Merge branch 'CLARI-128' into 'main'

Fixed Index out of bounds exception in blockificationpostprocessingservice -... See merge request fforesight/layout-parser!186
Fixed Index out of bounds exception in blockificationpostprocessingservice - this could should be documented btw, there are also probably other use-cases where the code doesnt work
2024-07-30 16:56:21 +02:00 · 2024-07-30 17:45:05 +03:00 · 2024-07-25 14:11:00 +02:00 · 2024-07-24 16:35:11 +03:00 · 2024-07-18 17:19:44 +02:00 · 2024-07-18 17:19:43 +02:00
118 changed files with 3611 additions and 1781 deletions
@@ -43,6 +43,8 @@ public class DocumentStructure implements Serializable {
        public static final String POSITION = "position";
        public static final String ID = "id";

+        public static final String REPRESENTATION_HASH = "representationHash";
+
    }

    @Schema(description = "Object containing the extra field names, a table cell has in its properties field.")
@@ -67,11 +69,25 @@ public class DocumentStructure implements Serializable {

    public static Rectangle2D parseRectangle2D(String bBox) {

-        List<Float> floats = Arrays.stream(bBox.split(RECTANGLE_DELIMITER)).map(Float::parseFloat).toList();
+        List<Float> floats = Arrays.stream(bBox.split(RECTANGLE_DELIMITER))
+                .map(Float::parseFloat)
+                .toList();
        return new Rectangle2D.Float(floats.get(0), floats.get(1), floats.get(2), floats.get(3));
    }


+    public static double[] parseRepresentationVector(String representationHash) {
+
+        String[] stringArray = representationHash.split("[,\\s]+");
+        double[] doubleArray = new double[stringArray.length];
+        for (int i = 0; i < stringArray.length; i++) {
+            doubleArray[i] = Double.parseDouble(stringArray[i]);
+        }
+
+        return doubleArray;
+    }
+
+
    public EntryData get(List<Integer> tocId) {

        if (tocId.isEmpty()) {
@@ -87,19 +103,24 @@ public class DocumentStructure implements Serializable {

    public Stream<EntryData> streamAllEntries() {

-        return Stream.concat(Stream.of(root), root.children.stream()).flatMap(DocumentStructure::flatten);
+        return Stream.concat(Stream.of(root), root.children.stream())
+                .flatMap(DocumentStructure::flatten);
    }


    public String toString() {

-        return String.join("\n", streamAllEntries().map(EntryData::toString).toList());
+        return String.join("\n",
+                           streamAllEntries().map(EntryData::toString)
+                                   .toList());
    }


    private static Stream<EntryData> flatten(EntryData entry) {

-        return Stream.concat(Stream.of(entry), entry.children.stream().flatMap(DocumentStructure::flatten));
+        return Stream.concat(Stream.of(entry),
+                             entry.children.stream()
+                                     .flatMap(DocumentStructure::flatten));
    }


@@ -2,5 +2,6 @@ package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

 public enum LayoutEngine {
    ALGORITHM,
-    AI
+    AI,
+    OUTLINE
 }
@@ -14,7 +14,7 @@ import lombok.NoArgsConstructor;
 public class SimplifiedSectionText {

    @Schema(description = "The number of this Section. This is used to map the simplified section text back to the original Section.")
-    private int sectionNumber;
+    private String sectionNumber;
    @Schema(description = "The text in this Section.")
    private String text;

@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.taas;

+import java.util.List;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AllArgsConstructor;
 import lombok.Builder;
@@ -13,6 +15,8 @@ public class StructureObject {

    @Schema(description = "The ID of this StructureObject.")
    Integer structureObjectNumber;
+    @Schema(description = "The Tree ID of this StructureObject.")
+    List<Integer> treeId;
    @Schema(description = "This value indicates the start of the string offsets in this Object, with respect to the reading order.")
    int page;
    @Schema(description = "This stringOffset indicates the start of the string offsets in this Object, with respect to the reading order of the entire document. It is equal to the previous' StructureObject stringOffset + its length.")
@@ -19,7 +19,6 @@ public record LayoutParsingRequest(
        @Schema(description = "Path to the original PDF file.")//
        @NonNull String originFileStorageId,//

-
        @Schema(description = "Optional Path to the table extraction file.")//
        Optional<String> tablesFileStorageId,//
        @Schema(description = "Optional Path to the image classification file.")//
@@ -37,9 +36,12 @@ public record LayoutParsingRequest(
        @NonNull String positionBlockFileStorageId,//
        @Schema(description = "Path where the Document Pages File will be stored.")//
        @NonNull String pageFileStorageId,//
+        @Schema(description = "Path where the Document Markdown File will be stored.")//
+        Optional<String> documentMarkdownFileStorageId,//
        @Schema(description = "Path where the Simplified Text File will be stored.")//
        @NonNull String simplifiedTextStorageId,//
        @Schema(description = "Path where the Viewer Document PDF will be stored.")//
-        @NonNull String viewerDocumentStorageId) {
+        @NonNull String viewerDocumentStorageId
+) {

 }
@@ -26,4 +26,6 @@ dependencies {
    implementation("org.springframework.boot:spring-boot-starter-web:3.1.3")
    implementation("org.jgrapht:jgrapht-core:1.5.2")
    implementation("org.tinspin:tinspin-indexes:2.1.3")
+    implementation("org.commonmark:commonmark:0.22.0")
+    implementation("org.commonmark:commonmark-ext-gfm-tables:0.22.0")
 }
@@ -24,6 +24,7 @@ import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.No
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingFinishedEvent;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
+import com.knecon.fforesight.service.layoutparser.processor.markdown.MarkdownMapper;
 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationDocument;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
@@ -116,29 +117,14 @@ public class LayoutParsingPipeline {
        log.info("Starting layout parsing for {}", layoutParsingRequest.identifier());

        File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId());
-//        File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId()).orElse(originFile);
-        File viewerDocumentFile = originFile;
+        File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId()).orElse(originFile);

-        VisualLayoutParsingResponse visualLayoutParsingResponse = new VisualLayoutParsingResponse();
-        if (layoutParsingRequest.visualLayoutParsingFileId()
-                .isPresent()) {
-            visualLayoutParsingResponse = layoutParsingStorageService.getVisualLayoutParsingFile(layoutParsingRequest.visualLayoutParsingFileId()
-                                                                                                         .get());
-        }
-
-        ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
-        if (layoutParsingRequest.imagesFileStorageId()
-                .isPresent()) {
-            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId()
-                                                                                     .get());
-        }
-
-        TableServiceResponse tableServiceResponse = new TableServiceResponse();
-        if (layoutParsingRequest.tablesFileStorageId()
-                .isPresent()) {
-            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId()
-                                                                                     .get());
-        }
+        VisualLayoutParsingResponse visualLayoutParsingResponse = layoutParsingRequest.visualLayoutParsingFileId()
+                .map(layoutParsingStorageService::getVisualLayoutParsingFile).orElse(new VisualLayoutParsingResponse());
+        ImageServiceResponse imageServiceResponse = layoutParsingRequest.imagesFileStorageId()
+                .map(layoutParsingStorageService::getImagesFile).orElse(new ImageServiceResponse());
+        TableServiceResponse tableServiceResponse = layoutParsingRequest.tablesFileStorageId()
+                .map(layoutParsingStorageService::getTablesFile).orElse(new TableServiceResponse());

        ClassificationDocument classificationDocument = parseLayout(settings.getLayoutParsingTypeOverride() == null //
                                                                            ? layoutParsingRequest.layoutParsingType() : settings.getLayoutParsingTypeOverride(),
@@ -151,7 +137,8 @@ public class LayoutParsingPipeline {
        log.info("Building document graph for {}", layoutParsingRequest.identifier());

        Document documentGraph = observeBuildDocumentGraph(settings.getLayoutParsingTypeOverride() == null //
-                                                                   ? layoutParsingRequest.layoutParsingType() : settings.getLayoutParsingTypeOverride(), classificationDocument);
+                                                                   ? layoutParsingRequest.layoutParsingType() : settings.getLayoutParsingTypeOverride(),
+                                                           classificationDocument);

        log.info("Creating viewer document for {}", layoutParsingRequest.identifier());

@@ -160,10 +147,13 @@ public class LayoutParsingPipeline {
        log.info("Storing resulting files for {}", layoutParsingRequest.identifier());

        layoutParsingStorageService.storeDocumentData(layoutParsingRequest, DocumentDataMapper.toDocumentData(documentGraph));
+        if(layoutParsingRequest.documentMarkdownFileStorageId().isPresent()) {
+            layoutParsingStorageService.storeMarkdownFile(layoutParsingRequest.documentMarkdownFileStorageId().get(), new MarkdownMapper().toMarkdownContent(documentGraph));
+        }
        layoutParsingStorageService.storeSimplifiedText(layoutParsingRequest, simplifiedSectionTextService.toSimplifiedText(documentGraph));
        layoutParsingStorageService.storeViewerDocument(layoutParsingRequest, viewerDocumentFile);

-        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.CLARIFYND)) {
+        if (layoutParsingRequest.researchDocumentStorageId() != null) {
            log.info("Building research document data for {}", layoutParsingRequest.identifier());
            var researchDocumentData = TaasDocumentDataMapper.fromDocument(documentGraph);
            layoutParsingStorageService.storeResearchDocumentData(layoutParsingRequest, researchDocumentData);
@@ -244,17 +234,18 @@ public class LayoutParsingPipeline {
        Map<Integer, List<TableCells>> pdfTableCells = cvTableParsingAdapter.buildCvParsedTablesPerPage(tableServiceResponse);
        Map<Integer, List<ClassifiedImage>> pdfImages = imageServiceResponseAdapter.buildClassifiedImagesPerPage(imageServiceResponse);
        Map<Integer, List<ClassifiedImage>> signatures = visualLayoutParsingAdapter.buildExtractedSignaturesPerPage(visualLayoutParsingResponse);
+
        ClassificationDocument classificationDocument = new ClassificationDocument();

        if (settings.isDebug() || identifier.containsKey("debug")) {
-            classificationDocument.getVisualizations().setActive(true);
+            classificationDocument.getLayoutDebugLayer().setActive(true);
        }

        List<ClassificationPage> classificationPages = new ArrayList<>();
        OutlineObject lastProcessedOutlineObject = null;

        // parsing the structure elements could be useful as well
-        if(layoutParsingType != LayoutParsingType.REDACT_MANAGER_OLD && layoutParsingType != LayoutParsingType.DOCUMINE_OLD) {
+        if (layoutParsingType != LayoutParsingType.REDACT_MANAGER_OLD && layoutParsingType != LayoutParsingType.DOCUMINE_OLD) {
            classificationDocument.setOutlineObjectTree(outlineExtractorService.getOutlineObjectTree(originDocument));
        }

@@ -285,7 +276,7 @@ public class LayoutParsingPipeline {
            }
            stripper.getText(originDocument);
            List<TextPositionSequence> words = stripper.getTextPositionSequences();
-            classificationDocument.getVisualizations().addTextVisualizations(words, pageNumber);
+            classificationDocument.getLayoutDebugLayer().addTextVisualizations(words, pageNumber);

            PDRectangle pdr = pdPage.getMediaBox();

@@ -293,42 +284,34 @@ public class LayoutParsingPipeline {
            boolean isLandscape = pdr.getWidth() > pdr.getHeight() && (rotation == 0 || rotation == 180) || pdr.getHeight() > pdr.getWidth() && (rotation == 90 || rotation == 270);

            PDRectangle cropbox = pdPage.getCropBox();
-            classificationDocument.getVisualizations().addRulingVisualization(stripper.getRulings(), pageNumber);
+            classificationDocument.getLayoutDebugLayer().addRulingVisualization(stripper.getRulings(), pageNumber);
            CleanRulings cleanRulings = rulingCleaningService.deduplicateAndStraightenRulings(pdfTableCells.get(pageNumber), stripper.getRulings());

            PageInformation pageInformation = PageInformation.fromPDPage(pageNumber, pdPage);
            List<Cell> emptyTableCells = TableExtractionService.findCells(cleanRulings.getHorizontals(), cleanRulings.getVerticals(), pageInformation);
-            classificationDocument.getVisualizations().addCellVisualizations(emptyTableCells, pageNumber);
+            classificationDocument.getLayoutDebugLayer().addCellVisualizations(emptyTableCells, pageNumber);

            TextRulingsClassifier.classifyUnderlinedAndStrikethroughText(words, cleanRulings);

-            List<Box> graphics = graphicExtractorService.extractPathElementGraphics(originDocument,
-                                                                              pdPage,
-                                                                              pageNumber,
-                                                                              cleanRulings,
-                                                                              stripper.getTextPositionSequences(),
-
-                                                                              false);
+            List<Box> graphics = graphicExtractorService.extractPathElementGraphics(originDocument, pdPage, pageNumber, cleanRulings, stripper.getTextPositionSequences(), false);

            pdfImages.computeIfAbsent(pageNumber, x -> new ArrayList<>())
                    .addAll(graphics.stream()
-                                    .map(g -> new ClassifiedImage(new Rectangle2D.Double(g.x1, g.y1, g.width(), g.height()), ImageType.GRAPHIC, false, stripper.getPageNumber()))
+                                    .map(g -> new ClassifiedImage(new Rectangle2D.Double(g.x1, g.y1, g.width(), g.height()),
+                                                                  ImageType.GRAPHIC,
+                                                                  false,
+                                                                  stripper.getPageNumber(),
+                                                                  ""))
                                    .toList());

            ClassificationPage classificationPage = switch (layoutParsingType) {
                case REDACT_MANAGER_OLD ->
-                        redactManagerBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings, classificationDocument.getVisualizations());
+                        redactManagerBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings, classificationDocument.getLayoutDebugLayer());
                case DOCUMINE_OLD -> docuMineBlockificationService.blockify(words, cleanRulings);
-                case DOCUMINE, REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG, REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH -> docstrumBlockificationService.blockify(words,
-                                                                                                                                                                    cleanRulings,
-                                                                                                                                                                    true,
-                                                                                                                                                                    classificationDocument.getVisualizations(),
-                                                                                                                                                                    layoutParsingType);
-                case CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG -> docstrumBlockificationService.blockify(words,
-                                                                                                    cleanRulings,
-                                                                                                    false,
-                                                                                                    classificationDocument.getVisualizations(),
-                                                                                                    layoutParsingType);
+                case DOCUMINE, REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG, REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH ->
+                        docstrumBlockificationService.blockify(words, cleanRulings, true, classificationDocument.getLayoutDebugLayer(), layoutParsingType);
+                case CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG ->
+                        docstrumBlockificationService.blockify(words, cleanRulings, false, classificationDocument.getLayoutDebugLayer(), layoutParsingType);
            };

            classificationPage.setCleanRulings(cleanRulings);
@@ -338,7 +321,7 @@ public class LayoutParsingPipeline {
            classificationPage.setPageWidth(cropbox.getWidth());
            classificationPage.setPageHeight(cropbox.getHeight());

-            if(layoutParsingType != LayoutParsingType.REDACT_MANAGER_OLD && layoutParsingType != LayoutParsingType.DOCUMINE_OLD) {
+            if (layoutParsingType != LayoutParsingType.REDACT_MANAGER_OLD && layoutParsingType != LayoutParsingType.DOCUMINE_OLD) {
                List<OutlineObject> outlineObjects = classificationDocument.getOutlineObjectTree().getOutlineObjectsPerPage().getOrDefault(pageNumber - 1, new ArrayList<>());

                OutlineObject notFoundOutlineObject = null;
@@ -352,7 +335,7 @@ public class LayoutParsingPipeline {
                }
            }

-            classificationDocument.getVisualizations().addMarkedContentVisualizations(stripper.getMarkedContents(), pageNumber);
+            classificationDocument.getLayoutDebugLayer().addMarkedContentVisualizations(stripper.getMarkedContents(), pageNumber);
            // MarkedContent needs to be converted at this point, otherwise it leads to GC Problems in Pdfbox.
            classificationPage.setMarkedContentBboxPerType(convertMarkedContents(stripper.getMarkedContents()));

@@ -383,12 +366,12 @@ public class LayoutParsingPipeline {
        log.info("Calculating BodyTextFrame for {}", identifier);
        bodyTextFrameService.setBodyTextFrames(classificationDocument, layoutParsingType);
        for (ClassificationPage page : classificationDocument.getPages()) {
-            classificationDocument.getVisualizations().addCleanRulingVisualization(page.getCleanRulings(), page.getPageNumber());
+            classificationDocument.getLayoutDebugLayer().addCleanRulingVisualization(page.getCleanRulings(), page.getPageNumber());
        }
        log.info("Classify TextBlocks for {}", identifier);
        switch (layoutParsingType) {
-            case REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG, REDACT_MANAGER_OLD, CLARIFYND_PARAGRAPH_DEBUG, REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH -> redactManagerClassificationService.classifyDocument(
-                    classificationDocument);
+            case REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG, REDACT_MANAGER_OLD, CLARIFYND_PARAGRAPH_DEBUG, REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH ->
+                    redactManagerClassificationService.classifyDocument(classificationDocument);
            case DOCUMINE_OLD, DOCUMINE -> docuMineClassificationService.classifyDocument(classificationDocument);
            case CLARIFYND -> clarifyndClassificationService.classifyDocument(classificationDocument);
        }
@@ -5,7 +5,7 @@ import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.ComponentScan;
 import org.springframework.context.annotation.Configuration;

-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+import com.knecon.fforesight.service.viewerdoc.service.PDFTronViewerDocumentService;

 import io.micrometer.observation.ObservationRegistry;

@@ -13,12 +13,11 @@ import io.micrometer.observation.ObservationRegistry;
@ComponentScan
 public class LayoutParsingServiceProcessorConfiguration {

-
    @Bean
    @Autowired
-    public ViewerDocumentService viewerDocumentService(ObservationRegistry registry) {
+    public PDFTronViewerDocumentService viewerDocumentService(ObservationRegistry registry) {

-        return new ViewerDocumentService(registry);
+        return new PDFTronViewerDocumentService(registry);
    }

 }
@@ -1,9 +1,11 @@
 package com.knecon.fforesight.service.layoutparser.processor;

+import java.io.ByteArrayInputStream;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.charset.StandardCharsets;
 import java.nio.file.Files;
 import java.nio.file.Path;
 import java.nio.file.Paths;
@@ -21,6 +23,7 @@ import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsi
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.image.ImageServiceResponse;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.VisualLayoutParsingResponse;
+import com.knecon.fforesight.service.viewerdoc.service.ViewerDocVersioningUtility;
 import com.knecon.fforesight.tenantcommons.TenantContext;

 import io.micrometer.observation.annotation.Observed;
@@ -36,6 +39,7 @@ public class LayoutParsingStorageService {
    private final StorageService storageService;
    private final ObjectMapper objectMapper;

+
    @Observed(name = "LayoutParsingStorageService", contextualName = "get-origin-file")
    public File getOriginFile(String storageId) throws IOException {

@@ -53,11 +57,18 @@ public class LayoutParsingStorageService {
        }
        File tempFile = createTempFile("viewerDocument", ".pdf");
        storageService.downloadTo(TenantContext.getTenantId(), storageId, tempFile);
+
+        if (!ViewerDocVersioningUtility.isCurrentVersion(tempFile)) {
+            assert tempFile.delete();
+            return Optional.empty();
+        }
+
        return Optional.of(tempFile);
    }


-    public ImageServiceResponse getImagesFile(String storageId) throws IOException {
+    @SneakyThrows
+    public ImageServiceResponse getImagesFile(String storageId) {

        try (InputStream inputStream = getObject(storageId)) {

@@ -68,7 +79,8 @@ public class LayoutParsingStorageService {
    }


-    public TableServiceResponse getTablesFile(String storageId) throws IOException {
+    @SneakyThrows
+    public TableServiceResponse getTablesFile(String storageId) {

        try (var tableClassificationStream = getObject(storageId)) {

@@ -78,11 +90,12 @@ public class LayoutParsingStorageService {
        }
    }

-    public VisualLayoutParsingResponse getVisualLayoutParsingFile(String storageId) throws IOException {
+
+@SneakyThrows
+    public VisualLayoutParsingResponse getVisualLayoutParsingFile(String storageId) {

        try (InputStream inputStream = getObject(storageId)) {
-            VisualLayoutParsingResponse visualLayoutParsingResponse = objectMapper.readValue(inputStream, VisualLayoutParsingResponse.class);
-            return visualLayoutParsingResponse;
+            return objectMapper.readValue(inputStream, VisualLayoutParsingResponse.class);
        }
    }

@@ -154,4 +167,16 @@ public class LayoutParsingStorageService {
        }
    }

+
+    @SneakyThrows
+    @Observed(name = "LayoutParsingStorageService", contextualName = "store-markdown-file")
+    public void storeMarkdownFile(String markdownFileStorageId, String markdownContent) {
+
+        try (InputStream inputStream = new ByteArrayInputStream(markdownContent.getBytes(StandardCharsets.UTF_8))) {
+
+            storageService.storeObject(TenantContext.getTenantId(), markdownFileStorageId, inputStream);
+        }
+
+    }
+
 }
@@ -17,4 +17,5 @@ public class LayoutparserSettings {

    boolean debug;
    LayoutParsingType layoutParsingTypeOverride;
+    String pdftronLicense;
 }
@@ -18,7 +18,7 @@ import com.knecon.fforesight.service.layoutparser.processor.model.table.CleanRul
 import com.knecon.fforesight.service.layoutparser.processor.model.text.RedTextPosition;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextDirection;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
-import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutparsingVisualizations;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutDebugLayer;

 import lombok.RequiredArgsConstructor;

@@ -33,7 +33,7 @@ public class DocstrumSegmentationService {
    private final ReadingOrderService readingOrderService;


-    public List<Zone> segmentPage(List<TextPositionSequence> textPositions, boolean xyOrder, CleanRulings usedRulings, LayoutparsingVisualizations visualizations) {
+    public List<Zone> segmentPage(List<TextPositionSequence> textPositions, boolean xyOrder, CleanRulings usedRulings, LayoutDebugLayer visualizations) {

        List<Zone> zones = new ArrayList<>();
        zones.addAll(computeZones(textPositions, usedRulings, visualizations, TextDirection.ZERO));
@@ -45,7 +45,7 @@ public class DocstrumSegmentationService {
    }


-    private List<Zone> computeZones(List<TextPositionSequence> textPositions, CleanRulings rulings, LayoutparsingVisualizations visualizations, TextDirection direction) {
+    private List<Zone> computeZones(List<TextPositionSequence> textPositions, CleanRulings rulings, LayoutDebugLayer visualizations, TextDirection direction) {

        List<RedTextPosition> positions = textPositions.stream()
                .filter(t -> t.getDir() == direction)
@@ -0,0 +1,331 @@
+package com.knecon.fforesight.service.layoutparser.processor.markdown;
+
+import java.util.ArrayList;
+import java.util.HashSet;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.TreeMap;
+
+import org.commonmark.Extension;
+import org.commonmark.ext.gfm.tables.TableBlock;
+import org.commonmark.ext.gfm.tables.TableBody;
+import org.commonmark.ext.gfm.tables.TableCell;
+import org.commonmark.ext.gfm.tables.TableHead;
+import org.commonmark.ext.gfm.tables.TableRow;
+import org.commonmark.ext.gfm.tables.TablesExtension;
+import org.commonmark.node.Block;
+import org.commonmark.node.Document;
+import org.commonmark.node.Emphasis;
+import org.commonmark.node.HardLineBreak;
+import org.commonmark.node.Heading;
+import org.commonmark.node.Node;
+import org.commonmark.node.Paragraph;
+import org.commonmark.node.StrongEmphasis;
+import org.commonmark.node.Text;
+import org.commonmark.renderer.markdown.MarkdownRenderer;
+
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.AbstractNodeVisitor;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
+
+public class MarkdownMapper extends AbstractNodeVisitor {
+
+    Document markdownDocument = new Document();
+
+
+    public String toMarkdownContent(com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document document) {
+
+        visit(document);
+
+        return buildRenderer().render(this.markdownDocument);
+    }
+
+
+    @Override
+    public void visit(Headline headline) {
+
+        markdownDocument.appendChild(parseHeadline(headline));
+    }
+
+
+    @Override
+    public void visit(com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph paragraph) {
+
+        markdownDocument.appendChild(parseParagraph(paragraph));
+    }
+
+
+    @Override
+    public void visit(Table table) {
+
+        markdownDocument.appendChild(parseTable(table));
+    }
+
+
+    private static MarkdownRenderer buildRenderer() {
+
+        List<Extension> extensions = List.of(TablesExtension.create());
+        return MarkdownRenderer.builder().extensions(extensions).build();
+    }
+
+
+    private Block parseTable(Table table) {
+
+//        if (table.getNumberOfRows() == 1 && table.getNumberOfCols() == 1) {
+//            org.commonmark.node.Paragraph markdownParagraph = new org.commonmark.node.Paragraph();
+//            parseTextBlock(table.getTextBlock(), true).forEach(markdownParagraph::appendChild);
+//            return markdownParagraph;
+//        }
+
+        TableBlock tableNode = new TableBlock();
+        TableHead head = new TableHead();
+        TableRow tableRow = createTableRow(table, 0);
+        head.appendChild(tableRow);
+        int row = 1;
+        tableNode.appendChild(head);
+        TableBody tableBody = new TableBody();
+        for (; row < table.getNumberOfRows(); row++) {
+            tableBody.appendChild(createTableRow(table, row));
+        }
+        tableNode.appendChild(tableBody);
+        return tableNode;
+    }
+
+
+    private TableRow createTableRow(Table table, int row) {
+
+        TableRow tableRow = new TableRow();
+        table.streamRow(row)
+                .map(this::createTableCell)
+                .forEach(tableRow::appendChild);
+        return tableRow;
+    }
+
+
+    private Node createTableCell(com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell tc) {
+
+        var cell = new TableCell();
+        List<SemanticNode> childNodes = tc.streamChildren()
+                .toList();
+        if (childNodes.isEmpty()) {
+            parseTextBlock(tc.getTextBlock(), false).forEach(cell::appendChild);
+        } else {
+            childNodes.forEach(semanticNode -> parseTextBlock(semanticNode.getTextBlock(), false).forEach(cell::appendChild));
+        }
+        return cell;
+    }
+
+
+    private Paragraph parseParagraph(com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph paragraph) {
+
+        org.commonmark.node.Paragraph markdownParagraph = new org.commonmark.node.Paragraph();
+        parseTextBlock(paragraph.getTextBlock(), true).forEach(markdownParagraph::appendChild);
+        return markdownParagraph;
+    }
+
+
+    private Heading parseHeadline(Headline headline) {
+
+        Heading heading = new Heading();
+        heading.setLevel(headline.getTreeId().size());
+        heading.appendChild(parseTextBlockAsText(headline.getTextBlock()));
+        return heading;
+
+    }
+
+
+    private Text parseTextBlockAsText(TextBlock textBlock) {
+
+        return new Text(textBlock.getSearchText());
+    }
+
+
+    private List<Node> parseTextBlock(TextBlock textBlock, boolean includeLineBreaks) {
+
+        LinkedList<Node> result = new LinkedList<>();
+        List<TextRangeWithTextType> textRanges = mergeTextStyles(textBlock);
+
+        String fullText = getText(textBlock, textBlock.getTextRange(), includeLineBreaks);
+        List<Integer> lineTextSizes = getLineTextSizes(fullText);
+        int idx = 0;
+        int charCount = 0;
+        for (TextRangeWithTextType textRange : textRanges) {
+            String text = getText(textBlock, textRange.textRange(), includeLineBreaks);
+            String[] lines = text.split("\n");
+            for (String line : lines) {
+                charCount += line.length();
+                switch (textRange.fontStyle()) {
+                    case REGULAR -> result.add(new Text(line));
+                    case BOLD -> {
+                        StrongEmphasis boldBlock = new StrongEmphasis();
+                        boldBlock.appendChild(new Text(line));
+                        result.add(boldBlock);
+                    }
+                    case ITALIC -> {
+                        Emphasis italicBlock = new Emphasis();
+                        italicBlock.appendChild(new Text(line));
+                        result.add(italicBlock);
+                    }
+                    case BOLD_ITALIC -> {
+                        Emphasis italicBlock = new Emphasis();
+
+                        StrongEmphasis boldBlock = new StrongEmphasis();
+                        boldBlock.appendChild(new Text(line));
+
+                        italicBlock.appendChild(boldBlock);
+                        result.add(italicBlock);
+                    }
+                }
+                if (includeLineBreaks && lineTextSizes.get(idx).equals(charCount)) {
+                    result.add(new HardLineBreak());
+                    idx++;
+                }
+
+            }
+        }
+        if (!result.isEmpty() && result.getLast() instanceof HardLineBreak) {
+            result.removeLast();
+        }
+        return result;
+    }
+
+
+    private static List<Integer> getLineTextSizes(String input) {
+
+        String[] parts = input.split("\n");
+        List<Integer> textSizes = new ArrayList<>();
+
+        int size = 0;
+        for (int i = 0; i < parts.length; i++) {
+            size += parts[i].length();
+            textSizes.add(size);
+        }
+
+        if (textSizes.isEmpty()) {
+            textSizes.add(0);
+        }
+
+        return textSizes;
+    }
+
+
+    private static String getText(TextBlock textBlock, TextRange textRange, boolean includeLineBreaks) {
+
+        return includeLineBreaks ? textBlock.subSequenceWithLineBreaks(textRange) : textBlock.subSequence(textRange).toString();
+    }
+
+
+    private List<TextRangeWithTextType> mergeTextStyles(TextBlock textBlock) {
+
+        List<TextRangeWithTextType> result = new ArrayList<>();
+
+        TreeMap<Integer, Set<FontStyleChange>> styleChanges = new TreeMap<>();
+
+        int start = textBlock.getTextRange().start();
+        int end = textBlock.getTextRange().end();
+
+        for (TextRange bold : textBlock.getBoldTextBoundaries()) {
+            styleChanges.computeIfAbsent(bold.start() + start, k -> new HashSet<>()).add(FontStyleChange.enter(FontStyle.BOLD));
+            styleChanges.computeIfAbsent(bold.end() + start, k -> new HashSet<>()).add(FontStyleChange.leave(FontStyle.BOLD));
+        }
+
+        for (TextRange italic : textBlock.getItalicTextBoundaries()) {
+            styleChanges.computeIfAbsent(italic.start() + start, k -> new HashSet<>()).add(FontStyleChange.enter(FontStyle.ITALIC));
+            styleChanges.computeIfAbsent(italic.end() + start, k -> new HashSet<>()).add(FontStyleChange.leave(FontStyle.ITALIC));
+        }
+
+        if (styleChanges.isEmpty()) {
+            result.add(new TextRangeWithTextType(new TextRange(start, end), FontStyle.REGULAR));
+            return result;
+        }
+
+        Set<FontStyle> currentStyles = new HashSet<>();
+        currentStyles.add(FontStyle.REGULAR);
+
+        for (Map.Entry<Integer, Set<FontStyleChange>> entry : styleChanges.entrySet()) {
+            int point = entry.getKey();
+            Set<FontStyleChange> changes = entry.getValue();
+
+            if (point > start) {
+                FontStyle style = determineFontStyle(currentStyles);
+                result.add(new TextRangeWithTextType(new TextRange(start, point), style));
+            }
+
+            changes.stream()
+                    .filter(FontStyleChange::leave)
+                    .map(FontStyleChange::style)
+                    .toList()
+                    .forEach(currentStyles::remove);
+
+            currentStyles.addAll(changes.stream()
+                                         .filter(FontStyleChange::enter)
+                                         .map(FontStyleChange::style)
+                                         .toList());
+
+            if (currentStyles.isEmpty()) {
+                currentStyles.add(FontStyle.REGULAR);
+            }
+
+            start = point;
+        }
+
+        if (start < end) {
+            FontStyle style = determineFontStyle(currentStyles);
+            result.add(new TextRangeWithTextType(new TextRange(start, textBlock.getTextRange().end()), style));
+        }
+
+        return result;
+    }
+
+
+    private FontStyle determineFontStyle(Set<FontStyle> styles) {
+
+        if (styles.contains(FontStyle.BOLD) && styles.contains(FontStyle.ITALIC)) {
+            return FontStyle.BOLD_ITALIC;
+        } else if (styles.contains(FontStyle.BOLD)) {
+            return FontStyle.BOLD;
+        } else if (styles.contains(FontStyle.ITALIC)) {
+            return FontStyle.ITALIC;
+        } else {
+            return FontStyle.REGULAR;
+        }
+    }
+
+
+    enum FontStyle {
+        REGULAR,
+        BOLD,
+        ITALIC,
+        BOLD_ITALIC;
+    }
+
+    record FontStyleChange(boolean enter, FontStyle style) {
+
+        public static FontStyleChange enter(FontStyle style) {
+
+            return new FontStyleChange(true, style);
+        }
+
+
+        public static FontStyleChange leave(FontStyle style) {
+
+            return new FontStyleChange(false, style);
+        }
+
+
+        public boolean leave() {
+
+            return !enter;
+        }
+
+    }
+
+    record TextRangeWithTextType(TextRange textRange, FontStyle fontStyle) {
+
+    }
+}
@@ -1,9 +1,10 @@
 package com.knecon.fforesight.service.layoutparser.processor.model;

