從Internet上抓取指定URL的源碼的方案(C#)
來源：互聯網
上載者：User
調用方式：
#region 測試擷取遠程網頁
GetPageCode gpc = new GetPageCode();
gpc.Url="http://ppcode.com/";
gpc.ProxyState=1;//使用Proxy 伺服器，0為不使用，設定為1後下面的代理設定才起作用
gpc.ProxyAddress="http://proxyName.com";//Proxy 伺服器地址
gpc.ProxyPort="80";//Proxy 伺服器的連接埠
gpc.ProxyAccount="proxy";//Proxy 伺服器帳號
gpc.ProxyPassword="password";//Proxy 伺服器密碼
gpc.ProxyDomain="bqc";//Proxy 伺服器域
gpc.OutFilePath=filePath;//設定輸出檔案路徑的地方，如果不設定，則返回字串
gpc.GetSource();//處理
string tempErr=gpc.NoteMessage;//如果出錯，這裡會提示
string tempCode=gpc.OutString;//返回的字串
#endregion
類代碼：
using System;
using System.Collections;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.IO;
using System.Net;
using System.Text;
using System.Web;
namespace Test.Com
{
/// <summary>
/// 功能：取得Internet上的URL頁的源碼
/// 建立：2004-03-22
/// 作者：Rexsp MSN:yubo@x263.net
/// </summary>
public class GetPageCode
{
#region 私人變數
/// <summary>
/// 網頁URL地址
/// </summary>
private string url=null;
/// <summary>
/// 是否使用代碼伺服器：0 不使用 1 使用Proxy 伺服器
/// </summary>
private int proxyState=0;
/// <summary>
/// Proxy 伺服器地址
/// </summary>
private string proxyAddress=null;
/// <summary>
/// Proxy 伺服器連接埠
/// </summary>
private string proxyPort=null;
/// <summary>
/// Proxy 伺服器使用者名稱
/// </summary>
private string proxyAccount=null;
/// <summary>
/// Proxy 伺服器密碼
/// </summary>
private string proxyPassword=null;
/// <summary>
/// Proxy 伺服器域
/// </summary>
private string proxyDomain=null;
/// <summary>
/// 輸出檔案路徑
/// </summary>
private string outFilePath=null;
/// <summary>
/// 輸出的字串
/// </summary>
private string outString=null;
/// <summary>
/// 提示資訊
/// </summary>
private string noteMessage;
#endregion
#region 公用屬性
/// <summary>
/// 欲讀取的URL地址
/// </summary>
public string Url
{
get{return url;}
set{url=value;}
}
/// <summary>
/// 是否使用Proxy 伺服器標誌
/// </summary>
public int ProxyState
{
get{return proxyState;}
set{proxyState=value;}
}
/// <summary>
/// Proxy 伺服器地址
/// </summary>
public string ProxyAddress
{
get{return proxyAddress;}
set{proxyAddress=value;}
}
/// <summary>
/// Proxy 伺服器連接埠
/// </summary>
public string ProxyPort
{
get{return proxyPort;}
set{proxyPort=value;}
}
/// <summary>
/// Proxy 伺服器帳號
/// </summary>
public string ProxyAccount
{
get{return proxyAccount;}
set{proxyAccount=value;}
}
/// <summary>
/// Proxy 伺服器密碼
/// </summary>
public string ProxyPassword
{
get{return proxyPassword;}
set{proxyPassword=value;}
}
/// <summary>
/// Proxy 伺服器域
/// </summary>
public string ProxyDomain
{
get{return proxyDomain;}
set{proxyDomain=value;}
}
/// <summary>
/// 輸出檔案路徑
/// </summary>
public string OutFilePath
{
get{return outFilePath;}
set{outFilePath=value;}
}
/// <summary>
/// 返回的字串
/// </summary>
public string OutString
{
get{return outString;}
}
/// <summary>
/// 返回提示資訊
/// </summary>
public string NoteMessage
{
get{return noteMessage;}
}
#endregion
#region 建構函式
public GetPageCode()
{
}
#endregion
#region 公用方法
/// <summary>
/// 讀取指定URL地址，存到指定檔案中
/// </summary>
public void GetSource()
{
WebRequest request = WebRequest.Create(this.url);
//使用Proxy 伺服器的處理
if(this.proxyState==1)
{
//預設讀取80連接埠的資料
if(this.proxyPort==null)
this.ProxyPort="80";
WebProxy myProxy=new WebProxy();
myProxy = (WebProxy)request.Proxy;
myProxy.Address = new Uri(this.ProxyAddress+":"+this.ProxyPort);
myProxy.Credentials = new NetworkCredential(this.proxyAccount, this.proxyPassword, this.ProxyDomain);
request.Proxy = myProxy;
}
try
{
//請求服務
WebResponse response = request.GetResponse();
//返回資訊
Stream resStream = response.GetResponseStream();
StreamReader sr = new StreamReader(resStream, System.Text.Encoding.Default);
string tempCode= sr.ReadToEnd();
resStream.Close();
sr.Close();
//如果輸出檔案路徑為空白，便將得到的內容賦給OutString屬性
if(this.outFilePath==null)
{
this.outString=tempCode;
}
else
{
FileInfo fi = new FileInfo(this.outFilePath);
//如果存在檔案則先幹掉
if(fi.Exists)
fi.Delete();
StreamWriter sw = new StreamWriter(this.outFilePath,true,Encoding.Default);
sw.Write(tempCode);
sw.Flush();
sw.Close();
}
}
catch
{
this.noteMessage="出錯了，請檢查網路是否連通;";
}
}
#endregion
}
}
