RED-7141: Implemented docstrum layout parsing

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/queue/LayoutParsingType.java

@@ -2,9 +2,9 @@ package com.knecon.fforesight.service.layoutparser.internal.api.queue;
 
 public enum LayoutParsingType {
     REDACT_MANAGER,
-    TAAS,
+    REDACT_MANAGER_OLD,
+    REDACT_MANAGER_PARAGRAPH_DEBUG,
     DOCUMINE,
-
-    DOCSTRUM,
-    DOCSTRUM_XY
+    CLARIFYND,
+    CLARIFYND_PARAGRAPH_DEBUG
 }

layoutparser-service/layoutparser-service-processor/build.gradle.kts

@@ -24,4 +24,5 @@ dependencies {
     implementation("com.fasterxml.jackson.module:jackson-module-afterburner:${jacksonVersion}")
     implementation("com.fasterxml.jackson.datatype:jackson-datatype-jsr310:${jacksonVersion}")
     implementation("org.springframework.boot:spring-boot-starter-web:3.1.3")
+    implementation("org.jgrapht:jgrapht-core:1.5.2")
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java

@@ -47,10 +47,8 @@ import com.knecon.fforesight.service.layoutparser.processor.services.TableExtrac
 import com.knecon.fforesight.service.layoutparser.processor.services.blockification.DocstrumBlockificationService;
 import com.knecon.fforesight.service.layoutparser.processor.services.blockification.DocuMineBlockificationService;
 import com.knecon.fforesight.service.layoutparser.processor.services.blockification.RedactManagerBlockificationService;
-import com.knecon.fforesight.service.layoutparser.processor.services.blockification.TaasBlockificationService;
 import com.knecon.fforesight.service.layoutparser.processor.services.classification.DocuMineClassificationService;
 import com.knecon.fforesight.service.layoutparser.processor.services.classification.RedactManagerClassificationService;
-import com.knecon.fforesight.service.layoutparser.processor.services.classification.TaasClassificationService;
 import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.DocumentDataMapper;
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.TaasDocumentDataMapper;
@@ -78,14 +76,12 @@ public class LayoutParsingPipeline {
     CvTableParsingAdapter cvTableParsingAdapter;
     LayoutParsingStorageService layoutParsingStorageService;
     SectionsBuilderService sectionsBuilderService;
-    TaasClassificationService taasClassificationService;
     RedactManagerClassificationService redactManagerClassificationService;
     DocuMineClassificationService docuMineClassificationService;
     SimplifiedSectionTextService simplifiedSectionTextService;
     BodyTextFrameService bodyTextFrameService;
     RulingCleaningService rulingCleaningService;
     TableExtractionService tableExtractionService;
-    TaasBlockificationService taasBlockificationService;
     DocuMineBlockificationService docuMineBlockificationService;
     RedactManagerBlockificationService redactManagerBlockificationService;
     DocstrumBlockificationService docstrumBlockificationService;
@@ -100,40 +96,33 @@ 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 = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId()).orElse(originFile);
 
         VisualLayoutParsingResponse visualLayoutParsingResponse = new VisualLayoutParsingResponse();
-        if (layoutParsingRequest.visualLayoutParsingFileId()
-                .isPresent()) {
-            visualLayoutParsingResponse = layoutParsingStorageService.getVisualLayoutParsingFile(layoutParsingRequest.visualLayoutParsingFileId()
-                                                                                                         .get());
+        if (layoutParsingRequest.visualLayoutParsingFileId().isPresent()) {
+            visualLayoutParsingResponse = layoutParsingStorageService.getVisualLayoutParsingFile(layoutParsingRequest.visualLayoutParsingFileId().get());
         }
 
         ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
-        if (layoutParsingRequest.imagesFileStorageId()
-                .isPresent()) {
-            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId()
-                                                                                     .get());
+        if (layoutParsingRequest.imagesFileStorageId().isPresent()) {
+            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId().get());
         }
 
         TableServiceResponse tableServiceResponse = new TableServiceResponse();
-        if (layoutParsingRequest.tablesFileStorageId()
-                .isPresent()) {
-            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId()
-                                                                                     .get());
+        if (layoutParsingRequest.tablesFileStorageId().isPresent()) {
+            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId().get());
         }
 
