enable downsampling of images to make faster communication

src/main/java/ai/nets/samj/models/AbstractSamJ.java

@@ -219,7 +219,7 @@ public void setUsinIJRoiManager(boolean isUsingIJRoiManager) {
 	 * @param designationOfTheScript
 	 * 	the name (or some string to design) of the text that is going to be printed
 	 */
-	public void printScript(final String script, final String designationOfTheScript) {
+	public <T extends RealType<T> & NativeType<T>> void printScript(final String script, final String designationOfTheScript) {
 		if (!isDebugging) return;
 		debugPrinter.printText("START: =========== "+designationOfTheScript+" ===========");
 		debugPrinter.printText(LocalDateTime.now().toString());
@@ -272,6 +272,8 @@ void setImage(RandomAccessibleInterval<T> rai) throws IOException, RuntimeExcept
 			this.targetDims = new long[] {0, 0, 0};
 			this.imageSmall = false;
 			return;
+		} else {
+			scale = 1;
 		}
 		this.script = "";
 		sendImgLib2AsNp();
@@ -329,7 +331,7 @@ else if (task.status == TaskStatus.CRASHED)
 
 	protected <T extends RealType<T> & NativeType<T>> 
 	void sendImgLib2AsNp() {
-		createSHMArray((RandomAccessibleInterval<T>) this.img);
+		createSHMArray(Cast.unchecked(this.img));
 	}
 
 	private <T extends RealType<T> & NativeType<T>> void sendCropAsNp(long[] cropSize) {
@@ -346,13 +348,19 @@ else if (cropSize.length == 3 && cropSize[2] != 3)
 		targetDims = crop.dimensionsAsLongArray();
 
 		scale = (int) (Math.min(targetDims[0], targetDims[1]) / MAX_IMG_SIZE);
-		RandomAccessibleInterval<T> subsampledCrop = Views.subsample(crop, 
-				new long[] {scale, scale, 1});
-		targetReescaledDims = crop.dimensionsAsLongArray();
-		createSHMArray(subsampledCrop);
+		scale = Math.max(scale, 1);
+		if (scale == 1) {
+			createSHMArray(crop);
+		} else {
+			RandomAccessibleInterval<T> subsampledCrop = Views.subsample(crop, 
+					new long[] {scale, scale, 1});
+			targetReescaledDims = subsampledCrop.dimensionsAsLongArray();
+			createSHMArray(subsampledCrop);
+		}
 
 	}
 
+	@SuppressWarnings("unchecked")
 	private List<Mask> processAndRetrieveContours(HashMap<String, Object> inputs) 
 			throws IOException, RuntimeException, InterruptedException {
 		Map<String, Object> results = null;
@@ -388,7 +396,6 @@ else if (task.outputs.get("rle") == null)
 		final Iterator<List<Number>> contours_y = ((List<List<Number>>)results.get("contours_y")).iterator();
 		final Iterator<List<Number>> rles = ((List<List<Number>>)results.get("rle")).iterator();
 		final List<Mask> masks = new ArrayList<Mask>(contours_x_container.size());
-		Rectangle cropRect = new Rectangle((int) encodeCoords[0], (int) encodeCoords[1], (int) targetDims[0], (int) targetDims[1]);
 		while (contours_x.hasNext()) {
 			int[] xArr = contours_x.next().stream().mapToInt(Number::intValue).toArray();
 			int[] yArr = contours_y.next().stream().mapToInt(Number::intValue).toArray();
@@ -565,8 +572,8 @@ public List<Mask> processPoints(List<int[]> pointsList, List<int[]> pointsNegLis
 	private List<int[]> adaptPointPrompts(List<int[]> pointsList) {
 		pointsList = pointsList.stream().map(pp -> {
 			int[] newPoint = new int[2];
-			newPoint[0] = (int) ((pp[0] - this.encodeCoords[0]) / scale);
-			newPoint[1] = (int) ((pp[1] - this.encodeCoords[1]) / scale);
+			newPoint[0] = (int) Math.ceil((pp[0] - this.encodeCoords[0]) / (double) scale);
+			newPoint[1] = (int) Math.ceil((pp[1] - this.encodeCoords[1]) / (double) scale);
 			return newPoint;
 			}).collect(Collectors.toList());
 		return pointsList;
@@ -619,9 +626,9 @@ public List<Mask> processBox(int[] boundingBox, boolean returnAll)
 				reencodeCrop();
 			}
 		}
-		int[] adaptedBoundingBox = new int[] {(int) ((boundingBox[0] - encodeCoords[0]) / scale), 
-				(int) ((boundingBox[1] - encodeCoords[1]) / scale),
-				(int) ((boundingBox[2] - encodeCoords[0]) / scale), (int) ((boundingBox[3] - encodeCoords[1]) / scale)};;
+		int[] adaptedBoundingBox = new int[] {(int) Math.ceil((boundingBox[0] - encodeCoords[0]) / (double) scale), 
+				(int) Math.ceil((boundingBox[1] - encodeCoords[1]) / (double) scale),
+				(int) Math.ceil((boundingBox[2] - encodeCoords[0]) / (double) scale), (int) Math.ceil((boundingBox[3] - encodeCoords[1]) / (double) scale)};;
 		this.script = "";
 		processBoxWithSAM(returnAll);
 		HashMap<String, Object> inputs = new HashMap<String, Object>();
@@ -761,7 +768,6 @@ private void evaluateReencodingNeeded(List<int[]> pointsList, List<int[]> points
 		Rectangle alreadyEncoded = getCurrentlyEncodedArea();
 		Rectangle neededArea = getApproximateAreaNeeded(pointsList, pointsNegList, rect);
 		if (rect.equals(alreadyEncoded)) neededArea = getApproximateAreaNeeded(pointsList, pointsNegList);
-		ArrayList<int[]> notInRect = getPointsNotInRect(pointsList, pointsNegList, rect);
 
