public void Start()
{
Task.Factory.StartNew(() =>
{
_action(string.Format("开始{0}to{1}....", _startIndex, _endIndex));
for (int i = _startIndex; i < _endIndex + 1; i++)
{
HttpLibSyncRequestItem item = new HttpLibSyncRequestItem();
item.Encoding = "gb2312";
item.Url = string.Format(_url, i);
if (i == 1) { item.Url = item.Url.Replace("_1.html", ".html"); }
var html = HttpLibSyncRequest.Get(item);
Regex regex = new Regex("<a.target=._blank..href=.(?<url>.*?)..alt=.(?<title>.*?).><img.src2=.(?<img>.*?)..alt=.*?></a>", RegexOptions.IgnoreCase);
MatchCollection mc = regex.Matches(html);
foreach (Match match in mc)
{
HtmlTemplate tpl = new HtmlTemplate();
item = new HttpLibSyncRequestItem();
item.Encoding = "gb2312";
item.Url = "http://sc.chinaz.com" + match.Groups["url"].ToString();
tpl.HtmlTemplateThumbnailUrl = match.Groups["img"].ToString();
tpl.HtmlTemplateName = match.Groups["title"].ToString();
html = HttpLibSyncRequest.Get(item);
regex = new Regex("http.*?.rar", RegexOptions.IgnoreCase);
mc = regex.Matches(html);
List<string> durls = new List<string>();
foreach (Match durl in mc)
{
durls.Add(durl.ToString());
}
tpl.HtmlTemplateDownloadUrls = string.Join("|", durls);
tpl.HtmlTemplateAlias = durls[0].GetFileNameWithoutExtension();
regex = new Regex("<span>类别.*?>(?<category>.*?)</a></span>", RegexOptions.IgnoreCase);
Match category = regex.Match(html);
tpl.HtmlTemplateCategoryName = category.Groups["category"].ToString();
tpl.HtmlTemplateCategoryAlias = SpellCodeHelper.GetFirstPYLetter(tpl.HtmlTemplateCategoryName);
regex = new Regex("<a href=.(?<bigimg>.*?)..title.*?class=\"image_gall\"><img", RegexOptions.IgnoreCase);
Match bimg = regex.Match(html);
tpl.HtmlTemplateImageUrl = bimg.Groups["bigimg"].ToString();
regex = new Regex("<em>(?<datetime>.*?)</em>", RegexOptions.IgnoreCase);
Match dt = regex.Match(html);
tpl.HtmlTemplateCreateTime = dt.Groups["datetime"].ToString().ToDataTime();
if (DMContext.Query<HtmlTemplate>().Where(p => p.HtmlTemplateName == tpl.HtmlTemplateName).Single() == null)
{
DMContext.Insert(tpl);
_action(string.Format("成功采集[{0}]{1}", tpl.HtmlTemplateCategoryName, tpl.HtmlTemplateName));
}
else
{
_action(string.Format("已经存在[{0}]{1}", tpl.HtmlTemplateCategoryName, tpl.HtmlTemplateName));
}
}
_action("ok");
}
_action(string.Format("完成{0}to{1}....", _startIndex, _endIndex));
});
}
public static string Get(HttpLibSyncRequestItem objhttpItem)
{
#region 初始设置
//默认的编码
Encoding encoding = Encoding.UTF8;
//读取流的对象
StreamReader reader = null;
//HttpWebRequest对象用来发起请求
HttpWebRequest request = null;
//获取影响流的数据对象
HttpWebResponse response = null;
//需要返回的数据对象
string returnData = "String Error";
#region 验证证书
if (!string.IsNullOrEmpty(objhttpItem.CerPath))
{
//这一句一定要写在创建连接的前面。使用回调的方法进行证书验证。
ServicePointManager.ServerCertificateValidationCallback =
new System.Net.Security.RemoteCertificateValidationCallback(CheckValidationResult);
//初始化对像，并设置请求的URL地址
request = (HttpWebRequest)WebRequest.Create(GetUrl(objhttpItem.Url));
//创建证书文件
X509Certificate objx509 = new X509Certificate(objhttpItem.CerPath);
//添加到请求里
request.ClientCertificates.Add(objx509);
}
else
{
//初始化对像，并设置请求的URL地址
request = (HttpWebRequest)WebRequest.Create(GetUrl(objhttpItem.Url));
}
#endregion 验证证书
#region 设置代理
if (string.IsNullOrEmpty(objhttpItem.ProxyUserName) && string.IsNullOrEmpty(objhttpItem.ProxyPwd) && string.IsNullOrEmpty(objhttpItem.ProxyIp))
{
//不需要设置
}
else
{
//设置代理服务器
WebProxy myProxy = new WebProxy(objhttpItem.ProxyIp, false);
//建议连接
myProxy.Credentials = new NetworkCredential(objhttpItem.ProxyUserName, objhttpItem.ProxyPwd);
//给当前请求对象
request.Proxy = myProxy;
//设置安全凭证
request.Credentials = CredentialCache.DefaultNetworkCredentials;
}
#endregion 设置代理
//请求方式Get或者Post
request.Method = objhttpItem.Method;
request.Timeout = objhttpItem.Timeout;
request.ReadWriteTimeout = objhttpItem.ReadWriteTimeout;
//Accept
request.Accept = objhttpItem.Accept;
//ContentType返回类型
request.ContentType = objhttpItem.ContentType;
//UserAgent客户端的访问类型，包括浏览器版本和操作系统信息
request.UserAgent = objhttpItem.UserAgent;
#region 编码
if (string.IsNullOrEmpty(objhttpItem.Encoding) || objhttpItem.Encoding.ToLower().Trim() == "null")
