C# 获取网页标题title的代码
2018年02月06日
⁄ 综合
⁄ 共 15082字 ⁄ 字号
小 中 大
using System;
using System.Net;
using System.Collections;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Web;
using System.Web.SessionState;
using System.Web.UI;
using System.Web.UI.WebControls;
using System.Web.UI.HtmlControls;
using System.Data;
using System.Data.OleDb;
using System.IO;
using System.Text;
using System.Globalization;
using System.Text.RegularExpressions;
namespace GETURL
...{
public class getText : System.Web.UI.Page
...{
protected System.Web.UI.WebControls.Label label_mess;
protected System.Web.UI.WebControls.Label content;
protected System.Web.UI.WebControls.Label MetaUrl;
protected System.Web.UI.WebControls.Label GetTitle;
protected System.Web.UI.WebControls.Label label_Title;
protected System.Web.UI.WebControls.Label startSpider;
protected System.Web.UI.WebControls.TextBox searchurl;
protected System.Web.UI.WebControls.Button urlButton;
protected System.Web.UI.WebControls.Label txtsiteurl;
private void Page_Load(Object sender,EventArgs e)
...{
if(Request.QueryString["url"]!=null && Request.QueryString["url"].ToString()!="")
...{
searchurl.Text=Request.QueryString["url"].ToString();
getHTTP(sender,e);
}
}
public void getHTTP(Object sender,EventArgs e)
...{
startSpider.Text="";
GetTitle.Text="";
MetaUrl.Text="";
//string host_all_Path=Request.ServerVariables["PATH_TRANSLATED"];
//string host_aPath=Request.ServerVariables["PATH_INFO"];
//string host_IP=Request.ServerVariables["REMOTE_ADDR"];
//string host_Host_IP=Request.ServerVariables["LOCAL_ADDR"];
//string host_NAME=Request.ServerVariables["SERVER_NAME"];
if(searchurl.Text!="")
...{
txtsiteurl.Text="--spider url'Result!";
string aUrl;
aUrl=searchurl.Text;
string[] txtspiderurl=aUrl.Split('/');
if(txtspiderurl[0].ToString()!="http:")
...{
label_mess.Text="<hr>Url form must match'http://' ahead!!";
content.Text="";
}
else
...{
if(txtspiderurl.Length>=3)
...{
//=========================url spider start==================
if(txtspiderurl[2].ToString()!="" && txtspiderurl[1].ToString()=="")
...{
string description;
string keywords;
string title_site;
//about the url to spider the hostname
//and his IP_ADDR
//and the the url's Title
//the url's description,and the keywords.
//and last to show the whole site's contents.
string hosturl=txtspiderurl[2].ToString();
string defaultLanguage;
//defaultLanguage=Request.ServerVariables["REMOTE_ADDR"];
//startSpider.Text="<br>d: "+defaultLanguage;
HttpWebRequest myReq=(HttpWebRequest)WebRequest.Create(aUrl);
//(([a-z0-9A-Z]|/-)+/.)+[a-z0-9A-Z]{1,4}$
Regex match_hosturl=new Regex("^(/w+(-/w+)*)(/.(/w+(-/w+)*))*(/?/S*)?(|/.)$");
try
...{
if(match_hosturl.IsMatch(hosturl))
...{
label_mess.Text="<hr>You spidering the site:<font color=red size='3'>"+aUrl+"</font><br>";
//-------------------------------look the domain is formed by NUMERIC-----
Regex a_UrlRegex=new Regex(@"^(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]).(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0).(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[1-9]|0).(25[0-5]|2[0-4][0-9]|[0-1]{1}[0-9]{2}|[1-9]{1}[0-9]{1}|[0-9])$");
if(a_UrlRegex.IsMatch(hosturl))
...{
IPHostEntry IP_To_HostName=Dns.GetHostByAddress(hosturl);
content.Text="Host Name: <font color=red size='3'> "+IP_To_HostName.HostName+"</font><hr color='red'>";
}
else
...{
string s="";
content.Text+="Host domain: <font color=red size='3'> "+hosturl+"</font><hr color='red'>";
System.Net.IPAddress[] addressList=Dns.GetHostByName(hosturl).AddressList;
for (int i = 0; i < addressList.Length; i ++)
...{
s += addressList[i].ToString();
}
content.Text+="Host IP: <font color=red size=3>"+s+"</font><hr color=red>";
}
//---------------------------------end--the --program application--------------------------------------
