public static string GetPageString(string url)
{
string res = null;
int t = 0;
do
{
t++;
try
{
var request = WebRequest.Create(url) as HttpWebRequest;
request.Method = "GET";
request.UserAgent = "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/60.0.3112.113 Safari/537.36";
using (var response = request.GetResponse())
{
using (var resStream = response.GetResponseStream())
{
using (var sr = new StreamReader(resStream, Encoding.GetEncoding("GBK")))
{
res = sr.ReadToEnd();
}
}
}
}
catch (Exception ex)
{
if (t == 2)
{
LogHelp.Log("=================================页面获取失败：" + url, true);
}
res = null;
}
} while ((res == null || res.Length == 0) && t <= 2);
if (res == null || res.Length == 0)
{
MyDbCOntextHelp.AddErroeProcess(new Data.ErroeProcess
{
CreateTime = DateTime.Now,
Type = 3,
Url = url
});
}
return(res);
}
public static async Task <string> GetPageStringAsync(string url)
{
string res = null;
int t = 0;
do
{
t++;
try
{
var respMsg = await client.GetAsync(url);
using (var resStream = await respMsg.Content.ReadAsStreamAsync())
{
using (var sr = new StreamReader(resStream, Encoding.GetEncoding("GBK")))
{
res = await sr.ReadToEndAsync();
}
}
}
catch (Exception ex)
{
if (t == 5)
{
LogHelp.Log("=================================页面获取失败：" + url, true);
LogHelp.Log("=================================" + ex.Message);
LogHelp.Log("=================================", true);
}
res = null;
}
} while ((res == null || res.Length == 0) && t <= 5);
if (res == null || res.Length == 0)
{
MyDbCOntextHelp.AddErroeProcess(new Data.ErroeProcess
{
CreateTime = DateTime.Now,
Type = 3,
Url = url
});
}
return(res);
}
private async void ProcessDetailAsync(object stat)
{
try
{
var item = stat as Match;
MoviePage pageModel = null;
string movieName = ValidFileName(item.Groups[2].Value);
var movieUrl = _uri.Scheme + "://" + _uri.Authority + "/bbs/" + item.Groups[1].Value;
string moviedir = null;
if (MyDbCOntextHelp.ExistMovie(movieUrl))
{
pageModel = MyDbCOntextHelp.QueryMovie(m => m.Url == movieUrl);
moviedir = Path.Combine(_baseDir, pageModel.Name);
}
else
{
pageModel = new MoviePage
{
CreateTime = DateTime.Now,
UpdateTime = DateTime.Now,
Name = movieName,
Url = movieUrl,
IsHandler = false,
Type = _typeName,
};
LogHelp.Log("影片：" + pageModel.Name);
moviedir = Path.Combine(_baseDir, pageModel.Name);
}
if (!Directory.Exists(moviedir))
{
Directory.CreateDirectory(moviedir);
}
if (pageModel.Id != 0 || MyDbCOntextHelp.AddPicturePage(pageModel))
{
bool res = false;
var detailPageString = await HttpHelp.GetPageStringAsync(pageModel.Url);
if (detailPageString == null || detailPageString.Length == 0 || detailPageString.Contains("您无权进行当前操作，这可能因以下原因之一造成"))
{
return;
}
//下载次数
var m = _regDownloadCount.Match(detailPageString);
if (m.Groups.Count == 2)
{
pageModel.DownloadCount = int.Parse(m.Groups[1].Value);
}
LogHelp.Log("影片下载次数：" + pageModel.DownloadCount);
//图片
var imgAreaStr = _regImgArea.Match(detailPageString).Value;
var mimgs = _regImg.Matches(imgAreaStr);
var resList = new List <Resource>(mimgs.Count + 1);
for (int j = 0; j < mimgs.Count; j++)
{
var imgitem = mimgs[j];
var r = new Resource
{
CreateTime = DateTime.Now,
UpdateTime = DateTime.Now,
PicturePageId = pageModel.Id,
IsHandler = false,
Type = 1,
Url = imgitem.Groups[1].Value
};
if (!r.Url.StartsWith("http"))
{
r.Url = _uri.Scheme + "://" + _uri.Authority + "/bbs/" + r.Url;
}
resList.Add(r);
}
