feat: 新增 hostloc #43

This commit is contained in:
imsyy
2024-06-06 11:36:28 +08:00
parent 93d90eb653
commit d73ca1170c
7 changed files with 110 additions and 12 deletions

View File

@@ -46,6 +46,8 @@
| 稀土掘金 | 热榜 | juejin | 🟢 |
| 腾讯新闻 | 热点榜 | qq-news | 🟢 |
| 网易新闻 | 热点榜 | netease-news | 🟢 |
| 吾爱破解 | 榜单 | 52pojie | 🟢 |
| 全球主机交流 | 榜单 | hostloc | 🟢 |
| 虎嗅 | 24小时 | huxiu | 🟢 |
| 爱范儿 | 快讯 | ifanr | 🟢 |
| 英雄联盟 | 更新公告 | lol | 🟢 |
@@ -59,7 +61,7 @@
| 中央气象台 | 全国气象预警 | weatheralarm | 🟢 |
| 中国地震台 | 地震速报 | earthquake | 🟢 |
## 使用
## ⚙️ 使用
本项目支持 `Node.js` 调用,可在安装完成后调用 `serveHotApi` 来开启服务器

View File

@@ -48,6 +48,7 @@
"md5": "^2.3.0",
"node-cache": "^5.1.2",
"puppeteer": "^22.10.0",
"puppeteer-cluster": "^0.24.0",
"rss-parser": "^3.13.0",
"winston": "^3.13.0"
},

15
pnpm-lock.yaml generated
View File

@@ -38,6 +38,9 @@ importers:
puppeteer:
specifier: ^22.10.0
version: 22.10.0(typescript@5.4.5)
puppeteer-cluster:
specifier: ^0.24.0
version: 0.24.0(puppeteer@22.10.0(typescript@5.4.5))
rss-parser:
specifier: ^3.13.0
version: 3.13.0
@@ -1082,6 +1085,11 @@ packages:
resolution: {integrity: sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg==}
engines: {node: '>=6'}
puppeteer-cluster@0.24.0:
resolution: {integrity: sha512-zHPoNsrwkFLKFtgJJv2aC13EbMASQsE048uZd7CyikEXcl+sc1Nf6PMFb9kMoZI7/zMYxvuP658o2mw079ZZyQ==}
peerDependencies:
puppeteer: '>=22.0.0'
puppeteer-core@22.10.0:
resolution: {integrity: sha512-I54J4Vy4I07UHsgB1QSmuFoF7KNQjJWcvFBPhtY+ezMdBfwgGDr8dzYrJa11aPgP9kxIUHjhktcMmmfJkOAtTw==}
engines: {node: '>=18'}
@@ -2410,6 +2418,13 @@ snapshots:
punycode@2.3.1: {}
puppeteer-cluster@0.24.0(puppeteer@22.10.0(typescript@5.4.5)):
dependencies:
debug: 4.3.5
puppeteer: 22.10.0(typescript@5.4.5)
transitivePeerDependencies:
- supports-color
puppeteer-core@22.10.0:
dependencies:
'@puppeteer/browsers': 2.2.3

View File

@@ -43,7 +43,7 @@ export type RouterType = {
source_id: number;
pubdate: string;
};
"52pojie": {
discuz: {
title: string;
link: string;
guid: string;

View File

@@ -52,7 +52,7 @@ const getList = async (options: Options, noCache: boolean) => {
return {
fromCache: result.fromCache,
updateTime: result.updateTime,
data: list.map((v: RouterType["52pojie"]) => ({
data: list.map((v: RouterType["discuz"]) => ({
id: v.guid,
title: v.title,
desc: v.content,

66
src/routes/hostloc.ts Normal file
View File

@@ -0,0 +1,66 @@
import type { RouterData, ListContext, Options } from "../types.js";
import type { RouterType } from "../router.types.js";
import { web } from "../utils/getData.js";
import { extractRss, parseRSS } from "../utils/parseRSS.js";
import getTime from "../utils/getTime.js";
export const handleRoute = async (c: ListContext, noCache: boolean) => {
const type = c.req.query("type") || "hot";
const { fromCache, data, updateTime } = await getList({ type }, noCache);
const routeData: RouterData = {
name: "hostloc",
title: "全球主机交流",
type: "榜单",
parameData: {
type: {
name: "榜单分类",
type: {
hot: "最新热门",
digest: "最新精华",
new: "最新回复",
newthread: "最新发表",
},
},
},
link: "https://hostloc.com/",
total: data?.length || 0,
updateTime,
fromCache,
data,
};
return routeData;
};
const getList = async (options: Options, noCache: boolean) => {
const { type } = options;
const url = `https://hostloc.com/forum.php?mod=guide&view=${type}&rss=1`;
const result = await web({
url,
noCache,
userAgent:
"Mozilla/5.0 (Linux; Android 6.0; Nexus 5 Build/MRA58N) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Mobile Safari/537.36",
});
const parseData = async () => {
if (typeof result?.data === "string") {
const rssContent = extractRss(result.data);
return await parseRSS(rssContent);
} else {
return [];
}
};
const list = await parseData();
return {
fromCache: result.fromCache,
updateTime: result.updateTime,
data: list.map((v: RouterType["discuz"]) => ({
id: v.guid,
title: v.title,
desc: v.content,
author: v.author,
timestamp: getTime(v.pubDate),
hot: null,
url: v.link,
mobileUrl: v.link,
})),
};
};

View File

@@ -1,7 +1,7 @@
import type { Get, Post, Web } from "../types.ts";
import { config } from "../config.js";
import { getCache, setCache, delCache } from "./cache.js";
import puppeteer from "puppeteer";
import { Cluster } from "puppeteer-cluster";
import logger from "./logger.js";
import axios from "axios";
@@ -12,6 +12,27 @@ const request = axios.create({
withCredentials: true,
});
// puppeteer-cluster
export const createCluster = async () => {
return await Cluster.launch({
concurrency: Cluster.CONCURRENCY_BROWSER,
maxConcurrency: 5,
});
};
// Cluster
const cluster = await createCluster();
// Cluster configuration
cluster.task(async ({ page, data: { url, userAgent } }) => {
if (userAgent) {
await page.setUserAgent(userAgent);
}
await page.goto(url, { waitUntil: 'networkidle0' });
const pageContent = await page.content();
return pageContent;
});
// 请求拦截
request.interceptors.request.use(
(request) => {
@@ -117,14 +138,7 @@ export const web = async (options: Web) => {
}
// 缓存不存在时使用 Puppeteer 请求页面
logger.info("启动浏览器请求页面", { url });
const browser = await puppeteer.launch();
const page = await browser.newPage();
// 其他 Puppeteer 操作
if (userAgent) page.setUserAgent(userAgent);
await page.goto(url, { waitUntil: "networkidle0" });
// 获取页面内容
const pageContent = await page.evaluate(() => document.body.innerHTML);
await browser.close();
const pageContent = await cluster.execute({ url, userAgent });
// 存储新获取的数据到缓存
const updateTime = new Date().toISOString();
setCache(url, { data: pageContent, updateTime }, ttl);