ParseFetchedFile(FetchedFile fetchedFile, IList <string> inLinkTexts)
{
UniversalFileParser parser = new UniversalFileParser();
parser.SetFile(fetchedFile.MimeType, fetchedFile.Url, fetchedFile.Charset, fetchedFile.FilePath);
var content = parser.GetContentAsync().GetAwaiter().GetResult();
var textContent = parser.GetTextContentAsync().GetAwaiter().GetResult();
var title = parser.GetTitleAsync().GetAwaiter().GetResult();
var headers = parser.GetHeadersAsync().GetAwaiter().GetResult();
var links = parser.GetLinksAsync().GetAwaiter().GetResult();
var tokens = parser.GetTokensAsync(inLinkTexts).GetAwaiter().GetResult();
var urlFile = new UrlFile
{
PublishDate = (ulong)parser.GetPublishDateAsync().GetAwaiter().GetResult().ToBinary(),
Charset = fetchedFile.Charset,
Content = content,
FileHash = fetchedFile.FileHash,
FilePath = fetchedFile.FilePath,
MimeType = fetchedFile.MimeType,
TextContent = textContent,
Title = title,
Url = fetchedFile.Url,
HeaderTotalLength = (uint)headers.Sum(o => o.Text.Length),
HeaderCount = (uint)headers.Sum(o => (6 - o.Level)),
InLinkCount = (uint)inLinkTexts.Count,
InLinkTotalLength = (uint)inLinkTexts.Sum(o => o.Length),
PageRank = 0.1,
};
return(urlFile, links, tokens);
}
public (string Url, string Content) JudgeContent(FetchedFile fetchedFile, string content)
{
var sameFiles = mConfig.UrlFileStore.GetUrlFilesByHash(fetchedFile.FileHash);
var host = UrlHelper.GetHost(fetchedFile.Url);
foreach (var file in sameFiles)
{
if (file.Url == fetchedFile.Url)
{
continue;
}
var currentHost = UrlHelper.GetHost(file.Url);
bool isSameDns = false;
try
{
isSameDns = Dns.GetHostAddresses(currentHost).SequenceEqual(Dns.GetHostAddresses(host));
}
catch (Exception)
{ }
if (content == file.Content &&
(currentHost == host || isSameDns))
{
mLogger.Log(nameof(SimpleSimilarContentManager), $"Find Same UrlFile for {fetchedFile.Url}: {file.Url}");
return(file.Url, file.Content);
}
}
return(null, null);
}
private string GetSameUrl(FetchedFile fetchedFile, string content)
{
// Judge if there are other files that have similar content as this
var(sameUrl, sameContent) = mSimilarContentJudger.JudgeContent(fetchedFile, content);
lock (mContentUrlDictSyncLock)
{
if (mContentUrlDict.ContainsKey(content))
{
return(mContentUrlDict[content]);
}
if (sameUrl == null)
{
mContentUrlDict.Add(content, fetchedFile.Url);
}
}
return(sameUrl);
}
protected void FetchUrlAsync(string url)
{
mFetchSemaphore.Wait();
mLogger.Log(nameof(Crawler), "Begin Crawl: " + url, true);
var t = Task.Run(() =>
{
FetchedFile fetchedFile = null;
try
{
// Fetch Url
fetchedFile = FetchUrl(url);
if (fetchedFile == null)
{
return;
}
// Get InLink Texts
var inLinkTexts = GetInLinkTexts(fetchedFile.Url);
// Parse File
var(urlFile, linkList, tokens) = ParseFetchedFile(fetchedFile, inLinkTexts);
if (GetSameUrl(fetchedFile, urlFile.Content) != null)
{
// Has Same UrlFile, Skip
return;
}
// Save New UrlFile
// Get Old id and New id
ulong oldUrlFileId;
(urlFile, oldUrlFileId) = mConfig.UrlFileStore.SaveUrlFileAndGetOldId(urlFile);
// Clear old data
if (oldUrlFileId != 0)
{
mConfig.PostingListStore.ClearPostingLists(oldUrlFileId);
mConfig.InvertedIndexStore.ClearIndices(oldUrlFileId);
mConfig.LinkStore.ClearLinks(oldUrlFileId);
}
lock (mContentUrlDictSyncLock)
{
mContentUrlDict.Remove(urlFile.Content);
}
// Filter Links
linkList = FilterLinks(linkList);
// Save links
mConfig.LinkStore.SaveLinks(urlFile.UrlFileId,
linkList.Select(o => new Link
{
Text = o.Text,
Url = o.Url,
UrlFileId = urlFile.UrlFileId,
}));
// Save Indices
SaveIndices(tokens, inLinkTexts, urlFile, oldUrlFileId);
// Add newly-found urls
var urls = linkList.Select(o => o.Url).Distinct();
mUrlFrontier.PushUrls(urls);
// Push Back This Url
mUrlFrontier.PushBackUrl(url, urlFile.UpdateInterval);
mLogger.Log(nameof(Crawler), "End Crawl: " + url);
}
catch (NotSupportedException e)
{
mLogger.LogException(nameof(Crawler), "Not supported file format: " + url, e, false);
mErrorLogger.LogException(nameof(Crawler), "Not supported file format: " + url, e);
// Retry
mUrlFrontier.PushBackUrl(url, 0, true);
}
catch (InvalidDataException e)
{
mLogger.LogException(nameof(Crawler), "Invalid data: " + url, e);
mErrorLogger.LogException(nameof(Crawler), "Invalid data: " + url, e);
// Retry
mUrlFrontier.PushBackUrl(url, 0, true);
}
catch (UriFormatException e)
{
mLogger.LogException(nameof(Crawler), "Invalid Uri: " + url, e);
mErrorLogger.LogException(nameof(Crawler), "Invalid Uri: " + url, e);
mUrlFrontier.RemoveUrl(url);
}
catch (IOException e)
{
mLogger.LogException(nameof(Crawler), "Failed to fetch: " + url, e, false);
mErrorLogger.LogException(nameof(Crawler), "Failed to fetch: " + url, e);
mUrlFrontier.RemoveUrl(url);
}
catch (Exception e) when(
e is OperationCanceledException ||
e is TaskCanceledException
)
{
mUrlFrontier.PushBackUrl(url, 0);
}
catch (Exception e)
{
mLogger.LogException(nameof(Crawler), "Failed to crawl url: " + url, e);
mErrorLogger.LogException(nameof(Crawler), "Failed to crawl url: " + url, e);
// Retry
mUrlFrontier.PushBackUrl(url, 0, true);
}
finally
{
mFetchSemaphore.Release();
if (fetchedFile != null && File.Exists(fetchedFile.FilePath))
{
File.Delete(fetchedFile.FilePath);
}
}
}).ContinueWith(task =>
{
mTasks.TryRemove(task, out bool v);
});
mTasks.TryAdd(t, true);
}
