.gitlab-ci.yml

@@ -1,7 +1,3 @@
-variables:
-  #  SONAR_PROJECT_KEY: 'ocr-service:ocr-service-server'
-  GIT_SUBMODULE_STRATEGY: recursive
-  GIT_SUBMODULE_FORCE_HTTPS: 'true'
 include:
   - project: 'gitlab/gitlab'
     ref: 'main'

.gitmodules

@@ -1,8 +0,0 @@
-[submodule "ocr-service-v1/ocr-service-server/src/test/resources/files/syngenta"]
-	path = ocr-service-v1/ocr-service-server/src/test/resources/files/syngenta
-	url = https://gitlab.knecon.com/fforesight/documents/syngenta.git
-	update = merge
-[submodule "ocr-service-v1/ocr-service-server/src/test/resources/files/basf"]
-	path = ocr-service-v1/ocr-service-server/src/test/resources/files/basf
-	url = https://gitlab.knecon.com/fforesight/documents/basf.git
-	update = merge

README.md

@@ -15,17 +15,11 @@ The service uses PDFTron to attempt the removal of invisible elements and waterm
   Extracts all images from the PDF using PDFBox
 3. Striped Image Detection and Stitching  
   Detects if images are striped and stitches them together using Ghostscript.
-4. Image Processing
-   - Convert to grayscale
-   - Upscale to target DPI
-   - Filter using Gauss kernel
-   - Binarizes the resulting images using Leptonica and the Otsu thresholding algorithm.
-   - Despeckle using various morphological operations 
+4. Binarization  
+  Binarizes the resulting images using Leptonica and the Otsu thresholding algorithm.
 5. OCR Processing  
 Runs Tesseract on the images to extract text.
-6. Font style detection  
-Detection of bold text using stroke width estimation
-7. Text Integration  
+6. Text Integration  
 Draws the resulting text onto the original PDF using PDFBox.
 
 Steps 2.-5. happen in parallel and communicate via a blocking queue to limit RAM usage. 
@@ -74,14 +68,12 @@ String languages = "deu+eng"; // Defines languages loaded into Tesseract as 3-ch
 ```
 ## Integration
 
-The OCR-service communicates via RabbitMQ and uses the queues `ocr_request_queue`, `ocr_response_queue`,
-`ocr_dead_letter_queue`, and `ocr_status_update_response_queue`. 
+The OCR-service communicates via RabbitMQ and uses the queues `ocrQueue`, `ocrDLQ`, and `ocr_status_update_response_queue`. 
 
-### ocr_request_queue
+### ocrQueue
 This queue is used to start the OCR process, a DocumentRequest must be passed as a message. The service will then download the PDF from the provided cloud storage.
-### ocr_response_queue
-This queue is also used to signal the end of processing.
-### ocr_dead_letter_queue
-This queue is used to signal an error has occurred during processing.
 ### ocr_status_update_response_queue
 This queue is used by the OCR service to give updates about the progress of the ongoing OCR on a image per image basis. The total amount may change, when less images are found than initially assumed.
+This queue is also used to signal the end of processing.
+### ocrDLQ
+This queue is used to signal an error has occurred during processing.

buildSrc/src/main/kotlin/com.iqser.red.service.java-conventions.gradle.kts

@@ -12,10 +12,6 @@ group = "com.knecon.fforesight.service"
 java.sourceCompatibility = JavaVersion.VERSION_17
 java.targetCompatibility = JavaVersion.VERSION_17
 
-pmd {
-    isConsoleOutput = true
-}
-
 tasks.pmdMain {
     pmd.ruleSetFiles = files("${rootDir}/config/pmd/pmd.xml")
 }
@@ -29,8 +25,6 @@ tasks.named<Test>("test") {
     reports {
         junitXml.outputLocation.set(layout.buildDirectory.dir("reports/junit"))
     }
-    minHeapSize = "512m"
-    maxHeapSize = "8192m"
 }
 
 tasks.test {

config/pmd/pmd.xml

@@ -9,12 +9,12 @@
     </description>
 
     <rule ref="category/java/errorprone.xml">
+        <exclude name="DataflowAnomalyAnalysis"/>
         <exclude name="MissingSerialVersionUID"/>
+        <exclude name="NullAssignment"/>
         <exclude name="AvoidLiteralsInIfCondition"/>
         <exclude name="AvoidDuplicateLiterals"/>
-        <exclude name="NullAssignment"/>
-        <exclude name="AssignmentInOperand"/>
-        <exclude name="BeanMembersShouldSerialize"/>
+        <exclude name="AvoidFieldNameMatchingMethodName"/>
     </rule>
 
-</ruleset>
+</ruleset>

config/pmd/test_pmd.xml

@@ -10,13 +10,14 @@
 
 
     <rule ref="category/java/errorprone.xml">
+        <exclude name="DataflowAnomalyAnalysis"/>
         <exclude name="MissingSerialVersionUID"/>
+        <exclude name="NullAssignment"/>
         <exclude name="AvoidLiteralsInIfCondition"/>
         <exclude name="AvoidDuplicateLiterals"/>
-        <exclude name="NullAssignment"/>
-        <exclude name="AssignmentInOperand"/>
+        <exclude name="AvoidFieldNameMatchingMethodName"/>
+        <exclude name="AvoidFieldNameMatchingTypeName"/>
         <exclude name="TestClassWithoutTestCases"/>
-        <exclude name="BeanMembersShouldSerialize"/>
     </rule>
 
-</ruleset>
+</ruleset>

ocr-service-v1/ocr-service-api/build.gradle.kts

@@ -1,7 +1,7 @@
 plugins {
     `maven-publish`
     id("com.iqser.red.service.java-conventions")
-    id("io.freefair.lombok") version "8.4"
+    id("io.freefair.lombok") version "8.2.2"
 }
 
 publishing {

ocr-service-v1/ocr-service-api/src/main/java/com/knecon/fforesight/service/ocr/v1/api/model/DocumentRequest.java

@@ -13,12 +13,5 @@ public class DocumentRequest {
 
     protected String dossierId;
     protected String fileId;
-    protected boolean removeWatermark;
 
-    public DocumentRequest(String dossierId, String fileId) {
-
-        this.dossierId = dossierId;
-        this.fileId = fileId;
-    }
-    
 }

ocr-service-v1/ocr-service-api/src/main/java/com/knecon/fforesight/service/ocr/v1/api/model/OCRStatusUpdateResponse.java

@@ -15,6 +15,5 @@ public class OCRStatusUpdateResponse {
     private int numberOfPagesToOCR;
     private int numberOfOCRedPages;
     private boolean ocrFinished;
-    private boolean ocrStarted;
 
 }

ocr-service-v1/ocr-service-processor/build.gradle.kts

@@ -1,6 +1,6 @@
 plugins {
     id("com.iqser.red.service.java-conventions")
-    id("io.freefair.lombok") version "8.4"
+    id("io.freefair.lombok") version "8.2.2"
 }
 
 configurations {
@@ -14,18 +14,15 @@ dependencies {
     api("net.sourceforge.tess4j:tess4j:5.8.0")
     api("com.iqser.red.commons:metric-commons:2.1.0")
     api("com.iqser.red.commons:storage-commons:2.45.0")
-    api("com.knecon.fforesight:tenant-commons:0.21.0")
-    api("com.knecon.fforesight:lifecycle-commons:0.6.0")
+    api("com.knecon.fforesight:tenant-commons:0.14.0")
     api("com.pdftron:PDFNet:10.5.0")
     api("org.apache.pdfbox:pdfbox:3.0.0")
     api("org.apache.pdfbox:jbig2-imageio:3.0.4")
     api("com.github.jai-imageio:jai-imageio-core:1.4.0")
     api("com.github.jai-imageio:jai-imageio-jpeg2000:1.4.0")
-    api("org.apache.commons:commons-math3:3.6.1")
-    api("io.github.karols:hocr4j:0.2.0")
+    api("io.github.karols:hocr4j:0.1.2")
     api("com.amazonaws:aws-java-sdk-kms:1.12.440")
     api("com.google.guava:guava:31.1-jre")
-    api("com.iqser.red.commons:pdftron-logic-commons:2.32.0")
-    api("com.knecon.fforesight:viewer-doc-processor:0.125.0")
+    api("com.iqser.red.commons:pdftron-logic-commons:2.20.0")
     testImplementation("org.junit.jupiter:junit-jupiter:5.8.1")
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/OcrServiceProcessorConfiguration.java

@@ -1,26 +1,14 @@
 package com.knecon.fforesight.service.ocr.processor;
 
-import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.boot.context.properties.EnableConfigurationProperties;
-import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.ComponentScan;
 import org.springframework.context.annotation.Configuration;
 
 import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
-
-import io.micrometer.observation.ObservationRegistry;
 
 @Configuration
 @ComponentScan
 @EnableConfigurationProperties(OcrServiceSettings.class)
 public class OcrServiceProcessorConfiguration {
 
-    @Bean
-    @Autowired
-    public ViewerDocumentService viewerDocumentService(ObservationRegistry registry) {
-
-        return new ViewerDocumentService(registry);
-    }
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/initializer/PDFNetInitializer.java

@@ -1,20 +1,17 @@
 package com.knecon.fforesight.service.ocr.processor.initializer;
 
-import org.springframework.beans.factory.annotation.Value;
-import org.springframework.stereotype.Component;
-
 import com.pdftron.pdf.PDFNet;
 import com.sun.jna.NativeLibrary;
 
 import jakarta.annotation.PostConstruct;
 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
-import lombok.extern.slf4j.Slf4j;
+import org.springframework.beans.factory.annotation.Value;
+import org.springframework.stereotype.Component;
 
-@Slf4j
 @Component
 @RequiredArgsConstructor
-public class NativeLibrariesInitializer {
+public class PDFNetInitializer {
 
     @Value("${pdftron.license:}")
     private String pdftronLicense;
@@ -25,25 +22,8 @@ public class NativeLibrariesInitializer {
     // Do not change back to application runner, if it is application runner it takes messages from the queue before PDFNet is initialized, that leads to UnsatisfiedLinkError.
     public void init() {
 
-        log.info("Initializing Native Libraries");
-        log.info("Setting pdftron license: {}", pdftronLicense);
         PDFNet.setTempPath("/tmp/pdftron");
         PDFNet.initialize(pdftronLicense);
-
-        log.info("Setting jna.library.path: {}", System.getenv("VCPKG_DYNAMIC_LIB"));
         System.setProperty("jna.library.path", System.getenv("VCPKG_DYNAMIC_LIB"));
-
-        log.info("Asserting Native Libraries loaded");
-
-        try (NativeLibrary leptonicaLib = NativeLibrary.getInstance("leptonica")) {
-            assert leptonicaLib != null;
-            log.info("Leptonica library loaded from {}", leptonicaLib.getFile().getAbsolutePath());
-        }
-
-        try (NativeLibrary tesseractLib = NativeLibrary.getInstance("tesseract")) {
-            assert tesseractLib != null;
-            log.info("Tesseract library loaded from {}", tesseractLib.getFile().getAbsolutePath());
-        }
     }
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/ExtractedOcrImage.java

@@ -58,4 +58,11 @@ public class ExtractedOcrImage implements OcrImage {
         return affineTransform;
     }
 
+
+    @Override
+    public int getOptimalPageSegmentationMode() {
+
+        return ITessAPI.TessPageSegMode.PSM_SINGLE_BLOCK;
+    }
+
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/OcrImage.java

@@ -11,7 +11,6 @@ import lombok.SneakyThrows;
 import net.sourceforge.lept4j.Leptonica1;
 import net.sourceforge.lept4j.Pix;
 import net.sourceforge.lept4j.util.LeptUtils;
-import net.sourceforge.tess4j.ITessAPI;
 
 public interface OcrImage {
 
@@ -31,19 +30,9 @@ public interface OcrImage {
     int getNumberOnPage();
 
 
-    /**
-     * Retrieves the height of the original image (not necessarily in pdf coordinates).
-     *
-     * @return the height of the image
-     */
     int getHeight();
 
 
-    /**
-     * Retrieves the width of the original image (not necessarily in pdf coordinates).
-     *
-     * @return the width of the image
-     */
     int getWidth();
 
 
@@ -54,7 +43,7 @@ public interface OcrImage {
      */
     default QuadPoint getImageBounds() {
 
-        // cannot be solved with a nice rotation matrix. After rotating the text coordinates in the image will always start at (0,0) and will therefore always start at (0,0) in the PDF.
+        // cannot be solved with a nice rotation matrix, since the after rotating the text coordinates in the image will always start at (0,0) and will therefore always start at (0,0) in the PDF.
         // So in order to mimic this behavior we need to start with (0,0) coordinates always.
         if (getRotationDegrees() == 90 || getRotationDegrees() == 270) {
             return new QuadPoint(new Point2D.Double(0, 0), new Point2D.Double(0, getWidth()), new Point2D.Double(getHeight(), getWidth()), new Point2D.Double(getHeight(), 0));
@@ -75,6 +64,13 @@ public interface OcrImage {
     }
 
 
+    @SneakyThrows
+    default BufferedImage getBufferedImage() {
+
+        return LeptUtils.convertPixToImage(getPix());
+    }
+
+
     /**
      * Retrieves the rotation degree of the OCR image.
      *
@@ -88,13 +84,7 @@ public interface OcrImage {
      *
      * @return The optimal page segmentation mode.
      */
-    default int getOptimalPageSegmentationMode() {
-
-        if (getWidth() < 200 || getHeight() < 200) {
-            return ITessAPI.TessPageSegMode.PSM_SINGLE_BLOCK;
-        }
-        return ITessAPI.TessPageSegMode.PSM_AUTO;
-    } // TODO: evaluate if PSM can be dynamically chosen to increase performance
+    int getOptimalPageSegmentationMode(); // TODO: evaluate if PSM can be dynamically chosen to increase performance
 
 
     /**
@@ -119,6 +109,17 @@ public interface OcrImage {
     AffineTransform getImageCTM();
 
 
+    /**
+     * Retrieves the size (width * height) of the image.
+     *
+     * @return The size of the image.
