Jsoup爬取JD工具类

导入所需依赖：

  <dependency>
      <groupId>org.jsoup</groupId>
      <artifactId>jsoup</artifactId>
      <version>1.10.2</version>
    </dependency>

 

pojo层：

package com.lzyuan.pojo;

/**
 * Encoding:utf-8
 * Author: liuzheyuan
 * Date:  2020-09-24 16:58
 * Version:1.0
 * Description：
 */



public class Content {
    private String title;//描述
    private String img;//图片
    private String price;//价格
    private String scroll;//缩略图
    private String stock;//发货地

    public Content() {
    }

    public Content(String title, String img, String price, String scroll, String stock) {
        this.title = title;
        this.img = img;
        this.price = price;
        this.scroll = scroll;
        this.stock = stock;
    }

    public String getTitle() {
        return title;
    }

    public void setTitle(String title) {
        this.title = title;
    }

    public String getImg() {
        return img;
    }

    public void setImg(String img) {
        this.img = img;
    }

    public String getPrice() {
        return price;
    }

    public void setPrice(String price) {
        this.price = price;
    }

    public String getScroll() {
        return scroll;
    }

    public void setScroll(String scroll) {
        this.scroll = scroll;
    }

    public String getStock() {
        return stock;
    }

    public void setStock(String stock) {
        this.stock = stock;
    }

    @Override
    public String toString() {
        return "Content{" +
                "title='" + title + '\'' +
                ", img='" + img + '\'' +
                ", price='" + price + '\'' +
                ", scroll='" + scroll + '\'' +
                ", stock='" + stock + '\'' +
                '}';
    }
}

 

utile包：

package com.lzyuan.util;

import com.lzyuan.pojo.Content;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.InputStream;
import java.net.MalformedURLException;
import java.net.URL;
import java.net.URLConnection;
import java.util.ArrayList;
import java.util.List;

/**
 * Encoding:utf-8
 * Author: liuzheyuan
 * Date:  2020-09-23 11:30
 * Version:1.0
 * Description：
 */

public class GetPic {
    /**
     * @author liuzheyuan
     * Description 主方法 
     */
    public static void main(String[] args) throws Exception {
       List<Content> java = new GetPic().parmJD("apple");
        for (Content content : java) {
            System.out.println(content);
        }

    }
    public List<Content> parmJD(String keyword) throws Exception {

        String url="https://search.jd.com/Search?keyword="+keyword;
        Document document = Jsoup.parse(new URL(url), 3000);//3000毫秒响应
        //js代码
        Element element = document.getElementById("J_goodsList");//最大的div
        //System.out.println(list);
        //拿到所有的li
        Elements lis = element.getElementsByTag("li");
        ArrayList<Content> contentList = new ArrayList<>();
        // System.out.println(lis);
        //继续拿里面信息
        for (Element li : lis) {
            //System.out.println(li);
            //source-data-lazy-img
            String img = li.getElementsByTag("img").eq(0).attr("data-lazy-img");//图
            String price = li.getElementsByClass("p-price").eq(0).text();
            String title = li.getElementsByClass("p-name").eq(0).text();

            Content content = new Content();
            content.setTitle(title);
            content.setImg(img);
            content.setPrice(price);
            contentList.add(content);

        }
        return contentList;
    }