java登录央行征信网站

 package com.entrym.crawler.test;

 import java.util.HashMap;
 import java.util.Map;

 import org.apache.commons.lang.StringUtils;
 import org.apache.http.client.methods.HttpGet;
 import org.apache.http.client.methods.HttpPost;
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Document;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;

 import com.entrym.crawler.constans.CommonConstants;
 import com.entrym.crawler.util.CommonUtil;
 import com.entrym.crawler.util.PowerHttpClient;
 import com.entrym.crawler.util.verifyCode.Captcha;

 /**
  * 央行征信报告获取爬虫
  * @ClassName: PbccrcCrawler
  * @Description: TODO(这里用一句话描述这个类的作用)
  *
  */
 public class PbccrcCrawler
 {

     private static Logger logger = LoggerFactory.getLogger(PbccrcCrawler.class);
     private static String loginUrl = "https://ipcrs.pbccrc.org.cn/page/login/loginreg.jsp";
     private static String loginPostUrl = "https://ipcrs.pbccrc.org.cn/login.do";
     private static String welcomeUrl = "https://ipcrs.pbccrc.org.cn/welcome.do";
     private static String reportUrl = "https://ipcrs.pbccrc.org.cn/reportAction.do";
     private static String summaryReportUrl = "https://ipcrs.pbccrc.org.cn/summaryReport.do";
     private static String simpleReportUrl = "https://ipcrs.pbccrc.org.cn/simpleReport.do";

     private static String host = "ipcrs.pbccrc.org.cn";
     private static String userAgent = "Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko";

     private PowerHttpClient powerHttpClient;
     private boolean isLogin=false;//登录状态
     private boolean report0=false;//个人信用信息提示
     private boolean report1=false;//个人信用信息概要
     private boolean report2=false;//个人信用报告

     private String task_id=null;

     /**
      * 登录
      * @Title: login
      * @Description: TODO(这里用一句话描述这个方法的作用)
      * @param @param loginname
      * @param @param password    参数
      * @return void    返回类型
      * @throws
      */
     public  String login(String loginname ,String password)
     {
         try
         {
             powerHttpClient = new PowerHttpClient();
             //打开首页
             String html = powerHttpClient.getToString(loginUrl);
             logger.debug("----首页---",html);

             Document doc = Jsoup.parse(html);
             String token = doc.select("input[name=org.apache.struts.taglib.html.TOKEN]").val();
             String method = doc.select("input[name=method]").val();
             String date = doc.select("input[name=date]").val();
             String imgUrl = doc.select("img[id=imgrc]").attr("src");
             String imgData = "";
             if (StringUtils.isNotBlank(imgUrl))
             {
                 //处理验证码
                 imgUrl = "https://ipcrs.pbccrc.org.cn"+imgUrl;
                 logger.info(task_id+"Captcha imgUrl = {}",imgUrl);
                 HttpGet httpGet = powerHttpClient.getHttpGet(imgUrl);
                 httpGet.addHeader("Host", host);
                 httpGet.addHeader("User-Agent", userAgent);
                 httpGet.addHeader("Referer", loginUrl);

                 byte[] imgByte = powerHttpClient.getToByteArray(httpGet);
                 Captcha captcha = new Captcha("img");
                 CommonUtil.writeByteArrayToFile(imgByte, CommonConstants.RUNTIME_TEMPIMG_FOLDER+captcha.getFilePath());
                 System.out.println(CommonConstants.RUNTIME_TEMPIMG_FOLDER+captcha.getFilePath());
                 imgData = CommonUtil.consoleScanner();
             }

             //提交登录
             Map<String, String> map = new HashMap<>();
             map.put("org.apache.struts.taglib.html.TOKEN", token);
             map.put("method", method);
             map.put("date", date);
             map.put("_@IMGRC@_", imgData);
             map.put("loginname", loginname);
             map.put("password", password);

             HttpPost httpPost = powerHttpClient.getHttpPost(loginPostUrl);
             httpPost.addHeader("Host", host);
             httpPost.addHeader("User-Agent", userAgent);
             httpPost.addHeader("Referer", loginUrl);
             String html1 = powerHttpClient.postWithMap(map,httpPost);
             logger.debug("----提交登录结果---",html1);

             HttpGet httpGet1 = powerHttpClient.getHttpGet(welcomeUrl);
             httpGet1.addHeader("Host", host);
             httpGet1.addHeader("User-Agent", userAgent);
             httpGet1.addHeader("Referer", loginPostUrl);
             String html2 = powerHttpClient.getToString(httpGet1, "");
             logger.info("----欢迎页面---{}",html2);
             parseLogin(html2);
             return html2;
         } catch (Exception e) {
             logger.error(task_id+"---登录出现异常：{}",CommonUtil.getExceptionTrace(e));
         }
         return null;
     }

     /**
      * 解析登录和报告状态
      * @Title: parseLogin
      * @Description: TODO(这里用一句话描述这个方法的作用)
      * @param @param html    参数
      * @return void    返回类型
      * @throws
      */
     private void parseLogin(String html)
     {
         if (StringUtils.isBlank(html)) {
             return;
         }
         if (html.contains("欢迎登录个人信用信息服务平台")) {
             isLogin = true;
             logger.info(task_id+"----恭喜，登录成功---");
         }

     }

     public static void main(String[] args) {
         PbccrcCrawler pbccrcCrawler = new PbccrcCrawler();
         pbccrcCrawler.login("88888", "9999999");
     }

 }