HtmlAgilityPack 学习和笔记

        static void Main(string[] args)
        {
            string content = GetHtmlStr("http://www.ip138.com/", Encoding.GetEncoding("gb2312"));
            var ulr = "";
            HtmlDocument document = new HtmlDocument();
            document.LoadHtml(content);

            HtmlNodeCollection targetNodeCollection = document.DocumentNode.SelectNodes(@"//iframe [1]");//取iframe DOM 第一个
            if (targetNodeCollection != null && targetNodeCollection.Count == 1)
            {
                ulr = targetNodeCollection[0].GetAttributeValue("src", "");//类似jq $(xxx).attr("scr") 对应javascript getattribute
            }
            content = GetHtmlStr(ulr, Encoding.GetEncoding("gb2312"));
            document.LoadHtml(content);
            HtmlNodeCollection tempText = document.DocumentNode.SelectNodes(@"//center [1]");//取center DOM 第一个
            Console.WriteLine(tempText[0].InnerText);//这里就没有过滤了
        }

        public static string GetHtmlStr(string url, Encoding en)
        {
            try
            {
                WebRequest rGet = WebRequest.Create(url);
                rGet.Timeout = 30000;
                using (WebResponse rSet = rGet.GetResponse())
                {
                    using (StreamReader reader = new StreamReader(rSet.GetResponseStream(), en))
                    {
                        return reader.ReadToEnd();
                    }
                }
            }
            catch (WebException)
            {
                //连接失败
                return null;
            }
        }