Minor old fixes, bumped to 2.0.0 release

build.gradle

@@ -23,7 +23,7 @@ apply plugin: 'io.codearte.nexus-staging'
 
 group 'com.uddernetworks.newocr'
 archivesBaseName = "NewOCR"
-version '2.0.0-SNAPSHOT'
+version '2.0.0'
 
 sourceCompatibility = 12
 

src/main/java/com/uddernetworks/newocr/recognition/DefaultScannedImage.java

@@ -20,12 +20,14 @@ public class DefaultScannedImage implements ScannedImage {
 
     private transient File originalFile;
 
+    private BufferedImage binarizedImage;
     private transient BufferedImage originalImage;
 
     private final Int2ObjectMap<List<ImageLetter>> grid = new Int2ObjectLinkedOpenHashMap<>();
 
-    public DefaultScannedImage(File originalFile, BufferedImage originalImage) {
+    public DefaultScannedImage(File originalFile, BufferedImage binarizedImage, BufferedImage originalImage) {
         this.originalFile = originalFile;
+        this.binarizedImage = binarizedImage;
         this.originalImage = originalImage;
     }
 
@@ -102,6 +104,11 @@ public Int2ObjectMap.Entry<List<ImageLetter>> getLineEntry(int y) {
         return grid.int2ObjectEntrySet().stream().skip(y).findFirst().orElse(null);
     }
 
+    @Override
+    public BufferedImage getBinarizedImage() {
+        return binarizedImage;
+    }
+
     @Override
     public BufferedImage getOriginalImage() {
         return originalImage;

src/main/java/com/uddernetworks/newocr/recognition/OCRScan.java

@@ -17,6 +17,7 @@
 import org.slf4j.LoggerFactory;
 
 import javax.imageio.ImageIO;
+import java.awt.image.BufferedImage;
 import java.io.File;
 import java.util.*;
 import java.util.concurrent.CompletableFuture;
@@ -103,6 +104,7 @@ public ScannedImage scanImage(File file) {
         var inputOptional = this.options.getImageReadMethod().apply(file);
         if (inputOptional.isEmpty()) throw new RuntimeException("Input file not found!");
         var input = inputOptional.get();
+        var originalInput = copyBufferedImage(input);
         var values = OCRUtils.createGrid(input);
         var searchCharacters = new ArrayList<SearchCharacter>();
 
@@ -172,7 +174,7 @@ public ScannedImage scanImage(File file) {
 
         // Sorts the lines again based on X values, to move spaces from the back to their proper locations in the line.
 
-        ScannedImage scannedImage = new DefaultScannedImage(file, input);
+        ScannedImage scannedImage = new DefaultScannedImage(file, input, originalInput);
 
         sortedLines.keySet().stream().sorted().forEach(y -> {
             List<ImageLetter> line = sortedLines.get(y.intValue());
@@ -234,4 +236,9 @@ public int spaceRound(double input) {
         known += OCRUtils.diff(extra, 1) < 0.2D ? 1 : 0;
         return known;
     }
+
+    private BufferedImage copyBufferedImage(BufferedImage image) {
+        var colorModel = image.getColorModel();
+        return new BufferedImage(colorModel, image.copyData(null), colorModel.isAlphaPremultiplied(), null);
+    }
 }

src/main/java/com/uddernetworks/newocr/recognition/ScannedImage.java

@@ -88,6 +88,13 @@ public interface ScannedImage {
      */
     Int2ObjectMap.Entry<List<ImageLetter>> getLineEntry(int y);
 
+    /**
+     * Gets the binarized image before any preprocessing
+     *
+     * @return The binarized image, which may be null if pulled from caches
+     */
+    BufferedImage getBinarizedImage();
+
     /**
      * Gets the original image scanned by the OCR.
      *

src/main/java/com/uddernetworks/newocr/train/ImageReadMethod.java

@@ -38,7 +38,7 @@ public Optional<BufferedImage> apply(File file) {
         @Override
         public Optional<BufferedImage> apply(File file) {
             try {
-                return Optional.ofNullable(OCRUtils.readImage(file));
+                return Optional.of(OCRUtils.readImage(file));
             } catch (IOException e) {
                 LOGGER.error("Error reading file with ImageReadMethod.IMAGE_ICON", e);
                 return Optional.empty();