-import java.awt.geom.Rectangle2D;
+import java.util.HashSet;
+import java.util.Set;

 import com.fasterxml.jackson.annotation.JsonIgnore;
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.Rectangle;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.docstrum.model.BoundingBox;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;

@@ -20,6 +21,9 @@ public abstract class AbstractPageBlock extends BoundingBox {

    @JsonIgnore
    protected PageBlockType classification;
+
+    Set<LayoutEngine> engines = new HashSet<>();
+
    @JsonIgnore
    protected int page;

@@ -7,7 +7,7 @@ import com.knecon.fforesight.service.layoutparser.processor.model.outline.Outlin
 import com.knecon.fforesight.service.layoutparser.processor.model.outline.TableOfContents;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.StringFrequencyCounter;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.UnclassifiedText;
-import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutparsingVisualizations;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutDebugLayer;

 import lombok.Data;
 import lombok.NoArgsConstructor;
@@ -25,7 +25,7 @@ public class ClassificationDocument {
    private FloatFrequencyCounter fontSizeCounter = new FloatFrequencyCounter();
    private StringFrequencyCounter fontCounter = new StringFrequencyCounter();
    private StringFrequencyCounter fontStyleCounter = new StringFrequencyCounter();
-    private LayoutparsingVisualizations visualizations = new LayoutparsingVisualizations();
+    private LayoutDebugLayer layoutDebugLayer = new LayoutDebugLayer();
    private boolean headlines;

    private long rulesVersion;
@@ -0,0 +1,94 @@
+package com.knecon.fforesight.service.layoutparser.processor.model.graph;
+
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Footer;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Header;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Section;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SuperSection;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
+
+public abstract class AbstractNodeVisitor implements NodeVisitor {
+
+    @Override
+    public void visit(Document document) {
+
+        visitChildren(document);
+    }
+
+
+    @Override
+    public void visit(SuperSection superSection) {
+
+        visitChildren(superSection);
+    }
+
+
+    @Override
+    public void visit(Section section) {
+
+        visitChildren(section);
+    }
+
+
+    @Override
+    public void visit(Headline headline) {
+
+        visitChildren(headline);
+    }
+
+
+    @Override
+    public void visit(Paragraph paragraph) {
+
+        visitChildren(paragraph);
+    }
+
+
+    @Override
+    public void visit(Footer footer) {
+
+        visitChildren(footer);
+    }
+
+
+    @Override
+    public void visit(Header header) {
+
+        visitChildren(header);
+    }
+
+
+    @Override
+    public void visit(Image image) {
+
+        visitChildren(image);
+    }
+
+
+    @Override
+    public void visit(Table table) {
+
+        visitChildren(table);
+    }
+
+
+    @Override
+    public void visit(TableCell tableCell) {
+
+        visitChildren(tableCell);
+    }
+
+
+    private void visitChildren(SemanticNode semanticNode) {
+
+        semanticNode.streamChildren()
+                .forEach(node -> node.accept(this));
+
+    }
+
+}
@@ -2,6 +2,7 @@ package com.knecon.fforesight.service.layoutparser.processor.model.graph;

 import static java.lang.String.format;

+import java.util.ArrayList;
 import java.util.Collections;
 import java.util.LinkedList;
 import java.util.List;
@@ -39,7 +40,10 @@ public class DocumentTree {

    public TextBlock buildTextBlock() {

-        return allEntriesInOrder().map(Entry::getNode).filter(SemanticNode::isLeaf).map(SemanticNode::getLeafTextBlock).collect(new TextBlockCollector());
+        return allEntriesInOrder().map(Entry::getNode)
+                .filter(SemanticNode::isLeaf)
+                .map(SemanticNode::getLeafTextBlock)
+                .collect(new TextBlockCollector());
    }


@@ -113,13 +117,16 @@ public class DocumentTree {

    public Stream<SemanticNode> childNodes(List<Integer> treeId) {

-        return getEntryById(treeId).children.stream().map(Entry::getNode);
+        return getEntryById(treeId).children.stream()
+                .map(Entry::getNode);
    }


    public Stream<SemanticNode> childNodesOfType(List<Integer> treeId, NodeType nodeType) {

-        return getEntryById(treeId).children.stream().filter(entry -> entry.node.getType().equals(nodeType)).map(Entry::getNode);
+        return getEntryById(treeId).children.stream()
+                .filter(entry -> entry.node.getType().equals(nodeType))
+                .map(Entry::getNode);
    }


@@ -156,26 +163,32 @@ public class DocumentTree {

    public Stream<Entry> allEntriesInOrder() {

-        return Stream.of(root).flatMap(DocumentTree::flatten);
+        return Stream.of(root)
+                .flatMap(DocumentTree::flatten);
    }


    public Stream<Entry> allSubEntriesInOrder(List<Integer> parentId) {

-        return getEntryById(parentId).children.stream().flatMap(DocumentTree::flatten);
+        return getEntryById(parentId).children.stream()
+                .flatMap(DocumentTree::flatten);
    }


    @Override
    public String toString() {

-        return String.join("\n", allEntriesInOrder().map(Entry::toString).toList());
+        return String.join("\n",
+                           allEntriesInOrder().map(Entry::toString)
+                                   .toList());
    }


    private static Stream<Entry> flatten(Entry entry) {

-        return Stream.concat(Stream.of(entry), entry.children.stream().flatMap(DocumentTree::flatten));
+        return Stream.concat(Stream.of(entry),
+                             entry.children.stream()
+                                     .flatMap(DocumentTree::flatten));
    }


@@ -0,0 +1,45 @@
+package com.knecon.fforesight.service.layoutparser.processor.model.graph;
+
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Footer;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Header;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Section;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SuperSection;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
+
+public interface NodeVisitor {
+
+    void visit(Document document);
+
+
+    void visit(SuperSection superSection);
+
+
+    void visit(Section section);
+
+
+    void visit(Headline headline);
+
+
+    void visit(Paragraph paragraph);
+
+
+    void visit(Footer footer);
+
+
+    void visit(Header header);
+
+
+    void visit(Image image);
+
+
+    void visit(Table table);
+
+
+    void visit(TableCell tableCell);
+
+}
@@ -13,13 +13,13 @@ import lombok.Setter;
@Setter
@EqualsAndHashCode
@SuppressWarnings("PMD.AvoidFieldNameMatchingMethodName")
-public class Boundary implements Comparable<Boundary> {
+public class TextRange implements Comparable<TextRange> {

    private int start;
    private int end;


-    public Boundary(int start, int end) {
+    public TextRange(int start, int end) {

        if (start > end) {
            throw new IllegalArgumentException(format("start: %d > end: %d", start, end));
@@ -47,15 +47,15 @@ public class Boundary implements Comparable<Boundary> {
    }


-    public boolean contains(Boundary boundary) {
+    public boolean contains(TextRange textRange) {

-        return start <= boundary.start() && boundary.end() <= end;
+        return start <= textRange.start() && textRange.end() <= end;
    }


-    public boolean containedBy(Boundary boundary) {
+    public boolean containedBy(TextRange textRange) {

-        return boundary.contains(this);
+        return textRange.contains(this);
    }


@@ -79,22 +79,33 @@ public class Boundary implements Comparable<Boundary> {

    public boolean contains(int index) {

+        return start <= index && index <= end;
+    }
+
+
+    public boolean containsExclusive(int index) {
+
        return start <= index && index < end;
    }


-    public boolean intersects(Boundary boundary) {
+    public boolean intersects(TextRange textRange) {

-        return boundary.start() < this.end && this.start < boundary.end();
+        return textRange.start() < this.end && this.start < textRange.end();
    }


-    public List<Boundary> split(List<Integer> splitIndices) {
+    public List<TextRange> split(List<Integer> splitIndices) {

-        if (splitIndices.stream().anyMatch(idx -> !this.contains(idx))) {
-            throw new IndexOutOfBoundsException(format("%s splitting indices are out of range for %s", splitIndices.stream().filter(idx -> !this.contains(idx)).toList(), this));
+        if (splitIndices.stream()
+                .anyMatch(idx -> !this.containsExclusive(idx))) {
+            throw new IndexOutOfBoundsException(format("%s splitting indices are out of range for %s",
+                                                       splitIndices.stream()
+                                                               .filter(idx -> !this.containsExclusive(idx))
+                                                               .toList(),
+                                                       this));
        }
-        List<Boundary> splitBoundaries = new LinkedList<>();
+        List<TextRange> splitBoundaries = new LinkedList<>();
        int previousIndex = start;
        for (int splitIndex : splitIndices) {

@@ -102,23 +113,31 @@ public class Boundary implements Comparable<Boundary> {
            if (splitIndex == previousIndex) {
                continue;
            }
-            splitBoundaries.add(new Boundary(previousIndex, splitIndex));
+            splitBoundaries.add(new TextRange(previousIndex, splitIndex));
            previousIndex = splitIndex;
        }
-        splitBoundaries.add(new Boundary(previousIndex, end));
+        splitBoundaries.add(new TextRange(previousIndex, end));
        return splitBoundaries;
    }

+
    public IntStream intStream() {

        return IntStream.range(start, end);
    }

-    public static Boundary merge(Collection<Boundary> boundaries) {

-        int minStart = boundaries.stream().mapToInt(Boundary::start).min().orElseThrow(IllegalArgumentException::new);
-        int maxEnd = boundaries.stream().mapToInt(Boundary::end).max().orElseThrow(IllegalArgumentException::new);
-        return new Boundary(minStart, maxEnd);
+    public static TextRange merge(Collection<TextRange> boundaries) {
+
+        int minStart = boundaries.stream()
+                .mapToInt(TextRange::start)
+                .min()
+                .orElseThrow(IllegalArgumentException::new);
+        int maxEnd = boundaries.stream()
+                .mapToInt(TextRange::end)
+                .max()
+                .orElseThrow(IllegalArgumentException::new);
+        return new TextRange(minStart, maxEnd);
    }


@@ -130,12 +149,12 @@ public class Boundary implements Comparable<Boundary> {


    @Override
-    public int compareTo(Boundary boundary) {
+    public int compareTo(TextRange textRange) {

-        if (end < boundary.end() && start < boundary.start()) {
+        if (end < textRange.end() && start < textRange.start()) {
            return -1;
        }
-        if (start > boundary.start() && end > boundary.end()) {
+        if (start > textRange.start() && end > textRange.end()) {
            return 1;
        }

@@ -11,7 +11,7 @@ import java.util.Map;
 import java.util.Set;

 import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.Engine;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.utils.IdBuilder;
@@ -28,11 +28,11 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
@EqualsAndHashCode(onlyExplicitlyIncluded = true)
-public class RedactionEntity {
+public class TextEntity {

    // initial values
    @EqualsAndHashCode.Include
-    final Boundary boundary;
+    final TextRange textRange;
    @EqualsAndHashCode.Include
    final String type;
    @EqualsAndHashCode.Include
@@ -47,7 +47,7 @@ public class RedactionEntity {
    boolean dictionaryEntry;
    boolean dossierDictionaryEntry;
    Set<Engine> engines;
-    Set<RedactionEntity> references;
+    Set<TextEntity> references;
    @Builder.Default
    Deque<Integer> matchedRules = new LinkedList<>();
    String redactionReason;
@@ -66,9 +66,9 @@ public class RedactionEntity {
    SemanticNode deepestFullyContainingNode;


-    public static RedactionEntity initialEntityNode(Boundary boundary, String type, EntityType entityType) {
+    public static TextEntity initialEntityNode(TextRange textRange, String type, EntityType entityType) {

-        return RedactionEntity.builder().type(type).entityType(entityType).boundary(boundary).engines(new HashSet<>()).references(new HashSet<>()).build();
+        return TextEntity.builder().type(type).entityType(entityType).textRange(textRange).engines(new HashSet<>()).references(new HashSet<>()).build();
    }


@@ -132,7 +132,7 @@ public class RedactionEntity {
    public List<RedactionPosition> getRedactionPositionsPerPage() {

        if (redactionPositionsPerPage == null || redactionPositionsPerPage.isEmpty()) {
-            Map<Page, List<Rectangle2D>> rectanglesPerLinePerPage = deepestFullyContainingNode.getTextBlock().getPositionsPerPage(boundary);
+            Map<Page, List<Rectangle2D>> rectanglesPerLinePerPage = deepestFullyContainingNode.getTextBlock().getPositionsPerPage(textRange);

            Page firstPage = rectanglesPerLinePerPage.keySet()
                    .stream()
@@ -155,21 +155,21 @@ public class RedactionEntity {
    }


-    public boolean containedBy(RedactionEntity redactionEntity) {
+    public boolean containedBy(TextEntity textEntity) {

-        return this.boundary.containedBy(redactionEntity.getBoundary());
+        return this.textRange.containedBy(textEntity.getTextRange());
    }


-    public boolean contains(RedactionEntity redactionEntity) {
+    public boolean contains(TextEntity textEntity) {

-        return this.boundary.contains(redactionEntity.getBoundary());
+        return this.textRange.contains(textEntity.getTextRange());
    }


-    public boolean intersects(RedactionEntity redactionEntity) {
+    public boolean intersects(TextEntity textEntity) {

-        return this.boundary.intersects(redactionEntity.getBoundary());
+        return this.textRange.intersects(textEntity.getTextRange());
    }


@@ -185,13 +185,13 @@ public class RedactionEntity {
    }


-    public void addReference(RedactionEntity reference) {
+    public void addReference(TextEntity reference) {

        references.add(reference);
    }


-    public void addReferences(List<RedactionEntity> references) {
+    public void addReferences(List<TextEntity> references) {

        this.references.addAll(references);
    }
@@ -210,7 +210,7 @@ public class RedactionEntity {
        sb.append("Entity[\"");
        sb.append(value);
        sb.append("\", ");
-        sb.append(boundary);
+        sb.append(textRange);
        sb.append(", pages[");
        pages.forEach(page -> {
            sb.append(page.getNumber());
@@ -8,7 +8,7 @@ import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.RedactionEntity;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -39,7 +39,7 @@ public abstract class AbstractSemanticNode implements GenericSemanticNode {

    @Builder.Default
    @EqualsAndHashCode.Exclude
-    Set<RedactionEntity> entities = new HashSet<>();
+    Set<TextEntity> entities = new HashSet<>();

    @EqualsAndHashCode.Exclude
    Map<Page, Rectangle2D> bBoxCache;
@@ -11,8 +11,9 @@ import java.util.stream.Stream;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
-import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutparsingVisualizations;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutDebugLayer;

 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -33,7 +34,7 @@ public class Document extends AbstractSemanticNode {
    Set<Page> pages;
    Integer numberOfPages;

-    LayoutparsingVisualizations visualizations;
+    LayoutDebugLayer layoutDebugLayer;


    @Override
@@ -60,8 +61,8 @@ public class Document extends AbstractSemanticNode {
     *
     * @return A list of main sections within the document
     * @deprecated This method is marked for removal.
-     *  Use {@link #streamChildrenOfType(NodeType)} instead,
-     *  or {@link #getChildrenOfTypeSectionOrSuperSection()} which returns children of type SECTION as well as SUPER_SECTION.
+     * Use {@link #streamChildrenOfType(NodeType)} instead,
+     * or {@link #getChildrenOfTypeSectionOrSuperSection()} which returns children of type SECTION as well as SUPER_SECTION.
     */
    @Deprecated(forRemoval = true)
    public List<Section> getMainSections() {
@@ -101,8 +102,7 @@ public class Document extends AbstractSemanticNode {
    public Headline getHeadline() {

        return streamAllSubNodesOfType(NodeType.HEADLINE).map(node -> (Headline) node)
-                .findFirst()
-                .orElse(Headline.builder().build());
+                .findFirst().orElse(Headline.builder().build());
    }


@@ -163,4 +163,11 @@ public class Document extends AbstractSemanticNode {
        return bBox;
    }

+
+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
 }
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -34,6 +35,12 @@ public class Footer extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+        visitor.visit(this);
+    }
+
+
    @Override
    public TextBlock getTextBlock() {

@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -27,6 +28,13 @@ public class Header extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
+
    @Override
    public NodeType getType() {

@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -34,6 +35,13 @@ public class Headline extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
+
    @Override
    public TextBlock getTextBlock() {

@@ -7,6 +7,7 @@ import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -28,6 +29,8 @@ public class Image extends AbstractSemanticNode {

    String id;

+    String representationHash;
+
    ImageType imageType;
    boolean transparent;
    Rectangle2D position;
@@ -84,10 +87,29 @@ public class Image extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
+
    @Override
    public boolean isLeaf() {

        return true;
    }

+
+    public double getArea() {
+
+        return position.getWidth() * position.getHeight();
+    }
+
+
+    public boolean isFullPageImage() {
+
+        return imageType.equals(ImageType.OCR) || getArea() >= 0.5 * page.getArea();
+    }
+
 }
@@ -6,7 +6,6 @@ public enum ImageType {
    LOGO,
    FORMULA,
    SIGNATURE,
-
    SIGNATURE_VISUAL,
    OTHER,
    OCR,
@@ -6,7 +6,7 @@ import java.util.List;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.RedactionEntity;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlockCollector;

@@ -39,7 +39,7 @@ public class Page {

    @Builder.Default
    @EqualsAndHashCode.Exclude
-    Set<RedactionEntity> entities = new HashSet<>();
+    Set<TextEntity> entities = new HashSet<>();

    @Builder.Default
    @EqualsAndHashCode.Exclude
@@ -60,7 +60,10 @@ public class Page {

    public TextBlock getMainBodyTextBlock() {

-        return mainBody.stream().filter(SemanticNode::isLeaf).map(SemanticNode::getLeafTextBlock).collect(new TextBlockCollector());
+        return mainBody.stream()
+                .filter(SemanticNode::isLeaf)
+                .map(SemanticNode::getLeafTextBlock)
+                .collect(new TextBlockCollector());
    }


@@ -84,4 +87,10 @@ public class Page {
        return o instanceof Page && o.hashCode() == this.hashCode();
    }

+
+    public double getArea() {
+
+        return height * width;
+    }
+
 }
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;

 import lombok.AccessLevel;
@@ -34,6 +35,13 @@ public class Paragraph extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
+
    @Override
    public TextBlock getTextBlock() {

@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;

 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -26,15 +27,20 @@ public class Section extends AbstractSemanticNode {
    public Headline getHeadline() {

        return streamChildrenOfType(NodeType.HEADLINE).map(node -> (Headline) node)
-                .findFirst()
-                .orElseGet(() -> getParent().getHeadline());
+                .findFirst().orElseGet(() -> getParent().getHeadline());
+    }
+
+
+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
    }


    public boolean hasTables() {

-        return streamAllSubNodesOfType(NodeType.TABLE).findAny()
-                .isPresent();
+        return streamAllSubNodesOfType(NodeType.TABLE).findAny().isPresent();
    }


@@ -14,14 +14,16 @@ import java.util.stream.Stream;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.EntityType;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.RedactionEntity;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.AtomicTextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlockCollector;
+import com.knecon.fforesight.service.layoutparser.processor.utils.BBoxMergingUtility;
 import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;

 public interface SemanticNode {

@@ -42,7 +44,9 @@ public interface SemanticNode {
     */
    default TextBlock getTextBlock() {

-        return streamAllSubNodes().filter(SemanticNode::isLeaf).map(SemanticNode::getTextBlock).collect(new TextBlockCollector());
+        return streamAllSubNodes().filter(SemanticNode::isLeaf)
+                .map(SemanticNode::getTextBlock)
+                .collect(new TextBlockCollector());
    }


@@ -52,7 +56,7 @@ public interface SemanticNode {
     *
     * @return Set of all Entities associated with this Node
     */
-    Set<RedactionEntity> getEntities();
+    Set<TextEntity> getEntities();


    /**
@@ -68,7 +72,9 @@ public interface SemanticNode {

    default Page getFirstPage() {

-        return getTextBlock().getPages().stream().min(Comparator.comparingInt(Page::getNumber)).orElseThrow(() -> new IllegalStateException("SemanticNode has no Page!"));
+        return getTextBlock().getPages()
+                .stream()
+                .min(Comparator.comparingInt(Page::getNumber)).orElseThrow(() -> new IllegalStateException("SemanticNode has no Page!"));
    }


@@ -77,18 +83,19 @@ public interface SemanticNode {
     *
     * @return Set of PageNodes this node appears on.
     */
-    default Set<Page> getPages(Boundary boundary) {
+    default Set<Page> getPages(TextRange textRange) {

-        if (!getBoundary().contains(boundary)) {
-            throw new IllegalArgumentException(format("%s which was used to query for pages is not contained in the %s of this node!", boundary, getBoundary()));
+        if (!getBoundary().contains(textRange)) {
+            throw new IllegalArgumentException(format("%s which was used to query for pages is not contained in the %s of this node!", textRange, getBoundary()));
        }
-        return getTextBlock().getPages(boundary);
+        return getTextBlock().getPages(textRange);
    }


    default boolean isOnPage(int pageNumber) {

-        return getPages().stream().anyMatch(page -> page.getNumber() == pageNumber);
+        return getPages().stream()
+                .anyMatch(page -> page.getNumber() == pageNumber);
    }


@@ -203,7 +210,9 @@ public interface SemanticNode {
     */
    default boolean hasEntitiesOfType(String type) {

-        return getEntities().stream().filter(entity -> entity.getEntityType().equals(EntityType.ENTITY)).anyMatch(redactionEntity -> redactionEntity.getType().equals(type));
+        return getEntities().stream()
+                .filter(entity -> entity.getEntityType().equals(EntityType.ENTITY))
+                .anyMatch(redactionEntity -> redactionEntity.getType().equals(type));
    }


@@ -213,9 +222,11 @@ public interface SemanticNode {
     * @param type string representing the type of entities to return
     * @return List of RedactionEntities of any the type
     */
-    default List<RedactionEntity> getEntitiesOfType(String type) {
+    default List<TextEntity> getEntitiesOfType(String type) {

-        return getEntities().stream().filter(redactionEntity -> redactionEntity.getType().equals(type)).toList();
+        return getEntities().stream()
+                .filter(redactionEntity -> redactionEntity.getType().equals(type))
+                .toList();
    }


@@ -225,9 +236,11 @@ public interface SemanticNode {
     * @param types A list of strings representing the types of entities to return
     * @return List of RedactionEntities of any provided type
     */
-    default List<RedactionEntity> getEntitiesOfType(List<String> types) {
+    default List<TextEntity> getEntitiesOfType(List<String> types) {

-        return getEntities().stream().filter(redactionEntity -> redactionEntity.isAnyType(types)).toList();
+        return getEntities().stream()
+                .filter(redactionEntity -> redactionEntity.isAnyType(types))
+                .toList();
    }


@@ -279,7 +292,8 @@ public interface SemanticNode {
     */
    default boolean containsStrings(List<String> strings) {

-        return strings.stream().allMatch(this::containsString);
+        return strings.stream()
+                .allMatch(this::containsString);
    }


@@ -303,7 +317,8 @@ public interface SemanticNode {
     */
    default boolean containsAnyString(List<String> strings) {

-        return strings.stream().anyMatch(this::containsString);
+        return strings.stream()
+                .anyMatch(this::containsString);
    }


@@ -315,7 +330,8 @@ public interface SemanticNode {
     */
    default boolean containsAnyStringIgnoreCase(List<String> strings) {

-        return strings.stream().anyMatch(this::containsStringIgnoreCase);
+        return strings.stream()
+                .anyMatch(this::containsStringIgnoreCase);
    }


@@ -323,19 +339,19 @@ public interface SemanticNode {
     * This function is used during insertion of EntityNodes into the graph, it checks if the boundary of the RedactionEntity intersects or even contains the RedactionEntity.
     * It sets the fields accordingly and recursively calls this function on all its children.
     *
-     * @param redactionEntity RedactionEntity, which is being inserted into the graph
+     * @param textEntity RedactionEntity, which is being inserted into the graph
     */
-    default void addThisToEntityIfIntersects(RedactionEntity redactionEntity) {
+    default void addThisToEntityIfIntersects(TextEntity textEntity) {

        TextBlock textBlock = getTextBlock();
-        if (textBlock.getBoundary().intersects(redactionEntity.getBoundary())) {
-            if (textBlock.containsBoundary(redactionEntity.getBoundary())) {
-                redactionEntity.setDeepestFullyContainingNode(this);
+        if (textBlock.getTextRange().intersects(textEntity.getTextRange())) {
+            if (textBlock.containsBoundary(textEntity.getTextRange())) {
+                textEntity.setDeepestFullyContainingNode(this);
            }

