获取淘宝/京东菜单,输出到txt文件

Maven依赖

<!--jsoup 是一款 Java 的HTML 解析器-->
<dependency>
    <groupId>org.jsoup</groupId>
    <artifactId>jsoup</artifactId>
    <version>1.11.3</version>
</dependency>

package com.sm.cn.test;

import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

import java.io.FileOutputStream;
import java.io.FileWriter;
import java.io.IOException;
import java.nio.charset.StandardCharsets;

/**
 * 下载商城菜单,输出到txt,淘宝为例子
 */
public class Test01 {
    public static void main(String[] args) throws IOException {
        f1();
        f2();
        f3();
    }

    /**
     * 淘宝
     * @throws IOException
     */
    public static void f1() throws IOException {
        StringBuffer stringBuffer = new StringBuffer();
        StringBuffer append = null;
        /*获取当前网页的document对象*/
        Document document = Jsoup.connect("https://www.taobao.com/").get();
        Elements elementsByClass = document.getElementsByClass("service-bd");
        for (Element byClass : elementsByClass) {
            Elements li = byClass.getElementsByTag("li");
            for (Element element : li) {
                Elements a = element.getElementsByTag("a");

                /*获取a标签中文件并换行输出*/
                //将可变字符串变为固定长度的字符串，方便下面的转码；
                append = stringBuffer.append(a.text() + "\r\n");
            }
        }
        String string = append.toString();
        /*使用字节流输出*/
        //因为中文可能会乱码，这里使用了转码，转成UTF-8；
        byte[] bytes = string.getBytes("utf-8");
        FileOutputStream out = new FileOutputStream("C://Users//Desktop//淘宝菜单目录.txt");
        out.write(bytes);
        out.close();
    }

    /**
     * 京东
     * @throws IOException
     */
    public static void f2() throws IOException {
        StringBuffer stringBuffer = new StringBuffer();
        StringBuffer append = null;
        /*获取当前网页的document对象*/
        Document document = Jsoup.connect("https://www.jd.com/").get();
        Elements menu = document.getElementsByClass("cate_menu");
        for (int i = 0; i < menu.size(); i++) {
            Element element = menu.get(i);

            Elements li = element.getElementsByTag("li");

            for (int j = 0; j < li.size(); j++) {

                Element element1 = li.get(j);
                Elements a = element1.getElementsByTag("a");
                
                /*获取a标签中文件并换行输出*/
                //将可变字符串变为固定长度的字符串，方便下面的转码；
                append = stringBuffer.append(a.text() + "\r\n");

            }
        }
        String string = append.toString();
        /*使用字符流输出*/
        FileWriter fileWriter = new FileWriter("C://Users//Desktop//京东商城目录.txt");
        fileWriter.write(string);
        fileWriter.close();
    }

    /**
     * 亚马逊
     * @throws IOException
     */
    public static void f3() throws IOException {
        StringBuffer stringBuffer = new StringBuffer();
        StringBuffer append = null;
        Document document = Jsoup.connect("https://www.amazon.cn/").get();
        Element searchDropdownBox = document.getElementById("searchDropdownBox");
        Elements option = searchDropdownBox.getElementsByTag("option");
        for (Element element : option) {
            append = stringBuffer.append(element.text() + "\r\n");
        }

        FileOutputStream stream = new FileOutputStream("C://Users//Desktop//亚马逊目录.txt");
        byte[] bytes = append.toString().getBytes(StandardCharsets.UTF_8);
        stream.write(bytes);
        stream.close();
    }
}