         ClassificationDocument classificationDocument = parseLayout(layoutParsingRequest.layoutParsingType(),
-                                                                    originFile,
-                                                                    imageServiceResponse,
-                                                                    tableServiceResponse,
-                                                                    visualLayoutParsingResponse,
+                originFile,
+                imageServiceResponse,
+                tableServiceResponse,
+                visualLayoutParsingResponse,
                 layoutParsingRequest.identifier().toString());
 
         log.info("Building document graph for {}", layoutParsingRequest.identifier());
 
-        Document documentGraph = observeBuildDocumentGraph(classificationDocument);
+        Document documentGraph = observeBuildDocumentGraph(layoutParsingRequest.layoutParsingType(), classificationDocument);
 
         log.info("Creating viewer document for {}", layoutParsingRequest.identifier());
 
@@ -145,7 +134,7 @@ public class LayoutParsingPipeline {
         layoutParsingStorageService.storeSimplifiedText(layoutParsingRequest, simplifiedSectionTextService.toSimplifiedText(documentGraph));
         layoutParsingStorageService.storeViewerDocument(layoutParsingRequest, viewerDocumentFile);
 
-        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.TAAS)) {
+        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.CLARIFYND)) {
             log.info("Building research document data for {}", layoutParsingRequest.identifier());
             var researchDocumentData = TaasDocumentDataMapper.fromDocument(documentGraph);
             layoutParsingStorageService.storeResearchDocumentData(layoutParsingRequest, researchDocumentData);
@@ -161,37 +150,37 @@ public class LayoutParsingPipeline {
                 .numberOfPages(documentGraph.getNumberOfPages())
                 .duration(System.currentTimeMillis() - start)
                 .message(format("""
-                                        Layout parsing has finished in %.02f s.
-                                        identifiers: %s
-                                        %s
-                                        Files have been saved with Ids:
-                                        Structure: %s
-                                        Text: %s
-                                        Positions: %s
-                                        PageData: %s
-                                        Simplified Text: %s
-                                        Viewer Doc: %s""",
-                                ((float) (System.currentTimeMillis() - start)) / 1000,
-                                layoutParsingRequest.identifier(),
-                                buildSemanticNodeCountMessage(documentGraph.getNumberOfPages(), documentGraph.buildSemanticNodeCounts()),
-                                layoutParsingRequest.structureFileStorageId(),
-                                layoutParsingRequest.textBlockFileStorageId(),
-                                layoutParsingRequest.positionBlockFileStorageId(),
-                                layoutParsingRequest.pageFileStorageId(),
-                                layoutParsingRequest.simplifiedTextStorageId(),
-                                layoutParsingRequest.viewerDocumentStorageId()))
+                                Layout parsing has finished in %.02f s.
+                                identifiers: %s
+                                %s
+                                Files have been saved with Ids:
+                                Structure: %s
+                                Text: %s
+                                Positions: %s
+                                PageData: %s
+                                Simplified Text: %s
+                                Viewer Doc: %s""",
+                        ((float) (System.currentTimeMillis() - start)) / 1000,
+                        layoutParsingRequest.identifier(),
+                        buildSemanticNodeCountMessage(documentGraph.getNumberOfPages(), documentGraph.buildSemanticNodeCounts()),
+                        layoutParsingRequest.structureFileStorageId(),
+                        layoutParsingRequest.textBlockFileStorageId(),
+                        layoutParsingRequest.positionBlockFileStorageId(),
+                        layoutParsingRequest.pageFileStorageId(),
+                        layoutParsingRequest.simplifiedTextStorageId(),
+                        layoutParsingRequest.viewerDocumentStorageId()))
                 .build();
 
     }
 
 
-    private Document observeBuildDocumentGraph(ClassificationDocument classificationDocument) {
+    private Document observeBuildDocumentGraph(LayoutParsingType layoutParsingType, ClassificationDocument classificationDocument) {
 
         AtomicReference<Document> documentReference = new AtomicReference<>();
 
