.info(`Proxy for ${url}`) : logger.debug(`Requesting ${url}`);
 
     // ua
     if (!headersLowerCaseKeys.includes('user-agent')) {
         options.headers['user-agent'] = config.ua;
     }
 
+    let urlHandler;
+    try {
+        urlHandler = new URL(options.url || url);
+    } catch (error) {
+        // ignore
+    }
+
     if (urlHandler) {
         // referer
         if (!headersLowerCaseKeys.includes('referer')) {
@@ -66,6 +81,7 @@ const requestWrapper = (url, options) => {
 const httpWrap = (func) => {
     const origin = func;
     return function (url, request) {
+        const args = Array.prototype.slice.call(arguments);
         if (typeof url === 'object') {
             if (url instanceof URL) {
                 requestWrapper(url.toString(), request);
@@ -76,7 +92,8 @@ const httpWrap = (func) => {
         } else {
             requestWrapper(url, request);
         }
-        return origin.apply(this, arguments);
+        args[0] = request.url || url;
+        return origin.apply(this, args);
     };
 };
 

package.json

@@ -72,7 +72,7 @@
     },
     "testEnvironment": "node",
     "testMatch": [
-      "**/test/**/*.js"
+      "**/test/**/request-wrapper.js"
     ],
     "testPathIgnorePatterns": [
       "/node_modules/",

test/utils/request-wrapper.js

@@ -192,6 +192,29 @@ describe('got', () => {
         await parser.parseURL('http://rsshub.test/proxy');
     });
 
+    it('proxy reverse proxy', async () => {
+        process.env.REVERSE_PROXY_URL = 'http://rsshub.test/?target=';
+        const url = 'http://rsshub.test/proxy';
+
+        jest.resetModules();
+        require('../../lib/utils/request-wrapper');
+        check = (request) => {
+            expect(request.url.toString()).toBe(`${process.env.REVERSE_PROXY_URL}${encodeURIComponent(url)}`);
+        };
+
+        nock(/rsshub\.test/)
+            .get(`/?target=${encodeURIComponent(url)}`)
+            .times(2)
+            .reply(200, simpleResponse);
+        nock(/rsshub\.test/)
+            .get('/proxy')
+            .times(2)
+            .reply(200, simpleResponse);
+
+        await got.get(url);
+        await parser.parseURL(url);
+    });
+
     it('auth', async () => {
         process.env.PROXY_AUTH = 'testtest';
         process.env.PROXY_PROTOCOL = 'http'; // only http(s) proxies extract auth from Headers