+     */
+    default int getImageSize() {
+
+        return getHeight() * getWidth();
+    }
+
+
     default void destroyPix() {
 
         LeptUtils.disposePix(getPix());

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/OcrResult.java

@@ -7,17 +7,27 @@ import com.knecon.fforesight.service.ocr.processor.service.HOcrPageParser;
 
 import io.github.karols.hocr4j.Word;
 
-public record OcrResult(OcrImage image, String tesseractOutputFilePath) {
+public record OcrResult(Image image, String hOcrPageAbsolutePath) {
 
     public static OcrResult create(OcrImage image, String tesseractResult) {
 
-        return new OcrResult(image, tesseractResult);
+        return new OcrResult(Image.fromOcrImage(image), tesseractResult);
     }
 
 
     public List<Word> getAllWords() {
 
-        return HOcrPageParser.extractHocrPage(tesseractOutputFilePath).getAllWords();
+        return HOcrPageParser.extractHocrPage(hOcrPageAbsolutePath).getAllWords();
+    }
+
+
+    public record Image(Integer pageNumber, AffineTransform ctm, QuadPoint position) {
+
+        public static Image fromOcrImage(OcrImage image) {
+
+            return new Image(image.getPageNumber(), image.getImageCTM(), image.getImageCoordinatesInInitialUserSpace());
+        }
+
     }
 
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/OcrResultToWrite.java

@@ -1,42 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.model;
-
-import java.util.List;
-import java.util.Map;
-import java.util.stream.Collectors;
-
-import com.knecon.fforesight.service.ocr.processor.model.scriptdetection.FontStyleDetectionModel;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontStyle;
-
-public record OcrResultToWrite(List<TextPositionInImage> textPositionInImage, QuadPoint imageBoundingBox) {
-
-    public static OcrResultToWrite fromFontStyleDetectionModel(FontStyleDetectionModel fontStyleDetectionModel) {
-
-        return new OcrResultToWrite(fontStyleDetectionModel.getTextPositionInImages(), fontStyleDetectionModel.getImageBounds());
-    }
-
-
-    public static Map<Integer, List<OcrResultToWrite>> buildOcrResultsToWrite(List<OcrResult> ocrResults, FontMetricsFactory fontMetricsFactory) {
-
-        return ocrResults.stream()
-                .collect(Collectors.groupingBy(ocrResult -> ocrResult.image().getPageNumber()))
-                .entrySet()
-                .stream()
-                .collect(Collectors.toMap(Map.Entry::getKey,
-                        entry -> entry.getValue()
-                                .stream()
-                                .map(ocrResult -> new OcrResultToWrite(toTextPositionInImage(ocrResult, fontMetricsFactory), ocrResult.image().getImageCoordinatesInInitialUserSpace()))
-                                .toList()));
-    }
-
-
-    private static List<TextPositionInImage> toTextPositionInImage(OcrResult ocrResult, FontMetricsFactory fontMetricsFactory) {
-
-        return ocrResult.getAllWords()
-                .stream()
-                .filter(word -> !word.isBlank())
-                .map(word -> new TextPositionInImage(word, ocrResult.image().getImageCTM(), fontMetricsFactory, FontStyle.REGULAR))
-                .toList();
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/PageInformation.java

@@ -1,42 +1,12 @@
 package com.knecon.fforesight.service.ocr.processor.model;
 
-import java.awt.geom.Rectangle2D;
-
 import org.apache.pdfbox.pdmodel.PDPage;
-import org.apache.pdfbox.pdmodel.common.PDRectangle;
 
-public record PageInformation(Rectangle2D mediabox, int number, int rotationDegrees) {
+public record PageInformation(int height, int width, int number, int rotationDegrees) {
 
     public static PageInformation fromPDPage(int pageNum, PDPage page) {
 
-        PDRectangle mediaBox = page.getMediaBox();
-        return new PageInformation(new Rectangle2D.Double(mediaBox.getLowerLeftX(), mediaBox.getLowerLeftY(), mediaBox.getWidth(), mediaBox.getHeight()),
-                                   pageNum,
-                                   page.getRotation());
-    }
-
-
-    public double height() {
-
-        return mediabox.getHeight();
-    }
-
-
-    public double width() {
-
-        return mediabox.getWidth();
-    }
-
-
-    public double minX() {
-
-        return mediabox.getX();
-    }
-
-
-    public double minY() {
-
-        return mediabox.getY();
+        return new PageInformation((int) page.getCropBox().getHeight(), (int) page.getCropBox().getWidth(), pageNum, page.getRotation());
     }
 
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/QuadPoint.java

@@ -1,6 +1,5 @@
 package com.knecon.fforesight.service.ocr.processor.model;
 
-import java.awt.Rectangle;
 import java.awt.geom.AffineTransform;
 import java.awt.geom.Line2D;
 import java.awt.geom.Point2D;
@@ -35,16 +34,6 @@ public record QuadPoint(Point2D a, Point2D b, Point2D c, Point2D d) {
                 new Point2D.Double(bounds.getRight(), bounds.getBottom()));
     }
 
-    public Rectangle2D getBounds2D() {
-
-        double minX = Math.min(Math.min(Math.min(a.getX(), b.getX()), c.getX()), d.getX());
-        double minY = Math.min(Math.min(Math.min(a.getY(), b.getY()), c.getY()), d.getY());
-        double maxX = Math.max(Math.max(Math.max(a.getX(), b.getX()), c.getX()), d.getX());
-        double maxY = Math.max(Math.max(Math.max(a.getY(), b.getY()), c.getY()), d.getY());
-
-        return new Rectangle2D.Double(minX, minY, maxX - minX, maxY - minY);
-    }
-
 
     public QuadPoint getTransformed(AffineTransform at) {
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/RenderedPageOcrImage.java

@@ -1,12 +1,20 @@
 package com.knecon.fforesight.service.ocr.processor.model;
 
 import java.awt.geom.AffineTransform;
+import java.awt.geom.Point2D;
+
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.PDPage;
 
 import lombok.AccessLevel;
 import lombok.Getter;
 import lombok.RequiredArgsConstructor;
+import lombok.Setter;
+import lombok.SneakyThrows;
 import lombok.experimental.FieldDefaults;
+import net.sourceforge.lept4j.Leptonica1;
 import net.sourceforge.lept4j.Pix;
+import net.sourceforge.tess4j.ITessAPI;
 
 @Getter
 @RequiredArgsConstructor
@@ -20,6 +28,13 @@ public class RenderedPageOcrImage implements OcrImage {
     int rotationDegrees;
 
 
+    @Override
+    public int getOptimalPageSegmentationMode() {
+
+        return ITessAPI.TessPageSegMode.PSM_SINGLE_BLOCK;
+    }
+
+
     @Override
     public AffineTransform getImageCTM() {
 
@@ -52,6 +67,17 @@ public class RenderedPageOcrImage implements OcrImage {
     }
 
 
+    @Override
+    public QuadPoint getImageBounds() {
+
+        if (rotationDegrees == 90 || rotationDegrees == 270) {
+            return new QuadPoint(new Point2D.Double(0, 0), new Point2D.Double(0, width), new Point2D.Double(height, width), new Point2D.Double(height, 0));
+        } else {
+            return new QuadPoint(new Point2D.Double(0, 0), new Point2D.Double(0, height), new Point2D.Double(width, height), new Point2D.Double(width, 0));
+        }
+    }
+
+
     @Override
     public int getPageNumber() {
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/TextPositionInImage.java

@@ -7,35 +7,29 @@ import org.apache.pdfbox.pdmodel.font.PDFont;
 import org.apache.pdfbox.util.Matrix;
 
 import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontStyle;
 
+import io.github.karols.hocr4j.Bounds;
 import io.github.karols.hocr4j.Word;
 import lombok.AccessLevel;
 import lombok.Getter;
-import lombok.Setter;
 import lombok.experimental.FieldDefaults;
 
 @Getter
-@FieldDefaults(level = AccessLevel.PRIVATE)
+@FieldDefaults(level = AccessLevel.PRIVATE, makeFinal = true)
 public class TextPositionInImage {
 
-    final QuadPoint position;
-    final String text;
-    final AffineTransform imageCTM;
-
-    @Setter
+    QuadPoint position;
+    String text;
+    AffineTransform imageCTM;
     FontMetricsFactory fontMetricsFactory;
-    @Setter
-    FontStyle fontStyle;
 
 
-    public TextPositionInImage(Word word, AffineTransform imageCTM, FontMetricsFactory fontMetricsFactory, FontStyle fontStyle) {
+    public TextPositionInImage(Word word, AffineTransform imageCTM, FontMetricsFactory fontMetricsFactory) {
 
         this.position = QuadPoint.fromBounds(word.getBounds());
         this.text = word.getText();
         this.imageCTM = imageCTM;
         this.fontMetricsFactory = fontMetricsFactory;
-        this.fontStyle = fontStyle;
     }
 
 
@@ -96,13 +90,6 @@ public class TextPositionInImage {
     }
 
 
-    public double getTextHeight() {
-
-        var metrics = fontMetricsFactory.calculateMetrics(text, getTransformedWidth(), getTransformedHeight());
-        return fontMetricsFactory.calculateFontSize(text, getTransformedWidth()) * metrics.getHeightScaling();
-    }
-
-
     public double getHeight() {
 
         return position.a().distance(position.b());

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/scriptdetection/FontStyleDetectionModel.java

@@ -1,58 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.model.scriptdetection;
-
-import java.util.List;
-
-import com.knecon.fforesight.service.ocr.processor.model.OcrImage;
-import com.knecon.fforesight.service.ocr.processor.model.OcrResult;
-import com.knecon.fforesight.service.ocr.processor.model.QuadPoint;
-import com.knecon.fforesight.service.ocr.processor.model.TextPositionInImage;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-
-import lombok.AccessLevel;
-import lombok.Getter;
-import lombok.RequiredArgsConstructor;
-import lombok.experimental.FieldDefaults;
-import net.sourceforge.lept4j.Leptonica1;
-import net.sourceforge.lept4j.Pix;
-import net.sourceforge.lept4j.util.LeptUtils;
-
-@Getter
-@RequiredArgsConstructor
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-public final class FontStyleDetectionModel {
-
-    QuadPoint imageBounds;
-    Pix image;
-    List<TextPositionAndWordImage> textPositionsAndWordImages;
-
-
-    public static FontStyleDetectionModel fromOcrResult(OcrResult ocrResult, FontMetricsFactory fontMetricsFactory, OcrServiceSettings settings) {
-
-        var image = Leptonica1.pixRead(ocrResult.tesseractOutputFilePath() + ".tiff");
-        var wordPixes = ocrResult.getAllWords().stream().filter(word -> !word.isBlank()).map(word -> TextPositionAndWordImage.create(ocrResult.image().getImageCTM(), word, image, settings, fontMetricsFactory)).toList();
-
-        return new FontStyleDetectionModel(ocrResult.image().getImageCoordinatesInInitialUserSpace(), image, wordPixes);
-    }
-
-
-    public List<TextPositionInImage> getTextPositionInImages() {
-
-        return textPositionsAndWordImages.stream().map(TextPositionAndWordImage::getTextPositionInImage).toList();
-    }
-
-
-    public List<WordImage> getWordImages() {
-
-        return textPositionsAndWordImages.stream().map(TextPositionAndWordImage::getWordImage).toList();
-    }
-
-
-    public void dispose() {
-
-        LeptUtils.disposePix(image);
-        getWordImages().forEach(WordImage::dispose);
-    }
-
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/scriptdetection/TextPositionAndWordImage.java

@@ -1,52 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.model.scriptdetection;
-
-import java.awt.geom.AffineTransform;
-import java.util.Objects;
-
-import org.apache.commons.math3.ml.clustering.Clusterable;
-
-import com.knecon.fforesight.service.ocr.processor.model.OcrResult;
-import com.knecon.fforesight.service.ocr.processor.model.TextPositionInImage;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontStyle;
-import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-
-import io.github.karols.hocr4j.Word;
-import lombok.Getter;
-import net.sourceforge.lept4j.Pix;
-
-@Getter
-public final class TextPositionAndWordImage implements Clusterable {
-
-    private final TextPositionInImage textPositionInImage;
-    private final WordImage wordImage;
-
-
-    public TextPositionAndWordImage(TextPositionInImage textPositionInImage, WordImage wordImage) {
-
-        this.textPositionInImage = textPositionInImage;
-        this.wordImage = wordImage;
-    }
-
-
-    public static TextPositionAndWordImage create(AffineTransform imageCTM, Word word, Pix image, OcrServiceSettings settings, FontMetricsFactory fontMetricsFactory) {
-
-        TextPositionInImage textPositionInImage = new TextPositionInImage(word, imageCTM, fontMetricsFactory, FontStyle.REGULAR);
-        WordImage wordImage = new WordImage(textPositionInImage.getTextHeight(), word, image, settings);
-        return new TextPositionAndWordImage(textPositionInImage, wordImage);
-    }
-
-
-    @Override
-    public double[] getPoint() {
-
-        return wordImage.getPoint();
-    }
-
-
-    public double getTextHeight() {
-
-        return wordImage.getTextHeight();
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/model/scriptdetection/WordImage.java

@@ -1,71 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.model.scriptdetection;
-
-import org.apache.commons.math3.ml.clustering.Clusterable;
-
-import com.knecon.fforesight.service.ocr.processor.service.scriptdetection.StrokeWidthCalculator;
-import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-import com.knecon.fforesight.service.ocr.processor.utils.ImageProcessingUtils;
-
-import io.github.karols.hocr4j.Word;
-import lombok.AccessLevel;
-import lombok.Getter;
-import lombok.RequiredArgsConstructor;
-import lombok.experimental.FieldDefaults;
-import net.sourceforge.lept4j.Box;
-import net.sourceforge.lept4j.Leptonica1;
-import net.sourceforge.lept4j.Pix;
-import net.sourceforge.lept4j.util.LeptUtils;
-
-@Getter
-@RequiredArgsConstructor
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-public class WordImage implements Clusterable {
-
-    Pix image;
-    String text;
-    double textHeight;
-    OcrServiceSettings settings;
-
-
-    public WordImage(double textHeight, Word word, Pix originalImage, OcrServiceSettings settings) {
-
-        Box box = new Box(word.getBounds().getLeft(), word.getBounds().getTop(), word.getBounds().getWidth(), word.getBounds().getHeight(), 1);
-        this.image = Leptonica1.pixClipRectangle(originalImage, box, null);
-        box.clear();
-        this.text = word.getText();
-        this.textHeight = textHeight;
-        this.settings = settings;
-    }
-
-
-    public boolean hasLargerStrokeWidth(double strokeWidth) {
-
-        int roundedStrokeWidth = (int) Math.round(strokeWidth);
-        double roundingError = (roundedStrokeWidth - strokeWidth) / strokeWidth;
-
-        // add 1 to open a bit bigger than the estimated regular stroke width
-        Pix openedPix = Leptonica1.pixOpenBrick(null, image, roundedStrokeWidth + 1, roundedStrokeWidth + 1);
-
-        double openedPixelDensity = ImageProcessingUtils.calculatePixelDensity(openedPix);
-
-        double pixelDensity = ImageProcessingUtils.calculatePixelDensity(image);
-
-        LeptUtils.disposePix(openedPix);
-
-        return (openedPixelDensity * (1 + roundingError)) / pixelDensity > (settings.getBoldThreshold());
-    }
-
-
-    @Override
-    public double[] getPoint() {
-
-        return new double[]{textHeight};
-    }
-
-
-    public void dispose() {
-
-        LeptUtils.disposePix(image);
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/FileStorageService.java

@@ -1,11 +1,13 @@
 package com.knecon.fforesight.service.ocr.processor.service;
 
+import java.io.ByteArrayInputStream;
 import java.io.File;
-import java.io.FileInputStream;
 import java.io.InputStream;
 import java.nio.file.Files;
-import java.nio.file.StandardCopyOption;
+import java.nio.file.Paths;
+import java.nio.file.StandardOpenOption;
 
+import org.apache.commons.io.IOUtils;
 import org.springframework.stereotype.Service;
 
 import com.iqser.red.service.persistence.service.v1.api.shared.model.dossiertemplate.dossier.file.FileType;
@@ -29,38 +31,47 @@ public class FileStorageService {
         return dossierId + "/" + fileId + "." + fileType.name() + fileType.getExtension();
     }
 
+
+    @SneakyThrows
+    public byte[] getOriginalFile(String dossierId, String fileId) {
+
+        try (InputStream inputStream = getInputStream(getStorageId(dossierId, fileId, FileType.ORIGIN))) {
+            return IOUtils.toByteArray(inputStream);
+        }
+    }
+
+
+    @SneakyThrows
+    public InputStream getOriginalFileAsStream(String dossierId, String fileId) {
+
+        return getInputStream(getStorageId(dossierId, fileId, FileType.ORIGIN));
+    }
+
+
+    public void storeOriginalFile(String dossierId, String fileId, InputStream stream) {
+
+        storageService.storeObject(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.ORIGIN), stream);
+    }
+
+
     public boolean untouchedFileExists(String dossierId, String fileId) {
 
         return storageService.objectExists(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.UNTOUCHED));
     }
 
-    @SneakyThrows
-    public void storeFiles(String dossierId, String fileId, File documentFile, File viewerDocumentFile) {
 
-        try (var in = new FileInputStream(documentFile)) {
-            storageService.storeObject(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.ORIGIN), in);
-        }
-        try (var in = new FileInputStream(viewerDocumentFile)) {
-            storageService.storeObject(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.VIEWER_DOCUMENT), in);
-        }
+    public void storeUntouchedFile(String dossierId, String fileId, byte[] data) {
+
+        storageService.storeObject(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.UNTOUCHED), new ByteArrayInputStream(data));
     }
 
 
     @SneakyThrows
-    public void downloadFiles(String dossierId, String fileId, File documentFile, File viewerDocumentFile) {
+    private InputStream getInputStream(String storageId) {
 
-        storageService.downloadTo(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.ORIGIN), documentFile);
-        if (storageService.objectExists(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.VIEWER_DOCUMENT))) {
-            storageService.downloadTo(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.VIEWER_DOCUMENT), viewerDocumentFile);
-        } else {
-            Files.copy(documentFile.toPath(), viewerDocumentFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
-        }
-
-        if (!untouchedFileExists(dossierId, fileId)) {
-            try (var in = new FileInputStream(documentFile)) {
-                storageService.storeObject(TenantContext.getTenantId(), getStorageId(dossierId, fileId, FileType.UNTOUCHED), in);
-            }
-        }
+        File tempFile = File.createTempFile("temp", ".data");
+        storageService.downloadTo(TenantContext.getTenantId(), storageId, tempFile);
+        return Files.newInputStream(Paths.get(tempFile.getPath()), StandardOpenOption.DELETE_ON_CLOSE);
     }
 
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/GhostScriptService.java

@@ -3,18 +3,24 @@ package com.knecon.fforesight.service.ocr.processor.service;
 import java.io.InputStream;
 import java.nio.file.Path;
 import java.util.ArrayList;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.LinkedList;
 import java.util.List;
 import java.util.Map;
 import java.util.concurrent.BlockingQueue;
 import java.util.concurrent.LinkedBlockingDeque;
+import java.util.concurrent.LinkedTransferQueue;
 import java.util.stream.Collectors;
 
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.springframework.stereotype.Service;
 
+import com.azure.core.implementation.GeoObjectHelper;
+import com.knecon.fforesight.service.ocr.processor.model.OcrImage;
+import com.knecon.fforesight.service.ocr.processor.model.PageInformation;
 import com.knecon.fforesight.service.ocr.processor.model.RenderedPageImageFile;
+import com.knecon.fforesight.service.ocr.processor.model.RenderedPageOcrImage;
 import com.knecon.fforesight.service.ocr.processor.model.UnprocessedImage;
 import com.knecon.fforesight.service.ocr.processor.service.threads.BlockingQueueFiller;
 import com.knecon.fforesight.service.ocr.processor.service.threads.GhostScriptOutputHandler;
@@ -26,6 +32,7 @@ import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
 import lombok.experimental.FieldDefaults;
 import lombok.extern.slf4j.Slf4j;
+import net.sourceforge.lept4j.Pix;
 
 @Slf4j
 @Service

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/IOcrMessageSender.java

@@ -7,10 +7,7 @@ public interface IOcrMessageSender {
 
     void sendUpdate(String fileId, int finishedImages, int totalImages);
 
-    void sendOCRStarted(String fileId);
 
     void sendOcrFinished(String fileId, int totalImages);
 
-    void sendOcrResponse(String dossierId, String fileId);
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/OCRService.java

@@ -1,15 +1,14 @@
 package com.knecon.fforesight.service.ocr.processor.service;
 
+import java.io.ByteArrayInputStream;
 import java.io.ByteArrayOutputStream;
 import java.io.File;
-import java.io.FileInputStream;
-import java.io.FileOutputStream;
-import java.nio.file.Files;
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
 import java.nio.file.Path;
-import java.nio.file.StandardCopyOption;
 import java.util.LinkedList;
 import java.util.List;
-import java.util.Map;
 import java.util.concurrent.ArrayBlockingQueue;
 import java.util.concurrent.BlockingQueue;
 import java.util.stream.IntStream;
@@ -21,16 +20,11 @@ import org.springframework.util.FileSystemUtils;
 
 import com.iqser.red.pdftronlogic.commons.InvisibleElementRemovalService;
 import com.iqser.red.pdftronlogic.commons.WatermarkRemovalService;
-import com.knecon.fforesight.service.ocr.processor.model.OcrResultToWrite;
 import com.knecon.fforesight.service.ocr.processor.model.OcrImage;
 import com.knecon.fforesight.service.ocr.processor.model.OcrResult;
-import com.knecon.fforesight.service.ocr.processor.service.scriptdetection.FontStyleDetector;
 import com.knecon.fforesight.service.ocr.processor.service.threads.OCRThread;
 import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-import com.pdftron.pdf.PDFDoc;
 
-import io.micrometer.observation.ObservationRegistry;
-import io.micrometer.observation.annotation.Observed;
 import lombok.AccessLevel;
 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
@@ -50,8 +44,6 @@ public class OCRService {
     InvisibleElementRemovalService invisibleElementRemovalService;
     OcrResultWriter ocrResultWriter;
     GhostScriptService ghostScriptService;
-    FontStyleDetector boldDetector;
-    ObservationRegistry registry;
 
 
     /**
@@ -59,66 +51,54 @@ public class OCRService {
      * looking for stitchedImages (if so converting the current page to an image with ghostscript and work on this instead),
      * perform tesseract-ocr on these images (via threads) and write the generated ocr-text as invisible elements.