         Observation.createNotStarted("LayoutParsingPipeline", observationRegistry)
                 .contextualName("build-document-graph")
-                .observe(() -> documentReference.set(DocumentGraphFactory.buildDocumentGraph(classificationDocument)));
+                .observe(() -> documentReference.set(DocumentGraphFactory.buildDocumentGraph(layoutParsingType, classificationDocument)));
 
         return documentReference.get();
     }
@@ -200,14 +189,14 @@ public class LayoutParsingPipeline {
     private String buildSemanticNodeCountMessage(int numberOfPages, Map<NodeType, Long> semanticNodeCounts) {
 
         return String.format("%d pages with %d sections, %d headlines, %d paragraphs, %d tables with %d cells, %d headers, and %d footers parsed",
-                             numberOfPages,
-                             semanticNodeCounts.get(NodeType.SECTION) == null ? 0 : semanticNodeCounts.get(NodeType.SECTION),
-                             semanticNodeCounts.get(NodeType.HEADLINE) == null ? 0 : semanticNodeCounts.get(NodeType.HEADLINE),
-                             semanticNodeCounts.get(NodeType.PARAGRAPH) == null ? 0 : semanticNodeCounts.get(NodeType.PARAGRAPH),
-                             semanticNodeCounts.get(NodeType.TABLE) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE),
-                             semanticNodeCounts.get(NodeType.TABLE_CELL) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE_CELL),
-                             semanticNodeCounts.get(NodeType.HEADER) == null ? 0 : semanticNodeCounts.get(NodeType.HEADER),
-                             semanticNodeCounts.get(NodeType.FOOTER) == null ? 0 : semanticNodeCounts.get(NodeType.FOOTER));
+                numberOfPages,
+                semanticNodeCounts.get(NodeType.SECTION) == null ? 0 : semanticNodeCounts.get(NodeType.SECTION),
+                semanticNodeCounts.get(NodeType.HEADLINE) == null ? 0 : semanticNodeCounts.get(NodeType.HEADLINE),
+                semanticNodeCounts.get(NodeType.PARAGRAPH) == null ? 0 : semanticNodeCounts.get(NodeType.PARAGRAPH),
+                semanticNodeCounts.get(NodeType.TABLE) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE),
+                semanticNodeCounts.get(NodeType.TABLE_CELL) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE_CELL),
+                semanticNodeCounts.get(NodeType.HEADER) == null ? 0 : semanticNodeCounts.get(NodeType.HEADER),
+                semanticNodeCounts.get(NodeType.FOOTER) == null ? 0 : semanticNodeCounts.get(NodeType.FOOTER));
     }
 
 
@@ -263,15 +252,16 @@ public class LayoutParsingPipeline {
             PDRectangle cropbox = pdPage.getCropBox();
             CleanRulings cleanRulings = rulingCleaningService.getCleanRulings(pdfTableCells.get(pageNumber), stripper.getRulings());
 
-            List<Cell> emptyTableCells = tableExtractionService.findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical());
+            List<Cell> emptyTableCells = TableExtractionService.findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical());
 
             ClassificationPage classificationPage = switch (layoutParsingType) {
-                case REDACT_MANAGER -> redactManagerBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
-                case TAAS -> taasBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
+                case REDACT_MANAGER_OLD ->
+                        redactManagerBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
                 case DOCUMINE -> docuMineBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
-                case DOCSTRUM -> docstrumBlockificationService.blockify(stripper.getTextPositionSequences(), emptyTableCells, false);
-                case DOCSTRUM_XY -> docstrumBlockificationService.blockify(stripper.getTextPositionSequences(), emptyTableCells, true);
+                case REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG -> docstrumBlockificationService.blockify(stripper.getTextPositionSequences(), emptyTableCells, true);
+                case CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG -> docstrumBlockificationService.blockify(stripper.getTextPositionSequences(), emptyTableCells, false);
             };
+
             classificationPage.setCleanRulings(cleanRulings);
             classificationPage.setRotation(rotation);
             classificationPage.setLandscape(isLandscape);
@@ -298,8 +288,10 @@ public class LayoutParsingPipeline {
 
             tableExtractionService.extractTables(emptyTableCells, classificationPage);
 
-            if (layoutParsingType == LayoutParsingType.DOCSTRUM || layoutParsingType == LayoutParsingType.DOCSTRUM_XY) {
+            if (layoutParsingType == LayoutParsingType.REDACT_MANAGER) {
                 docstrumBlockificationService.combineBlocks(classificationPage);
+            } else if (layoutParsingType == LayoutParsingType.CLARIFYND) {
+                docstrumBlockificationService.mergeZones(classificationPage.getTextBlocks());
             }
 
