private static void saveCrawlInDatabase(crawlingResult crawledResult)
{
lock (_lockDatabaseAdd)
{
using (var context = new DBCrawling())
{
crawledResult.UrlHash = crawledResult.Url.ComputeHashOfOurSystem();
crawledResult.HostUrl = crawledResult.Url.returnRootUrl();
var vrResult = context.tblMainUrls.SingleOrDefault(b => b.UrlHash == crawledResult.UrlHash);
crawledResult.ParentUrlHash = crawledResult.ParentUrlHash.ComputeHashOfOurSystem();
if (crawledResult.blcrawlSuccess == true)
{
crawledResult.IsCrawled = true;
if (!string.IsNullOrEmpty(crawledResult.SourceCode))
{
double dblOriginalSourceCodeLenght = crawledResult.SourceCode.Length;
crawledResult.SourceCode = crawledResult.SourceCode.CompressString();
crawledResult.CompressionPercent = Convert.ToByte(
Math.Floor(
((crawledResult.SourceCode.Length.ToDouble() / dblOriginalSourceCodeLenght) * 100))
);
}
crawledResult.CrawlTryCounter = 0;
}
tblMainUrl finalObject = crawledResult.converToBaseMainUrlClass();
//this approach brings extra overhead to the server with deleting from server first
//therefore will use copy properties of object to another object without changing reference
//if (vrResult != null)
//{
// context.tblMainUrls.Remove(vrResult);
// context.SaveChanges();
//}
if (vrResult != null)
{
finalObject.DiscoverDate = vrResult.DiscoverDate;
finalObject.LinkDepthLevel = vrResult.LinkDepthLevel;
finalObject.CrawlTryCounter = vrResult.CrawlTryCounter;
if (crawledResult.blcrawlSuccess == false)
{
finalObject.CrawlTryCounter++;
}
finalObject.CopyProperties(vrResult);
}
else
{
context.tblMainUrls.Add(finalObject);
}
var gg = context.SaveChanges();
}
}
}
private static void extractLinks(crawlingResult myCrawlingResult, HtmlDocument doc)
{
var baseUri = new Uri(myCrawlingResult.Url);
// extracting all links
var vrNodes = doc.DocumentNode.SelectNodes("//a[@href]");
if (vrNodes != null)
{
foreach (HtmlNode link in vrNodes)//xpath notation
{
HtmlAttribute att = link.Attributes["href"];
//this is used to convert from relative path to absolute path
var absoluteUri = new Uri(baseUri, att.Value.ToString().decodeUrl());
if (!absoluteUri.ToString().StartsWith("http://") && !absoluteUri.ToString().StartsWith("https://"))
{
continue;
}
myCrawlingResult.lstDiscoveredLinks.Add(absoluteUri.ToString().Split('#').FirstOrDefault());
}
}
myCrawlingResult.lstDiscoveredLinks = myCrawlingResult.lstDiscoveredLinks.Distinct().Where(pr => pr.Length < 201).ToList();
var vrDocTitle = doc.DocumentNode.SelectSingleNode("//title")?.InnerText.ToString().Trim();
vrDocTitle = System.Net.WebUtility.HtmlDecode(vrDocTitle);
myCrawlingResult.PageTile = vrDocTitle;
}
public static void crawlPage(string srUrlToCrawl, int irUrlDepthLevel, string _srParentUrl, DateTime _dtDiscoverDate)
{
var vrLocalUrl = srUrlToCrawl;
crawlingResult crawlResult = new crawlingResult();
crawlResult.Url = vrLocalUrl;
if (!string.IsNullOrEmpty(_srParentUrl))
{
crawlResult.ParentUrlHash = _srParentUrl;
}
if (_dtDiscoverDate != DateTime.MinValue)
{
crawlResult.DiscoverDate = _dtDiscoverDate;
}
Stopwatch swTimerCrawling = new Stopwatch();
swTimerCrawling.Start();
HtmlWeb wbClient = new HtmlWeb();//you should use httpwebrequest for more control and better performance
wbClient.AutoDetectEncoding = true;
wbClient.BrowserTimeout = new TimeSpan(0, 2, 0);
HtmlAgilityPack.HtmlDocument doc = new HtmlAgilityPack.HtmlDocument();
try
{
doc = wbClient.Load(crawlResult.Url);
crawlResult.SourceCode = doc.Text;
}
catch (Exception E)
{
crawlResult.blcrawlSuccess = false;
logError(E, "crawlPage");
}
Interlocked.Increment(ref irCrawledUrlCount);
swTimerCrawling.Stop();
crawlResult.FetchTimeMS = Convert.ToInt32(swTimerCrawling.ElapsedMilliseconds);
crawlResult.LastCrawlingDate = DateTime.Now;
saveCrawlInDatabase(crawlResult);
if (crawlResult.blcrawlSuccess)
{
extractLinks(crawlResult, doc);
saveDiscoveredLinksInDatabaseForFutureCrawling(crawlResult);
}
doc = null;
}
private static void saveDiscoveredLinksInDatabaseForFutureCrawling(crawlingResult crawlResult)
{
lock (_lockDatabaseAdd)
{
using (var context = new DBCrawling())
{
HashSet <string> hsProcessedUrls = new HashSet <string>();
foreach (var vrPerLink in crawlResult.lstDiscoveredLinks)
{
var vrHashedLink = vrPerLink.ComputeHashOfOurSystem();
if (hsProcessedUrls.Contains(vrHashedLink))
{
continue;
}
var vrResult = context.tblMainUrls.Any(databaseRecord => databaseRecord.UrlHash == vrHashedLink);
if (vrResult == false)
{
crawlingResult newLinkCrawlingResult = new crawlingResult();
newLinkCrawlingResult.Url = vrPerLink.normalizeUrl();
newLinkCrawlingResult.HostUrl = newLinkCrawlingResult.Url.returnRootUrl();
newLinkCrawlingResult.UrlHash = vrPerLink.ComputeHashOfOurSystem();
newLinkCrawlingResult.ParentUrlHash = crawlResult.UrlHash;
newLinkCrawlingResult.LinkDepthLevel = (short)(crawlResult.LinkDepthLevel + 1);
context.tblMainUrls.Add(newLinkCrawlingResult.converToBaseMainUrlClass());
hsProcessedUrls.Add(vrHashedLink);
Interlocked.Increment(ref irDiscoveredUrlCount);
}
}
context.SaveChanges();
}
}
}
