asp.Net(C#)データ収集クラス

10999 ワード

using System; 
using System.Data;
using System.Collections;
using System.Collections.Generic;
using System.Configuration; 
using System.Web; 
using MSXML2; //     COM - Microsoft XML V3.0-V6.0
using System.Text.RegularExpressions;

namespace EC
{
    /// <summary> 
    ///         
    /// </summary> 
    public class GetRemoteObj
    {

        #region        
        public GetRemoteObj()
        {
            // 
            // TODO:             
            // 
        }
        ~GetRemoteObj()
        {
            Dispose();
        }
        #endregion

        #region IDisposable   

        public void Dispose()
        {
            GC.SuppressFinalize(this);
        }

        #endregion

        #region       
        /// <summary> 
        ///        
        /// </summary> 
        /// <param name="filename">    </param> 
        /// <returns></returns> 
        public string GetFileExtends(string filename)
        {
            string ext = null;
            if (filename.IndexOf('.') > 0)
            {
                string[] fs = filename.Split('.');
                ext = fs[fs.Length - 1];
            }
            return ext;
        }
        #endregion

        #region          
        /********************************** 
         *     :GetRemoteHtmlCode 
         *     :          
         *       :Url:  url 
         *     : 
         *          GetRemoteObj o = new GetRemoteObj(); 
         *          string url = @"http://www.baidu.com"; 
         *          string s = o.GetRemoteHtmlCode(url); 
         *          Response.Write(s); 
         *          o.Dispose(); 
         * ********************************/
        /// <summary> 
        ///           
        /// </summary> 
        /// <param name="url">  url</param> 
        /// <returns></returns> 
        public string GetRemoteHtmlCode(string Url)
        {
            string s = "";
            MSXML2.XMLHTTP _xmlhttp = new MSXML2.XMLHTTP();
            _xmlhttp.open("GET", Url, false, null, null);
            _xmlhttp.send("");
            if (_xmlhttp.readyState == 4)
            {
                s = System.Text.Encoding.Default.GetString((byte[])_xmlhttp.responseBody);
            }
            return s;
        }

        #endregion

        #region       
        /********************************** 
         *     :RemoteSave 
         *     :       
         *       :Url:  url;Path:       
         *     : 
         *          GetRemoteObj o = new GetRemoteObj(); 
         *          string s = ""; 
         *          string url = @"http://www.baidu.com/img/logo.gif"; 
         *          string path =Server.MapPath("Html/"); 
         *          s = o.RemoteSave(url,path); 
         *          Response.Write(s); 
         *          o.Dispose();          
         * ******************************/
        /// <summary> 
        ///        
        /// </summary> 
        /// <param name="Url">  url</param> 
        /// <param name="Path">      </param> 
        /// <returns></returns> 
        public string RemoteSave(string Url, string Path)
        {
            Random ra = new Random();
            string StringFileName = DateTime.Now.ToString("yyyyMMddHHmmssffff") + "." + GetFileExtends(Url);
            string StringFilePath = Path + StringFileName;
            MSXML2.XMLHTTP _xmlhttp = new MSXML2.XMLHTTP();
            _xmlhttp.open("GET", Url, false, null, null);
            _xmlhttp.send("");
            if (_xmlhttp.readyState == 4)
            {
                if (System.IO.File.Exists(StringFilePath))
                    System.IO.File.Delete(StringFilePath);
                System.IO.FileStream fs = new System.IO.FileStream(StringFilePath, System.IO.FileMode.CreateNew);
                System.IO.BinaryWriter w = new System.IO.BinaryWriter(fs);
                w.Write((byte[])_xmlhttp.responseBody);
                w.Close();
                fs.Close();
            }
            else
                throw new Exception(_xmlhttp.statusText);
            return StringFileName;
        }
        #endregion

