FCS/FCS.Common/Helper/IPHelper.cs

using System;
using System.Collections.Generic;
using System.Configuration;
using System.Diagnostics;
using System.IO;
using System.IO.Compression;
using System.Linq;
using System.Net;
using System.Text;
using System.Threading;
using System.Threading.Tasks;
using HtmlAgilityPack;
using Newtonsoft.Json;
using static System.Net.Mime.MediaTypeNames;

namespace FCS.Common
{
    public class IPHelper
    {
        private List<Task> taskList = new List<Task>();
        static object locker = new object();
        private static List<string> list = new List<string>();
        private static List<string> urlList = new List<string>();
        private static string title;
        private static string iPAgencyAddress = AppDomain.CurrentDomain.BaseDirectory + "/XmlConfig/IPAgency.txt";
        /// <summary>
        /// 初始化方法
        /// </summary>
        public static void GetIp(IEnumerable<string> ipList = null, bool isIpAgency = true)
        {
            if (isIpAgency)
            {
                string configFile = AppDomain.CurrentDomain.BaseDirectory + "/XmlConfig/UrlData.xml";
                var urlDataList = CommonHelper.ConvertXMLToObject<IPEntity>(configFile, "UrlSetting");
                var list = from a in urlDataList
                           select a.Url;
                //代理IP池
                new IPHelper().GetIPData(list.ToList(), "ip", iPAgencyAddress);
            }

            //爬取ip池
            if (ipList == null || ipList.Count() <= 0)
                ipList = new List<string> {
            "http://fenxi.zgzcw.com/1766186/dxdb/zhishu?company_id=8&company=Bet365",//大小指数-实时
            "http://fenxi.zgzcw.com/1766186/bjop/zhishu?company_id=115&company=%E5%A8%81%E5%BB%89%E5%B8%8C%E5%B0%94",//欧赔指数-实时
            "http://fenxi.zgzcw.com/1766186/ypdb/zhishu?company_id=1&company=%E6%BE%B3%E9%97%A8",//亚盘指数-实时
            "http://fenxi.zgzcw.com/2413445/dxdb",//大小指数-固化
            "http://fenxi.zgzcw.com/2413445/bjop",//欧赔指数-固化
            "http://fenxi.zgzcw.com/2413445/ypdb",//亚盘指数-固化
            "http://saishi.zgzcw.com/soccer/team/25",//球队介绍
            "http://saishi.zgzcw.com/soccer",//赛事
            "http://saishi.zgzcw.com/soccer/league/36/2018-2019/sxpl",//上下盘
            "http://saishi.zgzcw.com/soccer/league/36/2018-2019/sxds",//上下单双
            "http://saishi.zgzcw.com/soccer/league/36/2018-2019/jqds",//总进球单双
            "http://saishi.zgzcw.com/soccer/league/36/2018-2019/ssb",//射手榜
            "http://saishi.zgzcw.com/soccer/league/36/2018-2019",//积分榜
            "http://saishi.zgzcw.com/soccer/league/36",//球队,分组，比赛
            "http://cp.zgzcw.com/lottery/zucai/14csfc/index.jsp",//14场
            "http://cp.zgzcw.com/lottery/zucai/6cbqc/index.jsp",//6场
            "http://cp.zgzcw.com/lottery/zucai/4cjqc/index.jsp",//4场
            "http://live.zgzcw.com/jz/",//竞彩
            "http://live.zgzcw.com/bd/",//北单
            "http://live.zgzcw.com/sfc/",//足彩
            "http://fenxi.zgzcw.com/2404696/zrtj",//阵容
            "http://saishi.zgzcw.com/soccer/player/4369",//球员
            "http://nba.nowscore.com/League/index_cn.aspx?SclassID=5&MatchSeason=2017-2018",//CBA
            "http://www.500.com"
                        };
            new IPHelper().GetIPData(ipList.ToList());
        }
        private void GetIPData(List<string> _urlList, string _title = "", string path = "")
        {
            list = new List<string>();
            if (_urlList.Count <= 0)
                urlList.Add(ConfigurationManager.AppSettings["TestUrl"].ToString());
            else
                urlList = _urlList;
            title = _title;
            Trace.WriteLine("开始更新IP代理池！请稍后");
            string configFile = AppDomain.CurrentDomain.BaseDirectory + "/XmlConfig/UrlData.xml";
            var urlData = CommonHelper.ConvertXMLToObject<IPEntity>(configFile, "UrlSetting");
            urlData.ForEach(p =>
            {
                taskList.Add(Task.Run(() => GetIp(p)));
            });
            Thread.Sleep(300000);
            Task.WaitAll(taskList.ToArray());