-            redactionEntity.addIntersectingNode(this);
-            streamChildren().filter(semanticNode -> semanticNode.getBoundary().intersects(redactionEntity.getBoundary()))
-                    .forEach(node -> node.addThisToEntityIfIntersects(redactionEntity));
+            textEntity.addIntersectingNode(this);
+            streamChildren().filter(semanticNode -> semanticNode.getBoundary().intersects(textEntity.getTextRange()))
+                    .forEach(node -> node.addThisToEntityIfIntersects(textEntity));
        }
    }

@@ -386,7 +402,8 @@ public interface SemanticNode {
     */
    default Stream<SemanticNode> streamAllSubNodes() {

-        return getDocumentTree().allSubEntriesInOrder(getTreeId()).map(DocumentTree.Entry::getNode);
+        return getDocumentTree().allSubEntriesInOrder(getTreeId())
+                .map(DocumentTree.Entry::getNode);
    }


@@ -397,7 +414,9 @@ public interface SemanticNode {
     */
    default Stream<SemanticNode> streamAllSubNodesOfType(NodeType nodeType) {

-        return getDocumentTree().allSubEntriesInOrder(getTreeId()).filter(entry -> entry.getType().equals(nodeType)).map(DocumentTree.Entry::getNode);
+        return getDocumentTree().allSubEntriesInOrder(getTreeId())
+                .filter(entry -> entry.getType().equals(nodeType))
+                .map(DocumentTree.Entry::getNode);
    }


@@ -406,9 +425,9 @@ public interface SemanticNode {
     *
     * @return Boundary of this Node's TextBlock
     */
-    default Boundary getBoundary() {
+    default TextRange getBoundary() {

-        return getTextBlock().getBoundary();
+        return getTextBlock().getTextRange();
    }


@@ -453,17 +472,19 @@ public interface SemanticNode {
     */
    private Map<Page, Rectangle2D> getBBoxFromChildren() {

-        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
-        List<Map<Page, Rectangle2D>> childrenBBoxes = streamChildren().map(SemanticNode::getBBox).toList();
-        Set<Page> pages = childrenBBoxes.stream().flatMap(map -> map.keySet().stream()).collect(Collectors.toSet());
-        for (Page page : pages) {
-            Rectangle2D bBoxOnPage = childrenBBoxes.stream()
-                    .filter(childBboxPerPage -> childBboxPerPage.containsKey(page))
-                    .map(childBboxPerPage -> childBboxPerPage.get(page))
-                    .collect(RectangleTransformations.collectBBox());
-            bBoxPerPage.put(page, bBoxOnPage);
+        List<Map<Page, Rectangle2D>> childrenBBoxes = streamChildren().filter(child -> !isFullPageImage(child))
+                .map(SemanticNode::getBBox)
+                .toList();
+        return BBoxMergingUtility.mergeBBoxes(childrenBBoxes);
+    }
+
+
+    private static boolean isFullPageImage(SemanticNode child) {
+
+        if (!child.getType().equals(NodeType.IMAGE)) {
+            return false;
        }
-        return bBoxPerPage;
+        return ((Image) child).isFullPageImage();
    }


@@ -473,9 +494,14 @@ public interface SemanticNode {
    private Map<Page, Rectangle2D> getBBoxFromLeafTextBlock() {

        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
-        Map<Page, List<AtomicTextBlock>> atomicTextBlockPerPage = getTextBlock().getAtomicTextBlocks().stream().collect(Collectors.groupingBy(AtomicTextBlock::getPage));
+        Map<Page, List<AtomicTextBlock>> atomicTextBlockPerPage = getTextBlock().getAtomicTextBlocks()
+                .stream()
+                .collect(Collectors.groupingBy(AtomicTextBlock::getPage));
        atomicTextBlockPerPage.forEach((page, atbs) -> bBoxPerPage.put(page, RectangleTransformations.bBoxUnionAtomicTextBlock(atbs)));
        return bBoxPerPage;
    }

+
+    void accept(NodeVisitor visitor);
+
 }
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;

 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -26,8 +27,14 @@ public class SuperSection extends AbstractSemanticNode {
    public Headline getHeadline() {

        return streamChildrenOfType(NodeType.HEADLINE).map(node -> (Headline) node)
-                .findFirst()
-                .orElseGet(() -> getParent().getHeadline());
+                .findFirst().orElseGet(() -> getParent().getHeadline());
+    }
+
+
+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
    }


@@ -15,7 +15,8 @@ import java.util.stream.Stream;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.RedactionEntity;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.TextEntity;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlockCollector;

@@ -43,7 +44,7 @@ public class Table implements SemanticNode {

    @Builder.Default
    @EqualsAndHashCode.Exclude
-    Set<RedactionEntity> entities = new HashSet<>();
+    Set<TextEntity> entities = new HashSet<>();

    @EqualsAndHashCode.Exclude
    Map<Page, Rectangle2D> bBoxCache;
@@ -54,7 +55,7 @@ public class Table implements SemanticNode {
     * @param strings Strings to check whether a row contains them
     * @return Stream of all entities in this table, that appear in a row, which contains any of the provided strings
     */
-    public Stream<RedactionEntity> streamEntitiesWhereRowContainsStringsIgnoreCase(List<String> strings) {
+    public Stream<TextEntity> streamEntitiesWhereRowContainsStringsIgnoreCase(List<String> strings) {

        return IntStream.range(0, numberOfRows).boxed()
                .filter(row -> rowContainsStringsIgnoreCase(row, strings))
@@ -88,7 +89,7 @@ public class Table implements SemanticNode {
     * @param value  the string which the table cell should contain
     * @return a stream of all entities, which appear in a row where at least one cell has the provided header and the provided value.
     */
-    public Stream<RedactionEntity> streamEntitiesWhereRowHasHeaderAndValue(String header, String value) {
+    public Stream<TextEntity> streamEntitiesWhereRowHasHeaderAndValue(String header, String value) {

        List<Integer> vertebrateStudyCols = streamHeaders().filter(headerNode -> headerNode.containsString(header))
                .map(TableCell::getCol)
@@ -107,7 +108,7 @@ public class Table implements SemanticNode {
     * @param values the strings which the table cell should contain
     * @return a stream of all entities, which appear in a row where at least one cell has the provided header and any provided value.
     */
-    public Stream<RedactionEntity> streamEntitiesWhereRowHasHeaderAndAnyValue(String header, List<String> values) {
+    public Stream<TextEntity> streamEntitiesWhereRowHasHeaderAndAnyValue(String header, List<String> values) {

        List<Integer> colsWithHeader = streamHeaders().filter(headerNode -> headerNode.containsString(header))
                .map(TableCell::getCol)
@@ -125,7 +126,7 @@ public class Table implements SemanticNode {
     * @param types type strings to check whether a row contains an entity like them
     * @return Stream of all entities in this table, that appear in a row, which contains at least one entity with any of the provided types.
     */
-    public Stream<RedactionEntity> streamEntitiesWhereRowContainsEntitiesOfType(List<String> types) {
+    public Stream<TextEntity> streamEntitiesWhereRowContainsEntitiesOfType(List<String> types) {

        List<Integer> rowsWithEntityOfType = IntStream.range(0, numberOfRows).boxed()
                .filter(rowNumber -> streamEntityTypesInRow(rowNumber).anyMatch(existingType -> types.stream()
@@ -145,7 +146,7 @@ public class Table implements SemanticNode {
     * @param types type strings to check whether a row doesn't contain an entity like it
     * @return Stream of all entities in this table, that appear in a row, which does not contain any entity with any of the provided types.
     */
-    public Stream<RedactionEntity> streamEntitiesWhereRowContainsNoEntitiesOfType(List<String> types) {
+    public Stream<TextEntity> streamEntitiesWhereRowContainsNoEntitiesOfType(List<String> types) {

        List<Integer> rowsWithNoEntityOfType = IntStream.range(0, numberOfRows).boxed()
                .filter(rowNumber -> streamEntityTypesInRow(rowNumber).noneMatch(existingType -> types.stream()
@@ -163,7 +164,7 @@ public class Table implements SemanticNode {

        return streamRow(rowNumber).map(TableCell::getEntities)
                .flatMap(Collection::stream)
-                .map(RedactionEntity::getType)
+                .map(TextEntity::getType)
                .distinct();
    }

@@ -304,12 +305,12 @@ public class Table implements SemanticNode {
     * Finds all entities of the provided type, which appear in the same row that the provided entity appears in.
     *
     * @param type            the type of entities to search for
-     * @param redactionEntity the entity, which appears in the row to search
+     * @param textEntity the entity, which appears in the row to search
     * @return List of all entities of the provided type, which appear in the same row that the provided entity appears in.
     */
-    public List<RedactionEntity> getEntitiesOfTypeInSameRow(String type, RedactionEntity redactionEntity) {
+    public List<TextEntity> getEntitiesOfTypeInSameRow(String type, TextEntity textEntity) {

-        return redactionEntity.getIntersectingNodes()
+        return textEntity.getIntersectingNodes()
                .stream()
                .filter(node -> node instanceof TableCell)
                .map(node -> (TableCell) node)
@@ -353,4 +354,10 @@ public class Table implements SemanticNode {
        return bBoxCache;
    }

+
+    @Override
+    public void accept(NodeVisitor visitor) {
+        visitor.visit(this);
+    }
+
 }
@@ -5,6 +5,7 @@ import java.util.HashMap;
 import java.util.Map;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.NodeVisitor;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlockCollector;

@@ -42,6 +43,13 @@ public class TableCell extends AbstractSemanticNode {
    }


+    @Override
+    public void accept(NodeVisitor visitor) {
+
+        visitor.visit(this);
+    }
+
+
    @Override
    public NodeType getType() {

@@ -10,10 +10,12 @@ import java.util.Collections;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
+import java.util.Set;
+import java.util.stream.Collectors;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentPositionData;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentTextData;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
@@ -36,14 +38,14 @@ public class AtomicTextBlock implements TextBlock {
    Page page;

    //string coordinates
-    Boundary boundary;
+    TextRange textRange;
    String searchText;
    @Builder.Default
    List<Integer> lineBreaks = new ArrayList<>();
    @Builder.Default
-    List<Boundary> boldTextBoundaries = new ArrayList<>();
+    List<TextRange> boldTextBoundaries = new ArrayList<>();
    @Builder.Default
-    List<Boundary> italicTextBoundaries = new ArrayList<>();
+    List<TextRange> italicTextBoundaries = new ArrayList<>();
    String orientation;
    int textDirection;

@@ -64,10 +66,44 @@ public class AtomicTextBlock implements TextBlock {
    }


+    @Override
+    public String subSequenceWithLineBreaks(TextRange stringTextRange) {
+
+        if (stringTextRange.length() == 0 || !getTextRange().contains(stringTextRange)) {
+            return "";
+        }
+
+        Set<Integer> lbInBoundary = lineBreaks.stream()
+                .map(i -> i + this.textRange.start())
+                .filter(stringTextRange::contains)
+                .collect(Collectors.toSet());
+        if (stringTextRange.end() == getTextRange().end()) {
+            lbInBoundary.add(getTextRange().end());
+        }
+        StringBuilder sb = new StringBuilder();
+        for (int i = stringTextRange.start(); i < stringTextRange.end(); i++) {
+            char character = this.charAt(i);
+            if (lbInBoundary.contains(i + 1)) {
+                // always plus one, due to the linebreaks being an exclusive end index
+                if (!Character.isWhitespace(character)) {
+                    lbInBoundary.remove(i + 1);
+                    lbInBoundary.add(i + 2);
+                    sb.append(character);
+                    continue;
+                }
+                sb.append("\n");
+            } else {
+                sb.append(character);
+            }
+        }
+        return sb.toString();
+    }
+
+
    public static AtomicTextBlock fromSearchTextWithTextPosition(String searchText,
                                                                 List<Integer> lineBreaks,
-                                                                 List<Boundary> boldTextBoundaries,
-                                                                 List<Boundary> italicTextBoundaries,
+                                                                 List<TextRange> boldTextBoundaries,
+                                                                 List<TextRange> italicTextBoundaries,
                                                                 List<Rectangle2D> positions,
                                                                 List<Integer> stringIdxToPositionIdx,
                                                                 long idx,
@@ -89,7 +125,7 @@ public class AtomicTextBlock implements TextBlock {
                .italicTextBoundaries(italicTextBoundaries)
                .positions(positions)
                .stringIdxToPositionIdx(stringIdxToPositionIdx)
-                .boundary(new Boundary(offset, offset + searchText.length()))
+                .textRange(new TextRange(offset, offset + searchText.length()))
                .textDirection(textDirection)
                .orientation(orientation)
                .build();
@@ -100,7 +136,7 @@ public class AtomicTextBlock implements TextBlock {

        return AtomicTextBlock.builder()
                .id(textBlockIdx)
-                .boundary(new Boundary(stringOffset, stringOffset))
+                .textRange(new TextRange(stringOffset, stringOffset))
                .searchText("")
                .page(page)
                .numberOnPage(numberOnPage)
@@ -118,7 +154,7 @@ public class AtomicTextBlock implements TextBlock {
                .id(documentTextData.getId())
                .numberOnPage(documentTextData.getNumberOnPage())
                .page(page)
-                .boundary(new Boundary(documentTextData.getStart(), documentTextData.getEnd()))
+                .textRange(new TextRange(documentTextData.getStart(), documentTextData.getEnd()))
                .searchText(documentTextData.getSearchText())
                .lineBreaks(Arrays.stream(documentTextData.getLineBreaks()).boxed().toList())
                .stringIdxToPositionIdx(Arrays.stream(documentPositionData.getStringIdxToPositionIdx()).boxed().toList())
@@ -140,11 +176,11 @@ public class AtomicTextBlock implements TextBlock {
            throw new IndexOutOfBoundsException(format("line %d out of range for AtomicTextBlock with %d lines", lineNumber, numberOfLines()));
        }
        if (lineNumber == 0) {
-            return subSequence(boundary.start(), lineBreaks.get(0) + boundary.start());
+            return subSequence(textRange.start(), lineBreaks.get(0) + textRange.start());
        } else if (lineNumber == numberOfLines() - 1) {
-            return subSequence(lineBreaks.get(lineBreaks.size() - 1) + boundary.start(), boundary.end());
+            return subSequence(lineBreaks.get(lineBreaks.size() - 1) + textRange.start(), textRange.end());
        }
-        return subSequence(lineBreaks.get(lineNumber - 1) + boundary.start(), lineBreaks.get(lineNumber) + boundary.start());
+        return subSequence(lineBreaks.get(lineNumber - 1) + textRange.start(), lineBreaks.get(lineNumber) + textRange.start());
    }


@@ -159,9 +195,9 @@ public class AtomicTextBlock implements TextBlock {
    public int getNextLinebreak(int fromIndex) {

        return lineBreaks.stream()//
-                .filter(linebreak -> linebreak > fromIndex - boundary.start()) //
+                .filter(linebreak -> linebreak > fromIndex - textRange.start()) //
                .findFirst() //
-                .orElse(searchText.length()) + boundary.start();
+                .orElse(searchText.length()) + textRange.start();
    }


@@ -169,43 +205,43 @@ public class AtomicTextBlock implements TextBlock {
    public int getPreviousLinebreak(int fromIndex) {

        return lineBreaks.stream()//
-                .filter(linebreak -> linebreak <= fromIndex - boundary.start())//
+                .filter(linebreak -> linebreak <= fromIndex - textRange.start())//
                .reduce((a, b) -> b)//
-                .orElse(0) + boundary.start();
+                .orElse(0) + textRange.start();
    }


    @Override
    public Rectangle2D getPosition(int stringIdx) {

-        return positions.get(stringIdxToPositionIdx.get(stringIdx - boundary.start()));
+        return positions.get(stringIdxToPositionIdx.get(stringIdx - textRange.start()));
    }


    @Override
-    public List<Rectangle2D> getPositions(Boundary stringBoundary) {
+    public List<Rectangle2D> getPositions(TextRange stringTextRange) {

-        if (!containsBoundary(stringBoundary)) {
-            throw new IndexOutOfBoundsException(format("%s is out of bounds for %s", stringBoundary, this.boundary));
+        if (!containsBoundary(stringTextRange)) {
+            throw new IndexOutOfBoundsException(format("%s is out of bounds for %s", stringTextRange, this.textRange));
        }
-        if (stringBoundary.length() == 0) {
+        if (stringTextRange.length() == 0) {
            return Collections.emptyList();
        }

-        int startPositionIdx = stringIdxToPositionIdx.get(stringBoundary.start() - this.boundary.start());
+        int startPositionIdx = stringIdxToPositionIdx.get(stringTextRange.start() - this.textRange.start());

-        if (stringBoundary.end() == this.boundary.end()) {
+        if (stringTextRange.end() == this.textRange.end()) {
            return positions.subList(startPositionIdx, positions.size());
        }

-        return positions.subList(startPositionIdx, stringIdxToPositionIdx.get(stringBoundary.end() - this.boundary.start()));
+        return positions.subList(startPositionIdx, stringIdxToPositionIdx.get(stringTextRange.end() - this.textRange.start()));

    }


-    public Map<Page, List<Rectangle2D>> getPositionsPerPage(Boundary stringBoundary) {
+    public Map<Page, List<Rectangle2D>> getPositionsPerPage(TextRange stringTextRange) {

-        List<Rectangle2D> rectanglesPerLine = stringBoundary.split(getAllLineBreaksInBoundary(stringBoundary))
+        List<Rectangle2D> rectanglesPerLine = stringTextRange.split(getAllLineBreaksInBoundary(stringTextRange))
                .stream()
                .map(this::getPositions)
                .map(RectangleTransformations::rectangleBBoxWithGaps)
@@ -217,9 +253,9 @@ public class AtomicTextBlock implements TextBlock {
    }


-    private List<Integer> getAllLineBreaksInBoundary(Boundary boundary) {
+    protected List<Integer> getAllLineBreaksInBoundary(TextRange textRange) {

-        return getLineBreaks().stream().map(linebreak -> linebreak + this.boundary.start()).filter(boundary::contains).toList();
+        return getLineBreaks().stream().map(linebreak -> linebreak + this.textRange.start()).filter(textRange::contains).toList();
    }


@@ -11,7 +11,7 @@ import java.util.List;
 import java.util.Map;
 import java.util.stream.Stream;

-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;

 import lombok.AccessLevel;
@@ -24,7 +24,7 @@ public class ConcatenatedTextBlock implements TextBlock {

    List<AtomicTextBlock> atomicTextBlocks;
    String searchText;
-    Boundary boundary;
+    TextRange textRange;


    public static ConcatenatedTextBlock empty() {
@@ -37,29 +37,30 @@ public class ConcatenatedTextBlock implements TextBlock {

        this.atomicTextBlocks = new LinkedList<>();
        if (atomicTextBlocks.isEmpty()) {
-            boundary = new Boundary(-1, -1);
+            textRange = new TextRange(-1, -1);
            return;
        }
        var firstTextBlock = atomicTextBlocks.get(0);
        this.atomicTextBlocks.add(firstTextBlock);
-        boundary = new Boundary(firstTextBlock.getBoundary().start(), firstTextBlock.getBoundary().end());
+        textRange = new TextRange(firstTextBlock.getTextRange().start(), firstTextBlock.getTextRange().end());

-        atomicTextBlocks.subList(1, atomicTextBlocks.size()).forEach(this::concat);
+        atomicTextBlocks.subList(1, atomicTextBlocks.size())
+                .forEach(this::concat);
    }


    public ConcatenatedTextBlock concat(TextBlock textBlock) {

-        int start = textBlock.getBoundary().start();
-        int end = textBlock.getBoundary().end();
+        int start = textBlock.getTextRange().start();
+        int end = textBlock.getTextRange().end();
        if (this.atomicTextBlocks.isEmpty()) {
-            boundary.setStart(start);
-            boundary.setEnd(end);
-        } else if (boundary.end() != start) {
-            throw new UnsupportedOperationException(format("Can only concat consecutive TextBlocks, trying to concat %s and %s", boundary, textBlock.getBoundary()));
+            textRange.setStart(start);
+            textRange.setEnd(end);
+        } else if (textRange.end() != start) {
+            throw new UnsupportedOperationException(format("Can only concat consecutive TextBlocks, trying to concat %s and %s", textRange, textBlock.getTextRange()));
        }
        this.atomicTextBlocks.addAll(textBlock.getAtomicTextBlocks());
-        boundary.setEnd(end);
+        textRange.setEnd(end);
        this.searchText = null;
        return this;
    }
@@ -67,13 +68,18 @@ public class ConcatenatedTextBlock implements TextBlock {

    private AtomicTextBlock getAtomicTextBlockByStringIndex(int stringIdx) {

-        return atomicTextBlocks.stream().filter(textBlock -> textBlock.getBoundary().contains(stringIdx)).findAny().orElseThrow(IndexOutOfBoundsException::new);
+        return atomicTextBlocks.stream()
+                .filter(textBlock -> textBlock.getTextRange().containsExclusive(stringIdx))
+                .findAny()
+                .orElseThrow(IndexOutOfBoundsException::new);
    }


-    private List<AtomicTextBlock> getAllAtomicTextBlocksPartiallyInStringBoundary(Boundary boundary) {
+    private List<AtomicTextBlock> getAllAtomicTextBlocksPartiallyInStringBoundary(TextRange textRange) {

-        return atomicTextBlocks.stream().filter(tb -> tb.getBoundary().intersects(boundary)).toList();
+        return atomicTextBlocks.stream()
+                .filter(tb -> tb.getTextRange().intersects(textRange))
+                .toList();
    }


@@ -92,7 +98,9 @@ public class ConcatenatedTextBlock implements TextBlock {
    @Override
    public int numberOfLines() {

-        return atomicTextBlocks.stream().map(AtomicTextBlock::getLineBreaks).mapToInt(List::size).sum();
+        return atomicTextBlocks.stream()
+                .map(AtomicTextBlock::getLineBreaks)
+                .mapToInt(List::size).sum();
    }


@@ -113,7 +121,10 @@ public class ConcatenatedTextBlock implements TextBlock {
    @Override
    public List<Integer> getLineBreaks() {

-        return getAtomicTextBlocks().stream().flatMap(atomicTextBlock -> atomicTextBlock.getLineBreaks().stream()).toList();
+        return getAtomicTextBlocks().stream()
+                .flatMap(atomicTextBlock -> atomicTextBlock.getLineBreaks()
+                        .stream())
+                .toList();
    }


@@ -125,47 +136,48 @@ public class ConcatenatedTextBlock implements TextBlock {


    @Override
-    public List<Rectangle2D> getPositions(Boundary stringBoundary) {
+    public List<Rectangle2D> getPositions(TextRange stringTextRange) {

-        List<AtomicTextBlock> textBlocks = getAllAtomicTextBlocksPartiallyInStringBoundary(stringBoundary);
+        List<AtomicTextBlock> textBlocks = getAllAtomicTextBlocksPartiallyInStringBoundary(stringTextRange);

        if (textBlocks.size() == 1) {
-            return textBlocks.get(0).getPositions(stringBoundary);
+            return textBlocks.get(0).getPositions(stringTextRange);
        }

        AtomicTextBlock firstTextBlock = textBlocks.get(0);
-        List<Rectangle2D> positions = new LinkedList<>(firstTextBlock.getPositions(new Boundary(stringBoundary.start(), firstTextBlock.getBoundary().end())));
+        List<Rectangle2D> positions = new LinkedList<>(firstTextBlock.getPositions(new TextRange(stringTextRange.start(), firstTextBlock.getTextRange().end())));

        for (AtomicTextBlock textBlock : textBlocks.subList(1, textBlocks.size() - 1)) {
            positions.addAll(textBlock.getPositions());
        }

        var lastTextBlock = textBlocks.get(textBlocks.size() - 1);
-        positions.addAll(lastTextBlock.getPositions(new Boundary(lastTextBlock.getBoundary().start(), stringBoundary.end())));
+        positions.addAll(lastTextBlock.getPositions(new TextRange(lastTextBlock.getTextRange().start(), stringTextRange.end())));

        return positions;
    }


    @Override
-    public Map<Page, List<Rectangle2D>> getPositionsPerPage(Boundary stringBoundary) {
+    public Map<Page, List<Rectangle2D>> getPositionsPerPage(TextRange stringTextRange) {

-        List<AtomicTextBlock> textBlocks = getAllAtomicTextBlocksPartiallyInStringBoundary(stringBoundary);
+        List<AtomicTextBlock> textBlocks = getAllAtomicTextBlocksPartiallyInStringBoundary(stringTextRange);

        if (textBlocks.size() == 1) {
-            return textBlocks.get(0).getPositionsPerPage(stringBoundary);
+            return textBlocks.get(0).getPositionsPerPage(stringTextRange);
        }

        AtomicTextBlock firstTextBlock = textBlocks.get(0);
-        Map<Page, List<Rectangle2D>> rectanglesPerLinePerPage = firstTextBlock.getPositionsPerPage(new Boundary(stringBoundary.start(), firstTextBlock.getBoundary().end()));
+        Map<Page, List<Rectangle2D>> rectanglesPerLinePerPage = firstTextBlock.getPositionsPerPage(new TextRange(stringTextRange.start(), firstTextBlock.getTextRange().end()));

        for (AtomicTextBlock textBlock : textBlocks.subList(1, textBlocks.size() - 1)) {
-            rectanglesPerLinePerPage = mergeEntityPositionsWithSamePageNode(rectanglesPerLinePerPage, textBlock.getPositionsPerPage(textBlock.getBoundary()));
+            rectanglesPerLinePerPage = mergeEntityPositionsWithSamePageNode(rectanglesPerLinePerPage, textBlock.getPositionsPerPage(textBlock.getTextRange()));
        }

        AtomicTextBlock lastTextBlock = textBlocks.get(textBlocks.size() - 1);
        rectanglesPerLinePerPage = mergeEntityPositionsWithSamePageNode(rectanglesPerLinePerPage,
-                lastTextBlock.getPositionsPerPage(new Boundary(lastTextBlock.getBoundary().start(), stringBoundary.end())));
+                                                                        lastTextBlock.getPositionsPerPage(new TextRange(lastTextBlock.getTextRange().start(),
+                                                                                                                        stringTextRange.end())));

        return rectanglesPerLinePerPage;
    }
@@ -174,11 +186,42 @@ public class ConcatenatedTextBlock implements TextBlock {
    private Map<Page, List<Rectangle2D>> mergeEntityPositionsWithSamePageNode(Map<Page, List<Rectangle2D>> map1, Map<Page, List<Rectangle2D>> map2) {

        Map<Page, List<Rectangle2D>> mergedMap = new HashMap<>(map1);
-        map2.forEach((pageNode, rectangles) -> mergedMap.merge(pageNode, rectangles, (l1, l2) -> Stream.concat(l1.stream(), l2.stream()).toList()));
+        map2.forEach((pageNode, rectangles) -> mergedMap.merge(pageNode,
+                                                               rectangles,
+                                                               (l1, l2) -> Stream.concat(l1.stream(), l2.stream())
+                                                                       .toList()));
        return mergedMap;
    }


+    @Override
+    public String subSequenceWithLineBreaks(TextRange stringTextRange) {
+
+        if (stringTextRange.length() == 0 || !getTextRange().contains(stringTextRange)) {
+            return "";
+        }
+
+        List<AtomicTextBlock> textBlocks = getAllAtomicTextBlocksPartiallyInStringBoundary(stringTextRange);
+
+        if (textBlocks.size() == 1) {
+            return textBlocks.get(0).subSequenceWithLineBreaks(stringTextRange);
+        }
+
+        StringBuilder sb = new StringBuilder();
+        AtomicTextBlock firstTextBlock = textBlocks.get(0);
+        sb.append(firstTextBlock.subSequenceWithLineBreaks(new TextRange(stringTextRange.start(), firstTextBlock.getTextRange().end())));
+
+        for (AtomicTextBlock textBlock : textBlocks.subList(1, textBlocks.size() - 1)) {
+            sb.append(textBlock.searchTextWithLineBreaks());
+        }
+
+        var lastTextBlock = textBlocks.get(textBlocks.size() - 1);
+        sb.append(lastTextBlock.subSequenceWithLineBreaks(new TextRange(lastTextBlock.getTextRange().start(), stringTextRange.end())));
+
+        return sb.toString();
+    }
+
+
    @Override
    public String toString() {

@@ -187,16 +230,22 @@ public class ConcatenatedTextBlock implements TextBlock {


    @Override
-    public List<Boundary> getBoldTextBoundaries() {
+    public List<TextRange> getBoldTextBoundaries() {

-        return getAtomicTextBlocks().stream().map(AtomicTextBlock::getBoldTextBoundaries).flatMap(Collection::stream).toList();
+        return getAtomicTextBlocks().stream()
+                .map(AtomicTextBlock::getBoldTextBoundaries)
+                .flatMap(Collection::stream)
+                .toList();
    }


