Searchable PDF App

• Download source code - 3.2 MB

Introduction

This tool is using Tesseract to do the OCR (Optical Character Recognition). The tool is also using Ghostscript to convert non-searchable PDFs to TIFF files so that can be converted to searchable PDFs. The tool also provides an option to merge converted PDF files into one.

The tool is using iTextSharp library to see if the PDF is searchable or not. The following image files are supported: BMP, PNM, PNG, JFIF, JPEG, JPG, TIFF, TIF, GIF, PDF

1. OCR PDF with text option will OCR PDF files regardless of whether they are already searchable or not
2. Merge - option to merge converted PDF files into one searchable PDF
3. Log Time - show time it takes to process each file

First, install Tesseract and Ghostscript. Next, select Tesseract.exe and GsWin64.exe locations in the tool. Finally, select input and output folders and click Process.

Using the Code

ProcessFile() is the key function that will OCR file and convert it based to PDF based on the file type. The key part of this function is RunDosCommandAsynch().

Private Sub ProcessFile(ByVal sFilePath As String, sFileName As String)

    Dim dtStart As DateTime = DateTime.Now
    Dim bTiff As Boolean = False
    Dim sProcessFile As String = sFilePath
    Dim sExt As String = GetExtFromFileName(sFileName)
    If sExt = "pdf" Then
        If chkOcrWithText.Checked = False AndAlso PdfHasText(sFilePath) Then
            'PDF already has txt - do not process
            Log(sFilePath, "PDF already has text")

            Dim sDoneFile As String = IO.Path.Combine(sFolderDone, sFileName)
            CreateFolderIfDoesNotExist(sDoneFile)
            IO.File.Move(sFilePath, sDoneFile) 'Move to Done so that
                                               'it will not be processed again
            Exit Sub
        Else
            bTiff = True
            sProcessFile = System.IO.Path.Combine(System.IO.Path.GetTempPath(), _
                           Guid.NewGuid().ToString("N") & ".tiff")
            PdfToTiff(sFilePath, sProcessFile)
            If IO.File.Exists(sProcessFile) = False Then
                Log(sFilePath, "TIFF file could not be generated")
                sProcessFile = ""
            End If
        End If
    End If

    Dim sTempOutFile As String = System.IO.Path.Combine_
        (System.IO.Path.GetTempPath(), Guid.NewGuid().ToString("N"))
    Dim sTempOutFilePath As String = sTempOutFile & ".pdf"

    Dim sError As String = ""

    If sProcessFile <> "" Then
        sError = RunDosCommandAsynch(sTesseractPath, """" & _
        sProcessFile & """ """ & sTempOutFile & """ pdf") '10 min timeout
    End If

    Dim bSuccess As Boolean = IO.File.Exists(sTempOutFilePath) _
                              AndAlso GetFileSize(sTempOutFilePath) > 10

    If bSuccess AndAlso bTiff AndAlso PdfHasText(sTempOutFilePath) = _
                        False AndAlso PdfHasText(sFilePath) Then
        'New PDF has no text but the old one did
        Log(sFilePath, "Could not extract text from images. " & sError)
        sTempOutFilePath = sFilePath
    End If

    If bSuccess Then
        sFileName = GetBaseNameFromFileName(sFileName) & ".pdf"
        Dim sDoneFile As String = IO.Path.Combine(sFolderDone, sFileName)
        CreateFolderIfDoesNotExist(sDoneFile)
        IO.File.Move(sTempOutFilePath, sDoneFile) 'Done

        Dim iProcessSeconds As Integer = DateTime.Now.Subtract(dtStart).TotalSeconds
        Log(sDoneFile, "Done in " & iProcessSeconds & " sec")

        TryDelete(sFilePath)
    Else
        Dim sFailedFile As String = IO.Path.Combine(sFolderFailed, sFileName)
        CreateFolderIfDoesNotExist(sFailedFile)
        IO.File.Move(sFilePath, sFailedFile)    'Failed

        Log(sFailedFile, sError)
        LogFile(sFailedFile, sError)
    End If

    'Delete temp files
    If bTiff Then TryDelete(sProcessFile) ' tiff file
    TryDelete(sTempOutFilePath)    'PDF
End Sub

PdfToTiff uses ghost script to convert non-searchable PDF to TIFF. It is the secret to converting non-searchable PDFs to searchable PDFs.