 		if (rectContainsRect(alreadyEncoded, neededArea)
 				&& alreadyEncoded.width * 0.7 < extendedRect.width && alreadyEncoded.height * 0.7 < extendedRect.height) {

src/main/java/ai/nets/samj/models/EfficientSamJ.java

@@ -201,7 +201,6 @@ protected void createEncodeImageScript() {
 		code += "im_shm = shared_memory.SharedMemory(name='"
 							+ shma.getNameForPython() + "', size=" + shma.getSize() 
 							+ ")" + System.lineSeparator();
-		code += "task.update('here')" + System.lineSeparator();
 		int size = 1;
 		for (long l : targetDims) {size *= l;}
 		code += "im = np.ndarray(" + size + ", dtype='" + CommonUtils.getDataTypeFromRAI(Cast.unchecked(shma.getSharedRAI()))
@@ -210,19 +209,15 @@ protected void createEncodeImageScript() {
 			code += ll + ", ";
 		code = code.substring(0, code.length() - 2);
 		code += "])" + System.lineSeparator();
-		code += "task.update('here')" + System.lineSeparator();
 		//code += "np.save('/home/carlos/git/crop.npy', im)" + System.lineSeparator();
 		code += "input_h = im.shape[1]" + System.lineSeparator();
 		code += "input_w = im.shape[0]" + System.lineSeparator();
 		code += "globals()['input_h'] = input_h" + System.lineSeparator();
 		code += "globals()['input_w'] = input_w" + System.lineSeparator();
-		code += "task.update('here')" + System.lineSeparator();
 		//code += "task.update(str(im.shape))" + System.lineSeparator();
 		code += "im = torch.from_numpy(np.transpose(im, (2, 1, 0)))" + System.lineSeparator();
-		code += "task.update('here')" + System.lineSeparator();
 		//code += "task.update('after ' + str(im.shape))" + System.lineSeparator();
 		code += "im_shm.unlink()" + System.lineSeparator();
-		code += "task.update('here')" + System.lineSeparator();
 		this.script += code;
 		this.script += ""
 				+ "_ = predictor.get_image_embeddings(im[None, ...])" + System.lineSeparator();

src/main/java/ai/nets/samj/models/Sam2.java

@@ -20,6 +20,7 @@
 package ai.nets.samj.models;
 
 import java.util.ArrayList;
+import java.util.Arrays;
 import java.util.List;
 
 import ai.nets.samj.install.Sam2EnvManager;
@@ -33,6 +34,7 @@
 
 import io.bioimage.modelrunner.tensor.shm.SharedMemoryArray;
 import io.bioimage.modelrunner.utils.CommonUtils;
+import net.imglib2.Cursor;
 import net.imglib2.RandomAccessibleInterval;
 import net.imglib2.converter.RealTypeConverters;
 import net.imglib2.img.array.ArrayImgs;
@@ -286,13 +288,15 @@ protected void createEncodeImageScript() {
 		script += "im_shm = shared_memory.SharedMemory(name='"
 				+ shma.getNameForPython() + "', size=" + shma.getSize() 
 				+ ")" + System.lineSeparator();
-		int size = 1;
-		for (long l : targetDims) {size *= l;}
+		int size = (int) targetDims[2];
+		for (int i = 0; i < targetDims.length - 1; i ++) {
+			size *= Math.ceil(targetDims[i] / (double) scale);
+			}
 		script += "im = np.ndarray(" + size + ", dtype='" + CommonUtils.getDataTypeFromRAI(Cast.unchecked(shma.getSharedRAI()))
 				+ "', buffer=im_shm.buf).reshape([";
-		for (long ll : targetDims)
-			script += ll + ", ";
-		script = script.substring(0, script.length() - 2);
+		for (int i = 0; i < targetDims.length - 1; i ++)
+			script += (int) Math.ceil(targetDims[i] / (double) scale) + ", ";
+		script += targetDims[2];
 		script += "])" + System.lineSeparator();
 		script += "im = np.transpose(im, (1, 0, 2))" + System.lineSeparator();
 		//code += "np.save('/home/carlos/git/aa.npy', im)" + System.lineSeparator();
@@ -436,9 +440,10 @@ public static List<String> getListOfSupportedVariants(){
 	private <T extends RealType<T> & NativeType<T>>
 	void adaptImageToModel(RandomAccessibleInterval<T> ogImg, RandomAccessibleInterval<T> targetImg) {
 		if (ogImg.numDimensions() == 3 && ogImg.dimensionsAsLongArray()[2] == 3) {
-			for (int i = 0; i < 3; i ++) 
+			for (int i = 0; i < 3; i ++) {
 				RealTypeConverters.copyFromTo( ImgLib2Utils.convertViewToRGB(Views.hyperSlice(ogImg, 2, i), this.debugPrinter), 
 						Views.hyperSlice(targetImg, 2, i) );
+			}
 		} else if (ogImg.numDimensions() == 3 && ogImg.dimensionsAsLongArray()[2] == 1) {
 			debugPrinter.printText("CONVERTED 1 CHANNEL IMAGE INTO 3 TO BE FEEDED TO SAMJ");
 			IntervalView<UnsignedByteType> resIm =