    @Override
-    public List<Boundary> getItalicTextBoundaries() {
+    public List<TextRange> getItalicTextBoundaries() {

-        return getAtomicTextBlocks().stream().map(AtomicTextBlock::getItalicTextBoundaries).flatMap(Collection::stream).toList();
+        return getAtomicTextBlocks().stream()
+                .map(AtomicTextBlock::getItalicTextBoundaries)
+                .flatMap(Collection::stream)
+                .toList();
    }


@@ -10,7 +10,7 @@ import java.util.Map;
 import java.util.Set;
 import java.util.stream.Collectors;

-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;

 public interface TextBlock extends CharSequence {
@@ -21,10 +21,10 @@ public interface TextBlock extends CharSequence {
    List<AtomicTextBlock> getAtomicTextBlocks();


-    List<Boundary> getBoldTextBoundaries();
+    List<TextRange> getBoldTextBoundaries();


-    List<Boundary> getItalicTextBoundaries();
+    List<TextRange> getItalicTextBoundaries();


    String getOrientation();
@@ -33,7 +33,7 @@ public interface TextBlock extends CharSequence {
    int getTextDirection();


-    Boundary getBoundary();
+    TextRange getTextRange();


    int getNextLinebreak(int fromIndex);
@@ -48,31 +48,41 @@ public interface TextBlock extends CharSequence {
    Rectangle2D getPosition(int stringIdx);


-    List<Rectangle2D> getPositions(Boundary stringBoundary);
+    List<Rectangle2D> getPositions(TextRange stringTextRange);


-    Map<Page, List<Rectangle2D>> getPositionsPerPage(Boundary stringBoundary);
+    Map<Page, List<Rectangle2D>> getPositionsPerPage(TextRange stringTextRange);


    int numberOfLines();


+    String subSequenceWithLineBreaks(TextRange stringTextRange);
+
+
+    default String searchTextWithLineBreaks() {
+
+        return subSequenceWithLineBreaks(getTextRange());
+    }
+
    default int indexOf(String searchTerm) {

-        return indexOf(searchTerm, getBoundary().start());
+        return indexOf(searchTerm, getTextRange().start());
    }


    default Set<Page> getPages() {

-        return getAtomicTextBlocks().stream().map(AtomicTextBlock::getPage).collect(Collectors.toUnmodifiableSet());
+        return getAtomicTextBlocks().stream()
+                .map(AtomicTextBlock::getPage)
+                .collect(Collectors.toUnmodifiableSet());
    }


-    default Set<Page> getPages(Boundary boundary) {
+    default Set<Page> getPages(TextRange textRange) {

        return getAtomicTextBlocks().stream()
-                .filter(atomicTextBlock -> atomicTextBlock.getBoundary().intersects(boundary))
+                .filter(atomicTextBlock -> atomicTextBlock.getTextRange().intersects(textRange))
                .map(AtomicTextBlock::getPage)
                .collect(Collectors.toUnmodifiableSet());
    }
@@ -80,38 +90,38 @@ public interface TextBlock extends CharSequence {

    default int indexOf(String searchTerm, int startOffset) {

-        int start = getSearchText().indexOf(searchTerm, startOffset - getBoundary().start());
+        int start = getSearchText().indexOf(searchTerm, startOffset - getTextRange().start());
        if (start == -1) {
            return -1;
        }
-        return start + getBoundary().start();
+        return start + getTextRange().start();
    }


    default CharSequence getFirstLine() {

-        return subSequence(getBoundary().start(), getNextLinebreak(getBoundary().start()));
+        return subSequence(getTextRange().start(), getNextLinebreak(getTextRange().start()));
    }


-    default boolean containsBoundary(Boundary boundary) {
+    default boolean containsBoundary(TextRange textRange) {

-        if (boundary.end() < boundary.start()) {
-            throw new IllegalArgumentException(format("Invalid %s, StartIndex must be smaller than EndIndex", boundary));
+        if (textRange.end() < textRange.start()) {
+            throw new IllegalArgumentException(format("Invalid %s, StartIndex must be smaller than EndIndex", textRange));
        }
-        return getBoundary().contains(boundary);
+        return getTextRange().contains(textRange);
    }


    default boolean containsIndex(int stringIndex) {

-        return getBoundary().contains(stringIndex);
+        return getTextRange().containsExclusive(stringIndex);
    }


-    default CharSequence subSequence(Boundary boundary) {
+    default CharSequence subSequence(TextRange textRange) {

-        return subSequence(boundary.start(), boundary.end());
+        return subSequence(textRange.start(), textRange.end());
    }


@@ -128,21 +138,21 @@ public interface TextBlock extends CharSequence {
    @Override
    default CharSequence subSequence(int start, int end) {

-        return getSearchText().substring(start - getBoundary().start(), end - getBoundary().start());
+        return getSearchText().substring(start - getTextRange().start(), end - getTextRange().start());
    }


    @Override
    default int length() {

-        return getBoundary().length();
+        return getTextRange().length();
    }


    @Override
    default char charAt(int index) {

-        return getSearchText().charAt(index - getBoundary().start());
+        return getSearchText().charAt(index - getTextRange().start());
    }

 }
@@ -24,5 +24,7 @@ public class ClassifiedImage {
    private boolean hasTransparency;
    @NonNull
    private int page;
+    @NonNull
+    private String representation;

 }
@@ -12,12 +12,14 @@ import org.springframework.stereotype.Service;

 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;

+import io.micrometer.observation.annotation.Observed;
 import lombok.extern.slf4j.Slf4j;

@Service
@Slf4j
 public class OutlineValidationService {

+    @Observed(name = "OutlineValidationService", contextualName = "create-toc")
    public TableOfContents createToC(List<TextPageBlock> headlines) {

        List<TableOfContentItem> mainSections = new ArrayList<>();
@@ -32,7 +32,7 @@ public class ImageServiceResponseAdapter {
                    .add(new ClassifiedImage(new Rectangle2D.Double(imageMetadata.getPosition().getX1(),
                            imageMetadata.getPosition().getY1(),
                            imageMetadata.getGeometry().getWidth(),
-                            imageMetadata.getGeometry().getHeight()), classification, imageMetadata.isAlpha(), imageMetadata.getPosition().getPageNumber()));
+                            imageMetadata.getGeometry().getHeight()), classification, imageMetadata.isAlpha(), imageMetadata.getPosition().getPageNumber(),imageMetadata.getRepresentation()));
        });

        // Currently This is a copy but, it will be changed later because i don' t think that we should unclassified images.
@@ -44,7 +44,7 @@ public class ImageServiceResponseAdapter {
                    .add(new ClassifiedImage(new Rectangle2D.Double(imageMetadata.getPosition().getX1(),
                            imageMetadata.getPosition().getY1(),
                            imageMetadata.getGeometry().getWidth(),
-                            imageMetadata.getGeometry().getHeight()), classification, imageMetadata.isAlpha(), imageMetadata.getPosition().getPageNumber()));
+                            imageMetadata.getGeometry().getHeight()), classification, imageMetadata.isAlpha(), imageMetadata.getPosition().getPageNumber(),imageMetadata.getRepresentation()));
        });

        return images;
@@ -79,7 +79,7 @@ public class VisualLayoutParsingAdapter {
                ClassifiedImage signature = new ClassifiedImage(new Rectangle2D.Float(t.getBox().getX1(),
                        t.getBox().getY1(),
                        t.getBox().getX2() - t.getBox().getX1(),
-                        t.getBox().getY2() - t.getBox().getY1()), ImageType.SIGNATURE, true, false, false, pageNumber);
+                        t.getBox().getY2() - t.getBox().getY1()), ImageType.SIGNATURE, true, false, false, pageNumber,"");

                signatures.add(signature);
            }
@@ -12,6 +12,7 @@ import lombok.NoArgsConstructor;
 public class ImageMetadata {

    private Classification classification;
+    private String representation;
    private Position position;
    private Geometry geometry;
    private Filters filters;
@@ -33,7 +33,7 @@ public class BodyTextFrameService {
        for (ClassificationPage page : classificationDocument.getPages()) {
            var updatedBodyTextFrame = getBodyTextFrameFromRulings(page, bodyTextFrame, landscapeBodyTextFrame);
            setBodyTextFrameAdjustedToPage(page, bodyTextFrame, landscapeBodyTextFrame);
-            classificationDocument.getVisualizations().addMainBodyVisualization(page.isLandscape() ? landscapeBodyTextFrame : bodyTextFrame, page.getPageNumber());
+            classificationDocument.getLayoutDebugLayer().addMainBodyVisualization(page.isLandscape() ? landscapeBodyTextFrame : bodyTextFrame, page.getPageNumber());
        }
    }

@@ -42,7 +42,9 @@ public class SimplifiedSectionTextService {

        return SimplifiedSectionText.builder()
                .sectionNumber(section.getTreeId()
-                                       .get(0))
+                                       .stream()
+                                       .map(String::valueOf)
+                                       .collect(Collectors.joining(".")))
                .text(section.getTextBlock().getSearchText())
                .build();
    }
@@ -12,6 +12,7 @@ import java.util.function.Function;
 import org.apache.commons.lang3.StringUtils;
 import org.springframework.stereotype.Service;

+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
 import com.knecon.fforesight.service.layoutparser.processor.model.SectionIdentifier;
@@ -204,21 +205,28 @@ public class BlockificationPostprocessingService {
            return false;
        }
        if (minDistance == distanceToDirectMatch) {
-            directMatch.setClassification(headlineType);
+            setClassificationAndAddOutlineEngine(directMatch, headlineType);
        } else if (minDistance == distanceToSplitCandidate) {
            SplitBlockResult splitBlockResult = splitBlock(classificationPage, splitCandidate, context.sectionIdentifier, outlineObject.getTitle());
            if (splitBlockResult.modifiedBlockToSplit) {
-                splitCandidate.setClassification(headlineType);
+                setClassificationAndAddOutlineEngine(splitCandidate, headlineType);
            }
            splitBlockResult.otherBlocks.forEach(other -> other.setClassification(null));
        } else {
            var merged = mergeBlocks(classificationPage, bestMergeCandidateCombination);
-            merged.setClassification(headlineType);
+            setClassificationAndAddOutlineEngine(merged, headlineType);
        }
        return true;
    }


+    private static void setClassificationAndAddOutlineEngine(TextPageBlock block, PageBlockType headlineType) {
+
+        block.setClassification(headlineType);
+        block.getEngines().add(LayoutEngine.OUTLINE);
+    }
+
+
    private SplitBlockResult splitBlock(ClassificationPage classificationPage, TextPageBlock blockToSplit, SectionIdentifier sectionIdentifier, String title) {

        List<TextPageBlock> otherBlocks = new ArrayList<>();
@@ -265,6 +273,10 @@ public class BlockificationPostprocessingService {
        List<TextPositionSequence> postSequence = new ArrayList<>();
        StringBuilder currentSequence = new StringBuilder();

+        if (target.isBlank()){
+            return new WordSequenceResult();
+        }
+
        for (TextPositionSequence sequence : textPositionSequences) {

            currentSequence.append(sanitizeString(sequence.toString()));
@@ -17,7 +17,7 @@ import com.knecon.fforesight.service.layoutparser.processor.model.table.CleanRul
 import com.knecon.fforesight.service.layoutparser.processor.model.table.TablePageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
-import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutparsingVisualizations;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutDebugLayer;

 import lombok.RequiredArgsConstructor;

@@ -35,7 +35,7 @@ public class DocstrumBlockificationService {
    public ClassificationPage blockify(List<TextPositionSequence> textPositions,
                                       CleanRulings rulings,
                                       boolean xyOrder,
-                                       LayoutparsingVisualizations visualizations,
+                                       LayoutDebugLayer visualizations,
                                       LayoutParsingType layoutParsingType) {

        CleanRulings usedRulings = rulings.withoutTextRulings();
@@ -1,9 +1,6 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.blockification;

-import static java.util.stream.Collectors.toSet;
-
 import java.util.ArrayList;
-import java.util.Comparator;
 import java.util.Iterator;
 import java.util.List;

@@ -11,13 +8,11 @@ import org.springframework.stereotype.Service;

 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
-import com.knecon.fforesight.service.layoutparser.processor.model.FloatFrequencyCounter;
 import com.knecon.fforesight.service.layoutparser.processor.model.Orientation;
 import com.knecon.fforesight.service.layoutparser.processor.model.table.CleanRulings;
-import com.knecon.fforesight.service.layoutparser.processor.model.text.StringFrequencyCounter;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
-import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutparsingVisualizations;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutDebugLayer;

@SuppressWarnings("all")
@Service
@@ -35,7 +30,7 @@ public class RedactManagerBlockificationService {
     * @param visualizations
     * @return Page object that contains the Textblock and text statistics.
     */
-    public ClassificationPage blockify(List<TextPositionSequence> textPositions, CleanRulings cleanRulings, LayoutparsingVisualizations visualizations) {
+    public ClassificationPage blockify(List<TextPositionSequence> textPositions, CleanRulings cleanRulings, LayoutDebugLayer visualizations) {

        CleanRulings usedRulings = cleanRulings.withoutTextRulings();

@@ -21,33 +21,31 @@ import lombok.extern.slf4j.Slf4j;
@RequiredArgsConstructor
 public class ClarifyndClassificationService {

-    private final HeadlineClassificationService headlineClassificationService;
-
    public void classifyDocument(ClassificationDocument document) {

        List<Float> headlineFontSizes = document.getFontSizeCounter().getHighterThanMostPopular();

        log.debug("Document FontSize counters are: {}", document.getFontSizeCounter().getCountPerValue());

-        headlineClassificationService.resetContext();
+        HeadlineClassificationService headlineClassificationService = new HeadlineClassificationService();

        for (ClassificationPage page : document.getPages()) {
-            classifyPage(page, document, headlineFontSizes);
+            classifyPage(headlineClassificationService, page, document, headlineFontSizes);
        }
    }


-    private void classifyPage(ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyPage(HeadlineClassificationService headlineClassificationService, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {

        for (AbstractPageBlock textBlock : page.getTextBlocks()) {
            if (textBlock instanceof TextPageBlock) {
-                classifyBlock((TextPageBlock) textBlock, page, document, headlineFontSizes);
+                classifyBlock(headlineClassificationService, (TextPageBlock) textBlock, page, document, headlineFontSizes);
            }
        }
    }


-    private void classifyBlock(TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyBlock(HeadlineClassificationService headlineClassificationService, TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {

        var bodyTextFrame = page.getBodyTextFrame();

@@ -24,10 +24,9 @@ import lombok.extern.slf4j.Slf4j;
@RequiredArgsConstructor
 public class DocuMineClassificationService {

-    private final HeadlineClassificationService headlineClassificationService;
-    private static final Pattern pattern = Pattern.compile("^([1-9]\\d?\\.){1,3}\\d{1,2}\\.?\\s[0-9A-Za-z \\[\\]]{2,50}", Pattern.CASE_INSENSITIVE);
-    private static final Pattern pattern2 = Pattern.compile("\\p{L}{3,}", Pattern.CASE_INSENSITIVE);
-    private static final Pattern pattern3 = Pattern.compile("^(\\d{1,1}\\.){1,3}\\d{1,2}\\.?\\s[a-z]{1,2}\\/[a-z]{1,2}.*");
+    private static final Pattern HEADLINE_WITH_IDENTIFER_PATTERN = Pattern.compile("^([1-9]\\d?\\.){1,3}\\d{1,2}\\.?\\s[0-9A-Za-z \\[\\]]{2,50}", Pattern.CASE_INSENSITIVE);
+    private static final Pattern AT_LEAST_3_PATTERN = Pattern.compile("\\p{L}{3,}", Pattern.CASE_INSENSITIVE);
+    private static final Pattern HEADLINE_PATTTERN_WITH_SLASHES = Pattern.compile("^(\\d{1,1}\\.){1,3}\\d{1,2}\\.?\\s[a-z]{1,2}\\/[a-z]{1,2}.*");


    public void classifyDocument(ClassificationDocument document) {
@@ -36,32 +35,39 @@ public class DocuMineClassificationService {

        log.debug("Document FontSize counters are: {}", document.getFontSizeCounter().getCountPerValue());

-        headlineClassificationService.resetContext();
+        HeadlineClassificationService headlineClassificationService = new HeadlineClassificationService();

        for (ClassificationPage page : document.getPages()) {
-            classifyPage(page, document, headlineFontSizes);
+            classifyPage(headlineClassificationService, page, document, headlineFontSizes);
        }
    }


-    private void classifyPage(ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyPage(HeadlineClassificationService headlineClassificationService,
+                              ClassificationPage page,
+                              ClassificationDocument document,
+                              List<Float> headlineFontSizes) {

        for (AbstractPageBlock textBlock : page.getTextBlocks()) {
            if (textBlock instanceof TextPageBlock) {
-                classifyBlock((TextPageBlock) textBlock, page, document, headlineFontSizes);
+                classifyBlock(headlineClassificationService, (TextPageBlock) textBlock, page, document, headlineFontSizes);
            }
        }
    }


-    private void classifyBlock(TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyBlock(HeadlineClassificationService headlineClassificationService,
+                               TextPageBlock textBlock,
+                               ClassificationPage page,
+                               ClassificationDocument document,
+                               List<Float> headlineFontSizes) {

        log.debug("headlineFontSizes: {}", headlineFontSizes);
        var bodyTextFrame = page.getBodyTextFrame();

-        Matcher matcher = pattern.matcher(textBlock.toString());
-        Matcher matcher2 = pattern2.matcher(textBlock.toString());
-        Matcher matcher3 = pattern3.matcher(textBlock.toString());
+        Matcher headlineWithIdentifierMatcher = HEADLINE_WITH_IDENTIFER_PATTERN.matcher(textBlock.toString());
+        Matcher atLeast3Matcher = AT_LEAST_3_PATTERN.matcher(textBlock.toString());
+        Matcher headlineWithSlashesMatcher = HEADLINE_PATTTERN_WITH_SLASHES.matcher(textBlock.toString());

        if (textBlock.getClassification() != null && textBlock.getClassification().isHeadline()) {
            headlineClassificationService.setLastHeadlineFromOutline(textBlock);
@@ -71,14 +77,10 @@ public class DocuMineClassificationService {
            textBlock.setClassification(PageBlockType.OTHER);
            return;
        }
-        if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.HEADER)
-            || (PositionUtils.isOverBodyTextFrame(bodyTextFrame,
-                                                  textBlock,
-                                                  page.getRotation()) && (document.getFontSizeCounter().getMostPopular()
-                                                                          == null
-                                                                          || textBlock.getHighestFontSize()
-                                                                             <= document.getFontSizeCounter()
-                                                                                     .getMostPopular()))) {
+        if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.HEADER) //
+            || (PositionUtils.isOverBodyTextFrame(bodyTextFrame, textBlock, page.getRotation()) //
+                && (document.getFontSizeCounter().getMostPopular() == null //
+                    || textBlock.getHighestFontSize() <= document.getFontSizeCounter().getMostPopular()))) {
            textBlock.setClassification(PageBlockType.HEADER);

        } else if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.FOOTER)
@@ -102,19 +104,25 @@ public class DocuMineClassificationService {
                       || textBlock.getMostPopularWordFontSize() > document.getFontSizeCounter().getMostPopular())
                   && PositionUtils.getApproxLineCount(textBlock) < 5.9

-                   && (textBlock.getMostPopularWordStyle().contains("bold") && Character.isDigit(textBlock.toString().charAt(0)) && matcher2.reset().find() && !textBlock.toString()
-                .contains(":")
-                       || textBlock.toString().equals(textBlock.toString().toUpperCase(Locale.ROOT)) && matcher2.reset().find() && !textBlock.toString().contains(":")
-                       || textBlock.toString().startsWith("APPENDIX")
-                       || textBlock.toString().startsWith("FIGURE")
+                   && (textBlock.getMostPopularWordStyle().contains("bold")
+                       && Character.isDigit(textBlock.toString().charAt(0))
+                       && atLeast3Matcher.reset().find()
+                       && !textBlock.toString().contains(":") //
+                       || textBlock.toString().equals(textBlock.toString().toUpperCase(Locale.ROOT)) && atLeast3Matcher.reset().find() && !textBlock.toString().contains(":") //
+                       || textBlock.toString().startsWith("APPENDIX") //
+                       || textBlock.toString().startsWith("FIGURE") //
+                       || textBlock.toString().startsWith("Continued TABLE") //
                       || textBlock.toString().startsWith("TABLE"))
                   && !textBlock.toString().endsWith(":")
-                   && matcher2.reset().find()) {
+                   && atLeast3Matcher.reset().find()) {
            PageBlockType headlineType = PageBlockType.getHeadlineType(1);
            headlineClassificationService.classifyHeadline(textBlock, headlineType);
            document.setHeadlines(true);

-        } else if (matcher.reset().find() && PositionUtils.getApproxLineCount(textBlock) < 2.9 && matcher2.reset().find() && !matcher3.reset().matches()) {
+        } else if (headlineWithIdentifierMatcher.reset().find()
+                   && PositionUtils.getApproxLineCount(textBlock) < 2.9
+                   && atLeast3Matcher.reset().find()
+                   && !headlineWithSlashesMatcher.reset().matches()) {
            PageBlockType headlineType = PageBlockType.getHeadlineType(2);
            headlineClassificationService.classifyHeadline(textBlock, headlineType);
            document.setHeadlines(true);
@@ -2,15 +2,12 @@ package com.knecon.fforesight.service.layoutparser.processor.services.classifica

 import static com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType.getHeadlineNumber;

-import org.springframework.stereotype.Service;
-
 import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;

 import lombok.Getter;
 import lombok.Setter;

-@Service
@Getter
@Setter
 public class HeadlineClassificationService {
@@ -19,13 +16,6 @@ public class HeadlineClassificationService {
    PageBlockType originalClassifiedBlockType;
    TextPageBlock lastHeadlineFromOutline;

-    public void resetContext() {
-        setLastHeadline(null);
-        setOriginalClassifiedBlockType(null);
-        setLastHeadlineFromOutline(null);
-    }
-
-
    public void setLastHeadlineFromOutline(TextPageBlock lastHeadlineFromOutline) {

        this.lastHeadlineFromOutline = lastHeadlineFromOutline;
@@ -22,8 +22,6 @@ import lombok.extern.slf4j.Slf4j;
@RequiredArgsConstructor
 public class RedactManagerClassificationService {

-    private final HeadlineClassificationService headlineClassificationService;
-

    public void classifyDocument(ClassificationDocument document) {

@@ -31,25 +29,25 @@ public class RedactManagerClassificationService {

        log.debug("Document FontSize counters are: {}", document.getFontSizeCounter().getCountPerValue());

-        headlineClassificationService.resetContext();
+        HeadlineClassificationService headlineClassificationService = new HeadlineClassificationService();

        for (ClassificationPage page : document.getPages()) {
-            classifyPage(page, document, headlineFontSizes);
+            classifyPage(headlineClassificationService, page, document, headlineFontSizes);
        }
    }


-    private void classifyPage(ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyPage(HeadlineClassificationService headlineClassificationService, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {

        for (AbstractPageBlock textBlock : page.getTextBlocks()) {
            if (textBlock instanceof TextPageBlock) {
-                classifyBlock((TextPageBlock) textBlock, page, document, headlineFontSizes);
+                classifyBlock(headlineClassificationService, (TextPageBlock) textBlock, page, document, headlineFontSizes);
            }
        }
    }


-    private void classifyBlock(TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
+    private void classifyBlock(HeadlineClassificationService headlineClassificationService, TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {

        var bodyTextFrame = page.getBodyTextFrame();

@@ -22,6 +22,7 @@ import com.knecon.fforesight.service.layoutparser.processor.model.Classification
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationHeader;
 import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationPage;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.AbstractSemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Footer;
@@ -31,7 +32,6 @@ import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.He
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.AbstractSemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.AtomicTextBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.image.ClassifiedImage;
 import com.knecon.fforesight.service.layoutparser.processor.model.outline.TableOfContentItem;
@@ -45,15 +45,17 @@ import lombok.Builder;
 import lombok.Getter;
 import lombok.experimental.FieldDefaults;
 import lombok.experimental.UtilityClass;
+import lombok.extern.slf4j.Slf4j;

@UtilityClass
+@Slf4j
 public class DocumentGraphFactory {

    public Document buildDocumentGraph(LayoutParsingType layoutParsingType, ClassificationDocument document) {

        Document documentGraph = new Document();

-        documentGraph.setVisualizations(document.getVisualizations());
+        documentGraph.setLayoutDebugLayer(document.getLayoutDebugLayer());

        Context context = new Context(documentGraph);

@@ -122,6 +124,7 @@ public class DocumentGraphFactory {
        List<Integer> treeId = context.documentTree.createNewChildEntryAndReturnId(parentNode, node);
        node.setLeafTextBlock(textBlock);
        node.setTreeId(treeId);
+        node.getEngines().addAll(originalTextBlock.getEngines());
    }


@@ -144,6 +147,7 @@ public class DocumentGraphFactory {
                .position(position)
                .transparent(image.isHasTransparency())
                .page(page)
+                .representationHash(image.getRepresentation())
                .documentTree(context.getDocumentTree())
                .build();
        page.getMainBody().add(imageNode);
@@ -276,7 +280,8 @@ public class DocumentGraphFactory {
            return pages.keySet()
                    .stream()
                    .filter(page -> page.getNumber() == pageIndex)
-                    .findFirst().orElseThrow(() -> new NoSuchElementException(format("ClassificationPage with number %d not found", pageIndex)));
+                    .findFirst()
+                    .orElseThrow(() -> new NoSuchElementException(format("ClassificationPage with number %d not found", pageIndex)));
        }

    }
@@ -4,7 +4,7 @@ import java.awt.geom.Rectangle2D;
 import java.util.Collections;
 import java.util.List;

-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;

 import lombok.AccessLevel;
 import lombok.Builder;
@@ -19,8 +19,8 @@ public class SearchTextWithTextPositionDto {
    String searchText;
    List<Integer> lineBreaks;
    List<Integer> stringIdxToPositionIdx;
-    List<Boundary> boldTextBoundaries;
-    List<Boundary> italicTextBoundaries;
+    List<TextRange> boldTextBoundaries;
+    List<TextRange> italicTextBoundaries;
    List<Rectangle2D> positions;


@@ -9,7 +9,7 @@ import java.util.List;
 import java.util.Locale;
 import java.util.Objects;

-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.RedTextPosition;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextDirection;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
@@ -26,6 +26,7 @@ public class SearchTextWithTextPositionFactory {
    // This is why, we need to initialize this to < -2, otherwise, if the very first symbol is a \n we would detect a hyphen linebreak that isn't there.
    // Also, Integer.MIN_VALUE is a bad idea due to potential overflow during arithmetic operations. This is why the default should be -3.
    public final int MAX_HYPHEN_LINEBREAK_DISTANCE = 3;
+    public static final double LINEBREAK_DELTA_TOLERANCE = 1.05;


    public SearchTextWithTextPositionDto buildSearchTextToTextPositionDto(List<TextPositionSequence> sequences) {
@@ -118,23 +119,23 @@ public class SearchTextWithTextPositionFactory {
    }


-    private static List<Boundary> mergeToBoundaries(List<Integer> integers) {
+    private static List<TextRange> mergeToBoundaries(List<Integer> integers) {

        if (integers.isEmpty()) {
            return Collections.emptyList();
        }
-        List<Boundary> boundaries = new LinkedList<>();
+        List<TextRange> boundaries = new LinkedList<>();
        int start = integers.get(0);
        int end = integers.get(0) + 1;
        for (int current : integers) {
            if (current > end + 1) {
-                boundaries.add(new Boundary(start, end));
+                boundaries.add(new TextRange(start, end));
                start = current;
            }
            end = current + 1;
        }
        if (boundaries.isEmpty()) {
-            boundaries.add(new Boundary(start, end));
+            boundaries.add(new TextRange(start, end));
        }
        return boundaries;
    }
@@ -160,8 +161,8 @@ public class SearchTextWithTextPositionFactory {
            return false;
        }

-        double deltaY = Math.abs(currentPosition.getYDirAdj() - previousPosition.getYDirAdj());
-        return deltaY >= currentPosition.getHeightDir();
+        double deltaY = Math.abs(currentPosition.getYDirAdj() - previousPosition.getYDirAdj()) * LINEBREAK_DELTA_TOLERANCE;
+        return deltaY >= currentPosition.getHeightDir() || deltaY >= previousPosition.getHeightDir();
    }


@@ -55,7 +55,8 @@ public class SectionNodeFactory {
                .collect(groupingBy(AbstractPageBlock::getPage));