        #region            
        /// <summary> 
        ///             
        /// </summary> 
        /// <param name="HtmlCode">HTML   </param> 
        /// <returns></returns> 
        public string ReplaceEnter(string HtmlCode)
        {
            string s = "";
            if (HtmlCode == null || HtmlCode == "")
                s = "";
            else
                s = HtmlCode.Replace("\"", "");
            s = s.Replace("\r
", ""); return s; } #endregion #region HTML /// <summary> /// HTML /// </summary> /// <param name="HtmlCode">html </param> /// <returns></returns> public string RemoveHTML(string HtmlCode) { string MatchVale = HtmlCode; foreach (Match s in Regex.Matches(HtmlCode, "<.+?>")) { MatchVale = MatchVale.Replace(s.Value, ""); } return MatchVale; } #endregion #region /// <summary> /// /// </summary> /// <param name="HtmlCode"></param> /// <returns></returns> public List<string> GetHref(string HtmlCode) { List<string> list = new List<string>(); string Reg = @"(h|H)(r|R)(e|E)(f|F) *= *('|"")?((\w|\\|\/|\.|:|-|_)+)('|""| *|>)?"; foreach (Match m in Regex.Matches(HtmlCode, Reg)) { list.Add((m.Value).ToLower().Replace("href=", "").Trim()); } return list; } #endregion #region /// <summary> /// /// </summary> /// <param name="HtmlCode"></param> /// <param name="imgHttp"> http:// </param> /// <returns></returns> public string GetImgSrc(string HtmlCode, string imgHttp) { string MatchVale = ""; string Reg = @"<img.+?>"; foreach (Match m in Regex.Matches(HtmlCode, Reg)) { MatchVale += GetImg((m.Value).ToLower().Trim(), imgHttp) + "||"; } return MatchVale; } /// <summary> /// <img src="" /> /// </summary> /// <param name="ImgString"><img src="" /> </param> /// <returns></returns> public string GetImg(string ImgString, string imgHttp) { string MatchVale = ""; string Reg = @"src=.+\.(bmp|jpg|gif|png|)"; foreach (Match m in Regex.Matches(ImgString.ToLower(), Reg)) { MatchVale += (m.Value).ToLower().Trim().Replace("src=", ""); } return (imgHttp + MatchVale); } #endregion #region Html /// <summary> /// Html : Dictionary key- value- /// </summary> /// <param name="HtmlCode"> HTML </param> /// <returns></returns> public Dictionary<string, string> GetCodeHref(string HtmlCode) { Dictionary<string, string> dic = new Dictionary<string, string>(); // HTML href <B> </B> string regex = @"\<a.*href\s*=\s*(?:""(?<url>[^""]*)""|'(?<url>[^']*)'|(?<url>[^\>^\s]+)).*\>(?<title>[^\<^\>]*)\<[^\</a\>]*/a\>"; Regex reg1 = new Regex(regex, RegexOptions.IgnoreCase);// MatchCollection ms1 = reg1.Matches(HtmlCode);// foreach (Match m1 in ms1)// { dic.Add(m1.Groups["title"].Value, m1.Groups["url"].Value); } return dic; } #endregion #region /// <summary> /// /// </summary> /// <param name="str"> </param> /// <param name="s"> </param> /// <param name="e"> </param> /// <returns></returns> public string GetRegValue(string str, string s, string e) { Regex rg = new Regex("(?<=(" + s + "))[.\\s\\S]*?(?=(" + e + "))", RegexOptions.Multiline | RegexOptions.Singleline); return rg.Match(str).Value; } /// <summary> /// /// </summary> /// <param name="str"> </param> /// <param name="s"> </param> /// <param name="e"> </param> /// <returns></returns> public List<string> GetRegValueArr(string str, string s, string e) { Match m; List<string> list = new List<string>(); Regex rg = new Regex("(?<=(" + s + "))[.\\s\\S]*?(?=(" + e + "))", RegexOptions.Multiline | RegexOptions.Singleline); for (m = rg.Match(str); m.Success; m = m.NextMatch()) { list.Add(m.Groups[0].Value); } return list; } #endregion #region public string RegReplace(string RegValue, string regStart, string regEnd) { string s = RegValue; if (RegValue != "" && RegValue != null) { if (regStart != "" && regStart != null) { s = s.Replace(regStart, ""); } if (regEnd != "" && regEnd != null) { s = s.Replace(regEnd, ""); } } return s; } #endregion } }