DocumentReaderPage Class

Public Sub DocumentReaderPageExample() 
   Dim documentFileName As String 
   Using dlg As New OpenFileDialog() 
      If dlg.ShowDialog() <> System.Windows.Forms.DialogResult.OK Then 
         Return 
      End If 
 
      documentFileName = dlg.FileName 
   End Using 
 
   ' Load the document using default options 
   Dim reader As DocumentReader = DocumentReader.Create(documentFileName, Nothing) 
 
   ' Show the document properties 
   Dim sb As New StringBuilder() 
   sb.AppendFormat("Reader used: {0}\n", reader.ReaderType) 
   sb.AppendFormat("Document has {0} pages\n", reader.Pages.Count) 
 
   ' Get the properties (meta data) 
   Dim props As IDictionary(Of String, String) = reader.GetProperties() 
   For Each prop As KeyValuePair(Of String, String) In props 
      sb.AppendFormat("{0}: {1}\n", prop.Key, prop.Value) 
   Next 
 
   MessageBox.Show(sb.ToString()) 
 
   ' Now show the pages sizes 
   sb = New StringBuilder() 
   For Each page As DocumentReaderPage In reader.Pages 
      sb.AppendFormat("Page: {0} size: {1}\n", page.PageNumber, page.Size) 
   Next 
   MessageBox.Show(sb.ToString()) 
 
   ' Now loop and show the text for each page till use cancels 
 
   ' If this is a Raster document such as TIFF or JPEG, we must use an OCR engine 
   Dim ocrEngine As IOcrEngine = Nothing 
 
   If reader.ReaderType = DocumentReaderType.Raster Then 
      ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.Advantage, False) 
      ocrEngine.Startup(Nothing, Nothing, Nothing, LEAD_VARS.OcrAdvantageRuntimeDir) 
   End If 
 
   reader.ObjectManager.BeginParse(ocrEngine) 
 
   For Each page As DocumentReaderPage In reader.Pages 
      ' Parse this page 
      Dim pageText As DocumentPageText = reader.ObjectManager.ParsePageText(page) 
      Dim text As String = pageText.BuildText() 
 
      If MessageBox.Show(text, String.Format("Page {0} text, continue to next page?", _ 
                                             page.PageNumber), MessageBoxButtons.YesNo) = _ 
                                             System.Windows.Forms.DialogResult.No Then 
         Exit For 
      End If 
   Next 
 
   reader.ObjectManager.EndParse() 
 
   If Not IsNothing(ocrEngine) Then 
      ocrEngine.Dispose() 
   End If 
 
   reader.Dispose() 
End Sub