        AbstractSemanticNode section;
-        if (isLeaf) {
+        boolean containsTablesAndTextBlocks = containsTablesAndTextBlocks(pageBlocks);
+        if (isLeaf && !containsTablesAndTextBlocks) {
            section = Section.builder().documentTree(context.getDocumentTree()).build();
        } else {
            section = SuperSection.builder().documentTree(context.getDocumentTree()).build();
@@ -68,7 +69,7 @@ public class SectionNodeFactory {
        section.setTreeId(getTreeId(parentNode, context, section));

        addFirstHeadlineDirectlyToSection(layoutParsingType, pageBlocks, context, section, document);
-        if (containsTablesAndTextBlocks(pageBlocks)) {
+        if (containsTablesAndTextBlocks) {
            splitPageBlocksIntoSubSections(pageBlocks).forEach(subSectionPageBlocks -> addSection(layoutParsingType,
                                                                                                  section,
                                                                                                  true,
@@ -116,8 +116,8 @@ public class DocumentDataMapper {
                .page(atomicTextBlock.getPage().getNumber().longValue())
                .searchText(atomicTextBlock.getSearchText())
                .numberOnPage(atomicTextBlock.getNumberOnPage())
-                .start(atomicTextBlock.getBoundary().start())
-                .end(atomicTextBlock.getBoundary().end())
+                .start(atomicTextBlock.getTextRange().start())
+                .end(atomicTextBlock.getTextRange().end())
                .lineBreaks(toPrimitiveIntArray(atomicTextBlock.getLineBreaks()))
                .build();
    }
@@ -24,6 +24,7 @@ public class PropertiesMapper {
        properties.put(DocumentStructure.ImageProperties.TRANSPARENT, String.valueOf(image.isTransparent()));
        properties.put(DocumentStructure.ImageProperties.POSITION, toString(image.getPosition()));
        properties.put(DocumentStructure.ImageProperties.ID, image.getId());
+        properties.put(DocumentStructure.ImageProperties.REPRESENTATION_HASH, image.getRepresentationHash());
        return properties;
    }

@@ -13,7 +13,7 @@ import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.Researc
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.RowData;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.StructureObject;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.taas.TableData;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
@@ -82,15 +82,15 @@ public class TaasDocumentDataMapper {
    }


-    private static Range toRange(Boundary boundary) {
+    private static Range toRange(TextRange textRange) {

-        return new Range(boundary.start(), boundary.end());
+        return new Range(textRange.start(), textRange.end());
    }


-    private static List<Range> toRange(List<Boundary> boundary) {
+    private static List<Range> toRange(List<TextRange> textRange) {

-        return boundary.stream().map(TaasDocumentDataMapper::toRange).toList();
+        return textRange.stream().map(TaasDocumentDataMapper::toRange).toList();
    }


@@ -99,6 +99,7 @@ public class TaasDocumentDataMapper {
        Page page = semanticNode.getFirstPage();
        Rectangle2D bBox = semanticNode.getBBox().get(page);
        return StructureObject.builder()
+                .treeId(semanticNode.getTreeId())
                .structureObjectNumber(structureObjectNumber)
                .boundingBox(toFloatArray(bBox))
                .stringOffset(semanticNode.getBoundary().start())
@@ -114,6 +115,7 @@ public class TaasDocumentDataMapper {
        Page page = table.getFirstPage();
        Rectangle2D bBox = table.getBBox().get(page);
        return StructureObject.builder()
+                .treeId(table.getTreeId())
                .structureObjectNumber(structureObjectNumber)
                .boundingBox(toFloatArray(bBox))
                .stringOffset(table.getBoundary().start())
@@ -82,6 +82,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {

    private int pageRotation;
    private PDRectangle pageSize;
+    private Matrix translateMatrix;
    private final GlyphList glyphList;
    private final Map<COSDictionary, Float> fontHeightMap = new WeakHashMap<COSDictionary, Float>();

@@ -133,6 +134,12 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
        this.pageRotation = page.getRotation();
        this.pageSize = page.getCropBox();

+        if (pageSize.getLowerLeftX() == 0 && pageSize.getLowerLeftY() == 0) {
+            translateMatrix = null;
+        } else {
+            // translation matrix for cropbox
+            translateMatrix = Matrix.getTranslateInstance(-pageSize.getLowerLeftX(), -pageSize.getLowerLeftY());
+        }
        super.processPage(page);
    }

@@ -257,13 +264,22 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
                return;
            }
        }
+        // adjust for cropbox if needed
+        Matrix translatedTextRenderingMatrix;
+        if (translateMatrix == null) {
+            translatedTextRenderingMatrix = textRenderingMatrix;
+        } else {
+            translatedTextRenderingMatrix = Matrix.concatenate(translateMatrix, textRenderingMatrix);
+            nextX -= pageSize.getLowerLeftX();
+            nextY -= pageSize.getLowerLeftY();
+        }

        // This is a hack for unicode letter with 2 chars e.g. RA see unicodeProblem.pdf
        if (unicodeMapping.length() == 2) {
            processTextPosition(new TextPosition(pageRotation,
                                                 pageSize.getWidth(),
                                                 pageSize.getHeight(),
-                                                 textRenderingMatrix,
+                                                 translatedTextRenderingMatrix,
                                                 nextX,
                                                 nextY,
                                                 Math.abs(dyDisplay),
@@ -277,7 +293,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
            processTextPosition(new TextPosition(pageRotation,
                                                 pageSize.getWidth(),
                                                 pageSize.getHeight(),
-                                                 textRenderingMatrix,
+                                                 translatedTextRenderingMatrix,
                                                 nextX,
                                                 nextY,
                                                 Math.abs(dyDisplay),
@@ -293,7 +309,7 @@ public class LegacyPDFStreamEngine extends PDFStreamEngine {
            processTextPosition(new TextPosition(pageRotation,
                                                 pageSize.getWidth(),
                                                 pageSize.getHeight(),
-                                                 textRenderingMatrix,
+                                                 translatedTextRenderingMatrix,
                                                 nextX,
                                                 nextY,
                                                 Math.abs(dyDisplay),
@@ -1,39 +1,17 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.visualization;

-import java.awt.Color;
-import java.awt.geom.Line2D;
-import java.awt.geom.Point2D;
-import java.awt.geom.Rectangle2D;
-import java.awt.geom.RectangularShape;
 import java.io.File;
-import java.util.ArrayList;
-import java.util.Collections;
-import java.util.Comparator;
 import java.util.List;
-import java.util.Map;
-import java.util.Optional;
-import java.util.stream.Collectors;
-import java.util.stream.Stream;

 import org.springframework.stereotype.Service;

-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
-import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
-import com.knecon.fforesight.service.viewerdoc.ContentStreams;
-import com.knecon.fforesight.service.viewerdoc.model.ColoredLine;
-import com.knecon.fforesight.service.viewerdoc.model.ColoredRectangle;
-import com.knecon.fforesight.service.viewerdoc.model.LayoutGrid;
-import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
-import com.knecon.fforesight.service.viewerdoc.model.Standard14EmbeddableFont;
-import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+import com.knecon.fforesight.service.layoutparser.processor.visualization.LayoutGrid;
+import com.knecon.fforesight.service.viewerdoc.service.PDFTronViewerDocumentService;

 import io.micrometer.observation.annotation.Observed;
 import lombok.AccessLevel;
@@ -46,359 +24,41 @@ import lombok.experimental.FieldDefaults;
@FieldDefaults(level = AccessLevel.PRIVATE, makeFinal = true)
 public class LayoutGridService {

-    ViewerDocumentService viewerDocumentService;
-
-    static float FONT_SIZE = 10f;
-    static float LINE_WIDTH = 1f;
-    static Standard14EmbeddableFont FONT = Standard14EmbeddableFont.helvetica();
-
-    static Color INNER_LINES_COLOR = new Color(255, 175, 175);
-    static Color PARAGRAPH_COLOR = new Color(70, 130, 180);
-
-    static Color DUPLICATE_PARAGRAPH_COLOR = new Color(70, 180, 101);
-    static Color TABLE_COLOR = new Color(102, 205, 170);
-    static Color SECTION_COLOR = new Color(50, 50, 50);
-    static Color HEADLINE_COLOR = new Color(162, 56, 56);
-    static Color HEADER_COLOR = new Color(171, 131, 6);
-    static Color IMAGE_COLOR = new Color(253, 63, 146);
+    PDFTronViewerDocumentService viewerDocumentService;


    @SneakyThrows
    @Observed(name = "ViewerDocumentService", contextualName = "create-viewer-document")
    public void addLayoutGrid(File originFile, Document document, File destinationFile, boolean layerVisibilityDefaultValue, boolean writeVisualLayoutParsingGrid) {

-        List<Visualizations> allVisualizations;
-        Visualizations layoutGrid = this.addLayoutGrid(document, layerVisibilityDefaultValue, false);
-        if (writeVisualLayoutParsingGrid) {
-            Visualizations visualLayoutGrid = this.addLayoutGrid(document, layerVisibilityDefaultValue, true);
-            allVisualizations = Stream.concat(Stream.of(layoutGrid, visualLayoutGrid), document.getVisualizations().streamAll())
-                    .toList();
+        LayoutGrid layoutGrid = createLayoutGrid(document);
+        layoutGrid.setVisibleByDefault(layerVisibilityDefaultValue);
+//        Visualizations visualLayoutGrid = this.addLayoutGrid(document, layerVisibilityDefaultValue, true);
+        if (document.getLayoutDebugLayer().isActive()) {
+            viewerDocumentService.addLayerGroups(originFile, destinationFile, List.of(layoutGrid, document.getLayoutDebugLayer()));
        } else {
-            allVisualizations = Stream.concat(Stream.of(layoutGrid), document.getVisualizations().streamAll())
-                    .toList();
+            viewerDocumentService.addLayerGroups(originFile, destinationFile, List.of(layoutGrid));
+
        }
-
-        viewerDocumentService.addVisualizationsOnPage(originFile, destinationFile, allVisualizations);
    }


-    @SneakyThrows
-    @Observed(name = "ViewerDocumentService", contextualName = "create-viewer-document")
-    public Visualizations addLayoutGrid(Document document, boolean layerVisibilityDefaultValue, boolean visualParsingGrid) {
+    private LayoutGrid createLayoutGrid(Document document) {

-        LayoutGrid layoutGrid = createLayoutGrid(document, visualParsingGrid);
-
-        return Visualizations.builder()
-                .layer(visualParsingGrid ? ContentStreams.KNECON_VISUAL_PARSING : ContentStreams.KNECON_LAYOUT)
-                .visualizationsOnPages(layoutGrid.getVisualizationsPerPages())
-                .layerVisibilityDefaultValue(layerVisibilityDefaultValue)
-                .build();
-    }
-
-
-    private LayoutGrid createLayoutGrid(Document document, boolean visualParsingGrid) {
-
-        LayoutGrid layoutGrid = new LayoutGrid(document.getNumberOfPages());
+        LayoutGrid layoutGrid = new LayoutGrid();
        document.streamAllSubNodes()
-                .filter(node -> (node.getEngines().contains(LayoutEngine.AI) && visualParsingGrid) || (node.getEngines().contains(LayoutEngine.ALGORITHM) && !visualParsingGrid))
+                .peek(layoutGrid::addTreeId)
                .forEach(semanticNode -> {
-                    Color color = switch (semanticNode.getType()) {
-                        case PARAGRAPH -> PARAGRAPH_COLOR;
-                        case TABLE -> TABLE_COLOR;
-                        case SECTION, SUPER_SECTION -> SECTION_COLOR;
-                        case HEADLINE -> HEADLINE_COLOR;
-                        case HEADER, FOOTER -> HEADER_COLOR;
-                        case IMAGE -> IMAGE_COLOR;
-                        default -> null;
-                    };
-
-                    if (semanticNode instanceof DuplicatedParagraph) {
-                        color = DUPLICATE_PARAGRAPH_COLOR;
-                    }
-
-                    if (isNotSectionOrTableCellOrDocument(semanticNode)) {
-                        addAsRectangle(semanticNode, layoutGrid, color);
-                    }
-                    if (semanticNode.getType().equals(NodeType.SECTION) || semanticNode.getType().equals(NodeType.SUPER_SECTION)) {
-                        addSection(semanticNode, layoutGrid, color);
-                    }
-                    if (semanticNode.getType().equals(NodeType.TABLE)) {
-                        Table table = (Table) semanticNode;
-                        addInnerTableLines(table, layoutGrid);
+                    switch (semanticNode.getType()) {
+                        case SECTION, SUPER_SECTION -> layoutGrid.addSection(semanticNode);
+                        case HEADLINE -> layoutGrid.addHeadline((Headline) semanticNode);
+                        case PARAGRAPH -> layoutGrid.addParagraph((Paragraph) semanticNode);
+                        case TABLE -> layoutGrid.addTable((Table) semanticNode);
+                        case IMAGE -> layoutGrid.addImage((Image) semanticNode);
+                        case HEADER, FOOTER -> layoutGrid.addHeaderOrFooter(semanticNode);
                    }
                });
        return layoutGrid;
    }

-
-    private void addInnerTableLines(Table table, LayoutGrid layoutGrid) {
-
-        if (table.getNumberOfCols() < 1 || table.getNumberOfRows() < 1) {
-            return;
-        }
-        for (Page page : table.getPages()) {
-
-            Optional<Integer> optionalFirstRowOnPage = table.streamCol(0)
-                    .filter(tableCell -> tableCell.isOnPage(page.getNumber()))
-                    .map(TableCell::getRow)
-                    .findFirst();
-            if (optionalFirstRowOnPage.isEmpty()) {
-                continue;
-            }
-            int firstRowOnPage = optionalFirstRowOnPage.get();
-
-            Stream<Double> xStream = switch (page.getRotation()) {
-                case 90 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMinX);
-                case 180 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMaxX);
-                case 270 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMaxX);
-                default -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMinX);
-            };
-            List<Double> xs = xStream.collect(Collectors.toList());
-            xs.remove(0);
-
-            Stream<Double> yStream = switch (page.getRotation()) {
-                case 90 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMinY);
-                case 180 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMinY);
-                case 270 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMaxY);
-                default -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMaxY);
-            };
-            List<Double> ys = yStream.collect(Collectors.toList());
-            ys.remove(0);
-
-            Rectangle2D tableBBox = table.getBBox()
-                    .get(page);
-            List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages()
-                    .get(page.getNumber() - 1).getColoredLines();
-            xs.forEach(x -> {
-                Line2D line = new Line2D.Double(new Point2D.Double(x, tableBBox.getMaxY()), new Point2D.Double(x, tableBBox.getMinY()));
-                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
-            });
-            ys.forEach(y -> {
-                Line2D line = new Line2D.Double(new Point2D.Double(tableBBox.getMinX(), y), new Point2D.Double(tableBBox.getMaxX(), y));
-                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
-            });
-        }
-    }
-
-
-    private static Stream<Rectangle2D> streamBBoxOfCellsOnPage(Stream<TableCell> table, Page page) {
-
-        return table.filter(tableCell -> tableCell.isOnPage(page.getNumber()))
-                .map(TableCell::getBBox)
-                .map(bBoxMap -> bBoxMap.get(page));
-    }
-
-
-    private void addSection(SemanticNode semanticNode, LayoutGrid layoutGrid, Color color) {
-
-        Map<Page, Rectangle2D> bBoxMap = semanticNode.getBBox();
-        List<SemanticNode> subSections = semanticNode.streamAllSubNodesOfType(NodeType.SECTION)
-                .toList();
-        float lineWidthModifier = (subSections.stream()
-                                           .map(node -> node.getTreeId().size())
-                                           .max(Integer::compareTo)
-                                           .orElse(semanticNode.getTreeId().size()) - semanticNode.getTreeId().size());
-
-        Page firstPage = semanticNode.getFirstPage();
-        String treeIdString = buildTreeIdString(semanticNode);
-        if (!subSections.isEmpty()) {
-            addPlacedText(firstPage, bBoxMap.get(firstPage), treeIdString, layoutGrid);
-        } else {
-            bBoxMap.forEach(((page, textBBox) -> addPlacedText(page, textBBox, treeIdString, layoutGrid)));
-        }
-        if (bBoxMap.values().size() == 1) {
-            Rectangle2D r = RectangleTransformations.pad(bBoxMap.get(firstPage), LINE_WIDTH * (1 + lineWidthModifier), LINE_WIDTH * (1 + lineWidthModifier));
-            List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages()
-                    .get(firstPage.getNumber() - 1).getColoredLines();
-            List<Line2D> lines = createLinesFromRectangle(r, firstPage.getRotation());
-            // add string to top line
-            var firstLine = lines.remove(0);
-            coloredLines.add(new ColoredLine(firstLine, color, LINE_WIDTH));
-            for (Line2D line : lines) {
-                coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
-            }
-            return;
-        }
-        List<Page> pagesInOrder = bBoxMap.keySet()
-                .stream()
-                .sorted(Comparator.comparingInt(Page::getNumber))
-                .collect(Collectors.toList());
-        pagesInOrder.remove(0);
-        addLinesForFirstPageOfSection(semanticNode, color, firstPage, layoutGrid, lineWidthModifier);
-        var lastPage = pagesInOrder.remove(pagesInOrder.size() - 1);
-        addLinesForLastPageOfSection(semanticNode, color, lastPage, layoutGrid, lineWidthModifier);
-        for (Page middlePage : pagesInOrder) {
-            addLinesForMiddlePageOfSection(semanticNode, color, middlePage, layoutGrid, lineWidthModifier);
-        }
-    }
-
-
-    @SneakyThrows
-    private void addPlacedText(Page page, Rectangle2D textBBox, String s, LayoutGrid layoutGrid) {
-
-        // translates text, such that its right edge is a bit to the left of the drawn box
-        float translationAmount = ((FONT.getStringWidth(s) / 1000) * FONT_SIZE + (2 * LINE_WIDTH) + 4);
-
-        Point2D upperLeftCorner;
-        Point2D translationVector;
-        switch (page.getRotation()) {
-            case 90 -> {
-                upperLeftCorner = new Point2D.Double(textBBox.getMinX(), textBBox.getMinY());
-                translationVector = new Point2D.Double(FONT_SIZE, -translationAmount);
-            }
-            case 180 -> {
-                upperLeftCorner = new Point2D.Double(textBBox.getMaxX(), textBBox.getMinY());
-                translationVector = new Point2D.Double(translationAmount, FONT_SIZE);
-            }
-            case 270 -> {
-                upperLeftCorner = new Point2D.Double(textBBox.getMaxX(), textBBox.getMaxY());
-                translationVector = new Point2D.Double(-FONT_SIZE, translationAmount);
-            }
-            default -> {
-                upperLeftCorner = new Point2D.Double(textBBox.getMinX(), textBBox.getMaxY());
-                translationVector = new Point2D.Double(-translationAmount, -FONT_SIZE);
-            }
-        }
-
-        upperLeftCorner = add(upperLeftCorner, translationVector);
-
-        var placedTexts = layoutGrid.getVisualizationsPerPages()
-                .get(page.getNumber() - 1).getPlacedTexts();
-        placedTexts.add(PlacedText.textFacingUp(s, upperLeftCorner, FONT_SIZE, Color.BLACK, FONT));
-    }
-
-
-    private void addLinesForMiddlePageOfSection(SemanticNode semanticNode, Color color, Page middlePage, LayoutGrid layoutGrid, float lineWidthModifier) {
-
-        List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages()
-                .get(middlePage.getNumber() - 1).getColoredLines();
-        Rectangle2D r = RectangleTransformations.pad(semanticNode.getBBox()
-                                                             .get(middlePage), LINE_WIDTH * (1 + lineWidthModifier), LINE_WIDTH * (1 + lineWidthModifier));
-        var midPageLines = createLinesFromRectangle(r, middlePage.getRotation());
-        // remove top line
-        midPageLines.remove(0);
-        // remove top line
-        midPageLines.remove(1);
-        // add string to left line
-        var leftLine = midPageLines.remove(1);
-        coloredLines.add(new ColoredLine(leftLine, color, LINE_WIDTH));
-        for (Line2D line : midPageLines) {
-            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
-        }
-    }
-
-
-    private void addLinesForLastPageOfSection(SemanticNode semanticNode, Color color, Page lastPage, LayoutGrid layoutGrid, float lineWidthModifier) {
-
-        List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages()
-                .get(lastPage.getNumber() - 1).getColoredLines();
-        Rectangle2D r = RectangleTransformations.pad(semanticNode.getBBox()
-                                                             .get(lastPage), LINE_WIDTH * (1 + lineWidthModifier), LINE_WIDTH * (1 + lineWidthModifier));
-        var lastPageLines = createLinesFromRectangle(r, lastPage.getRotation());
-        // remove top line
-        lastPageLines.remove(0);
-        // add string to left line
-        var leftLine = lastPageLines.remove(2);
-        coloredLines.add(new ColoredLine(leftLine, color, LINE_WIDTH));
-        for (Line2D line : lastPageLines) {
-            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
-        }
-    }
-
-
-    private void addLinesForFirstPageOfSection(SemanticNode semanticNode, Color color, Page firstPage, LayoutGrid layoutGrid, float lineWidthModifier) {
-
-        List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages()
-                .get(firstPage.getNumber() - 1).getColoredLines();
-        Rectangle2D r = RectangleTransformations.pad(semanticNode.getBBox()
-                                                             .get(firstPage), LINE_WIDTH * (1 + lineWidthModifier), LINE_WIDTH * (1 + lineWidthModifier));
-        var firstPageLines = createLinesFromRectangle(r, firstPage.getRotation());
-        // remove bottom line
-        firstPageLines.remove(2);
-        // add string to top line
-        var firstLine = firstPageLines.remove(0);
-        coloredLines.add(new ColoredLine(firstLine, color, LINE_WIDTH));
-        for (Line2D line : firstPageLines) {
-            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
-        }
-    }
-
-
-    private String buildTreeIdString(SemanticNode semanticNode) {
-
-        return semanticNode.getTreeId()
-                .stream()
-                .map(Object::toString)
-                .collect(Collectors.joining("."));
-    }
-
-
-    /*
-    A __________________ B
-     |                  |
-     |                  |
-     |                  |
-     |                  |
-    D|__________________| C
-    The returned List are the lines [AB, BC, DC, AD]
-    The List is reordered, such that the order of the returned lines are always as viewed on the page.
-     */
-    private List<Line2D> createLinesFromRectangle(Rectangle2D r, int pageRotation) {
-        // +0.5 to join the lines
-        List<Line2D> lines = new ArrayList<>(4);
-        float lineWidthCorrection = LINE_WIDTH * 0.5f;
-        Point2D.Float a = new Point2D.Float((float) r.getMinX(), (float) r.getMaxY());
-        Point2D.Float a1 = new Point2D.Float((float) r.getMinX() - lineWidthCorrection, (float) r.getMaxY());
-        Point2D.Float b = new Point2D.Float((float) r.getMaxX(), (float) r.getMaxY());
-        Point2D.Float b1 = new Point2D.Float((float) r.getMaxX() + lineWidthCorrection, (float) r.getMaxY());
-        Point2D.Float c = new Point2D.Float((float) r.getMaxX(), (float) r.getMinY());
-        Point2D.Float c1 = new Point2D.Float((float) r.getMaxX() + lineWidthCorrection, (float) r.getMinY());
-        Point2D.Float d = new Point2D.Float((float) r.getMinX(), (float) r.getMinY());
-        Point2D.Float d1 = new Point2D.Float((float) r.getMinX() - lineWidthCorrection, (float) r.getMinY());
-        lines.add(new Line2D.Float(a1, b1));
-        lines.add(new Line2D.Float(b, c));
-        lines.add(new Line2D.Float(d1, c1));
-        lines.add(new Line2D.Float(a, d));
-
-        return switch (pageRotation) {
-            case 90 -> {
-                Collections.rotate(lines, 1);
-                yield lines;
-            }
-            case 180 -> {
-                Collections.rotate(lines, 2);
-                yield lines;
-            }
-            case 270 -> {
-                Collections.rotate(lines, 3);
-                yield lines;
-            }
-
-            default -> lines;
-        };
-    }
-
-
-    private static boolean isNotSectionOrTableCellOrDocument(SemanticNode semanticNode) {
-
-        return !(semanticNode.getType().equals(NodeType.DOCUMENT)
-                 || semanticNode.getType().equals(NodeType.SECTION)
-                 || semanticNode.getType().equals(NodeType.SUPER_SECTION)
-                 || semanticNode.getType().equals(NodeType.TABLE_CELL));
-    }
-
-
-    private void addAsRectangle(SemanticNode semanticNode, LayoutGrid layoutGrid, Color color) {
-
-        semanticNode.getBBox()
-                .forEach((page, textBBox) -> layoutGrid.getVisualizationsPerPages()
-                        .get(page.getNumber() - 1).getColoredRectangles().add(new ColoredRectangle(textBBox, color, LINE_WIDTH)));
-    }
-
-
-    private Point2D add(Point2D a, Point2D b) {
-
-        return new Point2D.Double(a.getX() + b.getX(), a.getY() + b.getY());
-    }
-
 }
@@ -0,0 +1,34 @@
+package com.knecon.fforesight.service.layoutparser.processor.utils;
+
+import java.awt.geom.Rectangle2D;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
+
+import lombok.experimental.UtilityClass;
+
+@UtilityClass
+public class BBoxMergingUtility {
+
+    public Map<Page, Rectangle2D> mergeBBoxes(List<Map<Page, Rectangle2D>> bboxesToMerge) {
+
+        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
+        Set<Page> pages = bboxesToMerge.stream()
+                .flatMap(map -> map.keySet()
+                        .stream())
+                .collect(Collectors.toSet());
+        for (Page page : pages) {
+            Rectangle2D bBoxOnPage = bboxesToMerge.stream()
+                    .filter(childBboxPerPage -> childBboxPerPage.containsKey(page))
+                    .map(childBboxPerPage -> childBboxPerPage.get(page))
+                    .collect(RectangleTransformations.collectBBox());
+            bBoxPerPage.put(page, bBoxOnPage);
+        }
+        return bBoxPerPage;
+    }
+
+}
@@ -51,13 +51,48 @@ public class TextPositionSequenceComparator implements Comparator<TextPositionSe

        double yDifference = Math.abs(pos1YBottom - pos2YBottom);

-        // we will do a simple tolerance comparison
-        if (yDifference < .1 || pos2YBottom >= pos1YTop && pos2YBottom <= pos1YBottom || pos1YBottom >= pos2YTop && pos1YBottom <= pos2YBottom) {
-            return Double.compare(x1, x2);
-        } else if (pos1YBottom < pos2YBottom) {
-            return -1;
-        } else {
-            return 1;
+
+        // Adjust for text rotation
+        switch (pos1.getRotation()) {
+            case 0:
+                // 0 degrees (horizontal, top to bottom and left to right): Sort primarily by y-coordinates from top to bottom (pos1YBottom < pos2YBottom).
+                if (yDifference < .1 || (pos2YBottom >= pos1YTop && pos2YBottom <= pos1YBottom) || (pos1YBottom >= pos2YTop && pos1YBottom <= pos2YBottom)) {
+                    return Double.compare(x1, x2);
+                } else if (pos1YBottom < pos2YBottom) {
+                    return -1;
+                } else {
+                    return 1;
+                }
+            case 90:
+                // 90 degrees (vertical, right to left): Sort by x-coordinates first (x1 > x2), then by y-coordinates from top to bottom (pos1YBottom < pos2YBottom).
+                if (x1 > x2) {
+                    return -1;
+                } else if (x1 < x2) {
+                    return 1;
+                } else {
+                    return Double.compare(pos1YBottom, pos2YBottom);
+                }
+            case 180:
+                // 180 degrees (horizontal, bottom to top and right to left): Sort primarily by y-coordinates from bottom to top (pos1YBottom > pos2YBottom).
+                if (yDifference < .1 || (pos2YBottom >= pos1YTop && pos2YBottom <= pos1YBottom) || (pos1YBottom >= pos2YTop && pos1YBottom <= pos2YBottom)) {
+                    return Double.compare(x2, x1);
+
+                } else if (pos1YBottom > pos2YBottom) {
+                    return -1;
+                } else {
+                    return 1;
+                }
+            case 270:
+                // 270 degrees (vertical, left to right): Sort by x-coordinates in reverse (x2 > x1), then by y-coordinates from bottom to top (pos2YBottom > pos1YBottom).
+                if (x2 > x1) {
+                    return -1;
+                } else if (x2 < x1) {
+                    return 1;
+                } else {
+                    return Double.compare(pos2YBottom, pos1YBottom);
+                }
+            default:
+                throw new RuntimeException("Rotation not supported. Only 0/90/180/270 degree rotation is supported.");
        }
    }