HttpWebResponse urlHttpWebRes=(HttpWebResponse)myReq.GetResponse();
myReq.ContentType="application/x-www-form-urlencoded";
Stream urlStream=urlHttpWebRes.GetResponseStream();
Encoding encode=System.Text.Encoding.GetEncoding("utf-8");
StreamReader urlStreamReader;
//Encoding.Default
if (a_UrlRegex.IsMatch(hosturl))
...{
if(hosturl=="127.0.0.1" || hosturl=="172.19.23.14")
...{
if (aUrl.IndexOf("?")!=-1)
...{
string[] getStringAhead=aUrl.Split('?');
string[] getStringHtm=getStringAhead[0].ToString().Split('.');
if(getStringHtm[getStringHtm.Length-1].ToString().ToLower()=="htm" || getStringHtm[getStringHtm.Length-1].ToString().ToLower()=="html")
...{
urlStreamReader=new StreamReader(urlStream,Encoding.Default);
}
else
...{
urlStreamReader=new StreamReader(urlStream,encode);
}
}
else
...{
string[] getStringHtm=aUrl.Split('.');
if(getStringHtm[getStringHtm.Length-1].ToString().ToLower()=="htm" || getStringHtm[getStringHtm.Length-1].ToString().ToLower()=="html")
...{
urlStreamReader=new StreamReader(urlStream,Encoding.Default);
}
else
...{
urlStreamReader=new StreamReader(urlStream,encode);
}
}
//urlStreamReader=new StreamReader(urlStream,Encoding.Default);
}
else
...{
urlStreamReader=new StreamReader(urlStream,Encoding.Default);
}
}
//The a to z URL
else
...{
if(hosturl.ToLower()=="microsoft.com" || hosturl.ToLower()=="www.microsoft.com")
...{
urlStreamReader=new StreamReader(urlStream,encode);
}
else
...{
urlStreamReader=new StreamReader(urlStream,Encoding.Default);
}
}
//The a to z URL End
StringBuilder urlContent=new StringBuilder();
Char[] getChar=new Char[256];
int Count=urlStreamReader.Read(getChar,0,256);
while (Count>0)
...{
String line=new String(getChar,0,Count);
urlContent.Append(line);
Count=urlStreamReader.Read(getChar,0,256);
//Regex getTitle=new Regex(@"<title>",RegexOptions.IgnoreCase);
//Regex endTitle=new Regex(@"</title>",RegexOptions.IgnoreCase);
//***********************************Get the Site's Title Name*******Start*******************************************
//*****************************************Writed 2005-5-19*******Author WeisNet System Information************************
Regex title_aReg=new Regex(@"<title");
Regex title_bReg=new Regex(@"</title");
int thefirstTitleIndexOf;
String urlContentString;
String urlContentStringToLower;
urlContentString=urlContent.ToString();
urlContentStringToLower=urlContentString.ToLower();
if(title_aReg.IsMatch(urlContentStringToLower) && title_bReg.IsMatch(urlContentStringToLower))
...{
thefirstTitleIndexOf=urlContentStringToLower.IndexOf("title");
int thelastTitleIndexOf;
thelastTitleIndexOf=urlContentStringToLower.IndexOf("/title");
//**************Check first title left is have space *********88
string testIfthisIsTagleft;
string getStringfromfirstTitle;
//TOlower string 'urlContentStringToLower'->urlContentString 0xDDDIFCIFWEISNET____
getStringfromfirstTitle=urlContentString.Substring(thefirstTitleIndexOf);
testIfthisIsTagleft=getStringfromfirstTitle.Substring(5,1);
//**************Check first title left is have space ***End******
if(testIfthisIsTagleft==">")
...{
int getTitleNum;
getTitleNum=thelastTitleIndexOf-1-thefirstTitleIndexOf-6;
string getTitleName;
//TOlower string 'urlContentStringToLower'->urlContentString 0xDDDIFCIFWEISNET____
getTitleName=urlContentString.Substring(thefirstTitleIndexOf+6,getTitleNum);
GetTitle.Text="Spider Title :<font color='red' size='3'>"+getTitleName+"</font>--"+urlContentString.Length/1024+"KB<hr color=red>";
}
else
...{
int getTitleNum;
int isHaveSpacetitlerightTagNum;
int lastTitlefromElseNum;
isHaveSpacetitlerightTagNum=getStringfromfirstTitle.IndexOf(">");
lastTitlefromElseNum=getStringfromfirstTitle.IndexOf("</title");
getTitleNum=lastTitlefromElseNum-isHaveSpacetitlerightTagNum-1;
string getTitleName;
getTitleName=getStringfromfirstTitle.Substring(isHaveSpacetitlerightTagNum+1,getTitleNum);
GetTitle.Text="Spider Title :<font color='red' size='3'>"+getTitleName+"</font>--"+urlContentString.Length%1024+"KB<hr color=red>";
}
}
else
...{
GetTitle.Text="No Title Document!";
}
//***************************************Get the Site's Title Name*****End***2005-8-13***********************************
//****************************************Get The Site's Charset Code*******Start******2005-5-23*****************************
string getCharsetValue;
String bString;
bString=urlContent.ToString();
String toLowerString;
toLowerString=bString.ToLower();
Regex meta_first_last=new Regex("<meta");
int getFirstMetaNum;
int getLastMetaNum;
if(meta_first_last.IsMatch(toLowerString))
...{
getFirstMetaNum=toLowerString.IndexOf("<meta");
getLastMetaNum=toLowerString.LastIndexOf("<meta");
string getStringFromLastMeta;
getStringFromLastMeta=toLowerString.Substring(getLastMetaNum+1);
int getLastMetaleftTagNum;
getLastMetaleftTagNum=getStringFromLastMeta.IndexOf(">");
string MetaAllString;
//OK WeisNet 0XDCNDIIGMT 08:00IKABIDKF2005-9-20 16:26:37JWEISNET__
MetaAllString=bString.Substring(getFirstMetaNum,getLastMetaleftTagNum+2+getLastMetaNum-getFirstMetaNum);
string formatMetaStringNoSpace;
formatMetaStringNoSpace=Regex.Replace(Regex.Replace(Regex.Replace(Regex.Replace(Regex.Replace(Regex.Replace(MetaAllString," +",""),"<meta","<meta "),"http-equiv"," http-equiv"),"content"," content")," content-type","content-type")," +"," ");
if (formatMetaStringNoSpace.IndexOf("http-equiv=")!=-1 && formatMetaStringNoSpace.IndexOf("content=")!=-1)
MetaUrl.Text="OK";
else
MetaUrl.Text="NO";
}
else
...{
MetaUrl.Text="Not Charset Code";
}
//MetaUrl.Text="ok";
//***************************************Get The Site's CharsetCode****End**************************************************
}
//startSpider.Text="<hr>content: " + urlContent.ToString()+"<hr color='red'>";
urlStreamReader.Close();
urlStream.Close();
urlHttpWebRes.Close();
}
else
...{
label_mess.Text="";
content.Text="<hr>Host domain Wrong!";
}
}
catch(ArgumentException exp) //WEBSITE ARGUMENTEXCEPTION'S==WEISNET OX12ACEB0068-----Start---->
...{
Console.WriteLine("The Urls Spider has Time Out!,Try Again...");
Console.WriteLine(exp.Message);
//content.Text="The Urls Spider has Time Out!,Try Again...<br>";
//startSpider.Text="Or you Check the Url is true you want Test!<br>";
}
catch(WebException exp)
...{
Console.WriteLine("The Urls Spider has Time Out!,Try Again...");
Console.WriteLine(exp.Message);
Console.WriteLine(exp.Status);
//content.Text="The Urls Spider has Time Out!,Try Again...<br>";
//startSpider.Text="Or you Check the Url is true you want Test!<br>";
}
catch(Exception exp)
...{
Console.WriteLine("The Urls Spider has Time Out!,Try Again...");
Console.WriteLine(exp.Message);
//content.Text="The Urls Spider has Time Out!,Try Again...<br>";
//startSpider.Text="Or you Check the Url is true you want Test!<br>";
} //WEBSITE ARGUMENTEXCEPTION'S==WEISNET OX12ACEB0068-------------------------------End------>
//===============================================Url spider Ending=====================================================
}
else
...{
label_mess.Text="<hr>Host Domain must Exists!!!";
content.Text="";
}
}
else
...{
label_mess.Text="<hr>Please write the whole url text!";
content.Text="";
}
}
}
else
...{
label_mess.Text="<hr>Please enter the URL";
content.Text="";
}
}
//=========================Ending Application========================
WeisNet WebTools#region WeisNet WebTools
override protected void OnInit(EventArgs e)
...{
InitializeComponent();
base.OnInit(e);
}
private void InitializeComponent()
...{
this.Load += new System.EventHandler(this.Page_Load);
}
#endregion
}
}
/**//*---------------------------------------------------Power By WeisNet System-----------------------------------------------*/