- 전자공시 추가
- 타임아웃 적용
This commit is contained in:
135
Form1.cs
135
Form1.cs
@@ -49,17 +49,19 @@ namespace NewsCrawler
|
||||
{
|
||||
bool bHasNew = false;
|
||||
|
||||
try
|
||||
{
|
||||
string strServerURL = "http://www.asiae.co.kr";
|
||||
WebRequest request = WebRequest.Create("http://www.asiae.co.kr/news/sokbo/sokbo_left.htm");
|
||||
//request.Method = "POST";
|
||||
//request.
|
||||
request.Credentials=CredentialCache.DefaultCredentials;
|
||||
request.Timeout=2000;
|
||||
|
||||
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
|
||||
Console.WriteLine(response.StatusDescription);
|
||||
|
||||
Stream dataStream = response.GetResponseStream();
|
||||
StreamReader reader = new StreamReader(dataStream, Encoding.GetEncoding("EUC-KR"));
|
||||
using(HttpWebResponse response = (HttpWebResponse)request.GetResponse())
|
||||
{
|
||||
using(Stream dataStream = response.GetResponseStream())
|
||||
{
|
||||
using(StreamReader reader = new StreamReader(dataStream, Encoding.GetEncoding("EUC-KR")))
|
||||
{
|
||||
string responseFromServer = WebUtility.HtmlDecode(reader.ReadToEnd());
|
||||
|
||||
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
|
||||
@@ -73,20 +75,20 @@ namespace NewsCrawler
|
||||
string strURL = strServerURL+item.SelectSingleNode(".//a").GetAttributeValue("href", "");
|
||||
|
||||
|
||||
if (Regex.IsMatch(strTime, @"\d+/\d+") == true)
|
||||
if(Regex.IsMatch(strTime, @"\d+/\d+")==true)
|
||||
{
|
||||
//Console.WriteLine("어제 기사 : " + item.InnerHtml);
|
||||
continue;
|
||||
}
|
||||
|
||||
|
||||
if (m_ItemList.Any(c => c.m_strURL==strURL)==false)
|
||||
if(m_ItemList.Any(c => c.m_strURL==strURL)==false)
|
||||
{
|
||||
NEWS_ITEM news = new NEWS_ITEM();
|
||||
news.m_strTitle=strTitle;
|
||||
news.m_Time=DateTime.ParseExact(strTime, "HH:mm", CultureInfo.CurrentCulture);
|
||||
news.m_strURL=strURL;
|
||||
news.m_strReference = "아시아경제";
|
||||
news.m_strReference="아시아경제";
|
||||
|
||||
m_ItemList.Add(news);
|
||||
|
||||
@@ -94,9 +96,14 @@ namespace NewsCrawler
|
||||
}
|
||||
}
|
||||
|
||||
reader.Close();
|
||||
dataStream.Close();
|
||||
response.Close();
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
catch(Exception e)
|
||||
{
|
||||
Console.WriteLine(e.ToString());
|
||||
}
|
||||
|
||||
return bHasNew;
|
||||
}
|
||||
@@ -105,15 +112,19 @@ namespace NewsCrawler
|
||||
{
|
||||
bool bHasNew = false;
|
||||
|
||||
try
|
||||
{
|
||||
//string strServerURL = "http://www.etoday.co.kr";
|
||||
WebRequest request = WebRequest.Create("http://www.etoday.co.kr/news/flash/flash_list.php?tab=2");
|
||||
WebRequest request = WebRequest.Create("http://www.etoday.co.kr/news/flash/flash_list.php?tab=1");
|
||||
request.Credentials=CredentialCache.DefaultCredentials;
|
||||
request.Timeout=2000;
|
||||
|
||||
HttpWebResponse response = (HttpWebResponse)request.GetResponse();
|
||||
Console.WriteLine(response.StatusDescription);
|
||||
|
||||
Stream dataStream = response.GetResponseStream();
|
||||
StreamReader reader = new StreamReader(dataStream, Encoding.GetEncoding("utf-8"));
|
||||
using(HttpWebResponse response = (HttpWebResponse)request.GetResponse())
|
||||
{
|
||||
using(Stream dataStream = response.GetResponseStream())
|
||||
{
|
||||
using(StreamReader reader = new StreamReader(dataStream, Encoding.GetEncoding("utf-8")))
|
||||
{
|
||||
string responseFromServer = WebUtility.HtmlDecode(reader.ReadToEnd());
|
||||
|
||||
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
|
||||
@@ -125,31 +136,102 @@ namespace NewsCrawler
|
||||
string strTitle = item.SelectSingleNode(".//a").InnerText;
|
||||