@@ -0,0 +1,111 @@
+package com.knecon.fforesight.service.layoutparser.processor.visualization;
+
+import java.awt.geom.AffineTransform;
+import java.awt.geom.Line2D;
+import java.awt.geom.Point2D;
+import java.awt.geom.Rectangle2D;
+
+import lombok.experimental.UtilityClass;
+
+@UtilityClass
+public class ConnectionLineUtil {
+
+    public static Line2D[] splitRectangleIntoLines(Rectangle2D rect) {
+
+        double x = rect.getX();
+        double y = rect.getY();
+        double width = rect.getWidth();
+        double height = rect.getHeight();
+
+        Line2D[] lines = new Line2D[4];
+        lines[0] = new Line2D.Double(x, y, x + width, y);           // Top
+        lines[1] = new Line2D.Double(x + width, y, x + width, y + height); // Right
+        lines[2] = new Line2D.Double(x + width, y + height, x, y + height); // Bottom
+        lines[3] = new Line2D.Double(x, y + height, x, y);          // Left
+
+        return lines;
+    }
+
+
+    public static Line2D transform(Line2D line2D, AffineTransform affineTransform) {
+
+        var p1 = affineTransform.transform(line2D.getP1(), null);
+        var p2 = affineTransform.transform(line2D.getP2(), null);
+        return new Line2D.Double(p1, p2);
+    }
+
+
+    public static double length(Line2D line2D) {
+
+        return line2D.getP1().distance(line2D.getP2());
+    }
+
+
+    public static Line2D findClosestMidpointLine(Rectangle2D rect1, Rectangle2D rect2) {
+
+        Line2D[] lines1 = splitRectangleIntoLines(rect1);
+        Line2D[] lines2 = splitRectangleIntoLines(rect2);
+
+        Line2D closestLine1 = null;
+        Line2D closestLine2 = null;
+        double minDistance = Double.MAX_VALUE;
+
+        for (Line2D line1 : lines1) {
+            for (Line2D line2 : lines2) {
+                double distance = lineDistance(line1, line2);
+                if (distance < minDistance) {
+                    minDistance = distance;
+                    closestLine1 = line1;
+                    closestLine2 = line2;
+                }
+            }
+        }
+
+        if (closestLine1 == null || closestLine2 == null) {
+            throw new IllegalStateException("Could not find closest lines");
+        }
+
+        Point2D midpoint1 = getMidpoint(closestLine1);
+        Point2D midpoint2 = getMidpoint(closestLine2);
+
+        return new Line2D.Double(midpoint1, midpoint2);
+    }
+
+
+    private static double lineDistance(Line2D line1, Line2D line2) {
+
+        return Math.abs(getMidpoint(line1).distance(getMidpoint(line2)));
+    }
+
+
+    private static Point2D getMidpoint(Line2D line) {
+
+        double x = (line.getX1() + line.getX2()) / 2;
+        double y = (line.getY1() + line.getY2()) / 2;
+        return new Point2D.Double(x, y);
+    }
+
+
+    public static Line2D[] createArrowHead(Line2D line, double arrowLength) {
+
+        Point2D start = line.getP1();
+        Point2D end = line.getP2();
+
+        // Calculate the angle of the line
+        double angle = Math.atan2(end.getY() - start.getY(), end.getX() - start.getX());
+
+        // Calculate the points for the two arrow lines
+        double arrowHeadAngle = Math.PI / 6;
+        double x1 = end.getX() - arrowLength * Math.cos(angle - arrowHeadAngle);
+        double y1 = end.getY() - arrowLength * Math.sin(angle - arrowHeadAngle);
+        double x2 = end.getX() - arrowLength * Math.cos(angle + arrowHeadAngle);
+        double y2 = end.getY() - arrowLength * Math.sin(angle + arrowHeadAngle);
+
+        // Create and return the two arrow lines
+        Line2D arrow1 = new Line2D.Double(end, new Point2D.Double(x1, y1));
+        Line2D arrow2 = new Line2D.Double(end, new Point2D.Double(x2, y2));
+
+        return new Line2D[]{arrow1, arrow2};
+    }
+
+}
@@ -7,7 +7,6 @@ import java.awt.geom.Rectangle2D;
 import java.util.Collection;
 import java.util.List;
 import java.util.concurrent.atomic.AtomicInteger;
-import java.util.stream.Stream;

 import org.apache.pdfbox.pdmodel.documentinterchange.markedcontent.PDMarkedContent;

@@ -21,12 +20,10 @@ import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageB
 import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
 import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;
 import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
-import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.layers.LayoutDebugLayerConfig;
 import com.knecon.fforesight.service.viewerdoc.model.ColoredLine;
 import com.knecon.fforesight.service.viewerdoc.model.ColoredRectangle;
 import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
-import com.knecon.fforesight.service.viewerdoc.model.Standard14EmbeddableFont;
-import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
 import com.knecon.fforesight.service.viewerdoc.model.VisualizationsOnPage;

 import lombok.AccessLevel;
@@ -36,72 +33,15 @@ import lombok.NoArgsConstructor;
 import lombok.Setter;
 import lombok.experimental.FieldDefaults;

+@Setter
@Getter
@NoArgsConstructor
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
-public class LayoutparsingVisualizations {
+public class LayoutDebugLayer extends LayoutDebugLayerConfig {

-    static final Standard14EmbeddableFont FONT = Standard14EmbeddableFont.helvetica();
-
-    static final Color WORDS_COLOR = new Color(68, 84, 147);
-    static final Color LINES_COLOR = new Color(152, 45, 179);
-    static final Color ZONES_COLOR = new Color(131, 38, 38);
-
-    static final Color RULINGS_COLOR = new Color(21, 221, 174);
-    static final Color TABLE_RULINGS_COLOR = new Color(255, 175, 175);
-    static final Color HEADER_RULING_COLOR = new Color(171, 131, 6);
-    static final Color FOOTER_RULING_COLOR = new Color(106, 82, 2);
-    static final Color UNDERLINE_RULING_COLOR = new Color(6, 39, 171);
-    static final Color STRIKETROUGH_RULING_COLOR = new Color(171, 6, 6);
-
-    static final Color CELLS_COLOR = new Color(31, 214, 27);
-
-    static final Color MAIN_BODY_COLOR = new Color(171, 131, 6);
-    static final Color MARKED_CONTENT_COLOR = new Color(171, 131, 6);
-
-    static final List<Color> ROTATING_CHARACTER_COLOR = List.of(new Color(255, 87, 51),
-                                                                new Color(255, 195, 0),
-                                                                new Color(76, 175, 80),
-                                                                new Color(33, 150, 243),
-                                                                new Color(155, 89, 182),
-                                                                new Color(233, 30, 99),
-                                                                new Color(0, 188, 212),
-                                                                new Color(121, 85, 72));
-
-    @Setter
    boolean active;

-    final Visualizations words = Visualizations.builder().layer(ContentStreams.WORDS).build();
-    final Visualizations lines = Visualizations.builder().layer(ContentStreams.LINES).build();
-    final Visualizations zones = Visualizations.builder().layer(ContentStreams.ZONES).build();
-    final Visualizations mainBody = Visualizations.builder().layer(ContentStreams.MAIN_BODY).build();
-    final Visualizations clean_rulings = Visualizations.builder().layer(ContentStreams.CLEAN_RULINGS).build();
-    final Visualizations rulings = Visualizations.builder().layer(ContentStreams.RULINGS).build();
-    final Visualizations cells = Visualizations.builder().layer(ContentStreams.CELLS).build();
-    final Visualizations markedContent = Visualizations.builder().layer(ContentStreams.MARKED_CONTENT).build();
-    final Visualizations neighbours = Visualizations.builder().layer(ContentStreams.NEIGHBOURS).build();
-    final Visualizations characters = Visualizations.builder().layer(ContentStreams.CHARACTERS).build();
-
-
-    public Stream<Visualizations> streamAll() {
-
-        if (!active) {
-            return Stream.empty();
-        }
-        return Stream.of(characters, //
-                         neighbours,//
-                         words, //
-                         lines, //
-                         zones, //
-                         rulings, //
-                         clean_rulings, //
-                         cells, //
-                         mainBody, //
-                         markedContent //
-        );
-    }
-

    public void addTextVisualizations(List<TextPositionSequence> textPositionSequences, int pageNumber) {

@@ -130,6 +70,7 @@ public class LayoutparsingVisualizations {
                                .toList());
    }

+
    public void addRulingVisualization(List<Ruling> rulings, int pageNumber) {

        if (!active) {
@@ -137,8 +78,7 @@ public class LayoutparsingVisualizations {
        }
        VisualizationsOnPage visualizationsOnPage = getOrCreateVisualizationsOnPage(pageNumber, this.rulings);
        visualizationsOnPage.getColoredLines()
-                .addAll(rulings
-                                .stream()
+                .addAll(rulings.stream()
                                .map(ruling -> new ColoredLine(ruling, decideOnRulingColor(ruling), 0.5f))
                                .toList());
    }
@@ -295,16 +235,4 @@ public class LayoutparsingVisualizations {

    }

-
-    private VisualizationsOnPage getOrCreateVisualizationsOnPage(int page, Visualizations visualizations) {
-
-        if (visualizations.getVisualizationsOnPages().containsKey(page - 1)) {
-            return visualizations.getVisualizationsOnPages()
-                    .get(page - 1);
-        }
-        VisualizationsOnPage visualizationsOnPage = VisualizationsOnPage.builder().build();
-        visualizations.getVisualizationsOnPages().put(page - 1, visualizationsOnPage);
-        return visualizationsOnPage;
-    }
-
 }
@@ -0,0 +1,430 @@
+package com.knecon.fforesight.service.layoutparser.processor.visualization;
+
+import java.awt.Color;
+import java.awt.geom.Line2D;
+import java.awt.geom.Point2D;
+import java.awt.geom.Rectangle2D;
+import java.awt.geom.RectangularShape;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.Comparator;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Optional;
+import java.util.stream.Collectors;
+import java.util.stream.Stream;
+
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.DuplicatedParagraph;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Headline;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Paragraph;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SuperSection;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
+import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
+import com.knecon.fforesight.service.viewerdoc.layers.LayoutGridLayerConfig;
+import com.knecon.fforesight.service.viewerdoc.model.ColoredLine;
+import com.knecon.fforesight.service.viewerdoc.model.ColoredRectangle;
+import com.knecon.fforesight.service.viewerdoc.model.FilledRectangle;
+import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+import lombok.AccessLevel;
+import lombok.Getter;
+import lombok.Setter;
+import lombok.SneakyThrows;
+import lombok.experimental.FieldDefaults;
+
+@FieldDefaults(level = AccessLevel.PRIVATE)
+public class LayoutGrid extends LayoutGridLayerConfig {
+
+    @Getter
+    @Setter
+    boolean visibleByDefault;
+
+    final HashMap<RectangleIdentifier, Rectangle2D> rectangleMap = new HashMap<>();
+
+
+    public void addParagraph(Paragraph paragraph) {
+
+        if (paragraph instanceof DuplicatedParagraph) {
+            addAsRectangle(paragraph, paragraphs, DUPLICATE_PARAGRAPH_COLOR);
+        } else {
+            addAsRectangle(paragraph, paragraphs, PARAGRAPH_COLOR);
+        }
+    }
+
+
+    public void addImage(Image image) {
+
+        if (image.isFullPageImage()) {
+            addAsRectangle(image, images, IMAGE_COLOR);
+        } else {
+            addAsRectangle(image, figures, IMAGE_COLOR);
+        }
+    }
+
+
+    public void addHeadline(Headline headline) {
+
+        addAsRectangle(headline, headlines, HEADLINE_COLOR);
+    }
+
+
+    public void addHeaderOrFooter(SemanticNode header) {
+
+        addAsRectangle(header, headerFooter, HEADER_COLOR);
+    }
+
+
+    public void addTreeId(SemanticNode semanticNode) {
+
+        Page page = semanticNode.getFirstPage();
+        addPlacedText(page, semanticNode.getBBox().get(page), semanticNode.getBBox().get(page), buildTreeIdString(semanticNode), 1, treeIds, TREEID_COLOR);
+    }
+
+
+    public void addTable(Table table) {
+
+        addAsRectangle(table, tables, TABLE_COLOR);
+        addInnerTableLines(table);
+        addHeaderCells(table);
+    }
+
+
+    private void addHeaderCells(Table table) {
+
+        table.streamHeaders()
+                .map(TableCell::getBBox)
+                .forEach(map -> map.forEach((page, textBBox) -> getOrCreateVisualizationsOnPage(page.getNumber(), tables).getFilledRectangles()
+                        .add(new FilledRectangle(textBBox, HEADER_CELL_COLOR, 0.1f))));
+    }
+
+
+    public void addSection(SemanticNode section) {
+
+        Map<Page, Rectangle2D> bBoxMap = section.getBBox();
+
+        List<SemanticNode> subSections = section.streamAllSubNodesOfType(NodeType.SECTION)
+                .toList();
+        Integer maxChildDepth = subSections.stream()
+                .map(node -> node.getTreeId().size())
+                .max(Integer::compareTo).orElse(section.getTreeId().size());
+        int ownDepth = section.getTreeId().size();
+
+        Page firstPage = section.getFirstPage();
+        String treeIdString = buildTreeIdString(section);
+
+        if (bBoxMap.values().size() == 1) {
+            handleSinglePage(section, firstPage, bBoxMap.get(firstPage), treeIdString, maxChildDepth, ownDepth);
+            return;
+        }
+        List<Page> pagesInOrder = bBoxMap.keySet()
+                .stream()
+                .sorted(Comparator.comparingInt(Page::getNumber))
+                .collect(Collectors.toList());
+        pagesInOrder.remove(0);
+        handleFirstPageOfSection(section, firstPage, bBoxMap.get(firstPage), treeIdString, maxChildDepth, ownDepth);
+        if (section instanceof SuperSection) {
+            return;
+        }
+        for (Page middlePage : pagesInOrder.subList(0, pagesInOrder.size() - 1)) {
+            handleForMiddlePageOfSection(section, middlePage, bBoxMap.get(middlePage), treeIdString, maxChildDepth, ownDepth);
+        }
+        var lastPage = pagesInOrder.remove(pagesInOrder.size() - 1);
+        handleLastPageOfSection(section, lastPage, bBoxMap.get(lastPage), treeIdString, maxChildDepth, ownDepth);
+    }
+
+
+    private String buildTreeIdString(SemanticNode semanticNode) {
+
+        return semanticNode.getTreeId()
+                .stream()
+                .map(Object::toString)
+                .collect(Collectors.joining("."));
+    }
+
+
+    @SneakyThrows
+    private void addPlacedText(Page page, Rectangle2D textBBox, Rectangle2D highestParentRect, String s, Integer maxChildDepth, Visualizations visualizations, Color color) {
+
+        // translates text, such that its right edge is a bit to the left of the drawn box
+        float translationAmount = ((FONT.getStringWidth(s) / 1000) * FONT_SIZE + LINE_WIDTH + 2 * maxChildDepth);
+
+        Point2D upperLeftCorner;
+        Point2D translationVector;
+        switch (page.getRotation()) {
+            case 90 -> {
+                if (highestParentRect != null) {
+                    upperLeftCorner = new Point2D.Double(highestParentRect.getMinX(), textBBox.getMinY());
+                } else {
+                    upperLeftCorner = new Point2D.Double(textBBox.getMinX(), textBBox.getMinY());
+                }
+                translationVector = new Point2D.Double(FONT_SIZE, -translationAmount);
+            }
+            case 180 -> {
+                if (highestParentRect != null) {
+                    upperLeftCorner = new Point2D.Double(highestParentRect.getMaxX(), textBBox.getMinY());
+                } else {
+                    upperLeftCorner = new Point2D.Double(textBBox.getMaxX(), textBBox.getMinY());
+                }
+                translationVector = new Point2D.Double(translationAmount, FONT_SIZE);
+            }
+            case 270 -> {
+
+                if (highestParentRect != null) {
+                    upperLeftCorner = new Point2D.Double(highestParentRect.getMaxX(), textBBox.getMaxY());
+                } else {
+                    upperLeftCorner = new Point2D.Double(textBBox.getMaxX(), textBBox.getMaxY());
+                }
+                translationVector = new Point2D.Double(-FONT_SIZE, translationAmount);
+            }
+            default -> {
+
+                if (highestParentRect != null) {
+                    upperLeftCorner = new Point2D.Double(highestParentRect.getMinX(), textBBox.getMaxY());
+                } else {
+                    upperLeftCorner = new Point2D.Double(textBBox.getMinX(), textBBox.getMaxY());
+                }
+                translationVector = new Point2D.Double(-translationAmount, -FONT_SIZE);
+            }
+        }
+
+        upperLeftCorner = add(upperLeftCorner, translationVector);
+
+        List<PlacedText> placedTexts = getOrCreateVisualizationsOnPage(page.getNumber(), visualizations).getPlacedTexts();
+
+        PlacedText newText = PlacedText.textFacingUp(s, upperLeftCorner, FONT_SIZE, color, FONT);
+
+        Optional<PlacedText> conflictingText = placedTexts.stream()
+                .filter(pt -> Math.abs(pt.lineStart().getY() - newText.lineStart().getY()) <= FONT_SIZE)
+                .findFirst();
+
+        if (conflictingText.isPresent()) {
+            PlacedText existingText = conflictingText.get();
+            if (newText.text().length() > existingText.text().length()) {
+                placedTexts.remove(existingText);
+                placedTexts.add(newText);
+            }
+        } else {
+            placedTexts.add(newText);
+        }
+
+    }
+
+
+    private void handleSinglePage(SemanticNode semanticNode, Page page, Rectangle2D rectangle2D, String treeIdString, Integer maxChildDepth, Integer ownDepth) {
+
+        RectangleAndLinesResult result = createLinesAndPlaceText(semanticNode, page, rectangle2D, treeIdString, maxChildDepth, ownDepth);
+        // add string to top line
+        var firstLine = result.pageLines().remove(0);
+        result.coloredLines().add(new ColoredLine(firstLine, SECTION_COLOR, LINE_WIDTH));
+        for (Line2D line : result.pageLines()) {
+            result.coloredLines().add(new ColoredLine(line, SECTION_COLOR, LINE_WIDTH));
+        }
+    }
+
+
+    private void handleFirstPageOfSection(SemanticNode semanticNode, Page firstPage, Rectangle2D rectangle2D, String treeIdString, Integer maxChildDepth, Integer ownDepth) {
+
+        RectangleAndLinesResult result = createLinesAndPlaceText(semanticNode, firstPage, rectangle2D, treeIdString, maxChildDepth, ownDepth);
+        // remove bottom line
+        result.pageLines().remove(2);
+        // add string to top line
+        var firstLine = result.pageLines().remove(0);
+        result.coloredLines().add(new ColoredLine(firstLine, SECTION_COLOR, LINE_WIDTH));
+        for (Line2D line : result.pageLines()) {
+            result.coloredLines().add(new ColoredLine(line, SECTION_COLOR, LINE_WIDTH));
+        }
+    }
+
+
+    private void handleForMiddlePageOfSection(SemanticNode semanticNode, Page middlePage, Rectangle2D rectangle2D, String treeIdString, Integer maxChildDepth, Integer ownDepth) {
+
+        RectangleAndLinesResult result = createLinesAndPlaceText(semanticNode, middlePage, rectangle2D, treeIdString, maxChildDepth, ownDepth);
+        // remove top line
+        result.pageLines().remove(0);
+        // remove bottom line
+        result.pageLines().remove(1);
+        // add string to left line
+        var leftLine = result.pageLines().remove(1);
+        result.coloredLines().add(new ColoredLine(leftLine, SECTION_COLOR, LINE_WIDTH));
+        for (Line2D line : result.pageLines()) {
+            result.coloredLines().add(new ColoredLine(line, SECTION_COLOR, LINE_WIDTH));
+        }
+    }
+
+
+    private void handleLastPageOfSection(SemanticNode semanticNode, Page lastPage, Rectangle2D rectangle2D, String treeIdString, Integer maxChildDepth, Integer ownDepth) {
+
+        RectangleAndLinesResult result = createLinesAndPlaceText(semanticNode, lastPage, rectangle2D, treeIdString, maxChildDepth, ownDepth);
+        // remove top line
+        result.pageLines().remove(0);
+        // add string to left line
+        var leftLine = result.pageLines().remove(2);
+        result.coloredLines().add(new ColoredLine(leftLine, SECTION_COLOR, LINE_WIDTH));
+        for (Line2D line : result.pageLines()) {
+            result.coloredLines().add(new ColoredLine(line, SECTION_COLOR, LINE_WIDTH));
+        }
+    }
+
+
+    private RectangleAndLinesResult createLinesAndPlaceText(SemanticNode semanticNode,
+                                                            Page page,
+                                                            Rectangle2D rectangle2D,
+                                                            String treeIdString,
+                                                            Integer maxChildDepth,
+                                                            Integer ownDepth) {
+
+        List<ColoredLine> coloredLines = getOrCreateVisualizationsOnPage(page.getNumber(), sections).getColoredLines();
+        int lineWidthModifier = maxChildDepth - ownDepth;
+        Rectangle2D r = RectangleTransformations.pad(semanticNode.getBBox().get(page), LINE_WIDTH * (1 + lineWidthModifier), LINE_WIDTH * (1 + lineWidthModifier));
+
+        SemanticNode highestParent = semanticNode.getHighestParent();
+        Rectangle2D highestParentRect = rectangleMap.get(new RectangleIdentifier(highestParent.getTreeId(), page.getNumber()));
+        addPlacedText(page, rectangle2D, highestParentRect, treeIdString, maxChildDepth, sections, SECTION_COLOR);
+        var lastPageLines = createLinesFromRectangle(r, page.getRotation());
+
+        if (semanticNode instanceof SuperSection) {
+            rectangleMap.put(new RectangleIdentifier(semanticNode.getTreeId(), page.getNumber()), r);
+        }
+
+        return new RectangleAndLinesResult(coloredLines, r, lastPageLines);
+    }
+
+
+    private void addInnerTableLines(Table table) {
+
+        if (table.getNumberOfCols() < 1 || table.getNumberOfRows() < 1) {
+            return;
+        }
+        for (Page page : table.getPages()) {
+
+            Optional<Integer> optionalFirstRowOnPage = table.streamCol(0)
+                    .filter(tableCell -> tableCell.isOnPage(page.getNumber()))
+                    .map(TableCell::getRow)
+                    .findFirst();
+            if (optionalFirstRowOnPage.isEmpty()) {
+                continue;
+            }
+            int firstRowOnPage = optionalFirstRowOnPage.get();
+
+            Stream<Double> xStream = switch (page.getRotation()) {
+                case 90 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMinX);
+                case 180 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMaxX);
+                case 270 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMaxX);
+                default -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMinX);
+            };
+            List<Double> xs = xStream.collect(Collectors.toList());
+            xs.remove(0);
+
+            Stream<Double> yStream = switch (page.getRotation()) {
+                case 90 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMinY);
+                case 180 -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMinY);
+                case 270 -> streamBBoxOfCellsOnPage(table.streamRow(firstRowOnPage), page).map(RectangularShape::getMaxY);
+                default -> streamBBoxOfCellsOnPage(table.streamCol(0), page).map(RectangularShape::getMaxY);
+            };
+            List<Double> ys = yStream.collect(Collectors.toList());
+            ys.remove(0);
+
+            Rectangle2D tableBBox = table.getBBox().get(page);
+            List<ColoredLine> coloredLines = getOrCreateVisualizationsOnPage(page.getNumber(), tables).getColoredLines();
+
+            xs.forEach(x -> {
+                Line2D line = new Line2D.Double(new Point2D.Double(x, tableBBox.getMaxY()), new Point2D.Double(x, tableBBox.getMinY()));
+                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
+            });
+            ys.forEach(y -> {
+                Line2D line = new Line2D.Double(new Point2D.Double(tableBBox.getMinX(), y), new Point2D.Double(tableBBox.getMaxX(), y));
+                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
+            });
+        }
+    }
+
+
+    private static Stream<Rectangle2D> streamBBoxOfCellsOnPage(Stream<TableCell> table, Page page) {
+
+        return table.filter(tableCell -> tableCell.isOnPage(page.getNumber()))
+                .map(TableCell::getBBox)
+                .map(bBoxMap -> bBoxMap.get(page));
+    }
+
+
+    private void addAsRectangle(SemanticNode semanticNode, Visualizations visualizations, Color color) {
+
+        addAsRectangle(semanticNode.getBBox(), visualizations, color);
+    }
+
+
+    private void addAsRectangle(Map<Page, Rectangle2D> bbox, Visualizations visualizations, Color color) {
+
+        bbox.forEach((page, textBBox) -> getOrCreateVisualizationsOnPage(page.getNumber(), visualizations).getColoredRectangles()
+                .add(new ColoredRectangle(textBBox, color, LINE_WIDTH)));
+    }
+
+
+    private record RectangleAndLinesResult(List<ColoredLine> coloredLines, Rectangle2D rectangle, List<Line2D> pageLines) {
+
+    }
+
+    private record RectangleIdentifier(List<Integer> treeId, Integer pageNumber) {
+
+    }
+
+
+    /*
+    A __________________ B
+     |                  |
+     |                  |
+     |                  |
+     |                  |
+    D|__________________| C
+    The returned List are the lines [AB, BC, DC, AD]
+    The List is reordered, such that the order of the returned lines are always as viewed on the page.
+     */
+    private List<Line2D> createLinesFromRectangle(Rectangle2D r, int pageRotation) {
+        // +0.5 to join the lines
+        List<Line2D> lines = new ArrayList<>(4);
+        float lineWidthCorrection = LINE_WIDTH * 0.5f;
+        Point2D.Float a = new Point2D.Float((float) r.getMinX(), (float) r.getMaxY());
+        Point2D.Float a1 = new Point2D.Float((float) r.getMinX() - lineWidthCorrection, (float) r.getMaxY());
+        Point2D.Float b = new Point2D.Float((float) r.getMaxX(), (float) r.getMaxY());
+        Point2D.Float b1 = new Point2D.Float((float) r.getMaxX() + lineWidthCorrection, (float) r.getMaxY());
+        Point2D.Float c = new Point2D.Float((float) r.getMaxX(), (float) r.getMinY());
+        Point2D.Float c1 = new Point2D.Float((float) r.getMaxX() + lineWidthCorrection, (float) r.getMinY());
+        Point2D.Float d = new Point2D.Float((float) r.getMinX(), (float) r.getMinY());
+        Point2D.Float d1 = new Point2D.Float((float) r.getMinX() - lineWidthCorrection, (float) r.getMinY());
+        lines.add(new Line2D.Float(a1, b1));
+        lines.add(new Line2D.Float(b, c));
+        lines.add(new Line2D.Float(d1, c1));
+        lines.add(new Line2D.Float(a, d));
+
+        return switch (pageRotation) {
+            case 90 -> {
+                Collections.rotate(lines, 1);
+                yield lines;
+            }
+            case 180 -> {
+                Collections.rotate(lines, 2);
+                yield lines;
+            }
+            case 270 -> {
+                Collections.rotate(lines, 3);
+                yield lines;
+            }
+
+            default -> lines;
+        };
+    }
+
+
+    private Point2D add(Point2D a, Point2D b) {
+
+        return new Point2D.Double(a.getX() + b.getX(), a.getY() + b.getY());
+    }
+
+}
@@ -32,11 +32,13 @@ dependencies {
    implementation("com.iqser.red.commons:storage-commons:2.45.0")
    implementation("com.knecon.fforesight:tenant-commons:0.21.0")
    implementation("com.knecon.fforesight:tracing-commons:0.5.0")
+    implementation("com.knecon.fforesight:lifecycle-commons:0.6.0")
    implementation("org.springframework.boot:spring-boot-starter-actuator:${springBootStarterVersion}")
    implementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
    implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
    implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
    implementation("net.logstash.logback:logstash-logback-encoder:7.4")
+    implementation("com.pdftron:PDFNet:10.7.0")

    // for integration testing only
    testImplementation(project(":viewer-doc-processor"))
@@ -6,19 +6,22 @@ import org.springframework.boot.autoconfigure.ImportAutoConfiguration;
 import org.springframework.boot.autoconfigure.SpringBootApplication;
 import org.springframework.boot.autoconfigure.security.servlet.SecurityAutoConfiguration;
 import org.springframework.context.annotation.Bean;
+import org.springframework.context.annotation.EnableAspectJAutoProxy;
 import org.springframework.context.annotation.Import;

 import com.amazonaws.services.s3.model.metrics.MetricsConfiguration;
 import com.iqser.red.storage.commons.StorageAutoConfiguration;
+import com.knecon.fforesight.lifecyclecommons.LifecycleAutoconfiguration;
 import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingServiceProcessorConfiguration;
 import com.knecon.fforesight.tenantcommons.MultiTenancyAutoConfiguration;

 import io.micrometer.observation.ObservationRegistry;
 import io.micrometer.observation.aop.ObservedAspect;