            Trace.WriteLine("IP代理池更新完毕！可用IP:" + list.Count);
            if (path.IsEmpty())
                path = AppDomain.CurrentDomain.BaseDirectory + "/XmlConfig/IP.txt";
            if (list.Count > 0)
                CommonHelper.Write_IP(path, JsonConvert.SerializeObject(list.Distinct().ToList()));
        }
        /// <summary>
        /// 代理
        /// </summary>
        /// <param name="DATA"></param>
        private void GetIp(IPEntity model)
        {
            for (int i = 0; i < model.Number; i++)
            {
                var _url = model.Url;
                if (model.Number != 1)
                {
                    if (model.Url.Split('?').Length > 1)
                        _url = model.Url + "&" + model.PageName + "=" + (i + 1);
                    else
                        _url = model.Url + (i == 0 ? string.Empty : "/" + (i + 1));

                }
                var doc = new HtmlDocument();
                doc.LoadHtml(new HttpHelper().GetHtml(new HttpItem { Url = _url, Timeout = 10 * 1000, Method = "GET", ReadWriteTimeout = 10000, ProxyIp = CommonHelper.GetIp(iPAgencyAddress, false) }).Html);
                var number = 0;
                while (string.IsNullOrEmpty(doc.DocumentNode.InnerHtml))
                {
                    if (number >= 20)
                        break;
                    doc.LoadHtml(new HttpHelper().GetHtml(new HttpItem { Url = _url, Timeout = 10 * 1000, Method = "GET", ReadWriteTimeout = 10000, ProxyIp = CommonHelper.GetIp(iPAgencyAddress, false) }).Html);
                    number++;
                }

                var parents = model.SelectParents.Split(',').ToList();
                var trparents = doc.DocumentNode.SelectSingleNode(parents[0]);
                number = 0;
                while (trparents == null)
                {
                    if (number >= 20)
                        break;
                    doc.LoadHtml(new HttpHelper().GetHtml(new HttpItem { Url = _url, Timeout = 10 * 1000, Method = "GET", ReadWriteTimeout = 10000, ProxyIp = CommonHelper.GetIp(iPAgencyAddress, false) }).Html);
                    trparents = doc.DocumentNode.SelectSingleNode(parents[0]);
                    number++;
                }
                if (trparents == null)
                    return;
                var tr = trparents.SelectNodes(parents[1]);
                foreach (HtmlNode item in tr)
                {
                    if (tr.IndexOf(item) != 0)
                    {
                        var td = item.SelectNodes(model.SelectParent);
                        if (td!=null)
                        {
                            var path = model.SelectPath.Split(',').ToList();
                            var ip = td[int.Parse(path[0])].InnerText;
                            if (path.Count == 2)
                            {
                                ip = td[int.Parse(path[0])].InnerText + ":" + td[int.Parse(path[1])].InnerText;
                            }
                            SetData(ip);
                        }
                      
                    }
                }
            }
        }

        /// <summary>
        /// Http请求
        /// </summary>
        /// <param name="uri"></param>
        /// <param name="proxy"></param>
        /// <returns></returns>
        public static string TestIp(Uri uri, string proxy = null)
        {
            var pageSource = string.Empty;
            try
            {
                var watch = new Stopwatch();
                watch.Start();
                var request = (HttpWebRequest)WebRequest.Create(uri);
                request.Accept = "text/html, */*; q=0.01";
                request.ServicePoint.Expect100Continue = false;//加快载入速度
                request.ServicePoint.UseNagleAlgorithm = false;//禁止Nagle算法加快载入速度
                request.AllowWriteStreamBuffering = false;//禁止缓冲加快载入速度
                request.Headers.Add(HttpRequestHeader.AcceptEncoding, "gzip,deflate");//定义gzip压缩页面支持
                request.ContentType = "text/html;charset=UTF-8";//定义文档类型及编码
                request.AllowAutoRedirect = true;//禁止自动跳转
                                                 //设置User-Agent，伪装成Google Chrome浏览器
                request.UserAgent = "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.62 Safari/537.36";
                request.Timeout = 5000;//定义请求超时时间为5秒
                request.KeepAlive = true;//启用长连接
                request.Method = "GET";//定义请求方式为GET
                request.ReadWriteTimeout = 10000;

                if (proxy != null) request.Proxy = new WebProxy(proxy);//设置代理服务器IP，伪装请求地址
                                                                       //request.CookieContainer = this.CookiesContainer;//附加Cookie容器
                request.ServicePoint.ConnectionLimit = int.MaxValue;//定义最大连接数

                using (var response = (HttpWebResponse)request.GetResponse())
                {//获取请求响应

