httpClient 测试

摘至：http://hi.baidu.com/xbnh0217/blog/item/b1feee321e162549ad4b5f8d.html

 

package com.bonck.service.cook;

import java.io.IOException;
import java.util.HashMap;

import org.apache.commons.httpclient.Cookie;
import org.apache.commons.httpclient.DefaultMethodRetryHandler;
import org.apache.commons.httpclient.Header;
import org.apache.commons.httpclient.HttpClient;
import org.apache.commons.httpclient.HttpException;
import org.apache.commons.httpclient.HttpStatus;
import org.apache.commons.httpclient.NameValuePair;
import org.apache.commons.httpclient.methods.GetMethod;
import org.apache.commons.httpclient.methods.PostMethod;
import org.apache.commons.httpclient.params.HttpMethodParams;

import com.bonck.util.LogService;

public class CommonParserTest {

/**
* 此登录需要验证码
* 这里开始登陆
* map 参数设置
* String url 登陆地址
* String code 验证码名字
* String vcode 验证码
* String uname 用户名 实参
* String upwd 密码 实参
* String username 用户名名字
* String password 用户名名字
* String hosturl cook这只url
* String encode 编码种类 gb2312 utf-8
* 登录入口并返回Map 存有client信息和cookie信息
* @return
*/
   public HashMap vlogin(HashMap sm) {
   //String url, String vcode,String usrName,String pwd
   HashMap map=new HashMap();

   HttpClient client = new HttpClient();
   // 设置请求时间
   client.getParams().setConnectionManagerTimeout(10000l);

   client.getHostConfiguration().setHost("www.myshida.com", 80, "http");

   // 获得PostMethod 对象 并设置参数
   String loginUrl=sm.get("url")+"";
   PostMethod post = new PostMethod("http://www.myshida.com/club/logging.php?action=login" );
   // 设置登录参数
   NameValuePair[] para = new NameValuePair[] {

   new NameValuePair("username", "liuxue"),

   new NameValuePair("password", "liuxue123456") ,
 
   new NameValuePair("loginfield", "username")
 
//   new NameValuePair("formhash", "2a8c1766"),
// 
//   new NameValuePair("referer", "index.php"),
 
//   new NameValuePair(sm.get("vcode")+"", (sm.get("code")+"")
     //验证码
   };
   // 将参数设置到也面中
   post.setRequestBody(para);

   // 设置请求参数 默认请求三次
   post.getParams().setParameter(HttpMethodParams.RETRY_HANDLER,
     new DefaultMethodRetryHandler());

   // Httpclient 对于接受后继服务的请求，象post和put等不能自动转发
   // 正确状态码是 SC_OK =200 301 302
   // 获得状态码
   try {
    int status = client.executeMethod(post);
//HttpStatus.SC_OK
    if (status == 302) {

     Header locationHeader = post.getResponseHeader("location");

     String location = null;

     if (locationHeader != null) {

      location = locationHeader.getValue();

      System.out
        .println("the Page was redirected to " + location);
     } else {
      System.out.println("this location was null");
      }
     }
    else if(status==HttpStatus.SC_OK){
     // 读取相应信息
     byte[] r = post.getResponseBody();

    //转码 获取相应的数据
     String content = new String(r,"gb2312");
//     LogService.info("-------------------------------****************");
     LogService.info(content);

     Cookie[] cookie = client.getState().getCookies();
     // 将cook设置到client里面
     client.getState().addCookies(cookie);
     // 设置释放链接
     post.releaseConnection();
     map.put("cookie", cookie);
     map.put("client", client);
    }
   } catch (HttpException e) {
    e.printStackTrace();
   } catch (IOException e) {
    e.printStackTrace();
   }

   return map;
}

/**
* 该登陆不需要验证码
* 这里开始登陆
* map 参数设置
* String url 登陆地址
* String uname 用户名 实参
* String upwd 密码 实参
* String username 用户名名字
* String password 用户名名字
* String hosturl cook这只url
* String encode 编码
* 登录入口并返回Map 存有client信息和cookie信息
* @return
*/
@SuppressWarnings("unchecked")
public HashMap login(HashMap sm) {
   //String url, String vcode,String usrName,String pwd
   HashMap map=new HashMap();

   HttpClient client = new HttpClient();
   // 设置请求时间
   client.getParams().setConnectionManagerTimeout(5000l);

   client.getHostConfiguration().setHost(sm.get("hosturl")+"", 80, "http");

   // 获得PostMethod 对象 并设置参数
   PostMethod post = new PostMethod( sm.get("url")+"");
   // 设置登录参数
   NameValuePair[] para = new NameValuePair[] {

   new NameValuePair(sm.get("username")+"", sm.get("uname")+""),

   new NameValuePair(sm.get("password")+"", sm.get("upwd")+"")
 
 
   //, new NameValuePair(sm.get("code")+"", (sm.get("vcode")+""))//验证码
   };
   // 将参数设置到也面中
   post.setRequestBody(para);

   // 设置请求参数 默认请求三次
   post.getParams().setParameter(HttpMethodParams.RETRY_HANDLER,
     new DefaultMethodRetryHandler());

   // Httpclient 对于接受后继服务的请求，象post和put等不能自动转发
   // 正确状态码是 SC_OK =200 301 302
   // 获得状态码
   try {
    int status = client.executeMethod(post);

    if (status == HttpStatus.SC_OK) {

     Header locationHeader = post.getResponseHeader("location");

     String location = null;

     if (locationHeader != null) {

      location = locationHeader.getValue();
       String st=new String(location.getBytes(),"gb2312");
      System.out
        .println("the Page was redirected to " + st);
     } else {

      System.out.println("this location was null");

     }
     // 读取相应信息
     byte[] r = post.getResponseBody();

    //转码 获取相应的数据
     String content = new String(r, sm.get("encode")+"");
     LogService.info("-------------------------------****************");
     LogService.info(content);

     Cookie[] cookie = client.getState().getCookies();
     // 将cook设置到client里面
     client.getState().addCookies(cookie);
     // 设置释放链接
     post.releaseConnection();
     map.put("cookie", cookie);
     map.put("client", client);
    }
   } catch (HttpException e) {
    e.printStackTrace();
   } catch (IOException e) {
    e.printStackTrace();
   }

   return map;
}

/***************************************************************************
* @version 1.0
* 主要用来爬取要爬的信息
* 根据网址对页面信息进行采集
* @param url 指定爬去的网页
* @param map client客户端 cookie Cookie信息
* @param ,String encoding 对内容进行编码转换
* @return
*/
public String getContentByParser(String url,HashMap map,String encoding) {
 
   HttpClient client=(HttpClient)map.get("client");
 
   Cookie []cookie=(Cookie[])map.get("cookie");
 
   GetMethod get=new GetMethod(url);
 
   get.getParams().setParameter(HttpMethodParams.RETRY_HANDLER, new DefaultMethodRetryHandler());
   //设置cookie
   get.setRequestHeader("Cookie",cookie.toString());
 
   String content = null;
   try {
    int statusCode=client.executeMethod(get);
  
    if(statusCode==HttpStatus.SC_OK){
 
     byte[] responseBody = get.getResponseBody();
     content = new String(responseBody, encoding);//对内容进行编码转换
    }
   } catch (HttpException e) { 
    e.printStackTrace();
   } catch (IOException e) { 
    e.printStackTrace();
   }
   return content;
}
public static void main(String[] args) {
   String url="";
   String username="";
   String pwd="";
   CommonParserTest test=new CommonParserTest();
   HashMap map=new HashMap();
   test.vlogin(map);
}
}