-@ImportAutoConfiguration({MultiTenancyAutoConfiguration.class})
+@ImportAutoConfiguration({MultiTenancyAutoConfiguration.class, LifecycleAutoconfiguration.class})
@Import({MetricsConfiguration.class, StorageAutoConfiguration.class, LayoutParsingServiceProcessorConfiguration.class})
@SpringBootApplication(exclude = {SecurityAutoConfiguration.class, ManagementWebSecurityAutoConfiguration.class})
+@EnableAspectJAutoProxy
 public class Application {

    public static void main(String[] args) {
@@ -0,0 +1,47 @@
+package com.knecon.fforesight.service.layoutparser.server;
+
+import org.springframework.beans.factory.annotation.Value;
+import org.springframework.stereotype.Component;
+
+import com.google.common.base.Strings;
+import com.knecon.fforesight.service.layoutparser.processor.LayoutparserSettings;
+import com.pdftron.pdf.PDFNet;
+
+import jakarta.annotation.PostConstruct;
+import jakarta.annotation.PreDestroy;
+import lombok.RequiredArgsConstructor;
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;
+
+@Slf4j
+@Component
+@RequiredArgsConstructor
+public class PDFNetInitializer {
+
+    @Value("${pdftron.license:}")
+    private String pdftronLicense;
+
+
+    @SneakyThrows
+    @PostConstruct
+    // 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() {
+
+        if (Strings.isNullOrEmpty(pdftronLicense)) {
+            return;
+        }
+        log.info("Initializing Native Libraries");
+        log.info("Setting pdftron license: {}", pdftronLicense);
+        PDFNet.setTempPath("/tmp/pdftron");
+        PDFNet.initialize(pdftronLicense);
+
+    }
+
+
+    @PreDestroy
+    public void terminate() {
+
+        PDFNet.terminate();
+    }
+
+}
@@ -15,6 +15,9 @@ project.version: 1.0-SNAPSHOT
 server:
  port: 8080

+lifecycle:
+  base-package: com.knecon.fforesight.service.layoutparser
+
 spring:
  application:
    name: layoutparser-service
@@ -50,7 +50,7 @@ public class BdrJsonBuildTest extends AbstractTest {
    protected Document buildGraph(File file) {

        return DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.CLARIFYND,
-                layoutParsingPipeline.parseLayout(LayoutParsingType.CLARIFYND,
+                                                       layoutParsingPipeline.parseLayout(LayoutParsingType.CLARIFYND,
                        file,
                        new ImageServiceResponse(),
                        new TableServiceResponse(),
@@ -97,7 +97,7 @@ public class HeadlinesGoldStandardIntegrationTest {
        goldStandardLog.getRedactionLogEntry().forEach(e -> goldStandardHeadlines.add(new Headline(e.getPositions().get(0).getPage(), e.getValue())));

        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.DOCUMINE,
-                layoutParsingPipeline.parseLayout(LayoutParsingType.DOCUMINE,
+                                                                         layoutParsingPipeline.parseLayout(LayoutParsingType.DOCUMINE,
                        pdfFileResource.getFile(),
                        new ImageServiceResponse(),
                        new TableServiceResponse(),
@@ -30,11 +30,11 @@ public class LayoutparserEnd2EndTest extends AbstractTest {
    @Autowired
    private LayoutParsingPipeline layoutParsingPipeline;

-    @Disabled
    @Test
+    @Disabled
    public void testLayoutParserEndToEnd() {

-        String filePath = "files/Minimal Examples/RotateTextWithRulingsTestFile.pdf";
+        String filePath = "files/syngenta/CustomerFiles/Documine/Flora/425_F.1.1.1 - A13617AV - Acute Oral Toxicity Study.pdf";

        runForFile(filePath);
    }
@@ -44,7 +44,7 @@ public class LayoutparserEnd2EndTest extends AbstractTest {
    @SneakyThrows
    public void testLayoutParserEndToEndWithFolder() {

-        String folder = "/home/kschuettler/Dokumente/TestFiles/large number of prod files";
+        String folder = "/Users/maverickstuder/Documents/Fforesight/layoutparser/layoutparser-service/layoutparser-service-server/src/test";
        List<Path> pdfFiles = Files.walk(Path.of(folder))
                .filter(path -> path.getFileName().toString().endsWith(".pdf"))
                .sorted(Comparator.comparing(Path::getFileName))
@@ -82,6 +82,11 @@ public class LayoutparserEnd2EndTest extends AbstractTest {
        assert tmpFile.getParentFile().exists() || tmpFile.getParentFile().mkdirs();

        storageService.downloadTo(TENANT_ID, layoutParsingRequest.viewerDocumentStorageId(), tmpFile);
+
+        tmpFile = new File("/tmp/layoutparserEND2END/" + fileName + "_MARKDOWN.md");
+        assert tmpFile.getParentFile().exists() || tmpFile.getParentFile().mkdirs();
+
+        storageService.downloadTo(TENANT_ID, layoutParsingRequest.documentMarkdownFileStorageId().get(), tmpFile);
    }


@@ -12,6 +12,7 @@ import java.util.Map;
 import java.util.function.Predicate;

 import org.apache.commons.lang3.StringUtils;
+import org.junit.jupiter.api.BeforeEach;
 import org.junit.jupiter.api.Test;
 import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.core.io.ClassPathResource;
@@ -32,18 +33,29 @@ import com.knecon.fforesight.service.layoutparser.processor.python_api.model.tab
 import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
 import com.knecon.fforesight.service.layoutparser.server.utils.AbstractTest;
-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+import com.knecon.fforesight.service.viewerdoc.service.PDFTronViewerDocumentService;

+import jakarta.annotation.PostConstruct;
 import lombok.SneakyThrows;

 public class OutlineDetectionTest extends AbstractTest {

-    ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
+    PDFTronViewerDocumentService viewerDocumentService = new PDFTronViewerDocumentService(null);
    LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);

    @Autowired
    protected LayoutParsingPipeline layoutParsingPipeline;

+    @Autowired
+    PDFNetInitializer pdfNetInitializer;
+
+
+    @BeforeEach
+    public void init() {
+
+        pdfNetInitializer.init();
+    }
+

    @Test
    @SneakyThrows
@@ -60,28 +72,17 @@ public class OutlineDetectionTest extends AbstractTest {

        OutlineObjectTree outlineObjectTree = classificationDocument.getOutlineObjectTree();
        assertEquals(outlineObjectTree.getRootNodes().size(), 8);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(1).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(3).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(4).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(5).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(6).size(), 2);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(7).size(), 3);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(8).size(), 2);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(10).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(11).size(), 4);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(12).size(), 1);
-        assertEquals(outlineObjectTree.getOutlineObjectsPerPage()
-                             .get(13).size(), 2);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(1).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(3).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(4).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(5).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(6).size(), 2);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(7).size(), 3);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(8).size(), 2);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(10).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(11).size(), 4);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(12).size(), 1);
+        assertEquals(outlineObjectTree.getOutlineObjectsPerPage().get(13).size(), 2);
        assertTrue(outlineObjectTree.getOutlineObjectsPerPage().values()
                           .stream()
                           .flatMap(Collection::stream)
@@ -98,29 +99,15 @@ public class OutlineDetectionTest extends AbstractTest {
                             .stream()
                             .map(outlineObjectTreeNode -> sanitizeString(outlineObjectTreeNode.getOutlineObject().getTitle()))
                             .toList());
-        assertEquals(tableOfContents.getMainSections()
-                             .get(5).getChildren().size(), 6);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(7).getChildren().size(), 3);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(8).getChildren().size(), 3);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(8).getChildren()
-                             .get(2).getChildren().size(), 1);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(8).getChildren()
-                             .get(2).getChildren()
-                             .get(0).getChildren().size(), 3);
+        assertEquals(tableOfContents.getMainSections().get(5).getChildren().size(), 6);
+        assertEquals(tableOfContents.getMainSections().get(7).getChildren().size(), 3);
+        assertEquals(tableOfContents.getMainSections().get(8).getChildren().size(), 3);
+        assertEquals(tableOfContents.getMainSections().get(8).getChildren().get(2).getChildren().size(), 1);
+        assertEquals(tableOfContents.getMainSections().get(8).getChildren().get(2).getChildren().get(0).getChildren().size(), 3);

-        assertEquals(tableOfContents.getMainSections()
-                             .get(0).getImages().size(), 1);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(6).getImages().size(), 1);
-        assertEquals(tableOfContents.getMainSections()
-                             .get(8).getChildren()
-                             .get(2).getChildren()
-                             .get(0).getChildren()
-                             .get(2).getImages().size(), 1);
+        assertEquals(tableOfContents.getMainSections().get(0).getImages().size(), 1);
+        assertEquals(tableOfContents.getMainSections().get(6).getImages().size(), 1);
+        assertEquals(tableOfContents.getMainSections().get(8).getChildren().get(2).getChildren().get(0).getChildren().get(2).getImages().size(), 1);

        Document document = buildGraph(fileName, classificationDocument);

@@ -159,17 +146,14 @@ public class OutlineDetectionTest extends AbstractTest {
                             .count(), 3 + 1);
        assertEquals(childrenOfTypeSectionOrSuperSection.get(8).streamChildren()
                             .filter(isSectionOrSuperSection)
-                             .toList()
-                             .get(3).streamChildren()
+                             .toList().get(3).streamChildren()
                             .filter(isSectionOrSuperSection)
                             .count(), 1 + 1);
        assertEquals(childrenOfTypeSectionOrSuperSection.get(8).streamChildren()
                             .filter(isSectionOrSuperSection)
-                             .toList()
-                             .get(3).streamChildren()
+                             .toList().get(3).streamChildren()
                             .filter(isSectionOrSuperSection)
-                             .toList()
-                             .get(1).streamChildren()
+                             .toList().get(1).streamChildren()
                             .filter(isSectionOrSuperSection)
                             .count(), 3 + 1);

@@ -3,11 +3,8 @@ package com.knecon.fforesight.service.layoutparser.server;
 import static org.assertj.core.api.AssertionsForClassTypes.assertThat;

 import java.io.File;
-import java.util.ArrayList;
-import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
-import java.util.concurrent.atomic.AtomicReference;
 import java.util.stream.Collectors;

 import org.junit.jupiter.api.Test;
@@ -16,12 +13,8 @@ import org.springframework.core.io.ClassPathResource;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.SimplifiedSectionText;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.SimplifiedText;
-import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingFinishedEvent;
-import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingPipeline;
-import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingStorageService;
-import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationDocument;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.image.ImageServiceResponse;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
@@ -30,7 +23,6 @@ import com.knecon.fforesight.service.layoutparser.processor.services.SimplifiedS
 import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.server.utils.AbstractTest;

-import io.micrometer.observation.Observation;
 import lombok.SneakyThrows;
 import lombok.extern.slf4j.Slf4j;

@@ -1,71 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.server.graph;
-
-import static org.junit.jupiter.api.Assertions.assertEquals;
-import static org.junit.jupiter.api.Assertions.assertFalse;
-import static org.junit.jupiter.api.Assertions.assertThrows;
-import static org.junit.jupiter.api.Assertions.assertTrue;
-
-import java.util.Collections;
-import java.util.List;
-
-import org.junit.jupiter.api.BeforeEach;
-import org.junit.jupiter.api.Test;
-
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.Boundary;
-
-class BoundaryTest {
-
-    Boundary startBoundary;
-
-
-    @BeforeEach
-    void setUp() {
-
-        startBoundary = new Boundary(10, 100);
-    }
-
-
-    @Test
-    void testContains() {
-
-        assertTrue(startBoundary.contains(11));
-        assertTrue(startBoundary.contains(50));
-        assertFalse(startBoundary.contains(9));
-        assertFalse(startBoundary.contains(100));
-        assertFalse(startBoundary.contains(150));
-        assertFalse(startBoundary.contains(-123));
-        assertTrue(startBoundary.contains(new Boundary(11, 99)));
-        assertTrue(startBoundary.contains(new Boundary(10, 100)));
-        assertTrue(startBoundary.contains(new Boundary(11, 11)));
-        assertFalse(startBoundary.contains(9, 100));
-        assertTrue(startBoundary.contains(100, 100));
-        assertFalse(startBoundary.contains(100, 101));
-        assertFalse(startBoundary.contains(150, 151));
-    }
-
-
-    @Test
-    void testIntersects() {
-
-        assertTrue(startBoundary.intersects(new Boundary(1, 11)));
-        assertTrue(startBoundary.intersects(new Boundary(11, 12)));
-        assertTrue(startBoundary.intersects(new Boundary(11, 100)));
-        assertFalse(startBoundary.intersects(new Boundary(100, 101)));
-        assertTrue(startBoundary.intersects(new Boundary(99, 101)));
-    }
-
-
-    @Test
-    void testSplit() {
-
-        assertEquals(4, startBoundary.split(List.of(12, 40, 90)).size());
-        assertEquals(List.of(new Boundary(10, 12), new Boundary(12, 40), new Boundary(40, 90), new Boundary(90, 100)), startBoundary.split(List.of(12, 40, 90)));
-        assertEquals(List.of(new Boundary(10, 40), new Boundary(40, 100)), startBoundary.split(List.of(40)));
-        assertEquals(1, startBoundary.split(Collections.emptyList()).size());
-        assertEquals(1, startBoundary.split(List.of(startBoundary.start())).size());
-        assertThrows(IndexOutOfBoundsException.class, () -> startBoundary.split(Collections.singletonList(0)));
-        assertThrows(IndexOutOfBoundsException.class, () -> startBoundary.split(Collections.singletonList(100)));
-        assertThrows(IndexOutOfBoundsException.class, () -> startBoundary.split(List.of(12, 40, 100)));
-    }
-
-}
@@ -57,7 +57,7 @@ public class DocumentGraphJsonWritingTest extends BuildDocumentTest {
    private void writeJsons(Path filename) {

        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH,
-                layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH,
+                                                                         layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_WITHOUT_DUPLICATE_PARAGRAPH,
                        filename.toFile(),
                        new ImageServiceResponse(),
                        new TableServiceResponse(),
@@ -0,0 +1,84 @@
+package com.knecon.fforesight.service.layoutparser.server.graph;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertFalse;
+import static org.junit.jupiter.api.Assertions.assertThrows;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+
+import java.util.Collections;
+import java.util.List;
+
+import org.junit.jupiter.api.BeforeEach;
+import org.junit.jupiter.api.Test;
+
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.TextRange;
+
+class TextRangeTest {
+
+    TextRange startTextRange;
+
+
+    @BeforeEach
+    void setUp() {
+
+        startTextRange = new TextRange(10, 100);
+    }
+
+
+    @Test
+    void testContains() {
+
+        assertTrue(startTextRange.contains(11));
+        assertTrue(startTextRange.contains(50));
+        assertFalse(startTextRange.contains(9));
+        assertTrue(startTextRange.contains(100));
+        assertFalse(startTextRange.contains(101));
+        assertFalse(startTextRange.contains(150));
+        assertFalse(startTextRange.contains(-123));
+        assertTrue(startTextRange.contains(new TextRange(11, 99)));
+        assertTrue(startTextRange.contains(new TextRange(10, 100)));
+        assertTrue(startTextRange.contains(new TextRange(11, 11)));
+        assertFalse(startTextRange.contains(9, 100));
+        assertTrue(startTextRange.contains(100, 100));
+        assertFalse(startTextRange.contains(100, 101));
+        assertFalse(startTextRange.contains(150, 151));
+    }
+
+
+    @Test
+    void testContainsExclusive() {
+
+        assertTrue(startTextRange.containsExclusive(11));
+        assertTrue(startTextRange.containsExclusive(50));
+        assertFalse(startTextRange.containsExclusive(9));
+        assertFalse(startTextRange.containsExclusive(100));
+        assertFalse(startTextRange.containsExclusive(150));
+        assertFalse(startTextRange.containsExclusive(-123));
+    }
+
+
+    @Test
+    void testIntersects() {
+
+        assertTrue(startTextRange.intersects(new TextRange(1, 11)));
+        assertTrue(startTextRange.intersects(new TextRange(11, 12)));
+        assertTrue(startTextRange.intersects(new TextRange(11, 100)));
+        assertFalse(startTextRange.intersects(new TextRange(100, 101)));
+        assertTrue(startTextRange.intersects(new TextRange(99, 101)));
+    }
+
+
+    @Test
+    void testSplit() {
+
+        assertEquals(4, startTextRange.split(List.of(12, 40, 90)).size());
+        assertEquals(List.of(new TextRange(10, 12), new TextRange(12, 40), new TextRange(40, 90), new TextRange(90, 100)), startTextRange.split(List.of(12, 40, 90)));
+        assertEquals(List.of(new TextRange(10, 40), new TextRange(40, 100)), startTextRange.split(List.of(40)));
+        assertEquals(1, startTextRange.split(Collections.emptyList()).size());
+        assertEquals(1, startTextRange.split(List.of(startTextRange.start())).size());
+        assertThrows(IndexOutOfBoundsException.class, () -> startTextRange.split(Collections.singletonList(0)));
+        assertThrows(IndexOutOfBoundsException.class, () -> startTextRange.split(Collections.singletonList(100)));
+        assertThrows(IndexOutOfBoundsException.class, () -> startTextRange.split(List.of(12, 40, 100)));
+    }
+
+}
@@ -4,11 +4,18 @@ import java.io.File;
 import java.nio.file.Path;
 import java.util.Map;

+import org.junit.jupiter.api.AfterAll;
+import org.junit.jupiter.api.BeforeEach;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
+import org.mockito.MockitoAnnotations;
+import org.springframework.amqp.rabbit.core.RabbitTemplate;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.boot.test.mock.mockito.MockBean;
 import org.springframework.core.io.ClassPathResource;

 import com.iqser.red.commons.jackson.ObjectMapperFactory;
+import com.iqser.red.storage.commons.service.StorageService;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.image.ImageServiceResponse;
@@ -16,17 +23,30 @@ import com.knecon.fforesight.service.layoutparser.processor.python_api.model.tab
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.VisualLayoutParsingResponse;
 import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
+import com.knecon.fforesight.service.layoutparser.server.PDFNetInitializer;
 import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest;
-import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+import com.knecon.fforesight.service.viewerdoc.service.PDFTronViewerDocumentService;
+import com.knecon.fforesight.tenantcommons.TenantsClient;
+import com.pdftron.pdf.PDFNet;

+import jakarta.annotation.PostConstruct;
 import lombok.SneakyThrows;

 public class ViewerDocumentTest extends BuildDocumentTest {

-    ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
+    @Autowired
+    PDFNetInitializer pdfNetInitializer;
+    PDFTronViewerDocumentService viewerDocumentService = new PDFTronViewerDocumentService(null);
    LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);


+    @BeforeEach
+    public void init() {
+
+        pdfNetInitializer.init();
+    }
+
+
    @Test
    @SneakyThrows
    public void testViewerDocument() {
@@ -63,7 +83,7 @@ public class ViewerDocumentTest extends BuildDocumentTest {
                                                                       tableResponse,
                                                                       new VisualLayoutParsingResponse(),
                                                                       Map.of("file", Path.of(fileName).getFileName().toFile().toString()));
-        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
+        PDFTronViewerDocumentService viewerDocumentService = new PDFTronViewerDocumentService(null);
        LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);
        Document document = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.DOCUMINE_OLD, classificationDocument);

@@ -190,7 +190,8 @@ public class PdfSegmentationServiceTest extends AbstractTest {
                                                                        imageMetadata.getGeometry().getHeight()),
                                                 ImageType.valueOf(imageMetadata.getClassification().getLabel().toUpperCase(Locale.ROOT)),
                                                 imageMetadata.isAlpha(),
-                                                 imageMetadata.getPosition().getPageNumber())));
+                                                 imageMetadata.getPosition().getPageNumber(),
+                                                 "")));

        System.out.println("object");
    }
@@ -62,6 +62,7 @@ public abstract class AbstractTest {
    protected final static String TEXT_FILE_ID = "texts";
    protected final static String POSITION_FILE_ID = "positions";
    protected final static String PAGES_FILE_ID = "pages";
+    protected final static String MARKDOWN_FILE_ID = "markdown";
    protected final static String TENANT_ID = "tenant";
    protected final static String VIEWER_DOCUMENT_ID = "viewer";
    protected final static String SIMPLIFIED_ID = "simplified";
@@ -105,7 +106,7 @@ public abstract class AbstractTest {
    }


-    protected LayoutParsingRequest buildDefaultLayoutParsingRequest(String fileName, LayoutParsingType layoutParsingType, boolean debug) {
+    public static LayoutParsingRequest buildDefaultLayoutParsingRequest(String fileName, LayoutParsingType layoutParsingType, boolean debug) {

        var identifier = debug ? Map.of("fileId", fileName, "debug", "true") : Map.of("fileId", fileName);
        return LayoutParsingRequest.builder()
@@ -121,6 +122,7 @@ public abstract class AbstractTest {
                .pageFileStorageId(fileName + PAGES_FILE_ID)
                .simplifiedTextStorageId(fileName + SIMPLIFIED_ID)
                .viewerDocumentStorageId(fileName + VIEWER_DOCUMENT_ID)
+                .documentMarkdownFileStorageId(Optional.of(fileName + MARKDOWN_FILE_ID))
                .build();
    }

@@ -9,3 +9,5 @@ storage:
  key: minioadmin
  secret: minioadmin

+
+
@@ -28,6 +28,11 @@ spring:
          max-interval: 15000
        prefetch: 1

+layoutparser:
+  debug: true
+
+pdftron.license: demo:1650351709282:7bd235e003000000004ec28a6743e1163a085e2115de2536ab6e2cfe5a
+
 management:
  endpoint:
    metrics.enabled: ${monitoring.enabled:false}
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04f3fc00d7e0851c6ee0663ce749562234cc95123ffdd643df88d621e4323ede
+size 238546
@@ -12,6 +12,7 @@ dependencies {
    implementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
    implementation("org.slf4j:slf4j-api:1.7.25")
    implementation("com.knecon.fforesight:tracing-commons:0.5.0")
+    implementation("com.pdftron:PDFNet:10.7.0")