      *
-     * @param dossierId          Id of dossier
-     * @param fileId             Id of file
-     * @param tmpDir             working directory for all files
-     * @param documentFile       the file to perform ocr on, results are written invisibly
-     * @param viewerDocumentFile debugging file, results are written visibly in an optional content group
+     * @param dossierId Id of dossier
+     * @param fileId    Id of file
+     * @param out       OutputStream where to write to
      */
-    @Observed(name = "OCRService", contextualName = "run-ocr-on-document")
     @SneakyThrows
-    public void runOcrOnDocument(String dossierId, String fileId, boolean removeWatermark, Path tmpDir, File documentFile, File viewerDocumentFile) {
+    public void runOcrOnDocument(String dossierId, String fileId, OutputStream out) {
 
-        if (removeWatermark) {
-            removeWatermarkIfEnabled(documentFile);
+        try (InputStream fileStream = removeWatermarkIfEnabled(dossierId, fileId); ByteArrayOutputStream transferOutputStream = new ByteArrayOutputStream()) {
+
+            invisibleElementRemovalService.removeInvisibleElements(fileStream, transferOutputStream, false);
+
+            try (InputStream transferInputStream = new ByteArrayInputStream(transferOutputStream.toByteArray())) {
+                log.info("Starting OCR for file {}", fileId);
+                long ocrStart = System.currentTimeMillis();
+                Statistics stats = runOcr(transferInputStream, out, fileId, dossierId);
+                long ocrEnd = System.currentTimeMillis();
+                log.info("ocr successful for file with dossierId {} and fileId {}, took {}s", dossierId, fileId, String.format("%.1f", (ocrEnd - ocrStart) / 1000.0));
+                log.info("Runtime breakdown: {}", stats);
+            }
         }
-        removeInvisibleElements(documentFile);
+    }
 
-        log.info("Starting OCR for file {}", fileId);
-        long ocrStart = System.currentTimeMillis();
-        Statistics stats = runOcr(tmpDir, documentFile, viewerDocumentFile, fileId, dossierId);
-        long ocrEnd = System.currentTimeMillis();
-        log.info("ocr successful for file with dossierId {} and fileId {}, took {}s", dossierId, fileId, String.format("%.1f", (ocrEnd - ocrStart) / 1000.0));
-        log.info("Runtime breakdown: {}", stats);
 
+    private InputStream removeWatermarkIfEnabled(String dossierId, String fileId) throws IOException {
+
+        if (settings.isRemoveWatermark()) {
+            try (var in = fileStorageService.getOriginalFileAsStream(dossierId, fileId); var transferOutputStream = new ByteArrayOutputStream()) {
+                watermarkRemovalService.removeWatermarks(in, transferOutputStream);
+                return new ByteArrayInputStream(transferOutputStream.toByteArray());
+            }
+        }
+        return fileStorageService.getOriginalFileAsStream(dossierId, fileId);
     }
 
 
     @SneakyThrows
-    private void removeInvisibleElements(File originFile) {
-
-        Path tmpFile = Files.createTempFile("invisibleElements", ".pdf");
-        try (var in = new FileInputStream(originFile); var out = new FileOutputStream(tmpFile.toFile())) {
-            invisibleElementRemovalService.removeInvisibleElements(in, out, false, false);
-        }
-        Files.copy(tmpFile, originFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
-        assert tmpFile.toFile().delete();
-    }
-
-
-    @SneakyThrows
-    private void removeWatermarkIfEnabled(File originFile) {
-
-        Path tmpFile = Files.createTempFile("removeWatermarks", ".pdf");
-        try (var in = new FileInputStream(originFile); var out = new FileOutputStream(tmpFile.toFile())) {
-            watermarkRemovalService.removeWatermarks(in, out);
-        }
-        Files.copy(tmpFile, originFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
-        assert tmpFile.toFile().delete();
-    }
-
-
-    @SneakyThrows
-    public Statistics runOcr(Path tmpDir, File documentFile, File viewerDocumentFile, String fileId, String dossierId) {
+    public Statistics runOcr(InputStream in, OutputStream out, String fileId, String dossierId) {
 
         long timestamp;
-
+        Path tmpDir = Path.of(OsUtils.getTemporaryDirectory()).resolve(dossierId + "-" + fileId);
         Path tmpImageDir = tmpDir.resolve("images");
         Path tesseractOutputDir = tmpDir.resolve("tesseract_output");
 
         tesseractOutputDir.toFile().mkdirs();
         tmpImageDir.toFile().mkdirs();
 
+        File documentFile = OsUtils.writeFileToTmpFolder(in, tmpDir);
+
         Statistics stats;
         try (PDDocument document = Loader.loadPDF(documentFile)) {
             OcrProgressLogger logger = new OcrProgressLogger(document.getNumberOfPages(), ocrMessageSender, fileId);
@@ -155,18 +135,14 @@ public class OCRService {
                 ocrThread.join();
             }
 
-            log.info("Tesseract OCR has finished for file {} and dossier {}", fileId, dossierId);
-
+            log.info("OCR processing has finished, writing results");
             timestamp = System.currentTimeMillis();
-            Map<Integer, List<OcrResultToWrite>> imageWithTextPositionsPerPage = boldDetector.detectBold(ocrResults, document);
-            stats.increaseFontStyleDetectionDuration(System.currentTimeMillis() - timestamp);
-
-            timestamp = System.currentTimeMillis();
-            ocrResultWriter.drawOcrResultsToPdf(documentFile, viewerDocumentFile, imageWithTextPositionsPerPage);
-
+            var dictionariesToUpdate = ocrResultWriter.drawOcrResultsToPdf(document, ocrResults);
             log.info("Saving document");
+            document.saveIncremental(out, dictionariesToUpdate);
             stats.increaseWritingTextDuration(System.currentTimeMillis() - timestamp);
 
+            FileSystemUtils.deleteRecursively(tmpDir);
             logger.sendFinished();
             return stats;
         }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/OcrImageFactory.java

@@ -96,7 +96,6 @@ public class OcrImageFactory {
         }
 
         imageProcessingThread.setAllImagesExtracted(true);
-        imageProcessingThread.interrupt();
 
         imageProcessingThread.join();
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/OcrResultWriter.java

@@ -1,38 +1,31 @@
 package com.knecon.fforesight.service.ocr.processor.service;
 
 import java.awt.Color;
-import java.awt.geom.Line2D;
 import java.awt.geom.Point2D;
-import java.awt.geom.Rectangle2D;
-import java.io.File;
-import java.io.FileInputStream;
-import java.util.ArrayList;
-import java.util.Collection;
-import java.util.HashMap;
-import java.util.LinkedList;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Map;
-import java.util.Optional;
-import java.util.function.Function;
-import java.util.stream.Stream;
+import java.util.Set;
+import java.util.stream.Collectors;
 
+import org.apache.pdfbox.cos.COSDictionary;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.PDDocumentCatalog;
+import org.apache.pdfbox.pdmodel.PDPage;
+import org.apache.pdfbox.pdmodel.PDPageContentStream;
+import org.apache.pdfbox.pdmodel.PDResources;
+import org.apache.pdfbox.pdmodel.graphics.optionalcontent.PDOptionalContentGroup;
+import org.apache.pdfbox.pdmodel.graphics.optionalcontent.PDOptionalContentProperties;
 import org.apache.pdfbox.pdmodel.graphics.state.RenderingMode;
 import org.springframework.stereotype.Service;
 
-import com.iqser.red.pdftronlogic.commons.Converter;
-import com.knecon.fforesight.service.ocr.processor.model.OcrResultToWrite;
+import com.knecon.fforesight.service.ocr.processor.model.OcrResult;
 import com.knecon.fforesight.service.ocr.processor.model.QuadPoint;
 import com.knecon.fforesight.service.ocr.processor.model.TextPositionInImage;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontStyle;
-import com.knecon.fforesight.service.viewerdoc.ContentStreams;
-import com.knecon.fforesight.service.viewerdoc.model.ColoredLine;
-import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
-import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
-import com.knecon.fforesight.service.viewerdoc.model.VisualizationsOnPage;
-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
-import com.pdftron.pdf.PDFDoc;
-import com.pdftron.pdf.Page;
-import com.pdftron.pdf.TextExtractor;
+import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
+import com.knecon.fforesight.service.ocr.processor.service.fonts.Type0FontMetricsFactory;
+import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
 
 import lombok.AccessLevel;
 import lombok.RequiredArgsConstructor;
@@ -46,179 +39,185 @@ import lombok.extern.slf4j.Slf4j;
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class OcrResultWriter {
 
-    public static final Color REGULAR_TEXT_COLOR = Color.BLUE;
-    public static final Color BOLD_TEXT_COLOR = Color.CYAN;
-
-    public static final Color REGULAR_TEXT_IN_IGNORE_ZONE = Color.RED;
-    public static final Color BOLD_TEXT_IN_IGNORE_ZONE = Color.RED;
-
-    ViewerDocumentService viewerDocumentService;
+    static String ocrLayerName = "knecon OCR";
+    OcrServiceSettings settings;
 
 
     @SneakyThrows
-    public void drawOcrResultsToPdf(File document, File viewerDocument, Map<Integer, List<OcrResultToWrite>> imagesWithResultsPerPage) {
+    public Set<COSDictionary> drawOcrResultsToPdf(PDDocument document, List<OcrResult> ocrResults) {
 
-        Map<Integer, VisualizationsOnPage> ocrVisualizationsOnPages = new HashMap<>();
-        Map<Integer, VisualizationsOnPage> ocrTextDebugVisualizationsOnPages = new HashMap<>();
-        Map<Integer, VisualizationsOnPage> ocrBBoxDebugVisualizationsOnPages = new HashMap<>();
-
-        try (var in = new FileInputStream(document); PDFDoc doc = new PDFDoc(in)) {
-
-            for (Integer pageNumber : imagesWithResultsPerPage.keySet()) {
-
-                List<Rectangle2D> textBBoxes = getTextBBoxes(doc.getPage(pageNumber));
-
-                ocrVisualizationsOnPages.put(pageNumber - 1, createVisualizations(imagesWithResultsPerPage.get(pageNumber), textBBoxes));
-                ocrTextDebugVisualizationsOnPages.put(pageNumber - 1, createDebugTextVisualizations(imagesWithResultsPerPage.get(pageNumber), textBBoxes));
-                ocrBBoxDebugVisualizationsOnPages.put(pageNumber - 1, createDebugBBoxVisualizations(imagesWithResultsPerPage.get(pageNumber)));
-            }
-        }
-
-        Visualizations visualizations = new Visualizations(ContentStreams.KNECON_OCR, ocrVisualizationsOnPages, false);
-
-        List<Visualizations> debugVisualizations = List.of(visualizations,
-                                                           new Visualizations(ContentStreams.KNECON_OCR_TEXT_DEBUG, ocrTextDebugVisualizationsOnPages, false),
-                                                           new Visualizations(ContentStreams.KNECON_OCR_BBOX_DEBUG, ocrBBoxDebugVisualizationsOnPages, false));
-
-        viewerDocumentService.addVisualizationsOnPage(document, document, List.of(visualizations));
-        viewerDocumentService.addVisualizationsOnPage(viewerDocument, viewerDocument, debugVisualizations);
-    }
-
-
-    @SuppressWarnings("PMD")
-    private List<Rectangle2D> getTextBBoxes(Page page) {
-
-        List<Rectangle2D> textBBoxes = new ArrayList<>();
-        try (var textExtractor = new TextExtractor()) {
-            textExtractor.begin(page);
-            try {
-
-                for (TextExtractor.Line line = textExtractor.getFirstLine(); line.isValid(); line = getNextLine(line)) {
-                    for (TextExtractor.Word word = line.getFirstWord(); word.isValid(); word = getNextWord(word)) {
-                        textBBoxes.add(Converter.toRectangle2D(word.getBBox()));
-                    }
-                }
-            } catch (Exception e) {
-                log.warn("Could not get word dimension, {}", e.getMessage());
-            }
-            return textBBoxes;
-        }
-    }
-
-
-    private static TextExtractor.Word getNextWord(TextExtractor.Word word) {
-
-        TextExtractor.Word nextWord = word.getNextWord();
-        word.close();
-        return nextWord;
-    }
-
-
-    private static TextExtractor.Line getNextLine(TextExtractor.Line line) {
-
-        TextExtractor.Line newLine = line.getNextLine();
-        line.close();
-        return newLine;
-    }
-
-
-    private VisualizationsOnPage createVisualizations(List<OcrResultToWrite> ocrResultsToWrite, List<Rectangle2D> ignoreZones) {
-
-        List<TextPositionInImage> words = ocrResultsToWrite.stream()
-                .map(OcrResultToWrite::textPositionInImage)
-                .flatMap(Collection::stream)
-                .filter(word -> ignoreZones.stream()
-                        .noneMatch(ignoreZone -> word.getTransformedTextBBox().getBounds2D().intersects(ignoreZone)))
-                .toList();
-
-        List<PlacedText> placedTexts = words.stream()
-                .map(word -> new PlacedText(word.getText(),
-                                            null,
-                                            Color.BLACK,
-                                            (float) word.getFontSize(),
-                                            word.getFontMetricsFactory(),
-                                            Optional.of(word.getTextMatrix()),
-                                            Optional.of(RenderingMode.NEITHER)))
-                .toList();
-        return VisualizationsOnPage.builder().placedTexts(placedTexts).build();
-    }
-
-
-    private VisualizationsOnPage createDebugTextVisualizations(List<OcrResultToWrite> ocrResultsToWrite, List<Rectangle2D> textBBoxes) {
-
-        List<TextPositionInImage> wordsToDraw = new ArrayList<>();
-        List<TextPositionInImage> ignoredWords = new ArrayList<>();
-
-        for (OcrResultToWrite ocrResultToWrite : ocrResultsToWrite) {
-            for (TextPositionInImage textPositionInImage : ocrResultToWrite.textPositionInImage()) {
-                if (textBBoxes.stream()
-                        .anyMatch(ignoreZone -> textPositionInImage.getTransformedTextBBox().getBounds2D().intersects(ignoreZone))) {
-                    ignoredWords.add(textPositionInImage);
-                } else {
-                    wordsToDraw.add(textPositionInImage);
-                }
-            }
-        }
-        Stream<PlacedText> placedTexts = wordsToDraw.stream()
-                .map(word -> new PlacedText(word.getText(),
-                                            null,
-                                            word.getFontStyle().equals(FontStyle.REGULAR) ? REGULAR_TEXT_COLOR : BOLD_TEXT_COLOR,
-                                            (float) word.getFontSize(),
-                                            word.getFontMetricsFactory(),
-                                            Optional.of(word.getTextMatrix()),
-                                            Optional.of(RenderingMode.FILL)));
-
-        Stream<PlacedText> placedTexts2 = ignoredWords.stream()
-                .map(word -> new PlacedText(word.getText(),
-                                            null,
-                                            word.getFontStyle().equals(FontStyle.REGULAR) ? REGULAR_TEXT_IN_IGNORE_ZONE : BOLD_TEXT_IN_IGNORE_ZONE,
-                                            (float) word.getFontSize(),
-                                            word.getFontMetricsFactory(),
-                                            Optional.of(word.getTextMatrix()),
-                                            Optional.of(RenderingMode.FILL)));
-
-        return VisualizationsOnPage.builder()
-                .placedTexts(Stream.of(placedTexts, placedTexts2)
-                                     .flatMap(Function.identity())
-                                     .toList())
