Possibility of showing a progress bar when running an algorithm in a python plugin

Python SDK
1

Hi all,

When running an algorithm in a python plugin, let’s say an inference with an nnUNet which takes a minute or two, is it possible to show in the GUI a progress bar or any sign that it is working like a loading image? Or can we deactivate the compute button until the computation is finished?

Many thanks for your help,
Farid

2

Hi Farid,
I am afraid this is not possible at the moment, but this is already on our roadmap.
We will let you know when this feature is available.

3

Hi Raphael,

Thanks for the info. Good to know it is already on the roadmap.

4

we found a workaround for now that seems to do the job. the solution was to create and show another window in a separate thread. I put the code here in case someone else would need it:

    def show_loading_window(self, stop_event):
        """Function to create and display the loading window."""
        window = tk.Tk()
        window.title("Loading")
        window.geometry("200x100")
        
        # Set the background color of the window
        window.configure(bg="darkgray")
        
        # Create a label with white text and dark gray background
        label = tk.Label(
            window,
            text="Loading...",
            font=("Arial", 14),
            bg="darkgray",
            fg="white"  # Text color
        )
        label.pack(expand=True, fill="both")
        
        # Check periodically if the stop event is set
        def check_stop():
            if stop_event.is_set():
                window.destroy()  # Close the window
            else:
                window.after(100, check_stop)  # Check again after 100 ms

        # Start the periodic check
        check_stop()
        
        # Start the tkinter event loop
        window.mainloop()

    def compute(self):
        stop_event = threading.Event()  # Event to signal when to stop the loading window
        # Start the loading window in a separate thread
        loading_thread = threading.Thread(target=self.show_loading_window, args=(stop_event,), daemon=True)
        loading_thread.start()
        self.load_checkpoint()
        image_imfusion = self.image  # Assuming a single image is provided
        img_path = self.tmp_path
        imfusion.save(image_imfusion, img_path)
        img, props = SimpleITKIO().read_images([img_path])
        props = fix_properties(props)
        image_imfusion = image_imfusion[0]
        affine = image_imfusion.world_to_image_matrix
        spacing = image_imfusion.spacing
        output = self.predictor.predict_single_npy_array(
            img,
            props,
        )
        output = np.expand_dims(output, axis=-1)
        image_out = imfusion.SharedImage(output)
        image_out.world_to_image_matrix = affine
        image_out.spacing = spacing
        self.imageset_out.add(image_out)
        self.imageset_out.modality = imfusion.Data.Modality.LABEL

        # Once done, set the stop event to close the loading window
        print("Main task completed. Closing loading window...")
        stop_event.set()

        # Ensure the loading thread has time to finish
        loading_thread.join()
1 Like