    testImplementation("org.apache.logging.log4j:log4j-slf4j-impl:2.22.1")
    testImplementation("org.junit.jupiter:junit-jupiter")
@@ -1,72 +0,0 @@
-package com.knecon.fforesight.service.viewerdoc;
-
-import java.util.List;
-
-import org.apache.pdfbox.cos.COSName;
-
-import lombok.AccessLevel;
-import lombok.experimental.FieldDefaults;
-
-@FieldDefaults(makeFinal = true, level = AccessLevel.PUBLIC)
-public class ContentStreams {
-
-    public static Identifier KNECON_LAYOUT = new Identifier("Layout grid", COSName.getPDFName("KNECON_LAYOUT"), true);
-
-    public static Identifier KNECON_VISUAL_PARSING = new Identifier("Layout grid - visual", COSName.getPDFName("KNECON_VISUAL_PARSING"), true);
-
-    public static Identifier KNECON_OCR = new Identifier("OCR", COSName.getPDFName("KNECON_OCR"), false);
-
-    public static Identifier KNECON_OCR_TEXT_DEBUG = new Identifier("OCR Text", COSName.getPDFName("KNECON_OCR_TEXT_DEBUG"), true);
-
-    public static Identifier KNECON_OCR_BBOX_DEBUG = new Identifier("OCR Boxes", COSName.getPDFName("KNECON_OCR_BBOX_DEBUG"), true);
-
-    public static Identifier OTHER = new Identifier("other", COSName.getPDFName("OTHER"), false);
-
-    public static Identifier ESCAPE_START = new Identifier("escape start", COSName.getPDFName("ESCAPE_START"), false);
-
-    public static Identifier ESCAPE_END = new Identifier("escape start", COSName.getPDFName("ESCAPE_END"), false);
-
-    public static Identifier CLEAN_RULINGS = new Identifier("Cleaned Rulings", COSName.getPDFName("KNECON_CLEAN_RULINGS"), true);
-
-    public static Identifier RULINGS = new Identifier("Rulings", COSName.getPDFName("KNECON_RULINGS"), true);
-
-    public static Identifier WORDS = new Identifier("Words", COSName.getPDFName("KNECON_WORDS"), true);
-
-    public static Identifier ZONES = new Identifier("Text Zones", COSName.getPDFName("KNECON_ZONES"), true);
-
-    public static Identifier LINES = new Identifier("Text Lines", COSName.getPDFName("KNECON_LINES"), true);
-
-    public static Identifier CELLS = new Identifier("Cells", COSName.getPDFName("KNECON_CELLS"), true);
-
-    public static Identifier MAIN_BODY = new Identifier("Main Text Body", COSName.getPDFName("KNECON_MAIN_BODY"), true);
-
-    public static Identifier MARKED_CONTENT = new Identifier("Marked content", COSName.getPDFName("KNECON_MARKED_CONTENT"), true);
-
-    public static Identifier NEIGHBOURS = new Identifier("Neighbours", COSName.getPDFName("KNECON_NEIGHBOURS"), true);
-
-    public static Identifier CHARACTERS = new Identifier("Characters", COSName.getPDFName("KNECON_CHARACTERS"), true);
-
-    public static List<Identifier> allContentStreams = List.of(KNECON_LAYOUT,
-                                                               KNECON_VISUAL_PARSING,
-                                                               KNECON_OCR,
-                                                               KNECON_OCR_BBOX_DEBUG,
-                                                               KNECON_OCR_TEXT_DEBUG,
-                                                               OTHER,
-                                                               ESCAPE_START,
-                                                               ESCAPE_END,
-                                                               RULINGS,
-                                                               CLEAN_RULINGS,
-                                                               WORDS,
-                                                               ZONES,
-                                                               LINES,
-                                                               MAIN_BODY,
-                                                               MARKED_CONTENT,
-                                                               NEIGHBOURS,
-                                                               CHARACTERS,
-                                                               CELLS);
-
-    public record Identifier(String name, COSName cosName, boolean optionalContent) {
-
-    }
-
-}
@@ -0,0 +1,77 @@
+package com.knecon.fforesight.service.viewerdoc;
+
+import org.apache.pdfbox.cos.COSName;
+
+/*
+These identifiers are used to mark content in the pdf, such that it may be found later. The markedContentName must therefore be unique.
+The String "name" is only used to display optional content in the optional content view in the pdf.
+Therefore, it may be null, if optionalContent is false.
+If optionalContent is false, the layer will not be created as a OCG, and will not be listed in the OCG view.
+ */
+public record LayerIdentifier(String name, String markedContentName) {
+
+    public String markedContentName() {
+        // The prefix KNECON_ is used to identify marked contents as knecon contents later on
+        return KNECON_IDENTIFIER_PREFIX + markedContentName;
+    }
+
+
+    public COSName cosName() {
+
+        return COSName.getPDFName(markedContentName);
+    }
+
+
+    public static final String KNECON_IDENTIFIER_PREFIX = "KNECON_";
+
+    public static final LayerIdentifier KNECON_OCR = new LayerIdentifier(null, "OCR");
+    public static final LayerIdentifier KNECON_OCR_TEXT = new LayerIdentifier(null, "OCR_TEXT");
+    public static final LayerIdentifier KNECON_OCR_LINES = new LayerIdentifier(null, "OCR_LINES");
+
+    // layers
+    // layout grid
+    public static final LayerIdentifier KNECON_LAYOUT = new LayerIdentifier("Layout grid", "LAYOUT");
+    public static final LayerIdentifier KNECON_LAYOUT_SECTION = new LayerIdentifier("Section", "LAYOUT_SECTION");
+    public static final LayerIdentifier KNECON_LAYOUT_PARAGRAPH = new LayerIdentifier("Paragraph ", "LAYOUT_PARAGRAPH");
+    public static final LayerIdentifier KNECON_LAYOUT_KEY_VALUE = new LayerIdentifier("Key-Value Pairs ", "LAYOUT_KEY_VALUE");
+    public static final LayerIdentifier KNECON_LAYOUT_HEADLINE = new LayerIdentifier("Headline", "LAYOUT_HEADLINE");
+    public static final LayerIdentifier KNECON_LAYOUT_HEADER_FOOTER = new LayerIdentifier("Header/Footer", "LAYOUT_HEADER_FOOTER");
+    public static final LayerIdentifier KNECON_LAYOUT_TABLE = new LayerIdentifier("Tables", "LAYOUT_TABLE");
+    public static final LayerIdentifier KNECON_LAYOUT_FIGURES = new LayerIdentifier("Figures", "LAYOUT_FIGURES");
+    public static final LayerIdentifier KNECON_LAYOUT_IMAGES = new LayerIdentifier("Images", "LAYOUT_IMAGES");
+    public static final LayerIdentifier KNECON_LAYOUT_TREE_IDs = new LayerIdentifier("Tree IDs", "LAYOUT_TREE_IDs");
+
+    //layout grid debug
+    public static final LayerIdentifier KNECON_LAYOUT_DEBUG = new LayerIdentifier("Layout elements", "DEBUG_LAYOUT");
+    public static final LayerIdentifier CLEAN_RULINGS = new LayerIdentifier("Classified Rulings", "CLEAN_RULINGS");
+    public static final LayerIdentifier RULINGS = new LayerIdentifier("Rulings", "RULINGS");
+    public static final LayerIdentifier WORDS = new LayerIdentifier("Words", "WORDS");
+    public static final LayerIdentifier ZONES = new LayerIdentifier("Text Zones", "ZONES");
+    public static final LayerIdentifier LINES = new LayerIdentifier("Text Lines", "LINES");
+    public static final LayerIdentifier CELLS = new LayerIdentifier("Cells", "CELLS");
+    public static final LayerIdentifier MAIN_BODY = new LayerIdentifier("Main Text Body", "MAIN_BODY");
+    public static final LayerIdentifier MARKED_CONTENT = new LayerIdentifier("Marked content", "MARKED_CONTENT");
+    public static final LayerIdentifier NEIGHBOURS = new LayerIdentifier("Neighbours", "NEIGHBOURS");
+    public static final LayerIdentifier CHARACTERS = new LayerIdentifier("Characters", "CHARACTERS");
+
+    public static final LayerIdentifier KNECON_VISUAL_PARSING = new LayerIdentifier("Visual Layout Parser", "VISUAL_PARSING");
+
+    //ocr
+    public static final LayerIdentifier KNECON_OCR_DEBUG = new LayerIdentifier("OCR", "OCR_DEBUG");
+    public static final LayerIdentifier KNECON_OCR_TEXT_DEBUG = new LayerIdentifier("OCR Text", "OCR_TEXT_DEBUG");
+    public static final LayerIdentifier KNECON_OCR_BBOX_DEBUG = new LayerIdentifier("OCR Words", "OCR_BBOX_DEBUG");
+    public static final LayerIdentifier KNECON_OCR_LINE_DEBUG = new LayerIdentifier("OCR Lines", "OCR_LINE_DEBUG");
+    public static final LayerIdentifier KNECON_OCR_OVERLAPPED_TEXT = new LayerIdentifier("OCR overlapped Text", "OCR_OVERLAPPED_TEXT_DEBUG");
+
+    //azure idp
+    public static final LayerIdentifier KNECON_AZURE_IDP = new LayerIdentifier("IDP", "IDP");
+    public static final LayerIdentifier IDP_FIGURES = new LayerIdentifier("IDP Figures", "IDP_FIGURES");
+    public static final LayerIdentifier IDP_TABLES = new LayerIdentifier("IDP Tables", "IDP_TABLES");
+    public static final LayerIdentifier IDP_KV_PAIRS = new LayerIdentifier("IDP Key Value Pair", "IDP_KV_PAIRS");
+    public static final LayerIdentifier IDP_SECTIONS = new LayerIdentifier("IDP Sections", "IDP_SECTIONS");
+    public static final LayerIdentifier IDP_LINES = new LayerIdentifier("IDP Lines", "IDP_LINES");
+    public static final LayerIdentifier IDP_PARAGRAPHS = new LayerIdentifier("IDP Paragraphs", "IDP_PARAGRAPHS");
+    public static final LayerIdentifier IDP_LIST = new LayerIdentifier("IDP Lists", "IDP_LISTS");
+    public static final LayerIdentifier IDP_BARCODES = new LayerIdentifier("IDP Barcodes", "IDP_BARCODES");
+
+}
@@ -0,0 +1,19 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+import com.knecon.fforesight.service.viewerdoc.model.VisualizationsOnPage;
+
+public abstract class AbstractLayerGroup implements LayerGroup {
+
+    protected VisualizationsOnPage getOrCreateVisualizationsOnPage(int page, Visualizations visualizations) {
+
+        if (visualizations.getVisualizationsOnPages().containsKey(page)) {
+            return visualizations.getVisualizationsOnPages()
+                    .get(page);
+        }
+        VisualizationsOnPage visualizationsOnPage = VisualizationsOnPage.builder().build();
+        visualizations.getVisualizationsOnPages().put(page, visualizationsOnPage);
+        return visualizationsOnPage;
+    }
+
+}
@@ -0,0 +1,49 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.awt.Color;
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+import lombok.Getter;
+
+public class IdpLayerConfig extends AbstractLayerGroup {
+
+    @Getter
+    public final LayerIdentifier groupIdentifier = LayerIdentifier.KNECON_AZURE_IDP;
+
+    public static final LayerGroup CONFIG_INSTANCE = new IdpLayerConfig();
+
+    protected final Visualizations figures = Visualizations.builder().layer(LayerIdentifier.IDP_FIGURES).visibleByDefault(true).build();
+    protected final Visualizations tables = Visualizations.builder().layer(LayerIdentifier.IDP_TABLES).visibleByDefault(true).build();
+    protected final Visualizations keyValuePairs = Visualizations.builder().layer(LayerIdentifier.IDP_KV_PAIRS).visibleByDefault(true).build();
+    protected final Visualizations paragraphs = Visualizations.builder().layer(LayerIdentifier.IDP_PARAGRAPHS).build();
+    protected final Visualizations sections = Visualizations.builder().layer(LayerIdentifier.IDP_SECTIONS).build();
+    protected final Visualizations lines = Visualizations.builder().layer(LayerIdentifier.IDP_LINES).build();
+    protected final Visualizations lists = Visualizations.builder().layer(LayerIdentifier.IDP_LIST).visibleByDefault(true).build();
+    protected final Visualizations barcodes = Visualizations.builder().layer(LayerIdentifier.IDP_BARCODES).visibleByDefault(true).build();
+
+    protected static final Color TABLE_COLOR = new Color(102, 205, 170);
+    protected static final Color INNER_LINES_COLOR = new Color(255, 175, 175);
+    protected static final Color SECTION_COLOR = new Color(50, 50, 50);
+    protected static final Color SECTION_HEADING_COLOR = new Color(162, 56, 56);
+    protected static final Color TITLE_COLOR = new Color(221, 25, 25);
+    protected static final Color HEADER_FOOTER_COLOR = new Color(171, 131, 6);
+    protected static final Color FOOTNOTE_COLOR = new Color(6, 64, 171);
+    protected static final Color FORMULA_COLOR = new Color(80, 171, 6);
+    protected static final Color PARAGRAPH_COLOR = new Color(70, 130, 180);
+    protected static final Color IMAGE_COLOR = new Color(253, 63, 146);
+    protected static final Color KEY_VALUE_BBOX_COLOR = new Color(0, 39, 85);
+    protected static final Color KEY_COLOR = new Color(30, 92, 172);
+    protected static final Color VALUE_COLOR = new Color(30, 172, 146);
+    protected static final Color LINES_COLOR = new Color(152, 45, 179);
+
+
+    @Override
+    public List<Visualizations> getVisualizations() {
+
+        return List.of(paragraphs, sections, figures, tables, keyValuePairs, lines, lists, barcodes);
+    }
+
+}
@@ -0,0 +1,62 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+public interface LayerGroup {
+
+    LayerIdentifier getGroupIdentifier();
+
+
+    List<Visualizations> getVisualizations();
+
+
+    default List<LayerIdentifier> getSubLayers() {
+
+        return getVisualizations().stream()
+                .map(Visualizations::getLayer)
+                .toList();
+
+    }
+
+
+    default boolean isVisibleByDefault() {
+
+        return false;
+    }
+
+
+    /*
+    indicates the sub layers are all optional content
+     */
+    default boolean subLayersAreOptionalContent() {
+
+        return true;
+    }
+
+
+    /*
+    indicates the LayerGroup is also a optional content group, and should be displayed as such:
+    layer
+        - sublayer0
+        - sublayer1
+
+    see note in specification 8.11.4.3
+     */
+    default boolean isOptionalContent() {
+
+        return true;
+    }
+
+
+    default boolean isEmpty() {
+
+        return getVisualizations().isEmpty();
+    }
+
+
+
+
+}
@@ -0,0 +1,73 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.awt.Color;
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Standard14EmbeddableFont;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+import lombok.Getter;
+
+public class LayoutDebugLayerConfig extends AbstractLayerGroup {
+
+    @Getter
+    public final LayerIdentifier groupIdentifier = LayerIdentifier.KNECON_LAYOUT_DEBUG;
+
+    public static final LayerGroup CONFIG_INSTANCE = new LayoutDebugLayerConfig();
+
+    protected static final Standard14EmbeddableFont FONT = Standard14EmbeddableFont.helvetica();
+
+    protected static final Color WORDS_COLOR = new Color(68, 84, 147);
+    protected static final Color LINES_COLOR = new Color(152, 45, 179);
+    protected static final Color ZONES_COLOR = new Color(131, 38, 38);
+
+    protected static final Color RULINGS_COLOR = new Color(21, 221, 174);
+    protected static final Color TABLE_RULINGS_COLOR = new Color(255, 175, 175);
+    protected static final Color HEADER_RULING_COLOR = new Color(171, 131, 6);
+    protected static final Color FOOTER_RULING_COLOR = new Color(106, 82, 2);
+    protected static final Color UNDERLINE_RULING_COLOR = new Color(6, 39, 171);
+    protected static final Color STRIKETROUGH_RULING_COLOR = new Color(171, 6, 6);
+
+    protected static final Color CELLS_COLOR = new Color(31, 214, 27);
+
+    protected static final Color MAIN_BODY_COLOR = new Color(171, 131, 6);
+    protected static final Color MARKED_CONTENT_COLOR = new Color(171, 131, 6);
+
+    protected static final List<Color> ROTATING_CHARACTER_COLOR = List.of(new Color(255, 87, 51),
+                                                                          new Color(255, 195, 0),
+                                                                          new Color(76, 175, 80),
+                                                                          new Color(33, 150, 243),
+                                                                          new Color(155, 89, 182),
+                                                                          new Color(233, 30, 99),
+                                                                          new Color(0, 188, 212),
+                                                                          new Color(121, 85, 72));
+
+    protected final Visualizations words = Visualizations.builder().layer(LayerIdentifier.WORDS).visibleByDefault(true).build();
+    protected final Visualizations lines = Visualizations.builder().layer(LayerIdentifier.LINES).build();
+    protected final Visualizations zones = Visualizations.builder().layer(LayerIdentifier.ZONES).build();
+    protected final Visualizations mainBody = Visualizations.builder().layer(LayerIdentifier.MAIN_BODY).build();
+    protected final Visualizations clean_rulings = Visualizations.builder().layer(LayerIdentifier.CLEAN_RULINGS).build();
+    protected final Visualizations rulings = Visualizations.builder().layer(LayerIdentifier.RULINGS).build();
+    protected final Visualizations cells = Visualizations.builder().layer(LayerIdentifier.CELLS).build();
+    protected final Visualizations markedContent = Visualizations.builder().layer(LayerIdentifier.MARKED_CONTENT).build();
+    protected final Visualizations neighbours = Visualizations.builder().layer(LayerIdentifier.NEIGHBOURS).build();
+    protected final Visualizations characters = Visualizations.builder().layer(LayerIdentifier.CHARACTERS).build();
+
+
+    public List<Visualizations> getVisualizations() {
+
+        return List.of(characters, //
+                       neighbours,//
+                       words, //
+                       lines, //
+                       zones, //
+                       rulings, //
+                       clean_rulings, //
+                       cells, //
+                       mainBody, //
+                       markedContent //
+        );
+    }
+
+}
@@ -0,0 +1,55 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.awt.Color;
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Standard14EmbeddableFont;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+import lombok.Getter;
+
+public class LayoutGridLayerConfig extends AbstractLayerGroup {
+
+    @Getter
+    public final LayerIdentifier groupIdentifier = LayerIdentifier.KNECON_LAYOUT;
+
+    public static final LayerGroup CONFIG_INSTANCE = new LayoutGridLayerConfig();
+
+    protected static final float FONT_SIZE = 10f;
+    protected static final float LINE_WIDTH = 1f;
+    protected static final Standard14EmbeddableFont FONT = Standard14EmbeddableFont.helvetica();
+
+    protected static final Color INNER_LINES_COLOR = new Color(255, 175, 175);
+    protected static final Color HEADER_CELL_COLOR = new Color(156, 21, 48);
+    protected static final Color PARAGRAPH_COLOR = new Color(70, 130, 180);
+
+    protected static final Color DUPLICATE_PARAGRAPH_COLOR = new Color(70, 180, 101);
+    protected static final Color TABLE_COLOR = new Color(102, 205, 170);
+    protected static final Color SECTION_COLOR = new Color(50, 50, 50);
+    protected static final Color HEADLINE_COLOR = new Color(162, 56, 56);
+    protected static final Color HEADER_COLOR = new Color(171, 131, 6);
+    protected static final Color IMAGE_COLOR = new Color(253, 63, 146);
+    protected static final Color TREEID_COLOR = new Color(53, 53, 53);
+    protected static final Color KEY_VALUE_BBOX_COLOR = new Color(0, 39, 85);
+    protected static final Color KEY_COLOR = new Color(30, 92, 172);
+    protected static final Color VALUE_COLOR = new Color(30, 172, 146);
+
+    protected final Visualizations sections = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_SECTION).visibleByDefault(true).build();
+    protected final Visualizations paragraphs = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_PARAGRAPH).visibleByDefault(true).build();
+    protected final Visualizations headlines = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_HEADLINE).visibleByDefault(true).build();
+    protected final Visualizations tables = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_TABLE).visibleByDefault(true).build();
+    protected final Visualizations figures = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_FIGURES).visibleByDefault(true).build();
+    protected final Visualizations headerFooter = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_HEADER_FOOTER).visibleByDefault(true).build();
+    protected final Visualizations images = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_IMAGES).build();
+    protected final Visualizations keyValue = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_KEY_VALUE).build();
+    protected final Visualizations treeIds = Visualizations.builder().layer(LayerIdentifier.KNECON_LAYOUT_TREE_IDs).build();
+
+
+    @Override
+    public List<Visualizations> getVisualizations() {
+
+        return List.of(headlines, paragraphs, tables, sections, headerFooter, keyValue, figures, images, treeIds);
+    }
+
+}
@@ -0,0 +1,38 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.awt.Color;
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+import lombok.Getter;
+
+public class OcrDebugLayerConfig extends AbstractLayerGroup {
+
+    @Getter
+    public final LayerIdentifier groupIdentifier = LayerIdentifier.KNECON_OCR_DEBUG;
+
+    public static final LayerGroup CONFIG_INSTANCE = new OcrDebugLayerConfig();
+
+    protected static final Color REGULAR_COLOR = new Color(6, 39, 171);
+    protected static final Color BOLD_COLOR = new Color(50, 246, 246);
+    protected static final Color ITALIC_COLOR = new Color(171, 105, 6);
+    protected static final Color BOLD_ITALIC_COLOR = new Color(6, 171, 102);
+    protected static final Color HANDWRITTEN_COLOR = new Color(171, 64, 6);
+    protected static final Color OVERLAPPED_COLOR = new Color(142, 8, 8);
+    protected static final Color TABLE_LINES_COLOR = new Color(21, 221, 174);
+
+    protected final Visualizations debugText = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_TEXT_DEBUG).visibleByDefault(true).build();
+    protected final Visualizations tableLines = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_LINE_DEBUG).visibleByDefault(true).build();
+    protected final Visualizations overlappedText = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_OVERLAPPED_TEXT).visibleByDefault(false).build();
+    protected final Visualizations debugBBox = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_BBOX_DEBUG).visibleByDefault(false).build();
+
+
+    @Override
+    public List<Visualizations> getVisualizations() {
+
+        return List.of(debugText, tableLines, debugBBox, overlappedText);
+    }
+
+}
@@ -0,0 +1,40 @@
+package com.knecon.fforesight.service.viewerdoc.layers;
+
+import java.util.List;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+
+public class OcrTextLayerConfig extends AbstractLayerGroup {
+
+    protected final Visualizations ocrText = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_TEXT).build();
+    protected final Visualizations tableLines = Visualizations.builder().layer(LayerIdentifier.KNECON_OCR_LINES).build();
+
+
+    @Override
+    public LayerIdentifier getGroupIdentifier() {
+
+        return LayerIdentifier.KNECON_OCR;
+    }
+
+
+    @Override
+    public List<Visualizations> getVisualizations() {
+
+        return List.of(ocrText, tableLines);
+    }
+
+
+    @Override
+    public boolean subLayersAreOptionalContent() {
+
+        return false;
+    }
+
+    @Override
+    public boolean isOptionalContent() {
+
+        return false;
+    }
+
+}
@@ -3,8 +3,13 @@ package com.knecon.fforesight.service.viewerdoc.model;
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.pdmodel.font.PDFont;

+import com.pdftron.pdf.Font;
+import com.pdftron.pdf.PDFDoc;
+
 public interface EmbeddableFont {

    PDFont embed(PDDocument document);

+    Font embed(PDFDoc doc);
+
 }
@@ -1,27 +0,0 @@
-package com.knecon.fforesight.service.viewerdoc.model;
-
-import java.util.HashMap;
-import java.util.Map;
-
-import lombok.AccessLevel;
-import lombok.Getter;
-import lombok.experimental.FieldDefaults;
-
-@Getter
-@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
-public class LayoutGrid {
-
-    int numberOfPages;
-    Map<Integer, VisualizationsOnPage> visualizationsPerPages;
-
-
-    public LayoutGrid(int numberOfPages) {
-
-        this.numberOfPages = numberOfPages;
-        this.visualizationsPerPages = new HashMap<>();
-        for (int i = 0; i < numberOfPages; i++) {
-            this.visualizationsPerPages.put(i, VisualizationsOnPage.builder().build());
-        }
-    }
-
-}
@@ -0,0 +1,92 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import java.util.Deque;
+import java.util.Iterator;
+import java.util.LinkedList;
+import java.util.Set;
+
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;
+
+public class MarkedContentStack {
+
+    private final Deque<MarkedContent> stack = new LinkedList<>();
+
+
+    public void enterMarkedContent(String name) {
+
+        stack.push(new MarkedContent(name));
+    }
+
+
+    public void leaveMarkedContent() {
+
+        stack.pop();
+    }
+
+
+    public String currentMarkedContent() {
+
+        if (stack.isEmpty()) {
+            return "";
+        }
+        return stack.peek().name();
+    }
+
+
+    public boolean currentMarkedContentContains(String name) {
+
+        Iterator<MarkedContent> markedContentIterator = stack.descendingIterator();
+        while (markedContentIterator.hasNext()) {
+            var markedContent = markedContentIterator.next();
+            if (markedContent.name().equals(name)) {
+                return true;
+            }
+        }
+        return false;
+    }
+
+
+    public boolean currentMarkedContentContainsNone(Set<String> names) {
+
+        if (stack.isEmpty()) {
+            return true;
+        }
+        Iterator<MarkedContent> markedContentIterator = stack.descendingIterator();
+        while (markedContentIterator.hasNext()) {
+            var markedContent = markedContentIterator.next();
+            if (names.contains(markedContent.name())) {
+                return false;
+            }
+        }
+        return true;
+    }
+
+
+    public boolean currentMarkedContentIsKneconContent() {
+
+        if (stack.isEmpty()) {
+            return false;
+        }
+        Iterator<MarkedContent> markedContentIterator = stack.descendingIterator();
+        while (markedContentIterator.hasNext()) {
+            var markedContent = markedContentIterator.next();
+            if (markedContent.name().startsWith(LayerIdentifier.KNECON_IDENTIFIER_PREFIX)) {
+                return true;
+            }
+        }
+        return false;
+
+    }
+
+
+    public void clear() {
+
+        stack.clear();
+    }
+
+
+    private record MarkedContent(String name) {
+
+    }
+
+}
@@ -1,10 +0,0 @@
-package com.knecon.fforesight.service.viewerdoc.model;
-
-import java.util.List;
-
-import org.apache.pdfbox.contentstream.operator.Operator;
-import org.apache.pdfbox.cos.COSBase;
-
-public record OperatorWithArguments(Operator operator, List<COSBase> arguments) {
-
-}
@@ -1,13 +1,14 @@
 package com.knecon.fforesight.service.viewerdoc.model;

 import java.awt.Color;
+import java.awt.geom.AffineTransform;
 import java.awt.geom.Point2D;
 import java.util.Optional;

 import org.apache.pdfbox.pdmodel.graphics.state.RenderingMode;
 import org.apache.pdfbox.util.Matrix;

-public record PlacedText(String text, Point2D lineStart, Color color, float fontSize, EmbeddableFont font, Optional<Matrix> textMatrix, Optional<RenderingMode> renderingMode) {
+public record PlacedText(String text, Point2D lineStart, Color color, float fontSize, EmbeddableFont font, Optional<AffineTransform> textMatrix, Optional<RenderingMode> renderingMode) {

    public static PlacedText textFacingUp(String text, Point2D lineStart, float fontSize, Color color, EmbeddableFont font) {

@@ -1,10 +1,15 @@
 package com.knecon.fforesight.service.viewerdoc.model;

+import java.util.Objects;
+
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.pdmodel.font.PDFont;
 import org.apache.pdfbox.pdmodel.font.PDType1Font;
 import org.apache.pdfbox.pdmodel.font.Standard14Fonts;

+import com.pdftron.pdf.Font;
+import com.pdftron.pdf.PDFDoc;
+
 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;

@@ -13,10 +18,11 @@ public class Standard14EmbeddableFont implements EmbeddableFont {

    private final PDType1Font font;

+    private final int pdfTronIdentifier;

    public static Standard14EmbeddableFont helvetica() {

-        return new Standard14EmbeddableFont(new PDType1Font(Standard14Fonts.FontName.HELVETICA));
+        return new Standard14EmbeddableFont(new PDType1Font(Standard14Fonts.FontName.HELVETICA), Font.e_helvetica);
    }


@@ -34,4 +40,12 @@ public class Standard14EmbeddableFont implements EmbeddableFont {
        return font;
    }

+
+    @Override
+    @SneakyThrows
+    public Font embed(PDFDoc document) {
+
+        return Font.create(document, pdfTronIdentifier);
+    }
+
 }
@@ -3,7 +3,7 @@ package com.knecon.fforesight.service.viewerdoc.model;
 import java.util.LinkedHashMap;
 import java.util.Map;

-import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.LayerIdentifier;

 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -17,9 +17,10 @@ import lombok.experimental.FieldDefaults;
@FieldDefaults(level = AccessLevel.PRIVATE)
 public class Visualizations {

-    ContentStreams.Identifier layer;
+    LayerIdentifier layer;
    @Builder.Default
    Map<Integer, VisualizationsOnPage> visualizationsOnPages = new LinkedHashMap<>();
-    boolean layerVisibilityDefaultValue;
+
+    boolean visibleByDefault;

 }
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Timo Bejan	28d8ad0a3f	Merge branch 'CLARI-128' into 'main' Fixed Index out of bounds exception in blockificationpostprocessingservice -... See merge request fforesight/layout-parser!186	2024-07-30 16:56:21 +02:00
Timo Bejan	0c1583c1be	Fixed Index out of bounds exception in blockificationpostprocessingservice - this could should be documented btw, there are also probably other use-cases where the code doesnt work	2024-07-30 17:45:05 +03:00
Andrei Isvoran	7633566d9b	Merge branch 'RED-9607-fp' into 'main' RED-9607 - Correctly determine text position sequence based on file rotation See merge request fforesight/layout-parser!184	2024-07-25 14:11:00 +02:00
Andrei Isvoran	cc4f09711e	RED-9607 - Correctly determine text position sequence based on file rotation	2024-07-24 16:35:11 +03:00
Kilian Schüttler	370165dc59	Merge branch 'document-data-markdown' into 'main' CLARI: document-data-markdown See merge request fforesight/layout-parser!181	2024-07-18 17:19:44 +02:00
Maverick Studer	8c052c38d7	CLARI: document-data-markdown	2024-07-18 17:19:43 +02:00
Kilian Schüttler	ea18d3d307	Merge branch 'RED-8800' into 'main' RED-8800: adjust coordinates in BE to ignore cropbox See merge request fforesight/layout-parser!179	2024-07-15 17:45:13 +02:00
Kilian Schüttler	2726fc3fe1	RED-8800: adjust coordinates in BE to ignore cropbox	2024-07-15 17:45:13 +02:00
Kilian Schüttler	033279e261	Merge branch 'RED-9353' into 'main' RED-9353: refactor PDFTronViewerDocumentService See merge request fforesight/layout-parser!178	2024-07-15 12:54:17 +02:00
Kilian Schüttler	ec0dd032c9	RED-9353: refactor PDFTronViewerDocumentService	2024-07-15 12:54:17 +02:00
Andrei Isvoran	598fa7f1c7	Merge branch 'RED-9496-shutdown-fp' into 'main' RED-9496 - Implement graceful shutdown See merge request fforesight/layout-parser!176	2024-07-04 14:26:00 +02:00
Andrei Isvoran	65b1f7d179	RED-9496 - Implement graceful shutdown	2024-07-04 14:21:20 +03:00
Kilian Schüttler	3173610be5	Merge branch 'CLARI-003' into 'main' CLARI-003: add treeId to StructureObject See merge request fforesight/layout-parser!176	2024-07-02 11:37:22 +02:00
Kilian Schuettler	e920eb5a78	CLARI-003: add treeId to StructureObject	2024-07-01 13:56:16 +02:00
Kilian Schüttler	7e4baea7e5	Merge branch 'RED-9353' into 'main' RED-9353: use azure ocr service See merge request fforesight/layout-parser!175	2024-07-01 11:13:27 +02:00
Kilian Schüttler	66d3433e04	RED-9353: use azure ocr service	2024-07-01 11:13:26 +02:00
Yannik Hampe	a2f559af51	Merge branch 'RED-3813' into 'main' RED-3813: Recategorize same image as experimental feature See merge request fforesight/layout-parser!155	2024-06-26 13:42:42 +02:00
Yannik Hampe	39f527a57c	Merge branch 'main' into 'RED-3813' # Conflicts: # layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java	2024-06-26 09:10:59 +02:00
yhampe	5c2844fe31	RED-3813: Recategorize same image as experimental feature fixed failing test	2024-06-26 09:08:37 +02:00
Kilian Schüttler	b216f02e15	Merge branch 'RED-9194' into 'main' RED-9194: roll back single digit headline change See merge request fforesight/layout-parser!171	2024-06-21 15:13:40 +02:00
Kilian Schuettler	2e2f30ba35	RED-9194: roll back single digit headline change	2024-06-21 14:42:30 +02:00
Kilian Schuettler	9f7ed974ec	RED-9194: roll back single digit headline change	2024-06-21 14:41:30 +02:00
Kilian Schuettler	570a348a77	RED-9194: roll back single digit headline change	2024-06-21 14:39:47 +02:00
Maverick Studer	859dba2ecf	Merge branch 'RED-9374' into 'main' hotfix for table/paragraph section creation on document start before first headline See merge request fforesight/layout-parser!170	2024-06-18 17:36:04 +02:00
Maverick Studer	1c5d755111	hotfix for table/paragraph section creation on document start before first headline	2024-06-18 17:36:04 +02:00
Maverick Studer	133e06460f	Merge branch 'RED-9374' into 'main' RED-9374: Ner Entities are at wrong locations See merge request fforesight/layout-parser!169	2024-06-18 16:31:24 +02:00
Maverick Studer	da91fcff97	RED-9374: Ner Entities are at wrong locations	2024-06-18 16:31:24 +02:00
Kilian Schüttler	79795e408a	Merge branch 'RED-9194' into 'main' RED-9194: allow single digit headline identifiers See merge request fforesight/layout-parser!168	2024-06-07 09:09:25 +02:00
Kilian Schuettler	b719db86ab	RED-9194: allow single digit headline identifiers	2024-06-06 16:32:05 +02:00
Maverick Studer	797602e373	Merge branch 'thread-safe-hcs-fields' into 'main' fixed issue with thread-safety of local fields in the HeadlineClassificationService See merge request fforesight/layout-parser!167	2024-06-06 14:51:24 +02:00
maverickstuder	3d2f66cf10	fixed issue with thread-safety of local fields in the HeadlineClassificationService: * HeadlineClassificationService is no singleton anymore * instead initialize it in the ClassificationService and pass it to the classifyMethods as required	2024-06-06 14:39:23 +02:00
Kilian Schüttler	e304a9f2d7	Merge branch 'RED-7074-le' into 'main' RED-7074: Design Subsection section tree structure algorithm See merge request fforesight/layout-parser!166	2024-06-06 13:22:14 +02:00
Maverick Studer	c05f67cf44	RED-7074: Design Subsection section tree structure algorithm	2024-06-06 13:22:14 +02:00
yhampe	9ecf9ca19f	RED-3813: Recategorize same image as experimental feature now writing hash into structure	2024-06-05 14:20:33 +02:00
yhampe	9be672c728	RED-3813: Recategorize same image as experimental feature working on pushing properties to persistence service	2024-05-28 13:51:45 +02:00
yhampe	a5fcebce30	RED-3813: Recategorize same image as experimental feature added representation to image and DocumentStructure	2024-05-17 07:34:05 +02:00