public static string GetSummary(string html, int summaryLength)
![](https://www.cnblogs.com/Images/OutliningIndicators/ExpandedBlockStart.gif)
{
string text = GetHtmlText(html);
text = DelHTML(text);
text = Strtitle(text);
if (summaryLength >= text.Length)
return text;
return text.Substring(0, summaryLength)+"
";
}
![](https://www.cnblogs.com/Images/OutliningIndicators/None.gif)
public static string GetHtmlText(string html)
![](https://www.cnblogs.com/Images/OutliningIndicators/ExpandedBlockStart.gif)
{
return System.Text.RegularExpressions.Regex.Replace(html, @"<[^>]*>", "");
}
public static string DelHTML(string Htmlstring)//将HTML去除
![](https://www.cnblogs.com/Images/OutliningIndicators/ExpandedBlockStart.gif)
{
![](https://www.cnblogs.com/Images/OutliningIndicators/ContractedSubBlock.gif)
#region
//删除脚本
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"<script[^>]*?>.*?</script>", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
![](https://www.cnblogs.com/Images/OutliningIndicators/InBlock.gif)
//删除HTML
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"<(.[^>]*)>", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"([\r\n])[\s]+", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"-->", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"<!--.*", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
//Htmlstring =System.Text.RegularExpressions. Regex.Replace(Htmlstring,@"<A>.*</A>","");
//Htmlstring =System.Text.RegularExpressions. Regex.Replace(Htmlstring,@"<[a-zA-Z]*=\.[a-zA-Z]*\?[a-zA-Z]+=\d&\w=%[a-zA-Z]*|[A-Z0-9]","");
![](https://www.cnblogs.com/Images/OutliningIndicators/InBlock.gif)
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(quot|#34);", "\"", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(amp|#38);", "&", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(lt|#60);", "<", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(gt|#62);", ">", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(nbsp|#160);", " ", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(iexcl|#161);", "\xa1", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(cent|#162);", "\xa2", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(pound|#163);", "\xa3", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
![](https://www.cnblogs.com/Images/OutliningIndicators/InBlock.gif)
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&(copy|#169);", "\xa9", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
Htmlstring = System.Text.RegularExpressions.Regex.Replace(Htmlstring, @"&#(\d+);", "", System.Text.RegularExpressions.RegexOptions.IgnoreCase);
![](https://www.cnblogs.com/Images/OutliningIndicators/InBlock.gif)
Htmlstring.Replace("<", "");
Htmlstring.Replace(">", "");
Htmlstring.Replace("\r\n", "");
//Htmlstring=HttpContext.Current.Server.HtmlEncode(Htmlstring).Trim();
#endregion
return Htmlstring.Trim();
}
public static string DelWords(string strtitle)
![](https://www.cnblogs.com/Images/OutliningIndicators/ExpandedBlockStart.gif)
{
string strContent = strtitle;
strtitle = Regex.Replace(strContent, "(<H1.*?>|</H1>|</br>|<SPAN.*?>|</SPAN>|<FONT.*?>||<o:p></o:p>)", "");
strContent = strtitle;
strContent = Regex.Replace(strContent, "(<IMG.*?>)", "");
strContent = Regex.Replace(strContent, "(<P.*?>)", "");
strContent = Regex.Replace(strContent, "(<P>)", "");
strContent = Regex.Replace(strContent, "( )", "");
strContent = Regex.Replace(strContent, "(”)", "");
strContent = Regex.Replace(strContent, "(</P>)", "");
strContent = Regex.Replace(strContent, "(</BR>)", "");
strContent = Regex.Replace(strContent, "(<BR>)", "");
strContent = strContent.Replace("<?xml:namespace prefix = o ns = 'urn:schemas-microsoft-com:office:office'/>", "");
return strtitle;
}