public QuickDemo()
{
Console.Write("INIT...");
CorpusExplorerEcosystem.InitializeMinimal();
if (!Directory.Exists("corpus"))
{
Directory.CreateDirectory("corpus");
}
InitializeComponent();
#if DEBUG
if (!File.Exists(_corpusPath))
{
_corpusPath = "W:/eBooks-MFB/" + _corpusPath;
}
#endif
if (File.Exists(_corpusPath))
{
_quickIndex = new QuickIndex(_corpusPath);
_dict = Serializer.Deserialize <Dictionary <Guid, string> >("corpus/data.bin");
}
Console.WriteLine("OK!");
}
private void backgroundWorker1_DoWork(object sender, DoWorkEventArgs e)
{
try
{
btn_index_delete.Invoke((MethodInvoker) delegate { btn_index_delete.Enabled = false; });
btn_update.Invoke((MethodInvoker) delegate { btn_update.Enabled = false; });
var files = Directory.GetFiles(_rootPath, "*.pdf", SearchOption.AllDirectories);
progressBar1.Invoke((MethodInvoker) delegate
{
progressBar1.Maximum = 5;
progressBar1.Minimum = 0;
progressBar1.Value = 1;
Console.WriteLine($"1 / 5 = Read {files.Length} files");
});
var scraper = new TextSharpPdfScraper {
Strategy = TextSharpPdfScraper.TextSharpPdfScraperStrategy.Location
};
scraper.Input.Enqueue(files);
scraper.Execute();
progressBar1.Invoke((MethodInvoker) delegate
{
progressBar1.Maximum = 5;
progressBar1.Minimum = 0;
progressBar1.Value = 2;
Console.WriteLine($"2 / 5 = Cleaning");
});
// no cleanup
progressBar1.Invoke((MethodInvoker) delegate
{
progressBar1.Maximum = 5;
progressBar1.Minimum = 0;
progressBar1.Value = 3;
Console.WriteLine($"3 / 5 = Tagging");
});
var tagger = new RawTextTagger();
tagger.Input = scraper.Output;
tagger.Execute();
var corpus = tagger.Output.First();
corpus.Save(_corpusPath, false);
_dict = corpus.DocumentMetadata.ToDictionary(x => x.Key, x => (string)x.Value["Datei"]);
Serializer.Serialize(_dict, "corpus/data.bin", false);
progressBar1.Invoke((MethodInvoker) delegate
{
progressBar1.Maximum = 5;
progressBar1.Minimum = 0;
progressBar1.Value = 4;
Console.WriteLine($"4 / 5 = Build QuickIndex");
});
_quickIndex = new QuickIndex(_corpusPath);
btn_index_delete.Invoke((MethodInvoker) delegate { btn_index_delete.Enabled = true; });
btn_update.Invoke((MethodInvoker) delegate { btn_index_delete.Enabled = true; });
progressBar1.Invoke((MethodInvoker) delegate
{
progressBar1.Maximum = 5;
progressBar1.Minimum = 0;
progressBar1.Value = 4;
Console.WriteLine($"5 / 5 = COMPLETE!");
});
}
catch (Exception ex)
{
Console.WriteLine("########--------########");
Console.WriteLine(ex.Message);
Console.WriteLine("--------########--------");
Console.WriteLine(ex.StackTrace);
Console.WriteLine("########--------########");
Console.WriteLine();
}
}