                    // foreach (Cookie cookie in response.Cookies) this.CookiesContainer.Add(cookie);//将Cookie加入容器，保存登录状态

                    if (response.ContentEncoding.ToLower().Contains("gzip"))//解压
                    {
                        using (GZipStream stream = new GZipStream(response.GetResponseStream(), CompressionMode.Decompress))
                        {
                            using (StreamReader reader = new StreamReader(stream, Encoding.UTF8))
                            {
                                pageSource = reader.ReadToEnd();
                            }
                        }
                    }
                    else if (response.ContentEncoding.ToLower().Contains("deflate"))//解压
                    {
                        using (DeflateStream stream = new DeflateStream(response.GetResponseStream(), CompressionMode.Decompress))
                        {
                            using (StreamReader reader = new StreamReader(stream, Encoding.UTF8))
                            {
                                pageSource = reader.ReadToEnd();
                            }

                        }
                    }
                    else
                    {
                        using (Stream stream = response.GetResponseStream())//原始
                        {
                            using (StreamReader reader = new StreamReader(stream, Encoding.UTF8))
                            {
                                pageSource = reader.ReadToEnd();
                            }
                        }
                    }
                }
                request.Abort();
                watch.Stop();
                var threadId = System.Threading.Thread.CurrentThread.ManagedThreadId;//获取当前任务线程ID
                var milliseconds = watch.ElapsedMilliseconds;//获取请求执行时间
                return pageSource;

            }
            catch (Exception ex)
            {
                //Trace.WriteLine("连接失败：" + proxy + "||" + ex.Message);
                return string.Empty;
            }
        }

        /// <summary>
        /// 获取数据
        /// </summary>
        /// <param name="ip"></param>
        /// <returns></returns>
        private async void SetData(string ip)
        {
            taskList.Add(Task.Run(() =>
            {
                Thread.Sleep(100);
                try
                {

                    var url = urlList[new Random().Next(0, urlList.Count)];
                    Trace.WriteLine("url:" + url);
                    var html = "";
                    if (url.ToLower().Contains("nba.nowscore"))
                        html = new HttpHelper().GetHtml(new HttpItem { Url = url, Timeout = 10 * 1000, Method = "GET", Allowautoredirect = true, ReadWriteTimeout = 10000, ProxyIp =ip }).Html;
                    else
                        // var html = TestIp(new Uri(url), ip);
                        html = new HttpHelper().GetHtml(new HttpItem { Url = url, Timeout = 10 * 1000, Method = "GET",ReadWriteTimeout= 10000 ,ProxyIp = ip }).Html;
                    var doc = new HtmlDocument();
                    if (!string.IsNullOrEmpty(html))
                    {
                        var _title = title;
                        if (url.ToLower().Contains("nba.nowscore"))
                            _title = "捷豹";
                        doc.LoadHtml(html);
                        var body = doc.DocumentNode.SelectNodes(".//body");
                        if ((!url.ToLower().Contains("nba.nowscore") && (body == null || string.IsNullOrEmpty(body[0].InnerHtml)))
                        || (html.Contains("403") && html.ToLower().Contains("forbidden"))
                        || (html.Contains("404") && html.ToLower().Contains("not found"))
                        || (!string.IsNullOrEmpty(_title) && !html.ToLower().Contains(_title)))
                        {
                        }
                        else
                        {
                            lock (locker)
                            {
                                list.Add(ip);
                            }
                            Trace.WriteLine("连接成功：" + ip);
                            Console.WriteLine(ip);
                        }
                    }
                    else
                    {
                        //Trace.WriteLine("连接失败：" + ip);
                    }
                }
                finally
                {

                }
            }));

        }

    }

    public class IPEntity
    {
        /// <summary>
        /// 父节点，，也就是ip的父节点
        /// 最后一个为SelectNodes，之前的为SelectSingleNode
        /// </summary>
        public string SelectParents { get; set; }
        /// <summary>
        /// 路径
        /// </summary>
        public string Url { get; set; }

        /// <summary>
        /// 上一级的名称（td||th）
        /// </summary>
        public string SelectParent { get; set; } = "td";

        /// <summary>
        /// 最终Ip的路径 （ip和host）
        /// </summary>
        public string SelectPath { get; set; }
        /// <summary>
        /// 查询页数
        /// </summary>
        public int Number { get; set; }
        /// <summary>
        /// 匿名路径
        /// </summary>
        public int AnonymityPath { get; set; } = -1;
        /// <summary>
        /// 分页参数名称
        /// </summary>
        public string PageName { get; set; }
    }
}