Apache POI Mirroring Words in Arabic Language

Question

I'm developing an Arabic OCR application in java which extracts Arabic texts in images and then saving the text into a Microsoft Word file, for this purpose i use Apache-POI library.

My problem is that when i extract some text the order of the words are fine but when i save it in a Word file the order of the words are kinda messed up and looks mirrored

for example:

BUT after saving it as a Word:

and here is the code for saving the Word file:

public class SavingStringAsWordDoc {


    File f=theGUI.toBeSavedWord;

    public void saveAsWorddd (){
        String st=TesseractPerformer.toBeShown;

        try(FileOutputStream fout=new FileOutputStream(f);XWPFDocument docfile=new XWPFDocument()){

            XWPFParagraph paraTit=docfile.createParagraph();
            paraTit.setAlignment(ParagraphAlignment.LEFT);
            XWPFRun paraTitRun=paraTit.createRun();
            paraTitRun.setBold(true);
            paraTitRun.setFontSize(15);
            paraTit.setAlignment(ParagraphAlignment.RIGHT);
            docfile.createParagraph().createRun().setText(st);  //content to be written
            docfile.write(fout); //adding to output stream
        } catch(IOException e){
            System.out.println("IO ERROR:"+e);
        }
    }

i noticed one thing which might help understanding the problem: if i copy the messed up text in the word file and then paste it by choosing the (Keep Text Only) paste option it fixes the order of the paragraph

Answer 1

This needs bidirectional text direction support (bidi) and is not yet implemented in XWPF of apache poi per default. But the underlying object org.openxmlformats.schemas.wordprocessingml.x2006.main.CTPPr supports this. So we must get this underlying object from the XWPFParagraph and then set Bidi on.

Example:

import java.io.File;
import java.io.FileOutputStream;
import java.nio.charset.StandardCharsets;
import java.nio.file.Files;

import org.apache.poi.xwpf.usermodel.*;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTPPr;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.STOnOff;

public class CreateWord {

 public static void main(String[] args) throws Exception {

  String content = Files.readString(new File("ArabicTextFile.txt").toPath(), StandardCharsets.UTF_16);

  XWPFDocument document = new XWPFDocument();

  XWPFParagraph paragraph = document.createParagraph();

  // set bidirectional text support on
  CTP ctp = paragraph.getCTP();
  CTPPr ctppr = ctp.getPPr();
  if (ctppr == null) ctppr = ctp.addNewPPr();
  ctppr.addNewBidi().setVal(STOnOff.ON);

  XWPFRun run=paragraph.createRun(); 
  run.setBold(true);
  run.setFontSize(22);
  run.setText(content);

  FileOutputStream out = new FileOutputStream("CreateWord.docx");
  document.write(out);
  out.close();
  document.close();

 }
}

My ArabicTextFile.txt contains the text

هذا هو النص باللغة العربية لاختبار النص باللغة العربية

in UTF-16 encoding (Unicode).

Result in Word: