Windows based app with a numpad (numbers 0 to 9), I have a screen shot of 8 button on the numpad and Sikulixlibrary always find and clicks it, but when I try to us “SikuliLibrary.Click Text 8” it never finds the 8 button on the numpad.
What am I doing wrong?
I have also tried setting a region of interest wit this code
${appcoords}= SikuliLibrary.Get Image Coordinates ${IMAGE_DIR}app-image.png
SikuliLibrary.Set Roi ${appcoords} 5
SikuliLibrary.Select region ${appcoords}
SikuliLibrary.Click Text 8
I have also tried replacing Click Text with Region Click Text but still results in the 8 button not being found.
I’ve done very small stuff with tesseract in the past and main obstacle i had was that it didn’t recognize all of the fonts used in the toy stuff i was working on. Maybe in this case, it needs to be trained with the fonts in the images OCR is trying to read into actual text ? Eg, stuff like this: Tesseract training 101: How to prepare the data - Pretius
PS. that linked article might be a bit outdated as tesseract now ships with tools to do the process. Google is friend in this - as said, ive only dealt with toy stuff with tesseract many years ago..
I added “Set Ocr Text Read true” to the test but it didn’t fix the issue.
Prior to opening this question I tired it in the original Java Sikuli app and it does work.