java的简单获取url

太过爱你忘了你带给我的痛 2022-09-12 00:59 159阅读 0赞

现在获取数据大多都是python，但是java页可以

package test;
    
    import org.jsoup.Jsoup;
    import org.jsoup.nodes.Document;
    import org.jsoup.nodes.Element;
    import org.jsoup.select.Elements;
    
    import java.io.IOException;
    import java.util.ArrayList;
    import java.util.List;
    
    /**
     * @ProjectName: spider01
     * @Package: test
     * @ClassName: LYGGongGongPiingTai
     * @Author:
     * @Description:
     * @Date: 2021/9/18 9:52 上午
     * @Version: 1.0
     */
    public class LYGGongGongPiingTai {
        public static String getHtml(String url){
            String html = null;
            try {
                html = Jsoup.connect(url)
                        .ignoreContentType(true).timeout(5000).execute().body();
            } catch (IOException e) {
    
            }
            return html;
        }
        public static List<String> getList(String html){
            List<String> list = new ArrayList<String>();
            Document document =Jsoup.parse(html);
            Elements elements=document.select("a");
            for (Element element:elements){
                if ("_blank".equals(element.attr("target"))) {
    //                System.out.println(element.attr("href"));
                    String url = "http://spzx.lyg.gov.cn/" + element.attr("href");
                    list.add(url);
                }
            }
            return list;
        }
        private static void geturl() {
            for (int i = 1;i<5;i++) {
                System.out.println("当前运行页数" + i );
                String url = "http://spzx.lyg.gov.cn/lygweb/jyxx/001007/001007002/" + i + ".html";
                String html = getHtml(url);
    //            System.out.println(html);
                System.out.println(getList(html));
            }
        }
    
        private static  List<String>  getlygwebList(String html) {
            List<String> list = new ArrayList<String>();
            Document document =Jsoup.parse(html);
            Elements elements=document.select("p");
            for (Element element:elements){
    //            System.out.println(element.text());
                list.add(element.text());
            }
            return list;
        }
    
        private static void getlygweb() {
            String url = "http://spzx.lyg.gov.cn/lygweb/jyxx/001007/001007002/20210908/100e7667-6711-4f2a-9b57-383de53915e9.html";
            url = "http://spzx.lyg.gov.cn/lygweb/jyxx/001007/001007002/20210819/7570f01b-b069-437c-b343-c10988cfe082.html";
            String html = getHtml(url);
    //        System.out.println(html);
            System.out.println(getlygwebList(html));
        }
    
    
    
        public static void main(String[] args) {
            //1 获取链接
            geturl();
            //2简单的抽取html
            getlygweb();
        }
    
    
    
    
    }