Stop tess4j

[Alex Lopes]

How can i stop the Tess4j process or set a time out?

[Quan Nguyen]

See https://github.com/nguyenq/tess4j/issues/29

[Alex Lopes]

Thanks for reply, but I tested the code and still doesn't working. Sorry if it's a stupid question, I'm a beginner and trying to understand it. I have to set a diferent class for API? I'm using 'api = TessAPI.INSTANCE;'

[Quan Nguyen]

Not really. There is an example in the unit tests -- you may want to take a look.

[Alex Lopes]

Hi, i've tried this week make what you said, but i hadn't sucess. I made the code of TessAPITest but it doesn't stopped, and I tried to use the 'end_time' variable, and it doesn't make sense to me. I read the doc and it says "time to stop if not 0", but it ever stop at same time and with blank out put.

My code:

Tesseract.java

    ETEXT_DESC monitor = new ETEXT_DESC();
    ITessAPI.TimeVal timeout = new ITessAPI.TimeVal();
    timeout.tv_sec = new NativeLong(0L); 
    monitor.end_time = timeout;
    ProgressMonitor pmo = new ProgressMonitor(monitor);
    pmo.start();
    api.TessBaseAPIRecognize(handle, monitor);

ProgressMonitor.java

public class ProgressMonitor extends Thread {

    ITessAPI.ETEXT_DESC monitor;
    StringBuilder outputMessage = new StringBuilder();
    long startTime;

    private static final Logger logger = LoggerFactory.getLogger(new LoggHelper().toString());

    public ProgressMonitor(ITessAPI.ETEXT_DESC monitor) {
        this.monitor = monitor;
        this.startTime = System.currentTimeMillis();
    }

    public String getMessage() {
        return outputMessage.toString();
    }

    @Override
    public void run() {
        try {
            while (true) {
                logger.info("ocr alive: " + (monitor.ocr_alive == TRUE));
                logger.info("progress: " + monitor.progress);
                outputMessage.append(monitor.more_to_come);
                System.out.println((System.currentTimeMillis()-endTime)/1000);
                if (((System.currentTimeMillis()-tempoInicio)/1000) == 10 ) {
                    cancel();
                    break;
                }
                Thread.sleep(100);
            }
        } catch (Exception ioe) {
            ioe.printStackTrace();
        }
    }

    /**
     * Cancels OCR operation.
     */
    public void cancel() {
        monitor.cancel = new ITessAPI.CANCEL_FUNC() {
            @Override
            public boolean invoke(Pointer cancel_this, int words) {
                System.out.println("Calling with " + words  +" words");
                return true;
            }
        };
    }

    /**
     * Resets cancel flag.
     */
    public void reset() {
        monitor.cancel = null;
    }
}

[Quan Nguyen]

It looks almost correct. I would make a small correction:

if (((System.currentTimeMillis() - startTime) / 1000) >= 10 ) {

Nevertheless, it does not work anyway when I test now. If I remember correctly, it worked at one time when I first coded the cancellation. Things could have changed since inside Tesseract's native code.

[Alex Lopes]

Do you have any sugestion or another way to stop the tesseract process?

[Quan Nguyen]

Tess4J operates on individual images. If you perform OCR on multiple images, as in multi-page TIFF, you can use SwingWorker to control, and cancel, the OCR process, which would still complete on the current image but not subsequent ones.

[Alex Lopes]

If I send a pdf file with a sigle page, would it work?

[Quan Nguyen]

No, it would still finish that page. If the PDF has multiple pages, then yes.