webinfo.cs

可以抓取淘宝上的所有商家信息 一般情况下可以把商家信息抓取后

using System;

namespace BaseInfo
{
	/// <summary>
	/// webInfo 的摘要说明。
	/// </summary>
	public class webInfo
	{
		public webInfo()
		{
			
		}
		#region 根据url得到所需的web页面
		public static string getwebInfo(string url)
		{
			string em="";
			try
			{
				System.Net.HttpWebRequest hwr=(System.Net.HttpWebRequest)System.Net.HttpWebRequest.Create(url);;
				System.Net.HttpWebResponse res=(System.Net.HttpWebResponse)hwr.GetResponse();
				System.IO.Stream st=res.GetResponseStream();
				System.IO.StreamReader sr=new System.IO.StreamReader(st,System.Text.Encoding.Default);
				//.........
				em=sr.ReadToEnd();
				
				sr.Close();
				
			}
			catch
			{
				;
			}
			return em;
		}
		#endregion

		#region 去除里面的<font >
		public static string clearFontStr(string webStr)
		{
			string str1=webStr;
			string tmpS="";
			str1=str1.ToLower();
			str1=str1.Replace("</font>","");
			int len=str1.IndexOf("<font");
			while(len!=-1)
			{
				tmpS+=str1.Substring(0,len);
				str1=str1.Substring(len);
				int len1=str1.IndexOf(">");
				if(len1!=-1)
				{
					str1=str1.Substring(len1+1);
				}
				len=str1.IndexOf("<font");
			}
			tmpS+=str1;
			//。。。。。。。。。。。。。。。
			tmpS=tmpS.Replace("<em>","");
			tmpS=tmpS.Replace("</em>","");
			tmpS=tmpS.Replace("<strong>","");
			tmpS=tmpS.Replace("</strong>","");
			return tmpS;
		}
		#endregion

		#region 去除里面的Span
		public static string clearSpanStr(string tmpS)
		{
			string str1=tmpS.ToLower();
			string retStr="";
			str1=str1.Replace("</span>","");
			int len=str1.IndexOf("<span");
			while(len!=-1)
			{
				retStr+=str1.Substring(0,len);
				str1=str1.Substring(len+4);
				int len1=str1.IndexOf(">");
				if(len1!=-1)
				{
					str1=str1.Substring(len1+1);
				}
				len=str1.IndexOf("<span");
			}
			retStr+=str1;
			return retStr;
		}
		#endregion

		#region 得到字符串中的email列表
		public static string[] getEmailList(string strInfo)
		{
			string respHTML=strInfo;
			string[] retStr=new string[2000];
			string strRegex ="\\w+([-+.]\\w+)*@\\w+([-.]\\w+)*\\.\\w+([-.]\\w+)*";
			System.Text.RegularExpressions.MatchCollection m;
			System.Text.RegularExpressions.Regex r;
			r=new System.Text.RegularExpressions.Regex(strRegex,System.Text.RegularExpressions.RegexOptions.IgnoreCase);
			m=r.Matches(respHTML);
			for(int i=0;i<m.Count&&i<2000;i++)
			{
				retStr[i]=m[i].Value;
			}
			return retStr;
		}
		#endregion

		
	}
}