-                .build();
-    }
-
-
-    private VisualizationsOnPage createDebugBBoxVisualizations(List<OcrResultToWrite> ocrResultsToWrite) {
-
-        List<TextPositionInImage> words = ocrResultsToWrite.stream()
-                .map(OcrResultToWrite::textPositionInImage)
-                .flatMap(Collection::stream)
-                .toList();
-        List<ColoredLine> coloredLines = Stream.concat(//
-                                                       words.stream()
-                                                               .map(TextPositionInImage::getTransformedTextBBox)
-                                                               .map(this::quadPointAsLines),//
-                                                       ocrResultsToWrite.stream()
-                                                               .map(OcrResultToWrite::imageBoundingBox)
-                                                               .map(this::createGrid)//
-                )
-                .flatMap(Collection::stream)
-                .toList();
-        return VisualizationsOnPage.builder().coloredLines(coloredLines).build();
-    }
-
-
-    private List<ColoredLine> quadPointAsLines(QuadPoint rect) {
-
-        return List.of(new ColoredLine(new Line2D.Double(rect.a(), rect.b()), Color.ORANGE, 1),
-                       new ColoredLine(new Line2D.Double(rect.b(), rect.c()), Color.BLUE, 1),
-                       new ColoredLine(new Line2D.Double(rect.c(), rect.d()), Color.GREEN, 1),
-                       new ColoredLine(new Line2D.Double(rect.d(), rect.a()), Color.MAGENTA, 1));
+        FontMetricsFactory fontMetricsFactory = new Type0FontMetricsFactory(document);
+        Set<COSDictionary> dictionariesToUpdate = new HashSet<>();
+        Map<Integer, List<OcrResult>> resultsPerPage = ocrResults.stream().collect(Collectors.groupingBy(result -> result.image().pageNumber()));
+        resultsPerPage.keySet().forEach(pageNumber -> drawResultsPerPage(document, pageNumber, resultsPerPage, dictionariesToUpdate, fontMetricsFactory));
+        dictionariesToUpdate.add(document.getDocumentInformation().getCOSObject());
+        return dictionariesToUpdate;
     }
 
 
     @SneakyThrows
-    private List<ColoredLine> createGrid(QuadPoint rect) {
+    private void drawResultsPerPage(PDDocument document, Integer pageNumber, Map<Integer, List<OcrResult>> resultsPerPage, Set<COSDictionary> dictionariesToUpdate, FontMetricsFactory fontMetricsFactory) {
 
-        List<ColoredLine> lines = new LinkedList<>(quadPointAsLines(rect));
+        var pdPage = document.getPage(pageNumber - 1);
 
+        PDOptionalContentGroup textDebugLayer = new PDOptionalContentGroup(ocrLayerName);
+        PDOptionalContentGroup bBoxDebugLayer = new PDOptionalContentGroup(ocrLayerName + "BBox");
+        if (settings.isDebug()) {
+            textDebugLayer = addOptionalGroup(ocrLayerName, document, pdPage, dictionariesToUpdate);
+            bBoxDebugLayer = addOptionalGroup(ocrLayerName + " BBox", document, pdPage, dictionariesToUpdate);
+        }
+
+        escapeContentStreams(document, pdPage);
+
+        List<TextPositionInImage> words = buildTextPositionsOnPage(pageNumber, resultsPerPage, fontMetricsFactory);
+        try (var contentStream = new PDPageContentStream(document, pdPage, PDPageContentStream.AppendMode.APPEND, true)) {
+
+            // write invisible ocr text inside tagged content
+            contentStream.beginMarkedContent(settings.getOcrMarkedContentTag());
+            contentStream.saveGraphicsState();
+            contentStream.setNonStrokingColor(Color.BLUE);
+            contentStream.setStrokingColor(Color.BLUE);
+            contentStream.setLineWidth(1);
+            words.forEach(word -> drawInvisibleWord(word, contentStream));
+            contentStream.restoreGraphicsState();
+            contentStream.endMarkedContent();
+
+            if (settings.isDebug()) { // must not be written, as it will interfere with layout parsing
+                // write visible ocr text inside optional group
+                contentStream.beginMarkedContent(COSName.OC, textDebugLayer);
+                contentStream.saveGraphicsState();
+                contentStream.setNonStrokingColor(Color.BLUE);
+                words.forEach(word -> drawVisibleWord(word, contentStream));
+                contentStream.restoreGraphicsState();
+                contentStream.endMarkedContent();
+
+                // write word bounding boxes (tesseract output) inside optional group
+                contentStream.beginMarkedContent(COSName.OC, bBoxDebugLayer);
+                contentStream.saveGraphicsState();
+                resultsPerPage.get(pageNumber).stream().map(OcrResult::image).forEach(image -> drawGrid(contentStream, image.position()));
+                words.stream().map(TextPositionInImage::getTransformedTextBBox).forEach(word -> drawRectangle(contentStream, word));
+                contentStream.restoreGraphicsState();
+                contentStream.endMarkedContent();
+            }
+        }
+        dictionariesToUpdate.add(pdPage.getCOSObject());
+        dictionariesToUpdate.add(pdPage.getResources().getCOSObject());
+    }
+
+
+    private static List<TextPositionInImage> buildTextPositionsOnPage(Integer pageNumber, Map<Integer, List<OcrResult>> resultsPerPage, FontMetricsFactory fontMetricsFactory) {
+
+        return resultsPerPage.get(pageNumber)
+                .stream()
+                .flatMap(result -> result.getAllWords().stream().filter(word -> !word.isBlank()).map(word -> new TextPositionInImage(word, result.image().ctm(), fontMetricsFactory)))
+                .toList();
+    }
+
+
+    @SneakyThrows
+    private static void escapeContentStreams(PDDocument document, PDPage pdPage) {
+        // We need to append to the contentstream, otherwise the content could be overlapped by images
+        // But we also need to save the graphics state before, such that our appended content cannot be affected by previous contentstreams with side-effects, such as not escaped matrix transformations
+        try (var contentStream = new PDPageContentStream(document, pdPage, PDPageContentStream.AppendMode.PREPEND, false)) {
+            contentStream.saveGraphicsState();
+        }
+        try (var contentStream = new PDPageContentStream(document, pdPage, PDPageContentStream.AppendMode.APPEND, false)) {
+            contentStream.restoreGraphicsState();
+        }
+    }
+
+
+    private PDOptionalContentGroup addOptionalGroup(String ocrLayerName, PDDocument document, PDPage pdPage, Set<COSDictionary> dictionariesToUpdate) {
+
+        PDDocumentCatalog catalog = document.getDocumentCatalog();
+        PDOptionalContentProperties ocprops = catalog.getOCProperties();
+        if (ocprops == null) {
+            ocprops = new PDOptionalContentProperties();
+            catalog.setOCProperties(ocprops);
+        }
+        PDOptionalContentGroup layer = null;
+        if (ocprops.hasGroup(ocrLayerName)) {
+            layer = ocprops.getGroup(ocrLayerName);
+        } else {
+            layer = new PDOptionalContentGroup(ocrLayerName);
+            ocprops.addGroup(layer);
+        }
+
+        // enable debug layers by default only when DEBUG flag is set.
+        ocprops.setGroupEnabled(layer, settings.isDebug());
+        PDResources resources = pdPage.getResources();
+        if (resources == null) {
+            resources = new PDResources();
+            pdPage.setResources(resources);
+        }
+        dictionariesToUpdate.add(catalog.getCOSObject());
+        return layer;
+    }
+
+
+    @SneakyThrows
+    private void drawRectangle(PDPageContentStream contentStream, QuadPoint rect) {
+
+        contentStream.saveGraphicsState();
+        contentStream.setLineWidth(1);
+        contentStream.moveTo((float) rect.a().getX(), (float) rect.a().getY());
+        contentStream.lineTo((float) rect.b().getX(), (float) rect.b().getY());
+        contentStream.setStrokingColor(Color.ORANGE);
+        contentStream.stroke();
+        contentStream.moveTo((float) rect.b().getX(), (float) rect.b().getY());
+        contentStream.lineTo((float) rect.c().getX(), (float) rect.c().getY());
+        contentStream.setStrokingColor(Color.BLUE);
+        contentStream.stroke();
+        contentStream.moveTo((float) rect.c().getX(), (float) rect.c().getY());
+        contentStream.lineTo((float) rect.d().getX(), (float) rect.d().getY());
+        contentStream.setStrokingColor(Color.GREEN);
+        contentStream.stroke();
+        contentStream.moveTo((float) rect.d().getX(), (float) rect.d().getY());
+        contentStream.lineTo((float) rect.a().getX(), (float) rect.a().getY());
+        contentStream.setStrokingColor(Color.MAGENTA);
+        contentStream.stroke();
+        contentStream.restoreGraphicsState();
+    }
+
+
+    private void drawInvisibleWord(TextPositionInImage word, PDPageContentStream contentStream) {
+
+        drawWord(word, contentStream, RenderingMode.NEITHER);
+    }
+
+
+    private void drawVisibleWord(TextPositionInImage word, PDPageContentStream contentStream) {
+
+        drawWord(word, contentStream, RenderingMode.FILL);
+    }
+
+
+    //    @SneakyThrows
+    private void drawWord(TextPositionInImage position, PDPageContentStream contentStream, RenderingMode renderingMode) {
+
+        try {
+            contentStream.beginText();
+            contentStream.setRenderingMode(renderingMode);
+            contentStream.setFont(position.getFont(), (float) position.getFontSize());
+            contentStream.setTextMatrix(position.getTextMatrix());
+            contentStream.showText(position.getText());
+            contentStream.endText();
+
+        } catch (Exception e) {
+            log.error("Failed to write text {}", position.getText());
+            log.error(e.getMessage());
+        }
+    }
+
+
+    @SneakyThrows
+    private void drawGrid(PDPageContentStream contentStream, QuadPoint rect) {
+
+        drawRectangle(contentStream, rect);
+
+        contentStream.saveGraphicsState();
+        contentStream.setStrokingColor(Color.BLACK);
+        contentStream.setLineWidth(0.2F);
         int nRows = 8;
         int nCols = 8;
 
@@ -226,7 +225,7 @@ public class OcrResultWriter {
         Point2D start = add(rect.a(), abStep);
         Point2D end = add(rect.d(), abStep);
         for (int row = 0; row < nRows; ++row) {
-            lines.add(new ColoredLine(new Line2D.Double(start, end), Color.BLACK, 0.2f));
+            drawLine(start, end, contentStream);
             start = add(start, abStep);
             end = add(end, abStep);
         }
@@ -234,12 +233,21 @@ public class OcrResultWriter {
         start = add(rect.a(), adStep);
         end = add(rect.b(), adStep);
         for (int col = 0; col < nCols; ++col) {
-            lines.add(new ColoredLine(new Line2D.Double(start, end), Color.BLACK, 0.2f));
+            drawLine(start, end, contentStream);
             start = add(start, adStep);
             end = add(end, adStep);
         }
+        contentStream.restoreGraphicsState();
 
-        return lines;
+    }
+
+
+    @SneakyThrows
+    private void drawLine(Point2D a, Point2D b, PDPageContentStream contentStream) {
+
+        contentStream.moveTo((float) a.getX(), (float) a.getY());
+        contentStream.lineTo((float) b.getX(), (float) b.getY());
+        contentStream.stroke();
     }
 
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/Statistics.java

@@ -16,14 +16,12 @@ public class Statistics {
     AtomicLong pdf2ImgDuration;
     AtomicLong writingTextDuration;
     AtomicLong imageProcessingDuration;
-    AtomicLong fontStyleDetectionDuration;
 
 
     public Statistics(int numberOfExtractThreads, int numberOfOcrThreads) {
 
         this.imageExtraction = Collections.synchronizedList(new ArrayList<>(Collections.nCopies(numberOfExtractThreads, 0L)));
         this.tesseractDuration = Collections.synchronizedList(new ArrayList<>(Collections.nCopies(numberOfOcrThreads, 0L)));
-        this.fontStyleDetectionDuration = new AtomicLong(0);
         this.pdf2ImgDuration = new AtomicLong(0);
         this.writingTextDuration = new AtomicLong(0);
         this.imageProcessingDuration = new AtomicLong(0);
@@ -59,17 +57,12 @@ public class Statistics {
         writingTextDuration.addAndGet(duration);
     }
 
-    public void increaseFontStyleDetectionDuration(long duration) {
-
-        fontStyleDetectionDuration.addAndGet(duration);
-    }
-
 
     @Override
     public String toString() {
 
         return String.format(
-                "imageExtraction: mean %.2f s, max %.2f s, min %.2f, tesseract: mean %.2f s, max %.2f s, min %.2f, ImageProcessing=%.2f s, PDF2Img=%.2f s, writingText=%.2f s, FontstyleDetection=%.2f s",
+                "imageExtraction: mean %.2f s, max %.2f s, min %.2f, tesseract: mean %.2f s, max %.2f s, min %.2f, ImageProcessing=%.2f s, PDF2Img=%.2f s, writingText=%.2f s",
                 ((float) imageExtraction.stream().mapToLong(Long::longValue).average().orElse(0) / 1000),
                 ((float) imageExtraction.stream().mapToLong(Long::longValue).max().orElse(0) / 1000),
                 ((float) imageExtraction.stream().mapToLong(Long::longValue).min().orElse(0) / 1000),
@@ -78,8 +71,7 @@ public class Statistics {
                 ((float) tesseractDuration.stream().mapToLong(Long::longValue).min().orElse(0) / 1000),
                 (float) imageProcessingDuration.get() / 1000,
                 (float) pdf2ImgDuration.get() / 1000,
-                (float) writingTextDuration.get() / 1000,
-                (float) fontStyleDetectionDuration.get() / 1000);
+                (float) writingTextDuration.get() / 1000);
     }
 
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/fonts/FontMetricsFactory.java

@@ -4,12 +4,11 @@ import org.apache.pdfbox.pdmodel.font.PDFont;
 
 import com.knecon.fforesight.service.ocr.processor.model.FontMetrics;
 import com.knecon.fforesight.service.ocr.processor.model.HeightAndDescent;
-import com.knecon.fforesight.service.viewerdoc.model.EmbeddableFont;
 
 import lombok.SneakyThrows;
 import lombok.extern.slf4j.Slf4j;
 
-public interface FontMetricsFactory extends EmbeddableFont {
+public interface FontMetricsFactory {
 
     default FontMetrics calculateMetrics(String text, double textWidth, double textHeight) {
 
@@ -37,7 +36,6 @@ public interface FontMetricsFactory extends EmbeddableFont {
 
     PDFont getFont();
 
-
     HeightAndDescent calculateHeightAndDescent(String text);
 
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/fonts/FontStyle.java

@@ -1,5 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.service.fonts;
-
-public enum FontStyle {
-    REGULAR, BOLD, ITALIC
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/fonts/Type0FontMetricsFactory.java

@@ -1,7 +1,6 @@
 package com.knecon.fforesight.service.ocr.processor.service.fonts;
 
 import java.io.ByteArrayInputStream;
-import java.util.Set;
 
 import org.apache.fontbox.ttf.GlyphData;
 import org.apache.fontbox.ttf.TTFParser;
@@ -13,63 +12,26 @@ import org.apache.pdfbox.pdmodel.font.PDType0Font;
 
 import com.knecon.fforesight.service.ocr.processor.model.HeightAndDescent;
 
-import lombok.AllArgsConstructor;
-import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
 import lombok.extern.slf4j.Slf4j;
 
 @Slf4j
-@RequiredArgsConstructor
-@AllArgsConstructor
 public class Type0FontMetricsFactory implements FontMetricsFactory {
 
-    private final String resourcePath;
-    private PDType0Font type0Font;
-    private TrueTypeFont trueTypeFont;
-    private PDDocument documentThisIsEmbeddedIn;
-
-    // for this specific font back-/forward-slashes have a lot of descent screwing up the font size and therefore bold detection. So if we find such a character we ignore its descent.
