private bool ShouldCrawlNextPages()
{
CrawlQueueItem currentCrawler = null;
if (_pagesToCrawl.Count > 0)
{
currentCrawler = _pagesToCrawl.Peek();
}
return(!_contentFound &&
_crawledUri.Count + _currentTasks.Count < _configuration.MaxPages &&
(currentCrawler == null || currentCrawler.PageLevel < _configuration.MaxPageLevel));
}
private async Task <Tuple <CrawlPageResults, int> > CrawlPage(CrawlQueueItem crawlItem, CancellationToken cancellationToken)
{
CrawlPageResults results = null;
try
{
// Раскладываем в корень страницы.
using (FileStream pageStream = new FileStream(
Path.Combine(
_domainDirectory,
GetPageFileNameByUri(crawlItem.PageUri)),
FileMode.Create))
{
var pageCrawler = new PageCrawler(crawlItem.PageUri);
results =
await pageCrawler.StartCrawling(pageStream, cancellationToken);
// В прекрасном мире, здесь был бы стрим, который пропускал через себя страницу,
// которую по мере вычитывания ему давал PageCrawler и, тем самым проверялось
// бы и стоп условие и все шло бы без излишних проверок и прочего, но имеем то,
// что имеем.
if (!string.IsNullOrEmpty(_configuration.StopString) &&
results.PageContent.Contains(_configuration.StopString))
{
Logger.Log(
LogLevel.Info,
$"Stop string found at {crawlItem.PageUri}");
_contentFound = true;
}
}
await LoadContent(results);
}
catch (Exception e)
{
Logger.Log(
LogLevel.Error,
e,
$"Page crawl exception with uri {crawlItem.PageUri}");
}
finally
{
}
return(new Tuple <CrawlPageResults, int>(results, crawlItem.PageLevel));
}
private void EnqeueCrawlers(IEnumerable <Uri> nextPages, int parentLevel)
{
foreach (var page in nextPages)
{
if (!_queuedUri.Contains(page) && !_crawledUri.Contains(page))
{
var item = new CrawlQueueItem()
{
PageUri = page,
PageLevel = parentLevel + 1
};
_queuedUri.Add(page);
_pagesToCrawl.Enqueue(item);
}
}
}
public async Task <DomainCrawlStatistics> CrawlDomain(CancellationToken cancellationToken)
{
CancellationToken resultToken =
CancellationTokenSource.CreateLinkedTokenSource(
cancellationToken,
_internalCTS.Token).Token;
_contentFound = false;
await Task.Factory.StartNew(
async() =>
{
InitQueue();
Directory.CreateDirectory(_domainDirectory);
RobotsParams robotsParams = await _robotsReader.GetRobotsParams();
while ((_pagesToCrawl.Count > 0 && ShouldCrawlNextPages() || _currentTasks.Count > 0))
{
foreach (var task in _currentTasks.Where(t => t.IsCompleted).ToArray())
{
if (!task.IsFaulted && !task.IsCanceled)
{
var results = task.Result;
UpdateProgress(results.Item1);
if (results.Item2 < _configuration.MaxPageLevel)
{
EnqeueCrawlers(results.Item1.References, results.Item2);
}
_crawledUri.Add(task.Result.Item1.CrawledUri);
_queuedUri.Remove(task.Result.Item1.CrawledUri);
}
_currentTasks.Remove(task);
}
if (_currentTasks.Count >= _configuration.MaxTasks)
{
await Task.WhenAny(_currentTasks.ToArray());
continue;
}
if (_pagesToCrawl.Count > 0 &&
ShouldCrawlNextPages())
{
CrawlQueueItem currentCrawler = _pagesToCrawl.Dequeue();
_currentTasks.Add(CrawlPage(currentCrawler, resultToken));
if (robotsParams.CrawlDelay > 0)
{
await Task.Delay(robotsParams.CrawlDelay, resultToken);
}
}
}
SaveProgress();
},
TaskCreationOptions.LongRunning);
return(_statistics);
}
