pachong/get-cate.js

const webdriver = require('selenium-webdriver');
const chrome = require('selenium-webdriver/chrome');
const firefox = require('selenium-webdriver/firefox');
const { saveFile } = require('./tools/save-file');
const cheerio = require('cheerio');



(async function example() {
    // let driver = await new Builder().forBrowser('chrome').build();
    let driver = new webdriver.Builder()
        .forBrowser('firefox')
        .build();

    driver.get('https://lumn.tmall.com/search.htm?spm=a1z10.3-b-s.w5002-15438998724.1.248c3b51L3CXmJ&search=y');


    // 检查登录
    await driver.wait(webdriver.until.elementLocated(webdriver.By.className('j_Username')), 100000, '超时未登录');
    console.log('登陆成功');

    await driver.sleep(10000)
    console.log('等待');

    // 存储原始窗口的 ID
    const originalWindow = await driver.getWindowHandle();

    await driver.sleep(1500)

    // 获取分类列表
    let cateArr = [];
    const cateIndex = 0;

    // 判断是否有验证框
    try {
        const dialog = await driver.findElement(webdriver.By.className('baxia-dialog'));
        if (dialog) {
            const ele = await driver.findElement(webdriver.By.js(() => document.querySelector('.baxia-dialog .baxia-dialog-close')))
            ele.click();
        }
        console.log('有验证框');
    } catch (error) {
    }

    async function eachCate(cateIndex) {


        const cateList = await driver.findElements(webdriver.By.js(() => document.querySelectorAll('.cateAttrs li a')));
        // 确定当前分类
        const curCate = cateList[cateIndex];
        const cateName = await curCate.getText();
        let cateObj = {
            name: cateName,
            subType: []
        }

        const url = await curCate.getAttribute('href');


        // 切换新窗口打开分类
        await driver.switchTo().newWindow('tab');
        const cateWindow = await driver.getWindowHandle();

        await driver.get(url);

        await driver.sleep(2000)


        // 判断是否有验证框
        try {
            const dialog = await driver.findElement(webdriver.By.className('baxia-dialog'));
            if (dialog) {
                const ele = await driver.findElement(webdriver.By.js(() => document.querySelector('.baxia-dialog .baxia-dialog-close')))
                ele.click();
            }
            console.log('有验证框');
        } catch (error) {
        }


        // 判断是否有二级分类
        const cate2List = await driver.findElements(webdriver.By.js(() => document.querySelectorAll('.cateAttrs li a')));

        await driver.sleep(500)

        let arr = [];
        if (cate2List.length) {
            await driver.getPageSource().then(async htmlStr => {
                const $ = cheerio.load(htmlStr);
                $('.cateAttrs li a').each((i, e) => {
                    arr.push({
                        name: $(e).text()
                    });
                });
            });
        }
        await driver.sleep(2000)

        cateObj.subType = arr;
        cateArr.push(cateObj)
        await driver.close();
        await driver.switchTo().window(originalWindow)
        cateIndex++;
        console.log('开始' + cateIndex, cateList.length, JSON.stringify(cateObj))
        if (cateIndex < cateList.length) {
            eachCate(cateIndex);
        } else {
            saveFile('/cate/cate.json', JSON.stringify(cateArr,null,4))
        }
    }

    eachCate(cateIndex);


})();