-    private static final Set<Integer> slashGlyphIds = Set.of(18, 63);
+    private final PDType0Font type0Font;
+    private final TrueTypeFont trueTypeFont;
 
 
     @SneakyThrows
-    public static Type0FontMetricsFactory regular(PDDocument document) {
+    public Type0FontMetricsFactory(PDDocument document) {
 
-        String resourcePath = "fonts/cmu-regular.ttf";
-        return createFromResourcePath(resourcePath, document);
-    }
-
-
-    @SneakyThrows
-    public static Type0FontMetricsFactory bold(PDDocument document) {
-
-        String resourcePath = "fonts/cmu-bold.ttf";
-        return createFromResourcePath(resourcePath, document);
-    }
-
-
-    @SneakyThrows
-    @SuppressWarnings("PMD.CloseResource")
-    private static TrueTypeFont readFromResourcePath(String resourcePath) {
-
-        // The ttf is closed with the document, see PDType0Font line 134
-        try (var in = Thread.currentThread().getContextClassLoader().getResourceAsStream(resourcePath); var buffer = new RandomAccessReadBuffer(in)) {
-            return new TTFParser().parse(buffer);
+        try (var in = Thread.currentThread().getContextClassLoader().getResourceAsStream("fonts/cmu-regular.ttf"); var buffer = new RandomAccessReadBuffer(in)) {
+            this.trueTypeFont = new TTFParser().parse(buffer); // since Type0Font can be descendant from any font, we need to remember the original TrueTypeFont for the glyph information
+            this.type0Font = PDType0Font.load(document, this.trueTypeFont, false); // use Type0Font for unicode support
         }
     }
 
 
-    @SneakyThrows
-    @SuppressWarnings("PMD.CloseResource")
-    private static Type0FontMetricsFactory createFromResourcePath(String resourcePath, PDDocument document) {
-
-        TrueTypeFont trueTypeFont = readFromResourcePath(resourcePath);
-        // since Type0Font can be descendant from any font, we need to remember the original TrueTypeFont for the glyph information
-        return new Type0FontMetricsFactory(resourcePath, PDType0Font.load(document, trueTypeFont, true), trueTypeFont, document); // use Type0Font for unicode support)
-
-    }
-
-
     @SneakyThrows
     public HeightAndDescent calculateHeightAndDescent(String text) {
 
@@ -93,9 +55,8 @@ public class Type0FontMetricsFactory implements FontMetricsFactory {
                 if (glyph == null || glyph.getBoundingBox() == null) {
                     continue;
                 }
-                if (!slashGlyphIds.contains(glyphId)) {
-                    descent = Math.min(descent, glyph.getYMinimum());
-                }
+
+                descent = Math.min(descent, glyph.getYMinimum());
                 height = Math.max(height, glyph.getYMaximum());
             } catch (Exception e) {
                 log.warn("descent and height of string {} could not be parsed, using average fallback value!", text);
@@ -113,28 +74,4 @@ public class Type0FontMetricsFactory implements FontMetricsFactory {
         return type0Font;
     }
 
-
-    @Override
-    @SneakyThrows
-    public PDFont embed(PDDocument document) {
-
-        if (documentThisIsEmbeddedIn.equals(document)) {
-            return getFont();
-        }
-
-        // no need to close, the font will be closed with the document it is embedded in
-
-        this.trueTypeFont = readFromResourcePath(resourcePath);
-        this.type0Font = PDType0Font.load(document, trueTypeFont, true);
-        this.documentThisIsEmbeddedIn = document;
-        return getFont();
-    }
-
-
-    @SneakyThrows
-    public void close() {
-
-        trueTypeFont.close();
-    }
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/scriptdetection/FontStyleDetector.java

@@ -1,158 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.service.scriptdetection;
-
-import java.util.Comparator;
-import java.util.HashMap;
-import java.util.List;
-import java.util.Map;
-import java.util.Optional;
-import java.util.stream.Stream;
-
-import org.apache.commons.math3.ml.clustering.Cluster;
-import org.apache.commons.math3.ml.clustering.DBSCANClusterer;
-import org.apache.pdfbox.pdmodel.PDDocument;
-import org.springframework.stereotype.Service;
-
-import com.knecon.fforesight.service.ocr.processor.model.OcrResult;
-import com.knecon.fforesight.service.ocr.processor.model.OcrResultToWrite;
-import com.knecon.fforesight.service.ocr.processor.model.scriptdetection.FontStyleDetectionModel;
-import com.knecon.fforesight.service.ocr.processor.model.scriptdetection.TextPositionAndWordImage;
-import com.knecon.fforesight.service.ocr.processor.model.scriptdetection.WordImage;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.FontStyle;
-import com.knecon.fforesight.service.ocr.processor.service.fonts.Type0FontMetricsFactory;
-import com.knecon.fforesight.service.ocr.processor.settings.OcrServiceSettings;
-
-import lombok.AccessLevel;
-import lombok.RequiredArgsConstructor;
-import lombok.experimental.FieldDefaults;
-import lombok.extern.slf4j.Slf4j;
-
-@Slf4j
-@Service
-@RequiredArgsConstructor
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-public class FontStyleDetector {
-
-    OcrServiceSettings settings;
-    StrokeWidthCalculator strokeWidthCalculator;
-
-
-    /**
-     * Implementation of the MOBDoB algorithm, refer to the paper here:
-     * <a href="http://mile.ee.iisc.ac.in/publications/softCopy/DocumentAnalysis/Sai_NCVPRIPG2013.pdf">Script Independent Detection of Bold Words in Multi Font-size Documents</a>
-     * <p>
-     * As a high level overview: We cluster all text based on its font size. We determine the cluster with the most words. This is assumed to be regular text.
-     * We then estimate the average stroke width of that cluster by thinning all text to a single pixel and calculating the ratio of remaining pixels.
-     * (<a href="http://www.leptonica.org/papers/conn.pdf">Leptonica Documentation on thinning</a>)
-     * For each word we scale this average strokewidth based on its fontsize compared to the most common fontsize.
-     * Using the scaled strokewidth we do an opening operation.
-     * (<a href="https://en.wikipedia.org/wiki/Opening_(morphology)">Opening (Morphology)</a>).
-     * We then threshold the ratio of remaining pixels to determine whether a word is bold or not.
-     * <p>
-     * I did take some liberties though. Firstly, the paper uses text height without ascender/descender height for the clustering. I'm using the previously implemented font size estimation.
-     * But that is calculated based on text width. Thus, I'm also using the height scaling factor to scale the font size by the text height.
-     * The paper does not describe its clustering algorithm, so I've decided on DBSCAN due to its good runtime and readily available implementation by apache commons math.
-     * Moreover, the paper states that stroke width scales linearly with text height. I've come to the conclusion this is not the case.
-     * It seems it scales with the square root of the text height. Or at least this seemed to give the best results for me.
-     */
-    public Map<Integer, List<OcrResultToWrite>> detectBold(List<OcrResult> ocrResults, PDDocument document) {
-
-        FontMetricsFactory fontMetricsFactory = Type0FontMetricsFactory.regular(document);
-        if (!settings.isBoldDetection()) {
-            return OcrResultToWrite.buildOcrResultsToWrite(ocrResults, fontMetricsFactory);
-        }
-
-        Map<Integer, List<OcrResultToWrite>> ocrResultToWritePerPage = new HashMap<>();
-
-        DBSCANClusterer<TextPositionAndWordImage> clusterer = new DBSCANClusterer<>(0.5, 1);
-
-        FontMetricsFactory boldFontMetricsFactory = Type0FontMetricsFactory.bold(document);
-
-        for (OcrResult result : ocrResults) {
-            FontStyleDetectionModel fontStyleDetectionModel = FontStyleDetectionModel.fromOcrResult(result, fontMetricsFactory, settings);
-
-            List<Cluster<TextPositionAndWordImage>> clusters = clusterer.cluster(fontStyleDetectionModel.getTextPositionsAndWordImages());
-            Optional<Cluster<TextPositionAndWordImage>> largestCluster = clusters.stream().max(Comparator.comparingInt(cluster -> cluster.getPoints().size()));
-
-            if (largestCluster.isEmpty()) {
-                insertResultIntoMap(result.image().getPageNumber(), ocrResultToWritePerPage, fontStyleDetectionModel);
-                continue;
-            }
-
-            List<TextPositionAndWordImage> wordsWithMostCommonTextHeight = largestCluster.get().getPoints();
-
-            double standardTextHeight = calculateStandardTextheight(wordsWithMostCommonTextHeight);
-            double regularStrokeWidth = calculateRegularStrokeWidth(wordsWithMostCommonTextHeight);
-
-            for (TextPositionAndWordImage textPositionsAndWordImage : fontStyleDetectionModel.getTextPositionsAndWordImages()) {
-                decideOnFontStyle(textPositionsAndWordImage, regularStrokeWidth, standardTextHeight, boldFontMetricsFactory);
-            }
-
-            insertResultIntoMap(result.image().getPageNumber(), ocrResultToWritePerPage, fontStyleDetectionModel);
-            fontStyleDetectionModel.dispose();
-        }
-
-        log.info("Finished bold detection");
-        return ocrResultToWritePerPage;
-    }
-
-
-    private static double calculateStandardTextheight(List<TextPositionAndWordImage> wordsWithMostCommonTextHeight) {
-
-        return wordsWithMostCommonTextHeight.stream()
-                .map(TextPositionAndWordImage::getWordImage)
-                .mapToDouble(WordImage::getTextHeight)
-                .filter(Double::isFinite)
-                .average()
-                .orElseThrow();
-    }
-
-
-    private double calculateRegularStrokeWidth(List<TextPositionAndWordImage> wordsWithMostCommonTextHeight) {
-
-        return wordsWithMostCommonTextHeight.stream()
-                .mapToDouble(textPositionAndWordImage -> strokeWidthCalculator.calculate(textPositionAndWordImage.getWordImage().getImage()))
-                .filter(Double::isFinite)
-                .average()
-                .orElseThrow();
-    }
-
-
-    private static void insertResultIntoMap(int pageNumber, Map<Integer, List<OcrResultToWrite>> ocrResultToWritePerPage, FontStyleDetectionModel fontStyleDetectionModel) {
-
-        OcrResultToWrite ocrResult = OcrResultToWrite.fromFontStyleDetectionModel(fontStyleDetectionModel);
-
-        ocrResultToWritePerPage.compute(pageNumber, (key, existingList) -> {
-            if (existingList == null) {
-                return List.of(ocrResult);
-            } else {
-                return Stream.concat(existingList.stream(), Stream.of(ocrResult)).toList();
-            }
-        });
-    }
-
-
-    private void decideOnFontStyle(TextPositionAndWordImage textPositionsAndWordImage,
-                                   double standardStrokeWidth,
-                                   double standardTextHeight,
-                                   FontMetricsFactory boldFontMetricsFactory) {
-
-        double scaledStrokeWidth = scaleStrokeWidthByFontSize(textPositionsAndWordImage, standardStrokeWidth, standardTextHeight);
-
-        if (textPositionsAndWordImage.getWordImage().hasLargerStrokeWidth(scaledStrokeWidth)) {
-            textPositionsAndWordImage.getTextPositionInImage().setFontMetricsFactory(boldFontMetricsFactory);
-            textPositionsAndWordImage.getTextPositionInImage().setFontStyle(FontStyle.BOLD);
-        } else {
-            textPositionsAndWordImage.getTextPositionInImage().setFontStyle(FontStyle.REGULAR);
-        }
-    }
-
-
-    private static double scaleStrokeWidthByFontSize(TextPositionAndWordImage textPositionsAndWordImage, double standardStrokeWidth, double standardFontSize) {
-
-        double influenceOfFontSize = 1.0; // the paper states that stroke width scales exactly linearly with font size. This did not seem to be true for me. Maybe some of the preprocessing steps are affecting this.
-        double fontsizeScalingFactor = Math.sqrt(textPositionsAndWordImage.getWordImage().getTextHeight() / standardFontSize);
-        return standardStrokeWidth + (influenceOfFontSize * (fontsizeScalingFactor - 1) * standardStrokeWidth);
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/scriptdetection/ItalicDetector.java

@@ -1,57 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.service.scriptdetection;
-
-import com.knecon.fforesight.service.ocr.processor.utils.ImageProcessingUtils;
-
-import lombok.AccessLevel;
-import lombok.NoArgsConstructor;
-import lombok.experimental.FieldDefaults;
-import net.sourceforge.lept4j.Leptonica1;
-import net.sourceforge.lept4j.Pix;
-import net.sourceforge.lept4j.Sel;
-import net.sourceforge.lept4j.util.LeptUtils;
-
-/**
- * This code is a good start for detecting italic text, although it has a few issues especially with glyphs which are naturally slanted. E.g. z, 2, 7, /
- * If we want this maybe we should exclude these glyphs and then it might have less false positives. But in its current state i don't recommend using it.
- */
-@NoArgsConstructor
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-public class ItalicDetector {
-
-
-    static String italicKernel = "ooxxooxxooxxoxxooXxooxxoxxooxxooxxoo";
-    Sel italicSel = Leptonica1.selCreateFromString(italicKernel, 9, 4, "italicKernel");
-    Sel brickSel = Leptonica1.selCreateBrick(3, 4, 1, 2, 1);
-
-
-    public boolean isItalic(Pix pix) {
-
-        Pix preprocessed = preprocess(pix);
-        Pix flipped = Leptonica1.pixFlipLR(null, pix);
-        Pix flippedPreprocessed = preprocess(flipped);
-        Leptonica1.pixFlipLR(flippedPreprocessed, flippedPreprocessed);
-        double pixelDensity = ImageProcessingUtils.calculatePixelDensity(preprocessed);
-        double flippedPixelDensity = ImageProcessingUtils.calculatePixelDensity(flippedPreprocessed);
-        LeptUtils.disposePix(preprocessed);
-        LeptUtils.disposePix(flipped);
-        LeptUtils.disposePix(flippedPreprocessed);
-        return flippedPixelDensity / pixelDensity < 0.85;
-    }
-
-
-    private Pix preprocess(Pix pix) {
-
-        Pix eroded = Leptonica1.pixErode(null, pix, italicSel.getPointer());
-        Pix dilated = Leptonica1.pixDilate(null, eroded, brickSel.getPointer());
-        LeptUtils.disposePix(eroded);
-        return dilated;
-    }
-
-
-    public void dispose() {
-
-        LeptUtils.dispose(italicSel);
-        LeptUtils.dispose(brickSel);
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/scriptdetection/StrokeWidthCalculator.java

@@ -1,58 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.service.scriptdetection;
-
-import static net.sourceforge.lept4j.ILeptonica.L_THIN_FG;
-
-import java.nio.IntBuffer;
-
-import org.springframework.stereotype.Service;
-
-import lombok.AccessLevel;
-import lombok.NoArgsConstructor;
-import lombok.experimental.FieldDefaults;
-import net.sourceforge.lept4j.Leptonica1;
-import net.sourceforge.lept4j.Pix;
-import net.sourceforge.lept4j.Sela;
-import net.sourceforge.lept4j.util.LeptUtils;
-
-@Service
-@NoArgsConstructor
-@FieldDefaults(level = AccessLevel.PRIVATE)
-public class StrokeWidthCalculator {
-
-    Sela thinningSel;
-
-
-    /**
-     * Uses a series of sels to thin all connected lines to a single pixel. Then the pixel ratio is a good estimation of the stroke width in pixels.
-     * <a href="http://www.leptonica.org/papers/conn.pdf">Leptonica Documentation on thinning</a>
-     * Since the baseline is a strokewidth of exactly one, we need to add 1 to the result.
