StringUtil html regex

StringUtil

public class StringUtil
    {
        #region 截取中文字符串
        /// <summary>
        /// 获取 gb2312 字符串长度
        /// </summary>
        /// <param name="content">字符串,兼容Null和DbNull.Value</param>
        /// <param name="length">长度</param>
        /// <returns></returns>
        public static string GetGB2312StringLength(object obj, int length)
        {
            return GetStringLength(obj, length, "gb2312");
        }

        /// <summary>
        /// 获取 gb2312 字符串长度
        /// </summary>
        /// <param name="content">字符串,兼容Null和DbNull.Value</param>
        /// <param name="length">长度</param>
        /// <returns></returns>
        public static string GetGB2312StringLength(object obj, int length, string suffix)
        {
            return GetStringLength(obj, length, "gb2312", suffix);
        }

        /// <summary>
        /// 获取字符串长度
        /// </summary>
        /// <param name="content">字符串,兼容Null和DbNull.Value</param>
        /// <param name="length">长度</param>
        /// <param name="encode">编码</param>
        /// <returns></returns>
        public static string GetStringLength(object obj, int length, string encode)
        {
            return GetStringLength(obj, length, encode, null);
        }

        /// <summary>
        /// 获取字符串长度
        /// </summary>
        /// <param name="content">字符串,兼容Null和DbNull.Value</param>
        /// <param name="length">长度</param>
        /// <param name="encode">编码</param>
        /// <returns></returns>
        public static string GetStringLength(object obj, int length, string encode, string suffix)
        {
            if (obj == null || obj == DBNull.Value) return string.Empty;
            string content = obj.ToString();

            string tmp = null;
            if (!string.IsNullOrEmpty(suffix))
                tmp = content;

            Encoding encoding = Encoding.GetEncoding(encode);
            StringBuilder sb = new StringBuilder();
            int totalLength = 0;
            foreach (char contentChar in content)
            {
                int size = encoding.GetByteCount(new char[] { contentChar });
                if (totalLength + size > length)
                    break;

                sb.Append(contentChar);
                totalLength += size;
            }

            string result = sb.ToString();
            if (!string.IsNullOrEmpty(suffix))
            {
                if (tmp != result)
                    result += suffix;
            }

            return result;
        }
        #endregion

        #region HTML
        /// <summary>
        /// 删除HTML标记
        /// </summary>
        /// <param name="html"></param>
        /// <returns></returns>
        public static string ConvertHTML2TextRegex(string html)
        {
            string[] aryRegex ={@"<%=[\w\W]*?%>",    @"<script[\w\W]*?</script>",     @"<style[\w\W]*?</style>",   @"<[/]?[\w\W]*?>",   @"([\r\n])[\s]+",
                              @"&(nbsp|#160);",    @"&(iexcl|#161);",  @"&(cent|#162);",            @"&(pound|#163);",   @"&(copy|#169);",
                              @"&#(\d+);",         @"-->",                          @"<!--.*\n"};
            string[] aryReplacment = { "", "", "", "", "", " ", "", "", "", "", "", "", "" };

            for (int i = 0; i < aryRegex.Length; i++)
            {
                Regex regex = new Regex(aryRegex[i], RegexOptions.IgnoreCase);
                html = regex.Replace(html, aryReplacment[i]);
            }

            //Replace "\r\n \t" to an empty character.
            html.Replace("\r\n", "");
            html.Replace("\t", "");

            return html;
        }

        /// <summary>
        /// 过滤HTML中的不安全标签包括js,style,meta 等
        /// </summary>
        /// <param name="content"></param>
        /// <returns></returns>
        public static string RemoveUnsafeHtml(string content)
        {
            content = Regex.Replace(content, @"(\<|\s+)o([a-z]+\s?=)", "$1$2", RegexOptions.IgnoreCase);
            content = Regex.Replace(content, @"(script|frame|form|meta|behavior|style)([\s|:|>])+", "$1.$2", RegexOptions.IgnoreCase);
            return content;
        }
        #endregion

        #region MD5
        /// <summary>
        /// MD5函数
        /// </summary>
        /// <param name="str">原始字符串</param>
        /// <returns>MD5结果</returns>
        public static string MD5(string str)
        {
            byte[] b = Encoding.UTF8.GetBytes(str);
            b = new MD5CryptoServiceProvider().ComputeHash(b);

            return StringConvert.ByteArrayToHexString(b);
        }
        #endregion
    }