LogHelp.Log("截图{0}张.....bt文件一个", mimgs.Count);
//bt 检查重复
var btRes = new Resource
{
CreateTime = DateTime.Now,
UpdateTime = DateTime.Now,
PicturePageId = pageModel.Id,
Type = 2,
IsHandler = false,
};
var p1 = detailPageString.IndexOf("检查重复</a>");
if (p1 == -1)
{
return;
}
var startbt = detailPageString.IndexOf("<a href=\"", p1);
var endbt = detailPageString.IndexOf("</a>", startbt);
var bta = detailPageString.Substring(startbt, endbt - startbt + 4);
var mbt = _regBt.Match(bta);
btRes.Name = mbt.Groups[2].Value;
btRes.Url = _uri.Scheme + "://" + _uri.Authority + "/bbs/" + mbt.Groups[1].Value;
resList.Add(btRes);
bool isAllHandle = false;
MyDbCOntextHelp.AddResourceList(resList);
for (int i = 0; i < resList.Count - 1; i++)
{
res = await HttpHelp.DownloadImgAsync(resList[i].Url, moviedir);
if (res)
{
}
}
res = await HttpHelp.DownloadFileAsync(btRes.Url, Path.Combine(moviedir, btRes.Name));
}
LogHelp.Log("处理完毕：" + movieName);
return;
}
finally
{
_semaphore.Release();
}
}
public static async Task <bool> DownloadImgAsync(string url, string path)
{
bool res = false;
int t = 0;
do
{
t++;
try
{
var fileName = Path.Combine(path, Path.GetFileName(url));//如果文件存在，则不需要再接受，直接返回
if (File.Exists(fileName) && new FileInfo(fileName).Length != 0)
{
return(true);
}
var respMsg = await client.GetAsync(url);
using (var resStream = await respMsg.Content.ReadAsStreamAsync())
{
//如果文件存在，则不需要再接受，直接返回
if (File.Exists(fileName) && new FileInfo(fileName).Length == respMsg.Content.Headers.ContentLength)
{
return(true);
}
using (var FileStream = new FileStream(fileName, FileMode.Create, FileAccess.Write, FileShare.ReadWrite, 4096, true))
{
byte[] buf = new byte[1024 * 1024];
int blen = 0;
do
{
blen = await resStream.ReadAsync(buf, 0, buf.Length);
if (blen != 0)
{
await FileStream.WriteAsync(buf, 0, blen);
}
} while (blen != 0);
}
res = true;
}
}
catch (Exception ex)
{
if (t == 2)
{
LogHelp.Log("=================================图片获取失败：" + url, true);
}
res = false;
}
} while (res == false && t <= 2);
if (!res)
{
MyDbCOntextHelp.AddErroeProcess(new Data.ErroeProcess
{
CreateTime = DateTime.Now,
Type = 1,
Url = url
});
}
return(res);
}
public static bool DownloadImg(string url, string path)
{
bool res = false;
int t = 0;
do
{
t++;
try
{
var fileName = Path.Combine(path, Path.GetFileName(url));//如果文件存在，则不需要再接受，直接返回
if (File.Exists(fileName) && new FileInfo(fileName).Length != 0)
{
return(true);
}
var request = WebRequest.Create(url) as HttpWebRequest;
request.Method = "GET";
request.UserAgent = "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/60.0.3112.113 Safari/537.36";
using (var response = request.GetResponse())
{
using (var resStream = response.GetResponseStream())
{
//如果文件存在，则不需要再接受，直接返回
if (File.Exists(fileName) && new FileInfo(fileName).Length == response.ContentLength)
{
return(true);
}
using (var FileStream = new FileStream(fileName, FileMode.Create, FileAccess.Write))
{
byte[] buf = new byte[8192];
int blen = 0;
do
{
blen = resStream.Read(buf, 0, buf.Length);
if (blen != 0)
{
FileStream.Write(buf, 0, blen);
}
} while (blen != 0);
}
res = true;
}
}
}
catch (Exception ex)
{
if (t == 2)
{
LogHelp.Log("=================================图片获取失败：" + url, true);
}
res = false;
}
} while (res == false && t <= 2);
if (!res)
{
MyDbCOntextHelp.AddErroeProcess(new Data.ErroeProcess
{
CreateTime = DateTime.Now,
Type = 1,
Url = url
});
}
return(res);
}