-     *
-     * @param input binarized pix with text on it
-     * @return estimated stroke width in pixels
-     */
-    public double calculate(Pix input) {
-
-        init();
-
-        Pix thinned = Leptonica1.pixThinConnectedBySet(input, L_THIN_FG, thinningSel, 0);
-
-        IntBuffer thinnedPixelCount = IntBuffer.allocate(1);
-        Leptonica1.pixCountPixels(thinned, thinnedPixelCount, null);
-
-        IntBuffer pixelCount = IntBuffer.allocate(1);
-        Leptonica1.pixCountPixels(input, pixelCount, null);
-
-        LeptUtils.disposePix(thinned);
-
-        return (double) pixelCount.get() / thinnedPixelCount.get() + 1;
-    }
-
-
-    private void init() {
-
-        if (thinningSel == null) {
-            thinningSel = Leptonica1.selaMakeThinSets(1, 0);
-        }
-    }
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/threads/BlockingQueueFiller.java

@@ -17,6 +17,7 @@ import lombok.experimental.FieldDefaults;
 import lombok.extern.slf4j.Slf4j;
 import net.sourceforge.tess4j.TessAPI1;
 
+
 /*
 This just moves the Elements from the GhostScriptOutputListener into the ImageProcessing queue asynchronously
  */
@@ -37,17 +38,9 @@ public class BlockingQueueFiller extends Thread {
     public void run() {
 
         // Interrupting signals that the image extraction has finished
-        try {
-            while (!allImagesQueued) {
+        while (!allImagesQueued) {
                 final UnprocessedImage image = imageInputQueue.take();
-                try {
-                    imageOutputQueue.put(image);
-                } catch (InterruptedException e) {
-                    imageOutputQueue.put(image);
-                }
-            }
-        } catch (InterruptedException e) {
-            log.info("All images extracted, emptying processing queue and stopping");
+                imageOutputQueue.put(image);
         }
 
         // empty the queue
@@ -61,5 +54,4 @@ public class BlockingQueueFiller extends Thread {
         }
 
     }
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/threads/GhostScriptOutputHandler.java

@@ -109,6 +109,7 @@ public class GhostScriptOutputHandler extends Thread {
         if (imageFile == null) {
             throw new IllegalArgumentException(String.format("Page number %d does not exist in this thread. It only has pagenumbers %s", pageNumber, pagesToProcess.keySet()));
         }
+        assert new File(imageFile.absoluteFilePath()).isFile();
         renderedPageImageFileOutput.add(imageFile);
     }
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/threads/ImageProcessingThread.java

@@ -4,6 +4,8 @@ import static net.sourceforge.tess4j.ITessAPI.TRUE;
 
 import java.nio.FloatBuffer;
 import java.nio.IntBuffer;
+import java.util.ArrayList;
+import java.util.List;
 import java.util.NoSuchElementException;
 import java.util.concurrent.BlockingQueue;
 
@@ -27,8 +29,6 @@ import lombok.Setter;
 import lombok.SneakyThrows;
 import lombok.experimental.FieldDefaults;
 import lombok.extern.slf4j.Slf4j;
-import net.sourceforge.lept4j.L_Kernel;
-import net.sourceforge.lept4j.Leptonica1;
 import net.sourceforge.lept4j.Pix;
 import net.sourceforge.lept4j.util.LeptUtils;
 import net.sourceforge.tess4j.ITessAPI;
@@ -45,7 +45,6 @@ public class ImageProcessingThread extends Thread {
     final BlockingQueue<UnprocessedImage> imageInputQueue;
     final BlockingQueue<OcrImage> imageOutputQueue;
     final ITessAPI.TessBaseAPI detectionScriptHandle = initDetectionScriptHandle();
-    final L_Kernel gaussianKernel = Leptonica1.makeGaussianKernel(2, 2, 1.0f, 1);
     final Statistics stats;
     final OcrServiceSettings settings;
     final PDDocument document;
@@ -58,20 +57,13 @@ public class ImageProcessingThread extends Thread {
     @Override
     public void run() {
 
-        try {
-            while (!allImagesExtracted) {
-                final UnprocessedImage image = imageInputQueue.take();
-                var ocrImage = this.process(image);
-                try {
-                    imageOutputQueue.put(ocrImage);
-                } catch (InterruptedException e) {
-                    imageOutputQueue.put(ocrImage);
-                }
-            }
-        } catch (InterruptedException e) {
-            log.info("All images extracted, emptying processing queue and stopping");
+        while (!allImagesExtracted) {
+            final UnprocessedImage image = imageInputQueue.take();
+            var ocrImage = this.process(image);
+            imageOutputQueue.put(ocrImage);
         }
 
+
         try {
             while (true) {
                 final UnprocessedImage image = imageInputQueue.remove();
@@ -82,9 +74,7 @@ public class ImageProcessingThread extends Thread {
             log.debug("No images left in processing queue, stopping.");
         }
 
-        TessAPI1.TessBaseAPIEnd(this.detectionScriptHandle);
         TessAPI1.TessBaseAPIDelete(this.detectionScriptHandle);
-        LeptUtils.dispose(gaussianKernel);
     }
 
 
@@ -107,10 +97,9 @@ public class ImageProcessingThread extends Thread {
     }
 
 
-    @SuppressWarnings("PMD.CompareObjectsWithEquals")
     private OcrImage processRenderedPageImageFile(RenderedPageImageFile renderedPageImageFile) {
 
-        Pix pix = processPix(renderedPageImageFile.asPix(), settings.getDpi(), settings.getDpi());
+        Pix pix = binarize(renderedPageImageFile.asPix(), settings.getDpi(), settings.getDpi());
 
         int orientDegree = detectOrientation(pix, settings.getDpi(), detectionScriptHandle);
         Pix rotatedPix = ImageProcessingUtils.deRotatePix(orientDegree, pix);
@@ -128,12 +117,12 @@ public class ImageProcessingThread extends Thread {
         return ocrImage;
     }
 
-    @SuppressWarnings("PMD.CompareObjectsWithEquals")
+
     private OcrImage processExtractedImage(ExtractedImage extractedImage) {
 
         float imageDPI = Math.abs(extractedImage.image().getWidth() / (extractedImage.ctm().getScalingFactorX() / 72));
 
-        Pix pix = processPix(extractedImage.asPix(), imageDPI, settings.getDpi());
+        Pix pix = binarize(extractedImage.asPix(), imageDPI, settings.getDpi());
 
         int orientDegree = detectOrientation(pix, settings.getDpi(), detectionScriptHandle);
         Pix rotatedPix = ImageProcessingUtils.deRotatePix(orientDegree, pix);
@@ -167,7 +156,7 @@ public class ImageProcessingThread extends Thread {
 
         orientationDegreeResultBuffer = IntBuffer.allocate(1);
         orientationDegreeConfidenceBuffer = FloatBuffer.allocate(1);
-        scriptureNameBuffer = new PointerByReference(); // Is this memory being freed?
+        scriptureNameBuffer = new PointerByReference();
         scriptureConfidenceBuffer = FloatBuffer.allocate(1);
 
         int orientationDegree = 0;
@@ -187,58 +176,15 @@ public class ImageProcessingThread extends Thread {
 
 
     @SneakyThrows
-    private Pix processPix(Pix pix, float imageDpi, int targetDpi) {
+    private Pix binarize(Pix pix, float imageDpi, int targetDpi) {
 
-        Pix grayScale;
-        Pix scaledUp;
-        Pix gaussian;
-        Pix binarized;
+        Pix grayScale = ImageProcessingUtils.convertToGrayScale(pix);
+        Pix scaledUp = ImageProcessingUtils.scaleToTargetDpi(imageDpi, targetDpi, grayScale);
+        return ImageProcessingUtils.despecklePix(scaledUp);
 
-        //convert to grayscale
-        if (pix.d == 8) {
-            grayScale = pix;
-        } else if (pix.d == 32) {
-            grayScale = Leptonica1.pixConvertRGBToGrayFast(pix);
-            LeptUtils.disposePix(pix);
-        } else if (pix.d == 1) {
-            grayScale = Leptonica1.pixConvert1To8(null, pix, (byte) 0, (byte) 255);
-            LeptUtils.disposePix(pix);
-        } else {
-            throw new UnsupportedOperationException(String.format("Unknown pix format with bpp of %d", pix.d));
-        }
-
-        // scale up
-        float targetFactor = targetDpi / imageDpi;
-        if (targetFactor > 2.1) {
-            scaledUp = Leptonica1.pixScaleGray4xLI(grayScale);
-            LeptUtils.disposePix(grayScale);
-        } else if (targetFactor > 1.1) {
-            scaledUp = Leptonica1.pixScaleGray2xLI(grayScale);
-            LeptUtils.disposePix(grayScale);
-        } else {
-            scaledUp = grayScale;
-        }
-
-        // remove noise and prep for Otsu
-        gaussian = Leptonica1.pixConvolve(scaledUp, gaussianKernel, 8, 1);
-        LeptUtils.disposePix(scaledUp);
-
-        // Threshold to binary
-        if (pix.w < 100 || pix.h < 100) {
-            binarized = Leptonica1.pixThresholdToBinary(gaussian, 170);
-        } else {
-            binarized = Leptonica1.pixOtsuThreshOnBackgroundNorm(gaussian, null, 50, 50, 165, 10, 100, 5, 5, 0.1f, null);
-            if (binarized == null) { // Sometimes Otsu just fails, then we binarize directly
-                binarized = Leptonica1.pixThresholdToBinary(gaussian, 170);
-            }
-        }
-        LeptUtils.disposePix(gaussian);
-
-        return binarized;
     }
 
 
-
     private static ITessAPI.TessBaseAPI initDetectionScriptHandle() {
 
         ITessAPI.TessBaseAPI handle = TessAPI1.TessBaseAPICreate();

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/service/threads/OCRThread.java

@@ -116,11 +116,16 @@ public class OCRThread extends Thread {
     @SneakyThrows
     public void executeTesseract(int psm, int dpi, Pix pix, String tesseractOutputFileName) {
 
-        Leptonica1.pixWrite(tesseractOutputFileName + ".tiff", pix, 5); // write the used image for later bold detection
+        if (settings.isDebug()) {
+            String[] a = tesseractOutputFileName.split("/");
+            String folder = "/tmp/pixs/" + a[a.length - 3];
+            new File(folder).mkdirs();
+            Leptonica1.pixWrite(folder + "/pix_" + a[a.length - 1] + ".png", pix, 3);
+        }
+
         instance.setVariable("user_defined_dpi", String.valueOf(dpi));
         instance.setPageSegMode(psm);
         instance.createDocumentsWithResults(pix, null, tesseractOutputFileName, List.of(ITesseract.RenderedFormat.HOCR), ITessAPI.TessPageIteratorLevel.RIL_BLOCK);
-
     }
 
 

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/settings/OcrServiceSettings.java

@@ -19,10 +19,10 @@ public class OcrServiceSettings {
     int psmOverride = -1; // Overrides the page segmentation mode if > 0
     int minImageHeight = 20; // Minimum height for images to be processed
     int minImageWidth = 20; // Minimum width for images to be processed
-    float minRotationConfidence = 2; // Sets a lower bound for the confidence rating for rotated pages.
+    float minRotationConfidence = 2; //
     boolean debug; // If true, overlays OCR images with a grid and draws word bounding boxes
+    boolean removeWatermark; // If true, watermarks will be removed
     String languages = "deu+eng"; // Defines languages loaded into Tesseract as 3-char codes, additional languages must also be installed in the docker environment
     COSName ocrMarkedContentTag = COSName.getPDFName("KNECON_OCR");
-    boolean boldDetection = true; // if true, bold detection will be attempted
-    double boldThreshold = 0.5; // Words are opened with a brick of average stroke width, if the ratio of remaining pixels is higher the word is determined bold.
+
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/utils/ImageProcessingUtils.java

@@ -6,17 +6,14 @@ import java.awt.Graphics;
 import java.awt.Graphics2D;
 import java.awt.Transparency;
 import java.awt.image.BufferedImage;
-import java.nio.IntBuffer;
 
 import org.apache.pdfbox.pdmodel.graphics.color.PDDeviceGray;
 import org.apache.pdfbox.pdmodel.graphics.color.PDDeviceRGB;
 
 import com.knecon.fforesight.service.ocr.processor.model.ExtractedImage;
-import com.sun.jna.ptr.PointerByReference;
 
 import lombok.SneakyThrows;
 import lombok.experimental.UtilityClass;
-import net.sourceforge.lept4j.L_Kernel;
 import net.sourceforge.lept4j.Leptonica1;
 import net.sourceforge.lept4j.Pix;
 import net.sourceforge.lept4j.util.LeptUtils;
@@ -40,6 +37,67 @@ public class ImageProcessingUtils {
     }
 
 
+    public static Pix despecklePix(Pix pix) {
+
+        assert pix.d == 8;
+        Pix despeckled;
+        if (pix.w < 100 || pix.h < 100) {
+            // too small to properly despeckle, just binarize instead.
+            despeckled = Leptonica1.pixThresholdToBinary(pix, 180);
+        } else {
+            despeckled = LeptUtils.despeckle(pix,
+                    LeptUtils.SEL_STR3,
+                    3); // sometimes this fails and I can't figure out why. Then we skip the despeckling and just simply convert to binary. Might have something to do with Imagesize, not sure though...
+            if (despeckled == null) {
+                despeckled = Leptonica1.pixThresholdToBinary(pix, 180);
+            }
+        }
+        if (pix != despeckled) {
+            LeptUtils.disposePix(pix);
+        }
+        return despeckled;
+    }
+
+
+    public static Pix scaleToTargetDpi(float imageDpi, int targetDpi, Pix grayScale) {
+
+        float targetFactor = targetDpi / imageDpi;
+
+        if (targetFactor > 3) {
+            Pix scaledUp;
+            scaledUp = Leptonica1.pixScaleGray4xLI(grayScale);
+            LeptUtils.disposePix(grayScale);
+            return scaledUp;
+        } else if (targetFactor > 1.9) {
+            Pix scaledUp;
+            scaledUp = Leptonica1.pixScaleGray2xLI(grayScale);
+            LeptUtils.disposePix(grayScale);
+            return scaledUp;
+        } else {
+            return grayScale;
+        }
+    }
+
+
+    @SneakyThrows
+    public static Pix convertToGrayScale(Pix pix) {
+
+        if (pix.d == 8) {
+            return pix;
+        } else if (pix.d == 32) {
+            Pix grayScale = Leptonica1.pixConvertRGBToGrayFast(pix);
+            LeptUtils.disposePix(pix);
+            return grayScale;
+        } else if (pix.d == 1) {
+            Pix grayScale = Leptonica1.pixConvert1To8(null, pix, (byte) 0, (byte) 255);
+            LeptUtils.disposePix(pix);
+            return grayScale;
+        } else {
+            throw new UnsupportedOperationException(String.format("Unknown pix format with bpp of %d", pix.d));
+        }
+    }
+
+
     public Pix deRotatePix(int orientDegree, Pix pix) {
 
         return switch (360 - orientDegree) {
@@ -70,16 +128,4 @@ public class ImageProcessingUtils {
         }
     }
 
-
-    public static double calculatePixelDensity(Pix pix) {
-
-        IntBuffer pixelCount = IntBuffer.allocate(1);
-        int result = Leptonica1.pixCountPixels(pix, pixelCount, null);
-        if (result == 0) {
-            return (double) pixelCount.get() / (pix.h * pix.w);
-        } else {
-            return -1;
-        }
-    }
-
 }

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/utils/KernelUtils.java

@@ -1,73 +0,0 @@
-package com.knecon.fforesight.service.ocr.processor.utils;
-
-import lombok.experimental.UtilityClass;
-import net.sourceforge.lept4j.L_Kernel;
-import net.sourceforge.lept4j.Leptonica1;
-
-@UtilityClass
-public class KernelUtils {
-
-    /*
-    -1, -1, -1
-    -1, 8, -1
-    -1, -1, -1
-     */
-    public L_Kernel createFullLaplacianKernel() {
-
-        L_Kernel laplacianKernel = Leptonica1.kernelCreate(3, 3);
-        Leptonica1.kernelSetElement(laplacianKernel, 0, 0, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 0, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 0, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 0, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 0, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 1, 8);
-        return laplacianKernel;
-    }
-
-    /*
-    0, 0, -1, 0, 0
-    0, -1, -1, -1, 0
-    -1, -1, 12, -1, -1
-    0, -1, -1, -1, 0
-    0, 0, -1, 0, 0
-     */
-    public L_Kernel createLaplacianKernel5x5() {
-
-        L_Kernel laplacianKernel = Leptonica1.kernelCreate(5, 5);
-        Leptonica1.kernelSetElement(laplacianKernel, 0, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 3, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 0, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 3, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 4, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 3, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 3, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 3, 3, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 4, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 2, 12);
-        return laplacianKernel;
-    }
-
-    /*
-    0, -1, 0
-    -1, 4, -1
-    0, -1, 0
-     */
-    public L_Kernel createLaplacianKernel() {
-
-        L_Kernel laplacianKernel = Leptonica1.kernelCreate(3, 3);
-        Leptonica1.kernelSetElement(laplacianKernel, 0, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 0, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 2, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 2, 1, -1);
-        Leptonica1.kernelSetElement(laplacianKernel, 1, 1, 4);
-        return laplacianKernel;
-    }
-
-
-}

ocr-service-v1/ocr-service-processor/src/main/java/com/knecon/fforesight/service/ocr/processor/utils/PdfDraw.java

@@ -17,57 +17,58 @@ public class PdfDraw {
     @SneakyThrows
     public static void drawGrid(ElementWriter writer, Page page) {
 
-        try (var eb = new ElementBuilder()) {
