public static RobotsTextResponse ParseText(string text)
{
var response = new RobotsTextResponse();
var startedParsing = false;
UserAgent currentAgent = null;
int index = 0;
while (index < text.Length)
{
var line = GetLine(text, ref index);
if (line.Length == 0)
{
if (startedParsing)
{
startedParsing = false;
}
continue;
}
if (line.StartsWith("User-agent:") && line.Length > 12 && (line[12] == '*' || IsMatchingUserAgent(line)))
{
if (line[12] == '*')
{
response.AnyAgent = new UserAgent();
currentAgent = response.AnyAgent;
currentAgent.Name = "*";
startedParsing = true;
}
else if (IsMatchingUserAgent(line))
{
response.MyAgent = new UserAgent();
currentAgent = response.MyAgent;
currentAgent.Name = line.Substring(12);
startedParsing = true;
}
else
{
continue;
}
}
if (!startedParsing)
{
continue;
}
if (line.StartsWith("Disallow:") && line.Length > 10)
{
currentAgent.DisallowedUrls.Add(line.Substring(10));
}
else if (line.StartsWith("Allow:") && line.Length > 7)
{
currentAgent.AllowedUrls.Add(line.Substring(7));
}
}
return(response);
}
public static bool IsAllowed(RobotsTextResponse robots, Uri baseUrl)
{
if (robots == null)
{
return(true);
}
var allowed = true;
//check my agent first
if (robots.MyAgent != null)
{
allowed = IsAllowedUserAgent(robots.MyAgent, baseUrl);
}
else if (robots.AnyAgent != null)
{
allowed = IsAllowedUserAgent(robots.AnyAgent, baseUrl);
}
return(allowed);
}
public static RobotsTextResponse GetRobotsResponse(Uri uri, int retries = 0, string cacheKey = null)
{
if (cacheKey == null)
{
//in case of redirects, keep same key
cacheKey = uri.Authority;
}
if (retries == RunSettings.MaxRetries)
{
ApplicationCache.Robots.Add(cacheKey, null);
return(null);
}
retries++;
var robots = new Uri($"{uri.Scheme}://{uri.Authority}/robots.txt");
var page = Download(robots);
if (page == null)
{
ApplicationCache.Robots.Add(cacheKey, null);
return(null);
}
if (page.StatusCode != HttpStatusCode.OK && page.StatusCode != HttpStatusCode.NotModified)
{
if ((page.StatusCode == HttpStatusCode.Moved || page.StatusCode == HttpStatusCode.Redirect &&
page.StatusCode == HttpStatusCode.TemporaryRedirect) && page.Headers.ContainsKey("Location"))
{
return(GetRobotsResponse(new Uri(page.Headers["Location"]), retries, cacheKey));
}
ApplicationCache.Robots.Add(cacheKey, null);
return(null);
}
var parsedRobots = RobotsTextResponse.ParseText(page.Html);
ApplicationCache.Robots.Add(cacheKey, parsedRobots);
return(parsedRobots);
}
private void ProcessUrl(string siteUrl, int retries = 0)
{
if (retries == RunSettings.MaxRetries)
{
return;
}
retries++;
//Console.WriteLine($"Crawling {siteUrl}");
var baseUrl = new Uri(siteUrl);
//get robots
RobotsTextResponse robots = null;
if (!ApplicationCache.Robots.TryGetValue(baseUrl.Authority, out robots))
{
robots = TcpCrawlingClient.GetRobotsResponse(baseUrl);
}
if (ApplicationCache.Robots.Count > RunSettings.MaxRobotsSize)
{
ApplicationCache.TrimRobots();
}
//check robots
if (!RobotsTextResponse.IsAllowed(robots, baseUrl))
{
return;
}
var page = TcpCrawlingClient.Download(baseUrl);
if (ApplicationCache.VisitedUrls.Count > RunSettings.MaxVisitedUrls)
{
ApplicationCache.TrimVisitedUrls();
}
ApplicationCache.VisitedUrls.Add(siteUrl);
if (page == null)
{
return;
}
if ((page.StatusCode == HttpStatusCode.MovedPermanently || page.StatusCode == HttpStatusCode.TemporaryRedirect) && page.Headers.ContainsKey("Location"))
{
ProcessRedirectedLocation(page, baseUrl, retries);
return;
}
if (page.StatusCode != HttpStatusCode.OK)
{
return;
}
var doc = new HtmlDocument();
doc.LoadHtml(page.Html);
//no index is returned, no follow means we index this but don't take anchors
var metaTags = doc.DocumentNode.Descendants("meta").ToList();
bool containsNoIndex = false, containsNoFollow = false;
var robotsMeta = metaTags.FirstOrDefault(x => x.Attributes["name"]?.Value == "robots")?.Attributes["content"]?.Value;
if (robotsMeta != null)
{
containsNoFollow = robotsMeta.Contains("nofollow");
containsNoIndex = robotsMeta.Contains("noindex");
}
var anchors = doc.DocumentNode.Descendants("a");
if (!containsNoIndex)
{
//we can save it on the disk so the indexer can do its job
FileTreeWorker.CreateWebsiteTree(baseUrl, page.Html);
}
if (containsNoFollow)
{
//do not extract the links
return;
}
foreach (var a in anchors)
{
var currentAnchor = a.Attributes["href"]?.Value;
if (string.IsNullOrEmpty(currentAnchor))
{
continue;
}
var rel = a.Attributes["rel"]?.Value;
if (rel != null && rel == "nofollow")
{
continue;
}
UrlFrontier.Enqueue(baseUrl, currentAnchor);
}
}
