prepare data to train OCR

convert text into image.sh

for i in `seq 1001`;
 do
text2image --text $i.gt.txt --outputbase /home/owner/Downloads/output/im$i --xsize 680 --ysize 200 --fonts_dir /usr/share/fonts/truetype/Amiri/ --font 'Amiri'  --margin 0 --ptsize 30  --degrade_image     --char_spacing 1;
 done

fun main(){

    val rand = Random(1200)
    for (i in 0..1000){
        val file = File("${i}.gt.txt")
        val num = String.format(Locale.forLanguageTag("ar") , "%d",rand.nextInt(98785)).apply {
            padStart(5 , '0')
        }

        print(convertToArabic(num))
        file.writeText(convertToArabic(num))
    }

}


fun convertToArabic(arabicStr: String): String {
    return (arabicStr + "")
            .replace("1".toRegex(), "١").replace("2".toRegex(), "٢")
            .replace("3".toRegex(), "٣").replace("4".toRegex(), "٤")
            .replace("5".toRegex(), "٥").replace("6".toRegex(), "٦")
            .replace("7".toRegex(), "٧").replace("8".toRegex(), "٨")
            .replace("9".toRegex(), "٩").replace("0".toRegex(), "٠")
}

for i in `seq 100`;
 do
tesseract /home/owner/Downloads/output/$i.tif  /home/owner/Downloads/data/output$i  -l ara_number
 done