-            double dX = 15;
-            double dY = 15;
-            int nRows = (int) (page.getPageHeight() / dY) + 1;
-            int nCols = (int) (page.getPageWidth() / dX) + 1;
-            for (int row = 0; row < nRows; ++row) {
-                for (int col = 0; col < nCols; ++col) {
-                    Element cell = eb.createRect(col * dX, row * dY, dX, dY);
-                    cell.setPathStroke(true);
-                    cell.getGState().setLineWidth(1);
-                    cell.getGState().setStrokeOpacity(0.1);
-                    cell.getGState().setStrokeColorSpace(ColorSpace.createDeviceRGB());
-                    if (row == 0 && col == 0) {
-                        cell.getGState().setStrokeColor(new ColorPt(0, 0, 1));
-                        cell.setPathFill(true);
-                        cell.getGState().setFillOpacity(0.8);
-                        cell.getGState().setFillColorSpace(ColorSpace.createDeviceRGB());
-                        cell.getGState().setFillColor(new ColorPt(0, 0, 1));
-                    } else {
-                        cell.setPathFill(false);
-                        cell.getGState().setStrokeColor(new ColorPt(0.1, 0.1, 0.1));
-                    }
-                    writer.writePlacedElement(cell);
+        ElementBuilder eb = new ElementBuilder();
+        double dX = 15;
+        double dY = 15;
+        int nRows = (int) (page.getPageHeight() / dY) + 1;
+        int nCols = (int) (page.getPageWidth() / dX) + 1;
+        for (int row = 0; row < nRows; ++row) {
+            for (int col = 0; col < nCols; ++col) {
+                Element cell = eb.createRect(col * dX, row * dY, dX, dY);
+                cell.setPathStroke(true);
+                cell.getGState().setLineWidth(1);
+                cell.getGState().setStrokeOpacity(0.1);
+                cell.getGState().setStrokeColorSpace(ColorSpace.createDeviceRGB());
+                if (row == 0 && col == 0) {
+                    cell.getGState().setStrokeColor(new ColorPt(0, 0, 1));
+                    cell.setPathFill(true);
+                    cell.getGState().setFillOpacity(0.8);
+                    cell.getGState().setFillColorSpace(ColorSpace.createDeviceRGB());
+                    cell.getGState().setFillColor(new ColorPt(0, 0, 1));
+                } else {
+                    cell.setPathFill(false);
+                    cell.getGState().setStrokeColor(new ColorPt(0.1, 0.1, 0.1));
                 }
+                writer.writePlacedElement(cell);
             }
         }
+        eb.destroy();
     }
 
 
     @SneakyThrows
     public static void drawRectCollection(ElementWriter writer, RectCollection rectCollection) {
 
-        try (var colorPt = new ColorPt(1, 0, 0); var eb = new ElementBuilder()) {
-            for (int i = 0; i < rectCollection.getNumRects(); ++i) {
-                try(var r = rectCollection.getRectAt(i)) {
-                    Element rect = eb.createRect(r.getX1(), r.getY1(), r.getWidth(), r.getHeight());
+        ColorPt colorPt = new ColorPt(1, 0, 0);
+        ElementBuilder eb = new ElementBuilder();
+        for (int i = 0; i < rectCollection.getNumRects(); ++i) {
+            Rect r = rectCollection.getRectAt(i);
+            Element rect = eb.createRect(r.getX1(), r.getY1(), r.getWidth(), r.getHeight());
 
-                    rect.setPathStroke(true);
-                    rect.getGState().setLineWidth(5);
-                    rect.getGState().setStrokeColorSpace(ColorSpace.createDeviceRGB());
-                    rect.getGState().setStrokeColor(colorPt);
+            rect.setPathStroke(true);
+            rect.getGState().setLineWidth(5);
+            rect.getGState().setStrokeColorSpace(ColorSpace.createDeviceRGB());
+            rect.getGState().setStrokeColor(colorPt);
 
-                    rect.setPathFill(true);
-                    rect.getGState().setFillColorSpace(ColorSpace.createDeviceRGB());
-                    rect.getGState().setFillColor(colorPt);
-                    rect.getGState().setFillOpacity(0.5);
+            rect.setPathFill(true);
+            rect.getGState().setFillColorSpace(ColorSpace.createDeviceRGB());
+            rect.getGState().setFillColor(colorPt);
+            rect.getGState().setFillOpacity(0.5);
 
-                    writer.writePlacedElement(rect);
-                }
-            }
+            writer.writePlacedElement(rect);
         }
+        colorPt.destroy();
+        eb.destroy();
     }
 
 }

ocr-service-v1/ocr-service-processor/src/test/java/com/knecon/fforesight/service/ocr/processor/service/Type0FontMetricsFactoryTest.java

@@ -20,7 +20,7 @@ class Type0FontMetricsFactoryTest {
     public void testStringWidth() {
 
         try (PDDocument document = Loader.loadPDF(new File(Type0FontMetricsFactoryTest.class.getClassLoader().getResource("InvisibleText.pdf").getPath()))) {
-            Type0FontMetricsFactory metricsFactory = Type0FontMetricsFactory.regular(document);
+            Type0FontMetricsFactory metricsFactory = new Type0FontMetricsFactory(document);
             FontMetrics fontMetrics = metricsFactory.calculateMetrics("deine mutter", 100, 50);
         }
 

ocr-service-v1/ocr-service-server/build.gradle.kts

@@ -3,10 +3,10 @@ import org.springframework.boot.gradle.tasks.bundling.BootBuildImage
 plugins {
     application
     id("com.iqser.red.service.java-conventions")
-    id("org.springframework.boot") version "3.2.3"
+    id("org.springframework.boot") version "3.1.5"
     id("io.spring.dependency-management") version "1.1.3"
     id("org.sonarqube") version "4.3.0.3225"
-    id("io.freefair.lombok") version "8.4"
+    id("io.freefair.lombok") version "8.2.2"
 }
 
 configurations {
@@ -17,14 +17,14 @@ configurations {
     }
 }
 
-val springBootStarterVersion = "3.2.3"
+val springBootStarterVersion = "3.1.5"
 
 dependencies {
     implementation(project(":ocr-service-processor"))
     implementation(project(":ocr-service-api"))
 
-    implementation("com.knecon.fforesight:tracing-commons:0.7.0")
-    implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.1.1")
+    implementation("com.knecon.fforesight:tracing:0.5.0")
+    implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
     implementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
 
     implementation("net.logstash.logback:logstash-logback-encoder:7.4")
@@ -39,7 +39,7 @@ tasks.named<BootBuildImage>("bootBuildImage") {
 
     environment.put("BPE_DELIM_JAVA_TOOL_OPTIONS", " ")
     environment.put("BPE_APPEND_JAVA_TOOL_OPTIONS", "-Dfile.encoding=UTF-8")
-    environment.put("BPE_GS_LIB", "/layers/fagiani_apt/apt/usr/share/ghostscript/9.55.0/Resource/Init/") // set ghostscript lib path, version in path must match version in Aptfile
+    environment.put("BPE_GS_LIB", "/layers/fagiani_apt/apt/usr/share/ghostscript/9.26/Resource/Init/") // set ghostscript lib path
     environment.put("BPE_FONTCONFIG_PATH", "/layers/fagiani_apt/apt/etc/fonts/") // set ghostscript fontconfig path
 
     var aptfile = layout.projectDirectory.file("src/main/resources/Aptfile").toString()
@@ -53,7 +53,7 @@ tasks.named<BootBuildImage>("bootBuildImage") {
 
     buildpacks.set(
         listOf(
-            "ghcr.io/knsita/buildpacks/fagiani_apt@sha256:9771d4d27d8050aee62769490b8882fffc794745c129fb98e1f33196e2c93504",
+            "ghcr.io/fagiani/buildpacks/fagiani_apt@sha256:6471c8c70f32b749e29f65ae562ac0339fecad26aa9217628c00a6c31f197dae",
             "ghcr.io/kschuettler/knecon-vcpkg@sha256:ba5e967b124de4865ff7e8f565684f752dd6e97b302e2dcf651283f6a19b98b9",
             "ghcr.io/kschuettler/knecon-tessdata@sha256:9062f728aa0340ac963bcdd6f5e740d683823a81d3f480db894da15bff72691a",
             "urn:cnb:builder:paketo-buildpacks/java"

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/Application.java

@@ -5,28 +5,27 @@ import org.springframework.boot.actuate.autoconfigure.security.servlet.Managemen
 import org.springframework.boot.autoconfigure.ImportAutoConfiguration;
 import org.springframework.boot.autoconfigure.SpringBootApplication;
 import org.springframework.boot.autoconfigure.security.servlet.SecurityAutoConfiguration;
+import org.springframework.cloud.openfeign.EnableFeignClients;
 import org.springframework.context.annotation.Bean;
-import org.springframework.context.annotation.EnableAspectJAutoProxy;
 import org.springframework.context.annotation.Import;
 import org.springframework.scheduling.annotation.EnableAsync;
 
 import com.iqser.red.pdftronlogic.commons.InvisibleElementRemovalService;
 import com.iqser.red.pdftronlogic.commons.WatermarkRemovalService;
-import com.knecon.fforesight.lifecyclecommons.LifecycleAutoconfiguration;
 import com.knecon.fforesight.service.ocr.processor.OcrServiceProcessorConfiguration;
+import com.knecon.fforesight.service.ocr.v1.server.client.FileStatusProcessingUpdateClient;
 import com.knecon.fforesight.service.ocr.v1.server.queue.MessagingConfiguration;
 import com.iqser.red.storage.commons.StorageAutoConfiguration;
 import com.knecon.fforesight.tenantcommons.MultiTenancyAutoConfiguration;
-import com.knecon.fforesight.tracing.OpenTelemetryConfig;
 
 import io.micrometer.core.aop.TimedAspect;
 import io.micrometer.core.instrument.MeterRegistry;
 
 @EnableAsync
-@ImportAutoConfiguration({MultiTenancyAutoConfiguration.class, LifecycleAutoconfiguration.class})
+@ImportAutoConfiguration({MultiTenancyAutoConfiguration.class})
 @SpringBootApplication(exclude = {SecurityAutoConfiguration.class, ManagementWebSecurityAutoConfiguration.class})
-@Import({MessagingConfiguration.class, StorageAutoConfiguration.class, OcrServiceProcessorConfiguration.class, OpenTelemetryConfig.class})
-@EnableAspectJAutoProxy
+@Import({MessagingConfiguration.class, StorageAutoConfiguration.class, OcrServiceProcessorConfiguration.class})
+@EnableFeignClients(basePackageClasses = FileStatusProcessingUpdateClient.class)
 public class Application {
 
     /**

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/client/FileStatusProcessingUpdateClient.java

@@ -0,0 +1,10 @@
+package com.knecon.fforesight.service.ocr.v1.server.client;
+
+import org.springframework.cloud.openfeign.FeignClient;
+
+import com.iqser.red.service.persistence.service.v1.api.internal.resources.FileStatusProcessingUpdateResource;
+
+@FeignClient(name = "FileStatusProcessingUpdateResource", url = "${persistence-service.url}")
+public interface FileStatusProcessingUpdateClient extends FileStatusProcessingUpdateResource {
+
+}

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/queue/MessagingConfiguration.java

@@ -11,10 +11,35 @@ import lombok.RequiredArgsConstructor;
 @RequiredArgsConstructor
 public class MessagingConfiguration {
 
-    public static final String OCR_REQUEST_QUEUE = "ocr_request_queue";
-    public static final String OCR_RESPONSE_QUEUE = "ocr_response_queue";
+    public static final String OCR_QUEUE = "ocrQueue";
+    public static final String OCR_DLQ = "ocrDLQ";
+
+    public static final String X_DEAD_LETTER_EXCHANGE = "x-dead-letter-exchange";
+    public static final String X_DEAD_LETTER_ROUTING_KEY = "x-dead-letter-routing-key";
+    public static final String X_MAX_PRIORITY = "x-max-priority";
+
     public static final String OCR_STATUS_UPDATE_RESPONSE_QUEUE = "ocr_status_update_response_queue";
+
     public static final String X_ERROR_INFO_HEADER = "x-error-message";
     public static final String X_ERROR_INFO_TIMESTAMP_HEADER = "x-error-message-timestamp";
 
+
+    @Bean
+    public Queue ocrQueue() {
+
+        return QueueBuilder.durable(OCR_QUEUE)
+                .withArgument(X_DEAD_LETTER_EXCHANGE, "")
+                .withArgument(X_DEAD_LETTER_ROUTING_KEY, OCR_DLQ)
+                .withArgument(X_MAX_PRIORITY, 2)
+                .maxPriority(2)
+                .build();
+    }
+
+
+    @Bean
+    public Queue ocrDeadLetterQueue() {
+
+        return QueueBuilder.durable(OCR_DLQ).build();
+    }
+
 }

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/queue/NoStatusUpdateOcrMessageSender.java

@@ -1,42 +0,0 @@
-package com.knecon.fforesight.service.ocr.v1.server.queue;
-
-import org.springframework.amqp.rabbit.core.RabbitTemplate;
-import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty;
-import org.springframework.stereotype.Service;
-
-import com.knecon.fforesight.service.ocr.processor.service.IOcrMessageSender;
-import com.knecon.fforesight.service.ocr.v1.api.model.DocumentRequest;
-import lombok.AccessLevel;
-import lombok.RequiredArgsConstructor;
-import lombok.experimental.FieldDefaults;
-
-@Service
-@RequiredArgsConstructor
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-@ConditionalOnProperty(value = "ocrService.sendStatusUpdates", havingValue = "false")
-public class NoStatusUpdateOcrMessageSender implements IOcrMessageSender {
-
-    RabbitTemplate rabbitTemplate;
-
-
-    public void sendOcrFinished(String fileId, int totalImages) {
-
-    }
-
-
-    public void sendOCRStarted(String fileId) {
-
-    }
-
-
-    public void sendUpdate(String fileId, int finishedImages, int totalImages) {
-
-    }
-
-
-    public void sendOcrResponse(String dossierId, String fileId) {
-
-        rabbitTemplate.convertAndSend(MessagingConfiguration.OCR_RESPONSE_QUEUE, new DocumentRequest(dossierId, fileId));
-    }
-
-}

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/queue/OcrMessageReceiver.java

@@ -1,8 +1,8 @@
 package com.knecon.fforesight.service.ocr.v1.server.queue;
 
-import java.io.File;
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
 import java.io.IOException;
-import java.nio.file.Path;
 import java.time.OffsetDateTime;
 import java.time.temporal.ChronoUnit;
 
@@ -10,15 +10,17 @@ import org.springframework.amqp.AmqpRejectAndDontRequeueException;
 import org.springframework.amqp.core.Message;
 import org.springframework.amqp.rabbit.annotation.RabbitHandler;
 import org.springframework.amqp.rabbit.annotation.RabbitListener;
+import org.springframework.http.HttpStatus;
 import org.springframework.stereotype.Service;
-import org.springframework.util.FileSystemUtils;
 
 import com.fasterxml.jackson.databind.ObjectMapper;
+import com.knecon.fforesight.service.ocr.v1.server.client.FileStatusProcessingUpdateClient;
 import com.knecon.fforesight.service.ocr.processor.service.FileStorageService;
-import com.knecon.fforesight.service.ocr.processor.service.IOcrMessageSender;
 import com.knecon.fforesight.service.ocr.processor.service.OCRService;
-import com.knecon.fforesight.service.ocr.processor.service.OsUtils;
 import com.knecon.fforesight.service.ocr.v1.api.model.DocumentRequest;
+import com.iqser.red.service.persistence.service.v1.api.shared.model.dossiertemplate.dossier.file.FileErrorInfo;
+
+import feign.FeignException;
 import lombok.AccessLevel;
 import lombok.RequiredArgsConstructor;
 import lombok.experimental.FieldDefaults;
@@ -30,49 +32,71 @@ import lombok.extern.slf4j.Slf4j;
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class OcrMessageReceiver {
 
-    FileStorageService fileStorageService;
-    ObjectMapper objectMapper;
-    OCRService ocrService;
-    IOcrMessageSender ocrMessageSender;
+     FileStorageService fileStorageService;
+     ObjectMapper objectMapper;
+     FileStatusProcessingUpdateClient fileStatusProcessingUpdateClient;
+     OCRService ocrService;
 
 
     @RabbitHandler
-    @RabbitListener(queues = MessagingConfiguration.OCR_REQUEST_QUEUE, concurrency = "1")
+    @RabbitListener(queues = MessagingConfiguration.OCR_QUEUE, concurrency = "1")
     public void receiveOcr(Message in) throws IOException {
 
-        if (in.getMessageProperties().isRedelivered()) {
-            throw new AmqpRejectAndDontRequeueException("Redelivered OCR Request, aborting...");
-        }
-
         DocumentRequest ocrRequestMessage = objectMapper.readValue(in.getBody(), DocumentRequest.class);
-        String dossierId = ocrRequestMessage.getDossierId();
-        String fileId = ocrRequestMessage.getFileId();
-        Path tmpDir = Path.of(OsUtils.getTemporaryDirectory()).resolve(dossierId + "-" + fileId);
+        log.info("--------------------------------------------------------------------------");
+        log.info("Start ocr for file with dossierId {} and fileId {}", ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId());
 
         try {
-            log.info("--------------------------------------------------------------------------");
-            log.info("Start ocr for file with dossierId {} and fileId {}", dossierId, fileId);
+            setStatusOcrProcessing(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId());
 
-            ocrMessageSender.sendOCRStarted(fileId);
+            if (!fileStorageService.untouchedFileExists(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId())) {
+                byte[] originalFile = fileStorageService.getOriginalFile(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId());
+                fileStorageService.storeUntouchedFile(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId(), originalFile);
+            }
 
-            tmpDir.toFile().mkdirs();
-            File documentFile = tmpDir.resolve("document.pdf").toFile();
-            File viewerDocumentFile = tmpDir.resolve("viewerDocument.pdf").toFile();
+            try (var transferStream = new ByteArrayOutputStream()) {
