<html>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
<head><title>网站垂直搜索</title></head>
<script language=javascript>
function hero()
{
var herowidth=400;//改为你要的网页宽度
var heroheight=300;//改为你要的网页高度
window.resizeTo(herowidth,heroheight);
}
hero();
function getXML(geturl)
{
if(geturl.search(/(http\s*\:\s*\/\/)|(https\s*\:\s*\/\/)/gi) == -1)
{
geturl = "http://" + geturl;
}
var xh = new ActiveXObject("Microsoft.XMLHTTP");
xh.onreadystatechange = function()
{
if(xh.readyState==4 && xh.status==200)
{
var flname = GetTempPathFileName(Url2FileName(geturl));
var objStream;
var imgs;
imgs = xh.responseText;
/****************分析url****************/
if(flname.search(/(\.net$)|(\.com$)|(\.html$)/gi) != -1)
{
var href = imgs.match(/href\s*=\s*\"\s*.*?\s*\"/gi);
//alert(href.join("\r\n"))
if(typeof(href) == "object")
{
var arr = new Array();
for(var i = 0; i < href.length; ++i)
{
if(href[i].indexOf("\"/\"") != -1 || href[i].indexOf("\"\"") != -1 || href[i].substring(href[i].length-5, href[i].length-1) == ".css" ||
href[i].substring(href[i].length-4, href[i].length-1) == ".js" || href[i].search(/([#\'\+])|(javascript\:void)/gi) != -1)
{
continue;
}
else if(href[i].search(/http.*?\:\/\//gi) == -1)
{
var thref = href[i].match(/\".+?\"/gi);
var hosturl = geturl.replace(/\?.*/gi, "");
href[i] = hosturl + "/" + thref[0].substring(1, thref[0].length-1);
href[i] = href[i].replace(/([^\:])\/{2,3}/gi, "$1/");
}
else
{
var url = href[i].match(/\".+?\"/gi);
href[i] = url[0].substring(1, url[0].length-1);
}
arr[arr.length+1] = href[i];
}
//alert(arr.join("\r\n"));
for(var i = 0; i < arr.length; ++i)
{
if(arr[i] == null)
continue;
try
{
//getXML(url[0].substring(1, url[0].length-1));
writeFile("C:\\Users\\admin.admin-PC\\Desktop\\xxxxx.txt", arr[i]);
}
catch(e)
{
alert(arr[i])
}
}
}
//return true;
}
/***************************************/
objStream = new ActiveXObject("ADODB.Stream");
objStream.Type=2;
objStream.Mode=3;
//objStream.Charset="UTF-8";
objStream.open();
objStream.WriteText(imgs);
objStream.SaveToFile(flname)
alert("over");
return true;
}
else
{
return false;
}
};
xh.open("GET",geturl,true);
xh.send();
}
function GetTempPathFileName(flname)
{
var objFSO = new ActiveXObject("Scripting.FileSystemObject");
var strTempPath = objFSO.GetSpecialFolder(2);
var tmppath = objFSO.BuildPath(strTempPath, "xxx1119");
if(!objFSO.FolderExists(tmppath))
{
tmppath = objFSO.CreateFolder(tmppath)
}
tmppath = objFSO.BuildPath(tmppath, flname);
if(objFSO.FileExists(tmppath))
{
objFSO.DeleteFile(tmppath);
}
return tmppath;
}
function Url2FileName(url)
{
var ret = url.match(/\/[^/?]+/g);
var name = "";
if(typeof(ret) == "object" && (name = ret[ret.length-1].substring(1, ret[ret.length-1].length)) != "")
{
return name;
}
return "" + (new Date().getTime()) + ".html";
}
function readFile(filename){
var stm = new ActiveXObject("ADODB.Stream");
stm.LineSeparator=13;
stm.Type=2;
stm.Mode=3;
stm.Charset="UTF-8";
stm.Open();
stm.LoadFromFile(filename)
var s = stm.ReadText();
stm.Close();
return s;
}
function writeFile(filename, s)
{
var buff = readFile(filename);
var stm = new ActiveXObject("ADODB.Stream");
stm.LineSeparator=13;
stm.Type=2;
stm.Mode=3;
stm.Charset="UTF-8";
stm.Open();
stm.Position = 0;
//stm.SetEOS();
stm.WriteText(buff+"\r\n"+s);
stm.SaveToFile(filename, 2);
stm.Close();
}
</script>
<body οnresize=hero();>
<div style="width:400px;height:300px;background-color:White;">
请输入url:
<input type="text" name="yourname"><br>
<input type="submit" value="提交" OnClick="getXML(yourname.value);">
</div>
</body>
</html>