qry-python-js/routes/index.js

const express = require("express");
const router = express.Router();

const {jsdomFromText, browser} = require("sdenv");
const {Script} = require("vm");
const fs = require("node:fs");
const path = require("node:path");
const crypto = require("crypto")


let sleep = ms => new Promise(resolve => setTimeout(resolve, ms));


const storeCookie = new Map;

const areaJsMap = {
    "yunnan": path.resolve(__dirname, '../public/yunnan/MO5zzCMcub4d.b4c45da.js'),
    "hubei": path.resolve(__dirname, '../public/hubei/5PXGXoOF7eGJ.ed63b8f.js'),
}


router.post('/rsCookie', async (req, res) => {
    console.time('rsCookie')
    let url = req.body['url'];
    let htmlStr = req.body['htmlStr'];
    let jsStr = req.body['jsText'];
    let areaName = req.body['areaName'];
    let cookie = req.body['cookie'];
    let userAgent = req.body['userAgent'];

    if (url == null || url === '') {
        return res.send('error url')
    }
    if (htmlStr == null || htmlStr === '') {
        return res.send('error html')
    }

    try {
        let jsText;
        if (jsStr == null) {
            let jsPath = areaJsMap[areaName];
            if (jsPath == null) {
                console.error('未找到js文件')
                return res.send('未找到js文件')
            }
            jsText = fs.readFileSync(jsPath).toString('utf8');
        } else {
            jsText = Buffer.from(jsStr, 'base64').toString('utf-8')
        }


        let cookies = await handle(url,
            Buffer.from(htmlStr, 'base64').toString('utf-8'),
            jsText,
            cookie != null ? Buffer.from(cookie, 'base64').toString('utf-8') : null,
            userAgent != null ? Buffer.from(userAgent, 'base64').toString('utf-8') : null)
        res.send(cookies);
    } catch (e) {
        console.error(e)
        return res.send(e.toString())
    } finally {
        console.timeEnd('rsCookie')
    }

})

async function handle(url, htmlStr, jsText, cookie, userAgent) {
    let uuid = crypto.randomUUID()

    // 获取 origin
    let baseUrl = new URL(url).origin;

    const [jsDom, cookieJar] = await jsdomFromText({
        url: url,
        referrer: url,
        userAgent: userAgent,
        contentType: "text/html",
        runScripts: "outside-only", // runScripts: 'dangerously'/'outside-only'
    })
    // 设置 cookie
    if (cookie != null) {
        cookieJar.setCookieSync(cookie, baseUrl);
    }
    // 加载dom
    const dom = await jsDom(htmlStr);

    window = dom.window
    // 加载 页面上的js
    const allScript = window.document.querySelectorAll('script[r="m"]');

    // let $ = cheerio.load(htmlStr);
    // let nextAll = $('script[r=m]');

    let js = '';
    allScript.forEach(script=>{
        let attr = script.textContent;
        if (attr) {
            js += attr
        } else {
            js += jsText
        }
        js += ";\n"
    })
    // allScript.each((script) => {
    //
    // })

    // 执行完成后的钩子
    window.onbeforeunload = async (url) => {
        const cookies = cookieJar.getCookieStringSync(baseUrl);
        // console.debug(`${url} 生成cookie：`, cookies);
        storeCookie.set(uuid, cookies)
        // window.close();
        return cookies;
    }

    browser(dom.window, 'chrome');

    let script = new Script(js);
    let internalVMContext = dom.getInternalVMContext();
    // 执行 js
    script.runInContext(internalVMContext);

    for (let i = 0; i < 10; i++) {
        await sleep(100)
        let hasCookie = storeCookie.has(uuid)
        if (hasCookie) {
            let newVar = storeCookie.get(uuid);
            storeCookie.delete(uuid)
            return newVar;
        }
    }
    throw new Error('执行超时')
}


module.exports = router