string strTime = item.SelectSingleNode(".//span[contains(@class, 'flash_press')]").InnerText;
|
||||
string strURL = item.SelectSingleNode(".//a").GetAttributeValue("href", "");
|
||||
strURL = "http://www.etoday.co.kr/news/section/newsview.php?idxno="+Regex.Replace(strURL, @"\D", "");
|
||||
strURL="http://www.etoday.co.kr/news/section/newsview.php?idxno="+Regex.Replace(strURL, @"\D", "");
|
||||
|
||||
if (Regex.IsMatch(strTime, @"\d+/\d+") == true)
|
||||
if(Regex.IsMatch(strTime, @"\d+/\d+")==true)
|
||||
{
|
||||
//Console.WriteLine("어제 기사 : " + item.InnerHtml);
|
||||
continue;
|
||||
}
|
||||
|
||||
if (m_ItemList.Any(c => c.m_strURL==strURL)==false)
|
||||
if(m_ItemList.Any(c => c.m_strURL==strURL)==false)
|
||||
{
|
||||
NEWS_ITEM news = new NEWS_ITEM();
|
||||
news.m_strTitle=strTitle;
|
||||
news.m_Time=DateTime.ParseExact(strTime, "HH:mm", CultureInfo.CurrentCulture);
|
||||
news.m_strURL=strURL;
|
||||
news.m_strReference = "이투데이";
|
||||
news.m_strReference="이투데이";
|
||||
|
||||
m_ItemList.Add(news);
|
||||
|
||||
bHasNew=true;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
catch(Exception e)
|
||||
{
|
||||
Console.WriteLine(e.ToString());
|
||||
}
|
||||
|
||||
reader.Close();
|
||||
dataStream.Close();
|
||||
response.Close();
|
||||
return bHasNew;
|
||||
}
|
||||
|
||||
bool ReadDart()
|
||||
{
|
||||
bool bHasNew = false;
|
||||
|
||||
try
|
||||
{
|
||||
string strServerURL = "https://dart.fss.or.kr";
|
||||
WebRequest request = WebRequest.Create("https://dart.fss.or.kr/dsac001/mainAll.do");
|
||||
request.Credentials=CredentialCache.DefaultCredentials;
|
||||
request.Timeout=2000;
|
||||
|
||||
using(HttpWebResponse response = (HttpWebResponse)request.GetResponse())
|
||||
{
|
||||
using(Stream dataStream = response.GetResponseStream())
|
||||
{
|
||||
using(StreamReader reader = new StreamReader(dataStream, Encoding.GetEncoding("utf-8")))
|
||||
{
|
||||
string responseFromServer = WebUtility.HtmlDecode(reader.ReadToEnd());
|
||||
|
||||
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
|
||||
doc.LoadHtml(responseFromServer);
|
||||
|
||||
string strXPath = "//div[@id='listContents']/div[contains(@class, 'table_list')]/table/tr";
|
||||
var lists = doc.DocumentNode.SelectNodes(strXPath);
|
||||
foreach(var item in lists)
|
||||
{
|
||||
var rows = item.SelectNodes(".//td");
|
||||
|
||||
string strTitle = rows[2].InnerText;
|
||||
strTitle=strTitle.Trim();
|
||||
string strTime = item.SelectSingleNode(".//td[contains(@class, 'cen_txt')]").InnerText;
|
||||
strTime=strTime.Trim();
|
||||
string strURL = rows[2].SelectSingleNode(".//a").GetAttributeValue("href", "");
|
||||
strURL=strURL.Trim();
|
||||
strURL =strServerURL+strURL;
|
||||
|
||||
if(Regex.IsMatch(strTime, @"\d+/\d+")==true)
|
||||
{
|
||||
//Console.WriteLine("어제 기사 : " + item.InnerHtml);
|
||||
continue;
|
||||
}
|
||||
|
||||
if(m_ItemList.Any(c => c.m_strURL==strURL)==false)
|
||||
{
|
||||
NEWS_ITEM news = new NEWS_ITEM();
|
||||
news.m_strTitle=strTitle;
|
||||
news.m_Time=DateTime.ParseExact(strTime, "HH:mm", CultureInfo.CurrentCulture);
|
||||
news.m_strURL=strURL;
|
||||
news.m_strReference="전자공시";
|
||||
|
||||
m_ItemList.Add(news);
|
||||
|
||||
bHasNew=true;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
catch(Exception e)
|
||||
{
|
||||
Console.WriteLine(e.ToString());
|
||||
}
|
||||
|
||||
return bHasNew;
|
||||
}
|
||||
@@ -160,6 +242,7 @@ namespace NewsCrawler
|
||||
|
||||
if(chAutoReload.Checked == true)
|
||||
{
|
||||
ReadDart();
|
||||
ReadAsiaE();
|
||||
ReadEtoday();
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user