Sub PdfToTiff(ByVal sInPdf As String, ByVal sOutTiff As String)
     Dim sError As String = ""
     'https://ghostscript.com/doc/9.21/Devices.htm
     RunDosCommand(sGhostscriptPath, "-dNOPAUSE -q -r300 -sDEVICE=tiff24nc _
     -dBATCH -sOutputFile=""" & sOutTiff & """ """ & sInPdf & """ -c quit", sError)
 End Sub

RunDosCommandAsynch does the action of running DOS command:

Function RunDosCommandAsynch(sExeFilePath As String, sArguments As String) As String

    Dim sError As String = ""
    Dim oProcess As Process = New Process()
    oProcess.StartInfo.UseShellExecute = False
    oProcess.StartInfo.RedirectStandardOutput = True
    oProcess.StartInfo.RedirectStandardError = True
    oProcess.StartInfo.FileName = sExeFilePath
    oProcess.StartInfo.Arguments = sArguments
    oProcess.StartInfo.WindowStyle = ProcessWindowStyle.Hidden
    oProcess.StartInfo.CreateNoWindow = True

    oProcess.Start()
    oProcess.WaitForExit(1000 * iTimeOutSec)
    If oProcess.HasExited = False Then
        oProcess.Kill()
        sError = "Timeout after " + iTimeOutSec + " seconds."
    End If

    If String.IsNullOrEmpty(sError) Then
        sError = oProcess.StandardError.ReadToEnd()
    End If

    'sRet = oProcess.StandardOutput.ReadToEnd()

    If oProcess.ExitCode <> 0 AndAlso String.IsNullOrEmpty(sError) Then
        sError = "ExitCode: " + oProcess.ExitCode
    End If

    oProcess.Close()

    Return sError
End Function

PdfHasText() function uses iTextSharp library to see if the PDF is searchable or not.

Function PdfHasText(ByVal sInPdf As String) As Boolean
       Dim doc As New PdfReader(sInPdf)
       'Dim sb As New StringBuilder()
       Dim bRet As Boolean = False

       For iPage = 1 To doc.NumberOfPages
           Dim pg As PdfDictionary = doc.GetPageN(iPage)
           Dim ir As Object = pg.Get(PdfName.CONTENTS)
           Dim value As PdfObject = doc.GetPdfObject(ir.Number)
           If value.IsStream() Then
               Dim stream As PRStream = value
               Dim streamBytes As Byte() = PdfReader.GetStreamBytes(stream)
               Dim tokenizer = New PRTokeniser(New RandomAccessFileOrArray(streamBytes))

               Try
                   While tokenizer.NextToken()
                       If tokenizer.TokenType = PRTokeniser.TK_STRING Then
                           Dim str As String = tokenizer.StringValue
                           If str <> "" Then
                               bRet = True
                               Exit While
                           End If
                           'sb.Append(str)
                       End If
                   End While
               Catch ex As Exception
                   'Ignore
               Finally
                   tokenizer.Close()
               End Try

               If bRet Then
                   Exit For
               End If
           End If
       Next
       doc.Close()

       Return bRet
       'Return sb.ToString()
   End Function

Merge() function uses iTextSharp library to merge PDF files.

Sub Merge(ByVal sInSubFolder As String)
        Dim sSubFolderDone As String = GetTargetSubFolder(sFolderDone, sInSubFolder)

        If IO.Directory.Exists(sSubFolderDone) = False Then
            Exit Sub
        End If

        Dim oOutFiles As String() = IO.Directory.GetFiles(sSubFolderDone)
        If oOutFiles.Length = 0 Then
            Exit Sub
        End If

        Dim sMergeOutFilePath As String = GetMergeOutFilePath(sSubFolderDone)

        Dim oDeleteFiles As New ArrayList()
        Dim oPdfDoc As New iTextSharp.text.Document()
        Dim oPdfWriter As PdfWriter = PdfWriter.GetInstance_
            (oPdfDoc, New FileStream(sMergeOutFilePath, FileMode.Create))
        oPdfDoc.Open()

        System.Array.Sort(Of String)(oOutFiles)
        For i As Integer = 0 To oOutFiles.Length - 1
            Dim sFromFilePath As String = oOutFiles(i)
            Dim oFileInfo As New FileInfo(sFromFilePath)
            Dim sBookmarkTitle As String = GetBaseNameFromFileName(oFileInfo.Name)
            AddPdf(sFromFilePath, oPdfDoc, oPdfWriter, sBookmarkTitle)

            oDeleteFiles.Add(sFromFilePath)
        Next

        Try
            oPdfDoc.Close()
            oPdfWriter.Close()
        Catch ex As Exception
            Log(ex.Message)
        End Try

        If chkDeleteSourceFiles.Checked Then
            For Each sFromFilePath As String In oDeleteFiles
                If IO.File.Exists(sFromFilePath) Then
                    Try
                        IO.File.Delete(sFromFilePath)
                    Catch ex As Exception
                        Log(sFromFilePath, "Could not delete, " & ex.Message)
                    End Try
                End If
            Next
        End If
    End Sub
 
    Sub AddPdf(ByVal sInFilePath As String, ByRef oPdfDoc As iTextSharp.text.Document, _
               ByRef oPdfWriter As PdfWriter, ByVal sBookmarkTitle As String)

        AddBookmark(oPdfDoc, sBookmarkTitle)

        Dim oDirectContent As iTextSharp.text.pdf.PdfContentByte = _
                                              oPdfWriter.DirectContent
        Dim oPdfReader As iTextSharp.text.pdf.PdfReader = _
                          New iTextSharp.text.pdf.PdfReader(sInFilePath)
        Dim iNumberOfPages As Integer = oPdfReader.NumberOfPages
        Dim iPage As Integer = 0

