关于getHTML()方法和getHtmlAjax()方法 GetHttpLength， 清除HTML标签

public string getHtml(string Url, string type = "UTF-8")
{
try
{
System.Net.WebRequest wReq = System.Net.WebRequest.Create(Url);
System.Net.WebResponse wResp = wReq.GetResponse();
System.IO.Stream respStream = wResp.GetResponseStream();
using (System.IO.StreamReader reader = new System.IO.StreamReader(respStream, Encoding.GetEncoding(type)))
{
return reader.ReadToEnd().Replace("\n", "").Replace("\r", "");
}
}
catch (System.Exception ex)
{
}
return "";
}

 

 

 

//清除HTML标签

public String ClearHtml(String str)
{
return (String.IsNullOrEmpty(str)) ? String.Empty : System.Text.RegularExpressions.Regex.Replace(str, @"<[^>]*>", String.Empty, System.Text.RegularExpressions.RegexOptions.IgnoreCase);
}

 

 

public string getHtmlAjax(string Url, string type = "UTF-8")
{
wb = new WebBrowser();
wb.Navigate(Url);
while (wb.ReadyState != WebBrowserReadyState.Complete)
{
Application.DoEvents();
}

System.Timers.Timer timer = new System.Timers.Timer();
var isComplete = false;
timer.Elapsed += new System.Timers.ElapsedEventHandler((sender, e) =>
{
//加载完毕
isComplete = true;
timer.Stop();
});
timer.Interval = 1000 * 5;
timer.Start();
while (!isComplete)
Application.DoEvents();
var htmldocument = (mshtml.HTMLDocument)wb.Document.DomDocument;
string Content = htmldocument.documentElement.outerHTML.Replace("\n", "").Replace("\r", "");
Regex reg = new Regex("<DIV class="panelContentWrap".*?下一页</A>");
Content = reg.Match(Content).Value.ToString();
reg = new Regex("<TABLE class=\"ID_table stocks-info-table\".*?下一页</A>");
Content = reg.Match(Content).Value.ToString();
return Content;
}

 

 

 

 

public long GetHttpLength(string url)
{
var length = 0l;
try
{
var req = (HttpWebRequest)WebRequest.CreateDefault(new Uri(url));
req.Method = "HEAD";
req.Timeout = 5000;
var res = (HttpWebResponse)req.GetResponse();
if (res.StatusCode == HttpStatusCode.OK)
{
length = res.ContentLength;
}

res.Close();
return length;
}
catch (WebException wex)
{
return 0;
}
}