private async Task SearchForPlagiarismInWeb(string fileName) { richTextBox1.Text = string.Empty; textBox2.Text = string.Empty; var simplifiedText = TextDocumentManager.SimplifiedTextFromFile(fileName); var plagiarismInWeb = await PlagiarismInWebFinder.Find(simplifiedText); richTextBox1.Text += simplifiedText + Environment.NewLine; richTextBox1.Text += $"Words count: {plagiarismInWeb.WordCount}" + Environment.NewLine; ShowUrlToWordIndexes(plagiarismInWeb.OrderedUrls, plagiarismInWeb.OrderedUrlToWordsIndexes); _webResults = plagiarismInWeb.OrderedWebResults; for (int i = 0; i < _webResults.Count; i++) { textBox2.Text += _webResults[i].Url; textBox2.Text += Environment.NewLine; textBox2.Text += String.Format("{0:P2} text found", _webResults[i].CharactersPercentage); textBox2.Text += Environment.NewLine; foreach (var commonTextPart in _webResults[i].CommonTextParts) { textBox2.Text += commonTextPart; textBox2.Text += Environment.NewLine; } textBox2.Text += Environment.NewLine; } }
public async Task <ActionResult> Main(HttpPostedFileBase file) { if (file?.ContentLength > 0 && (Path.GetFileName(file.FileName).Split('.')[1] == "doc" || Path.GetFileName(file.FileName).Split('.')[1] == "docx")) { var types = file.ContentType; var fileName = Path.GetFileName(file.FileName); var path = Path.Combine(Server.MapPath("~/App_Data/uploads"), fileName); file.SaveAs(path); var initialText = TextDocumentManager.TextFromFile(path); TextManager.PrepareText(initialText, out string[] initialWords, out Dictionary <int, string> initialDocIndexToSimplifiedWord, out int[] initialDocIndexes, out string[] simplifiedWords, out int wordCount); var plagiarismInWebSearch = PlagiarismInWebFinder.FindAsync(initialWords, initialDocIndexToSimplifiedWord, initialDocIndexes, simplifiedWords, Server.MapPath("~/App_Data/uploads")); var plagiarismInLocalDBSearch = PlagiarismInLocalDBFinder.FindAsync(initialWords, initialDocIndexToSimplifiedWord, initialDocIndexes, simplifiedWords); await Task.WhenAll(plagiarismInWebSearch, plagiarismInLocalDBSearch); var plagiarismInWeb = plagiarismInWebSearch.Result; var plagiarismInLocalDB = plagiarismInLocalDBSearch.Result; var DBPlagiarizedIndexes = plagiarismInLocalDB.InitialWordIndexToSourceIds.Keys.ToList(); DBPlagiarizedIndexes.AddRange(plagiarismInWeb.InitialWordIndexToSourceIds.Keys.ToList()); var allPlagiarismHtmlText = TextManager.ComposeHtmlText(initialWords, DBPlagiarizedIndexes); var(vodnost, toshnotnost) = TextAnalyzer.Analyze(simplifiedWords); PlagiarismResult plagiarism = new PlagiarismResult { InitialWords = initialWords, SimplifiedWords = simplifiedWords, WordCount = wordCount, Vodnost = vodnost, Toshnotnost = toshnotnost, PlagiarismWeb = plagiarismInWeb, PlagiarismDB = plagiarismInLocalDB, AllPlagiarismHtmlText = allPlagiarismHtmlText, DebugLogs = plagiarismInWeb.DebugLogs + plagiarismInLocalDB.DebugLogs }; ViewBag.DebugLogs = plagiarism.DebugLogs; foreach (KeyValuePair <int, HashSet <int> > lists in plagiarismInLocalDB.SourceIdToSourceWordsIndexes) { Session["Doc" + lists.Key] = lists.Value; } foreach (KeyValuePair <string, HashSet <int> > lists in plagiarismInWeb.SourceIdToSourceWordsIndexes) { Session[lists.Key] = lists.Value; } return(View("Main", plagiarism)); } return(RedirectToAction("Main")); }