+                ocrService.runOcrOnDocument(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId(), transferStream);
+                try (var inputStream = new ByteArrayInputStream(transferStream.toByteArray())) {
+                    fileStorageService.storeOriginalFile(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId(), inputStream);
+                }
+            } catch (IOException e) {
+                log.error("Failed to store file with dossierId {} and fileId {}", ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId());
+                throw new RuntimeException(e);
+            }
 
-            fileStorageService.downloadFiles(dossierId, fileId, documentFile, viewerDocumentFile);
-
-            ocrService.runOcrOnDocument(dossierId, fileId, ocrRequestMessage.isRemoveWatermark(), tmpDir, documentFile, viewerDocumentFile);
-
-            fileStorageService.storeFiles(dossierId, fileId, documentFile, viewerDocumentFile);
-
-            ocrMessageSender.sendOcrResponse(dossierId, fileId);
+            fileStatusProcessingUpdateClient.ocrSuccessful(ocrRequestMessage.getDossierId(), ocrRequestMessage.getFileId());
         } catch (Exception e) {
             log.warn("An exception occurred in ocr file stage: {}", e.getMessage());
             in.getMessageProperties().getHeaders().put(MessagingConfiguration.X_ERROR_INFO_HEADER, e.getMessage());
             in.getMessageProperties().getHeaders().put(MessagingConfiguration.X_ERROR_INFO_TIMESTAMP_HEADER, OffsetDateTime.now().truncatedTo(ChronoUnit.MILLIS));
             throw new RuntimeException(e);
-        } finally {
-            FileSystemUtils.deleteRecursively(tmpDir);
+        }
+    }
+
+
+    @RabbitHandler
+    @RabbitListener(queues = MessagingConfiguration.OCR_DLQ, concurrency = "1")
+    public void receiveOcrDLQ(Message failedMessage) throws IOException {
+
+        DocumentRequest ocrRequestMessage = objectMapper.readValue(failedMessage.getBody(), DocumentRequest.class);
+        log.info("OCR DQL received: {}", ocrRequestMessage);
+        String errorMessage = failedMessage.getMessageProperties().getHeader(MessagingConfiguration.X_ERROR_INFO_HEADER);
+        OffsetDateTime timestamp = failedMessage.getMessageProperties().getHeader(MessagingConfiguration.X_ERROR_INFO_TIMESTAMP_HEADER);
+        timestamp = timestamp != null ? timestamp : OffsetDateTime.now().truncatedTo(ChronoUnit.MILLIS);
+        fileStatusProcessingUpdateClient.ocrFailed(ocrRequestMessage.getDossierId(),
+                ocrRequestMessage.getFileId(),
+                new FileErrorInfo(errorMessage, MessagingConfiguration.OCR_DLQ, "ocr-service", timestamp));
+    }
+
+
+    private void setStatusOcrProcessing(String dossierId, String fileId) {
+
+        try {
+            fileStatusProcessingUpdateClient.ocrProcessing(dossierId, fileId);
+        } catch (FeignException e) {
+            if (e.status() == HttpStatus.CONFLICT.value()) {
+                throw new AmqpRejectAndDontRequeueException(e.getMessage());
+            }
         }
     }
 

ocr-service-v1/ocr-service-server/src/main/java/com/knecon/fforesight/service/ocr/v1/server/queue/OcrMessageSender.java

@@ -1,23 +1,17 @@
 package com.knecon.fforesight.service.ocr.v1.server.queue;
 
 import org.springframework.amqp.rabbit.core.RabbitTemplate;
-import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty;
 import org.springframework.stereotype.Service;
 
 import com.knecon.fforesight.service.ocr.processor.service.IOcrMessageSender;
-import com.knecon.fforesight.service.ocr.v1.api.model.DocumentRequest;
 import com.knecon.fforesight.service.ocr.v1.api.model.OCRStatusUpdateResponse;
 
-import jakarta.annotation.PostConstruct;
 import lombok.AccessLevel;
 import lombok.RequiredArgsConstructor;
 import lombok.experimental.FieldDefaults;
-import lombok.extern.slf4j.Slf4j;
 
-@Slf4j
 @Service
 @RequiredArgsConstructor
-@ConditionalOnProperty(value = "ocrService.sendStatusUpdates", havingValue = "true")
 @FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class OcrMessageSender implements IOcrMessageSender {
 
@@ -31,14 +25,6 @@ public class OcrMessageSender implements IOcrMessageSender {
     }
 
 
-    public void sendOCRStarted(String fileId) {
-
-        rabbitTemplate.convertAndSend(MessagingConfiguration.OCR_STATUS_UPDATE_RESPONSE_QUEUE,
-                OCRStatusUpdateResponse.builder().fileId(fileId).ocrStarted(true).build());
-
-    }
-
-
     public void sendUpdate(String fileId, int finishedImages, int totalImages) {
 
         rabbitTemplate.convertAndSend(MessagingConfiguration.OCR_STATUS_UPDATE_RESPONSE_QUEUE,
@@ -46,10 +32,4 @@ public class OcrMessageSender implements IOcrMessageSender {
 
     }
 
-
-    public void sendOcrResponse(String dossierId, String fileId) {
-
-        rabbitTemplate.convertAndSend(MessagingConfiguration.OCR_RESPONSE_QUEUE, new DocumentRequest(dossierId, fileId));
-    }
-
 }

ocr-service-v1/ocr-service-server/src/main/resources/Aptfile

@@ -1,5 +1,5 @@
 # you can list packages
-ghostscript=9.55.0~dfsg1-0ubuntu5.9
+ghostscript
 pkg-config
 zip
 unzip
@@ -11,7 +11,6 @@ libk5crypto3
 libkrb5support0
 libkeyutils1
 libkrb5-3
-libbrotli1
 
 # or include links to specific .deb files
 # http://ftp.debian.org/debian/pool/contrib/m/msttcorefonts/ttf-mscorefonts-installer_3.8_all.deb

ocr-service-v1/ocr-service-server/src/main/resources/application.yml

@@ -5,6 +5,8 @@ persistence-service.url: "http://persistence-service-v1:8080"
 tenant-user-management-service.url: "http://tenant-user-management-service:8080/internal"
 fforesight.tenants.remote: true
 
+logging.pattern.level: "%5p [${spring.application.name},%X{traceId:-},%X{spanId:-}]"
+
 logging.type: ${LOGGING_TYPE:CONSOLE}
 kubernetes.namespace: ${NAMESPACE:default}
 project.version: 1.0-SNAPSHOT
@@ -12,9 +14,6 @@ project.version: 1.0-SNAPSHOT
 server:
   port: 8080
 
-lifecycle:
-  base-package: com.knecon.fforesight.service.ocr
-
 spring:
   application:
     name: ocr-service
@@ -42,7 +41,6 @@ fforesight:
     ignored-endpoints: [ '/actuator/health', '/actuator/health/**' ]
     enabled: true
 
-logging.pattern.level: "%5p [${spring.application.name},%X{traceId:-},%X{spanId:-}]"
 
 management:
   endpoint:
@@ -60,6 +58,3 @@ management:
       endpoint: ${OTLP_ENDPOINT:http://otel-collector-opentelemetry-collector.otel-collector:4318/v1/traces}
 
 pdftron.license: ${PDFTRON_LICENSE}
-
-ocrService:
-  sendStatusUpdates: true

ocr-service-v1/ocr-service-server/src/main/resources/vcpkg.json

@@ -6,7 +6,7 @@
   "overrides": [
     {
       "name": "tesseract",
-      "version": "5.3.3"
+      "version": "5.3.2"
     },
     {
       "name": "leptonica",

ocr-service-v1/ocr-service-server/src/test/java/com/knecon/fforesight/service/ocr/v1/server/AbstractTest.java

@@ -24,10 +24,10 @@ import org.springframework.context.annotation.Primary;
 import org.springframework.test.context.junit.jupiter.SpringExtension;
 
 import com.iqser.red.commons.jackson.ObjectMapperFactory;
+import com.knecon.fforesight.service.ocr.processor.initializer.PDFNetInitializer;
 import com.iqser.red.storage.commons.StorageAutoConfiguration;
 import com.iqser.red.storage.commons.service.StorageService;
 import com.iqser.red.storage.commons.utils.FileSystemBackedStorageService;
-import com.knecon.fforesight.service.ocr.processor.initializer.NativeLibrariesInitializer;
 import com.knecon.fforesight.tenantcommons.TenantsClient;
 import com.pdftron.pdf.PDFNet;
 
@@ -36,7 +36,7 @@ import lombok.SneakyThrows;
 
 @ExtendWith({SpringExtension.class, MockitoExtension.class})
 @SpringBootTest(classes = Application.class, webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)
-@Import({AbstractTest.TestConfiguration.class, NativeLibrariesInitializer.class})
+@Import({AbstractTest.TestConfiguration.class, PDFNetInitializer.class})
 @AutoConfigureObservability
 public class AbstractTest {
 

ocr-service-v1/ocr-service-server/src/test/java/com/knecon/fforesight/service/ocr/v1/server/OcrServiceIntegrationTest.java

@@ -9,7 +9,6 @@ import java.io.FileInputStream;
 import java.io.FileOutputStream;
 import java.nio.file.Files;
 import java.nio.file.Path;
-import java.nio.file.StandardCopyOption;
 import java.util.Comparator;
 import java.util.List;
 import java.util.concurrent.TimeUnit;
@@ -26,7 +25,6 @@ import com.fasterxml.jackson.databind.ObjectMapper;
 import com.knecon.fforesight.service.ocr.processor.service.FileStorageService;
 import com.knecon.fforesight.service.ocr.processor.service.OCRService;
 import com.iqser.red.service.persistence.service.v1.api.shared.model.dossiertemplate.dossier.file.FileType;
-import com.knecon.fforesight.service.ocr.processor.service.OsUtils;
 import com.knecon.fforesight.tenantcommons.TenantContext;
 
 import io.micrometer.prometheus.PrometheusMeterRegistry;
@@ -50,9 +48,9 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     @Test
     public void testOCRMetrics() {
 
-        testOCR("files/syngenta/CustomerFiles/SinglePages/Watermark_Page1_10.SYN524464 FS (A16148C) - Absorção cutânea.pdf");
-        testOCR("files/syngenta/CustomerFiles/SinglePages/Watermark_Page1_10.SYN524464 FS (A16148C) - Absorção cutânea.pdf");
-        testOCR("files/syngenta/CustomerFiles/SinglePages/Watermark_Page1_10.SYN524464 FS (A16148C) - Absorção cutânea.pdf");
+        testOCR("files/Watermark.pdf");
+        testOCR("files/Watermark.pdf");
+        testOCR("files/Watermark.pdf");
 
         var ocrOnDocumentMeter = registry.getMeters().stream().filter(m -> m.getId().getName().equalsIgnoreCase("redactmanager_runOcrOnDocument")).findAny();
         assertThat(ocrOnDocumentMeter.isPresent()).isTrue();
@@ -66,7 +64,7 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     @SneakyThrows
     public void testOcr() {
 
-        String text = testOCR("files/402Study.pdf");
+        String text = testOCR("files/2009-1048395_50pages_tables.pdf");
     }
 
 
@@ -81,7 +79,7 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     @Test
     public void testMergeImages() {
         // check visually for most of the images containing text, the resulting text is kind of nonsense, just ensure it is there
-        String text = testOCR("files/syngenta/CustomerFiles/SinglePages/merge_images - Page241_18 Chlorothalonil RAR 08 Volume 3CA B 6a Oct 2017.pdf");
+        String text = testOCR("files/merge_images.pdf");
         assertThat(text).contains("Bodyweight change of dams with live young - group mean values",
                 "Control",
                 "mg/g day",
@@ -101,7 +99,7 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     @Test
     public void testOCRWatermark() {
 
-        assertThat(testOCR("files/syngenta/CustomerFiles/SinglePages/Watermark_Page1_10.SYN524464 FS (A16148C) - Absorção cutânea.pdf")).contains("syngenta");
+        assertThat(testOCR("files/Watermark.pdf")).contains("syngenta");
     }
 
 
@@ -118,17 +116,18 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     private String testOCR(String fileName) {
 
         ClassPathResource pdfFileResource = new ClassPathResource(fileName);
-        Path tmpDir = Path.of(OsUtils.getTemporaryDirectory()).resolve("OCR_TEST").resolve(Path.of(fileName).getFileName());
-        tmpDir.toFile().mkdirs();
-        var documentFile = tmpDir.resolve(Path.of("document.pdf"));
-        var viewerDocumentFile = tmpDir.resolve(Path.of("viewerDocument.pdf"));
-        Files.copy(pdfFileResource.getFile().toPath(), documentFile, StandardCopyOption.REPLACE_EXISTING);
-        Files.copy(pdfFileResource.getFile().toPath(), viewerDocumentFile, StandardCopyOption.REPLACE_EXISTING);
+        var originId = FileStorageService.getStorageId(TEST_DOSSIER_ID, "file", FileType.ORIGIN);
+        try (var fileStream = pdfFileResource.getInputStream()) {
+            storageService.storeObject(TenantContext.getTenantId(), originId, fileStream);
+        }
 
-        ocrService.runOcrOnDocument(TEST_DOSSIER_ID, "file", false, tmpDir, documentFile.toFile(), viewerDocumentFile.toFile());
-        System.out.println("File:" + documentFile);
+        Path tmpFileName = Path.of(getTemporaryDirectory()).resolve(Path.of(fileName).getFileName());
+        try (var out = new FileOutputStream(tmpFileName.toFile())) {
+            ocrService.runOcrOnDocument(TEST_DOSSIER_ID, "file", out);
+            System.out.println("File:" + tmpFileName);
+        }
 
-        try (var fileStream = new FileInputStream(documentFile.toFile())) {
+        try (var fileStream = new FileInputStream(tmpFileName.toFile())) {
             return extractAllTextFromDocument(fileStream);
         }
     }
@@ -163,22 +162,30 @@ public class OcrServiceIntegrationTest extends AbstractTest {
     @SneakyThrows
     public void testOcrForSpecificFile() {
 
-        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/syn-dm-testfiles2/A16361B - Acute Dermal Toxicity Study in Rats.pdf"));
+        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/syn-dm-testfiles/F.2. A16003E - Acute Inhalation Study.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/syn-dm-testfiles/A23220A - 404 - Skin Irritation in vivo.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/OcrMassTest/G.1.2 - 1768300_MMNA_A13617AV_report.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/OcrMassTest/SOLICITA_VICTRATO-GOLD-II_Item 17_Toxicidade Inalatoria Aguda.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/OcrMassTest/SOLICITA_VICTRATO-GOLD-II_Item 20_Sensibilizacao_02.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/OcrMassTest/ITEM 23_A15149W - Dermal absorption of formulated product.pdf"));
+//        testOCRForFile(new File("/home/kschuettler/Dokumente/TestFiles/OcrMassTest/SOLICITA_VICTRATO-GOLD-II_Item 16_Toxicidade Cutanea Aguda.pdf"));
     }
 
 
+
     @SneakyThrows
     private void testOCRForFile(File file) {
 
-        Path tmpDir = Path.of(OsUtils.getTemporaryDirectory()).resolve("OCR_TEST").resolve(file.toPath().getFileName());
-        tmpDir.toFile().mkdirs();
-        var documentFile = tmpDir.resolve(Path.of("document.pdf"));
-        var viewerDocumentFile = tmpDir.resolve(Path.of("viewerDocument.pdf"));
-        Files.copy(file.toPath(), documentFile, StandardCopyOption.REPLACE_EXISTING);
-        Files.copy(file.toPath(), viewerDocumentFile, StandardCopyOption.REPLACE_EXISTING);
+        var originId = FileStorageService.getStorageId(TEST_DOSSIER_ID, "file", FileType.ORIGIN);
+        try (var fileStream = new FileInputStream(file)) {
+            storageService.storeObject(TenantContext.getTenantId(), originId, fileStream);
+        }
 
-        ocrService.runOcrOnDocument(TEST_DOSSIER_ID, "file", false, tmpDir, documentFile.toFile(), viewerDocumentFile.toFile());
-        System.out.println("File:" + documentFile);
+        Path tmpFileName = Path.of(getTemporaryDirectory()).resolve(Path.of(file.getAbsolutePath()).getFileName());
+        try (var out = new FileOutputStream(tmpFileName.toFile())) {
+            ocrService.runOcrOnDocument(TEST_DOSSIER_ID, "file", out);
+            System.out.println("File:" + tmpFileName);
+        }
         System.out.println("\n\n");
     }
 

ocr-service-v1/ocr-service-server/src/test/resources/application.yml

@@ -15,10 +15,3 @@ management:
     health.enabled: true
   endpoints.web.exposure.include: prometheus, health, metrics
   metrics.export.prometheus.enabled: true
-  tracing:
-    enabled: ${TRACING_ENABLED:false}
-    sampling:
-      probability: ${TRACING_PROBABILITY:1.0}
-  otlp:
-    tracing:
-      endpoint: ${OTLP_ENDPOINT:http://otel-collector-opentelemetry-collector.otel-collector:4318/v1/traces}

ocr-service-v1/ocr-service-server/src/test/resources/files/basf

@@ -1 +0,0 @@
-Subproject commit 9dc6c2337dea32e63aef53271dba0692537c6605

ocr-service-v1/ocr-service-server/src/test/resources/files/syngenta

@@ -1 +0,0 @@
-Subproject commit 21fefb64bf27ca2b3329a6c69d90a27450b17930

publish-custom-image.sh

@@ -1,9 +1,5 @@
 #!/bin/bash
-
-set -e
-
 dir=${PWD##*/}
-
 gradle assemble
 
 # Get the current Git branch
@@ -15,32 +11,5 @@ commit_hash=$(git rev-parse --short=5 HEAD)
 # Combine branch and commit hash
 buildName="${USER}-${branch}-${commit_hash}"
 
-gradle bootBuildImage --publishImage -PbuildbootDockerHostNetwork=true -Pversion=${buildName}
-
-newImageName="nexus.knecon.com:5001/ff/ocr-service-server:$buildName"
-
-echo "full image name:"
-echo ${newImageName}
-echo ""
-
-if [ -z "$1" ]; then
-    exit 0
-fi
-
-namespace=${1}
-deployment_name="ocr-service-v1"
-
-echo "deploying to ${namespace}"
-
-oldImageName=$(rancher kubectl -n ${namespace} get deployment ${deployment_name} -o=jsonpath='{.spec.template.spec.containers[*].image}')
-
-if [ "${newImageName}" = "${oldImageName}" ]; then
-    echo "Image tag did not change, redeploying..."
-    rancher kubectl rollout restart deployment ${deployment_name} -n ${namespace}
-else
-    echo "upgrading the image tag..."
-    rancher kubectl set image deployment/${deployment_name} ${deployment_name}=${newImageName} -n ${namespace}
-fi
-rancher kubectl rollout status deployment ${deployment_name} -n ${namespace}
-echo "Built ${deployment_name}:${buildName} and deployed to ${namespace}"
-
+gradle bootBuildImage --publishImage -PbuildbootDockerHostNetwork=true -Pversion=$buildName --no-build-cache
+echo "nexus.knecon.com:5001/ff/${dir}-server:$buildName"