        Do While (iPage < iNumberOfPages)
            iPage += 1

            Dim iRotation As Integer = oPdfReader.GetPageRotation(iPage)
            Dim oPdfImportedPage As iTextSharp.text.pdf.PdfImportedPage = _
                                    oPdfWriter.GetImportedPage(oPdfReader, iPage)

            If chkResize.Checked Then
                If (oPdfImportedPage.Width <= oPdfImportedPage.Height) Then
                    oPdfDoc.SetPageSize(iTextSharp.text.PageSize.LETTER)
                Else
                    oPdfDoc.SetPageSize(iTextSharp.text.PageSize.LETTER.Rotate())
                End If

                oPdfDoc.NewPage()

                Dim iWidthFactor As Single = _
                    oPdfDoc.PageSize.Width / oPdfReader.GetPageSize(iPage).Width
                Dim iHeightFactor As Single = _
                    oPdfDoc.PageSize.Height / oPdfReader.GetPageSize(iPage).Height
                Dim iFactor As Single = Math.Min(iWidthFactor, iHeightFactor)

                Dim iOffsetX As Single = (oPdfDoc.PageSize.Width - _
                                         (oPdfImportedPage.Width * iFactor)) / 2
                Dim iOffsetY As Single = (oPdfDoc.PageSize.Height - _
                                         (oPdfImportedPage.Height * iFactor)) / 2

                oDirectContent.AddTemplate(oPdfImportedPage, iFactor, _
                                           0, 0, iFactor, iOffsetX, iOffsetY)

            Else
                oPdfDoc.SetPageSize(oPdfReader.GetPageSizeWithRotation(iPage))
                oPdfDoc.NewPage()

                If (iRotation = 90) Or (iRotation = 270) Then
                    oDirectContent.AddTemplate(oPdfImportedPage, 0, -1.0F, _
                    1.0F, 0, 0, oPdfReader.GetPageSizeWithRotation(iPage).Height)
                Else
                    oDirectContent.AddTemplate(oPdfImportedPage, 1.0F, 0, 0, 1.0F, 0, 0)
                End If
            End If
        Loop

    End Sub

Bonus VBS (For Processing PDF Files in Parallel)

As a bonus, I included a PdfOcr.vbs VB Script file that will convert a folder or a file to searchable PDF if you drag-and-drop folder or file on it. The advantage of using the approach is that you can process many PDF/image files in parallel. This way, only files that do not have corresponding searchable PDF will be processed. You should be able to run this on one server or on many servers running against same folder or file share.

Make sure that Tesseract and Ghostscript are installed. Next, make sure that sGhostscriptPath (line 7) and sTesseractPath (Line 8) point to location of Tesseract.exe and GsWin64.exe files.

• Download PdfOcrVbs - 1.9 KB

OcrImgFile sub does processing of the image files to searchable PDFs.

Sub OcrImgFile(sInFile)
  iPos = InStrRev(sInFile,".")
  sFileBase = Mid(sInFile,1,iPos - 1)
  sOutPdf = sFileBase & sFileSuffix
  
  If fso.FileExists(sOutPdf & ".pdf") Then    
    Msg sOutPdf & ".pdf already exists"
    Exit Sub
  End If

  oShell.run """" & sTesseractPath & """ """ & sInFile & """ """ & sOutPdf & """ pdf", 1 , True

  If fso.FileExists(sOutPdf & ".pdf") Then
    iCount = iCount + 1
  Else
    Msg sOutPdf & ".pdf could not be created"
  End If  
    
End Sub 

OcrPdfFile sub converts non-searchable PDFs to searchable PDFs.

Sub OcrPdfFile(sInPdf)
  If Right(sInPdf, Len(sFileSuffix) + 4) = sFileSuffix & ".pdf" Then
      Msg sInPdf & " in an output file and will not be processed"
    'Ignore Generated OCR PDFs
    Exit Sub
  End If

  iPos = InStrRev(sInPdf,".")
  sFileBase = Mid(sInPdf,1,iPos - 1)
  sOutPdf = sFileBase & sFileSuffix
  
  If fso.FileExists(sOutPdf & ".pdf") Then    
    Msg sOutPdf & ".pdf already exists"
    Exit Sub
  End If

  sOutTiff = sFileBase & sFileSuffix & ".tiff"
  oShell.run """" & sGhostscriptPath & """ -dNOPAUSE -q -r300 -sDEVICE=tiff24nc _
  -dBATCH -sOutputFile=""" & sOutTiff & """ """ & sInPdf & """ -c quit", 1 , True
  
  If fso.FileExists(sOutTiff) Then
    oShell.run """" & sTesseractPath & """ """ & sOutTiff & """ """ & _
                      sOutPdf & """ pdf", 1 , True
    fso.DeleteFile sOutTiff

    If fso.FileExists(sOutPdf & ".pdf") Then
      iCount = iCount + 1
    Else
      Msg sOutPdf & ".pdf could not be created"
    End If             

  Else
    Msg sOutTiff & " could not be created"
  End If    
End Sub  

History

• 2^nd August, 2023: Version 1