{
//读取数据时的编码方式
encoding = Encoding.UTF8;
}
else
{
//读取数据时的编码方式
encoding = System.Text.Encoding.GetEncoding(objhttpItem.Encoding);
}
#endregion 编码
#region Cookie
if (!string.IsNullOrEmpty(objhttpItem.Cookie))
{
//Cookie
request.Headers[HttpRequestHeader.Cookie] = objhttpItem.Cookie;
}
//设置Cookie
if (objhttpItem.CookieCollection != null)
{
if (request.CookieContainer.Count == 0)
{
request.CookieContainer.Add(objhttpItem.CookieCollection);
}
else
{
request.CookieContainer = new CookieContainer();
request.CookieContainer.Add(objhttpItem.CookieCollection);
}
}
#endregion Cookie
//来源地址
request.Referer = objhttpItem.Referer;
//是否执行跳转功能
request.AllowAutoRedirect = objhttpItem.Allowautoredirect;
#region Post数据
//验证在得到结果时是否有传入数据
if (!string.IsNullOrEmpty(objhttpItem.Postdata) && request.Method.Trim().ToLower().Contains("post"))
{
byte[] buffer = encoding.GetBytes(objhttpItem.Postdata);
request.ContentLength = buffer.Length;
request.GetRequestStream().Write(buffer, 0, buffer.Length);
}
#endregion Post数据
//设置最大连接
if (objhttpItem.Connectionlimit > 0)
{
request.ServicePoint.ConnectionLimit = objhttpItem.Connectionlimit;
}
#endregion
#region 读取数据
try
{
using (response = (HttpWebResponse)request.GetResponse())
{
if (response.Cookies != null)
{
objhttpItem.CookieCollection = response.Cookies;
}
if (response.Headers["set-cookie"] != null)
{
objhttpItem.Cookie = response.Headers["set-cookie"];
}
objhttpItem.Response = response;
objhttpItem.Request = request;
//从这里开始我们要无视编码了
if (encoding == null)
{
MemoryStream stream = new MemoryStream();
if (response.ContentEncoding != null && response.ContentEncoding.Equals("gzip", StringComparison.InvariantCultureIgnoreCase))
{
objhttpItem.Reader = reader;
//开始读取流并设置编码方式
//new GZipStream(response.GetResponseStream(), CompressionMode.Decompress).CopyTo(stream, 10240);
//.net4.0以下写法
stream = GetMemoryStream(response.GetResponseStream());
}
else
{
objhttpItem.Reader = reader;
//response.GetResponseStream().CopyTo(stream, 10240);
// .net4.0以下写法
stream = GetMemoryStream(response.GetResponseStream());
}
byte[] RawResponse = stream.ToArray();
string temp = Encoding.Default.GetString(RawResponse, 0, RawResponse.Length);
//<meta(.*?)charset([\s]?)=[^>](.*?)>
Match meta = Regex.Match(temp, "<meta([^<]*)charset=([^<]*)[\"']", RegexOptions.IgnoreCase | RegexOptions.Multiline);
string charter = (meta.Groups.Count > 2) ? meta.Groups[2].Value : string.Empty;
charter = charter.Replace("\"", string.Empty).Replace("'", string.Empty).Replace(";", string.Empty);
if (charter.Length > 0)
{
charter = charter.ToLower().Replace("iso-8859-1", "gbk");
encoding = Encoding.GetEncoding(charter);
}
else
{
if (response.CharacterSet.ToLower().Trim() == "iso-8859-1")
{
encoding = Encoding.GetEncoding("gbk");
}
else
{
if (string.IsNullOrEmpty(response.CharacterSet.Trim()))
{
encoding = Encoding.UTF8;
}
else
{
encoding = Encoding.GetEncoding(response.CharacterSet);
}
}
}
returnData = encoding.GetString(RawResponse);
}
else
{
if (response.ContentEncoding != null && response.ContentEncoding.Equals("gzip", StringComparison.InvariantCultureIgnoreCase))
{
//开始读取流并设置编码方式
using (reader = new StreamReader(new GZipStream(response.GetResponseStream(), CompressionMode.Decompress), encoding))
{
objhttpItem.Reader = reader;
returnData = reader.ReadToEnd();
}
}
else
{
//开始读取流并设置编码方式
using (reader = new StreamReader(response.GetResponseStream(), encoding))
{
objhttpItem.Reader = reader;
returnData = reader.ReadToEnd();
}
}
}
}
}
catch (WebException ex)
{
//这里是在发生异常时返回的错误信息
returnData = "String Error";
response = (HttpWebResponse)ex.Response;
objhttpItem.Response = response;
}
if (objhttpItem.IsToLower)
{
returnData = returnData.ToLower();
}
return returnData;
#endregion
}
private void GetAccessToken()
{
HttpLibSyncRequestItem item = new HttpLibSyncRequestItem();
item.Url = string.Format("{0}token?grant_type=client_credential&appid={1}&secret={2}", _config.ApiUrl, _config.AppId, _config.AppSecret);
string json = HttpLibSyncRequest.Get(item);
var data = new WeChatData(json);
_config.AccessToken = data["access_token"];
_config.ExpiresIn = DateTime.Now.AddSeconds(data.Get<int>("expires_in"));
}