             buildPageStatistics(classificationPage);
@@ -314,30 +306,21 @@ public class LayoutParsingPipeline {
         bodyTextFrameService.setBodyTextFrames(classificationDocument, layoutParsingType);
         log.info("Classify TextBlocks for {}", identifier);
         switch (layoutParsingType) {
-            case TAAS -> taasClassificationService.classifyDocument(classificationDocument);
+            case REDACT_MANAGER, REDACT_MANAGER_PARAGRAPH_DEBUG, REDACT_MANAGER_OLD, CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG ->
+                    redactManagerClassificationService.classifyDocument(classificationDocument);
             case DOCUMINE -> docuMineClassificationService.classifyDocument(classificationDocument);
-            case REDACT_MANAGER -> redactManagerClassificationService.classifyDocument(classificationDocument);
-            case DOCSTRUM_XY -> redactManagerClassificationService.classifyDocument(classificationDocument);
-            case DOCSTRUM -> redactManagerClassificationService.classifyDocument(classificationDocument);
         }
 
         log.info("Building Sections for {}", identifier);
 
-//        if (layoutParsingType == DOCSTRUM || layoutParsingType == DOCSTRUM_XY) {
-//            // Currently for debugging return paragraphs as sections, because there is a merging logic in sectionBuilder
-//            List<ClassificationSection> sections = new ArrayList<>();
-//            for (var page : classificationPages) {
-//                page.getTextBlocks().forEach(block -> {
-//                    block.setPage(page.getPageNumber());
-//                    var section = sectionsBuilderService.buildTextBlock(List.of(block), "a");
-//                    sections.add(section);
-//                });
-//            }
-//            classificationDocument.setSections(sections);
-//        } else {
-        sectionsBuilderService.buildSections(classificationDocument);
-        sectionsBuilderService.addImagesToSections(classificationDocument);
-//        }
+        switch (layoutParsingType) {
+            case CLARIFYND_PARAGRAPH_DEBUG, REDACT_MANAGER_PARAGRAPH_DEBUG -> sectionsBuilderService.buildParagraphDebugSections(classificationDocument);
+            default -> {
+                sectionsBuilderService.buildSections(classificationDocument);
+                sectionsBuilderService.addImagesToSections(classificationDocument);
+            }
+        }
+
         return classificationDocument;
     }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/BodyTextFrameService.java

@@ -25,6 +25,7 @@ public class BodyTextFrameService {
     private static final float RULING_HEIGHT_THRESHOLD = 0.15f; // multiplied with page height. Header/Footer Rulings must be within that border of the page.
     private static final float RULING_WIDTH_THRESHOLD = 0.75f; // multiplied with page width. Header/Footer Rulings must be at least that wide.
 
+
     public void setBodyTextFrames(ClassificationDocument classificationDocument, LayoutParsingType layoutParsingType) {
 
         Rectangle bodyTextFrame = calculateBodyTextFrame(classificationDocument.getPages(), classificationDocument.getFontSizeCounter(), false, layoutParsingType);
@@ -132,12 +133,7 @@ public class BodyTextFrameService {
                                                boolean landscape,
                                                LayoutParsingType layoutParsingType) {
 
-        float approximateHeaderLineCount;
-        if (layoutParsingType.equals(LayoutParsingType.TAAS)) {
-            approximateHeaderLineCount = 3.3f;
-        } else {
-            approximateHeaderLineCount = 2.9f;
-        }
+        float approximateHeaderLineCount = 2.9f;
 
         BodyTextFrameExpansionsRectangle expansionsRectangle = new BodyTextFrameExpansionsRectangle();
 
@@ -155,8 +151,9 @@ public class BodyTextFrameService {
                         continue;
                     }
 
-                    if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.HEADER)
-                            || MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.FOOTER)) {
+                    if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.HEADER) || MarkedContentUtils.intersects(textBlock,
+                            page.getMarkedContentBboxPerType(),
+                            MarkedContentUtils.FOOTER)) {
                         continue;
                     }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/SectionsBuilderService.java

@@ -7,6 +7,7 @@ import java.util.List;
 import java.util.Map;
 import java.util.stream.Collectors;
 
+import org.apache.logging.log4j.util.Strings;
 import org.springframework.stereotype.Service;
 
 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
@@ -110,6 +111,20 @@ public class SectionsBuilderService {
     }
 
 
+    public void buildParagraphDebugSections(ClassificationDocument document) {
+
+        List<ClassificationSection> sections = new ArrayList<>();
+        for (var page : document.getPages()) {
+            page.getTextBlocks().forEach(block -> {
+                block.setPage(page.getPageNumber());
+                var section = buildTextBlock(List.of(block), Strings.EMPTY);
+                sections.add(section);
+            });
+        }
+        document.setSections(sections);
+    }
+
+
     public void addImagesToSections(ClassificationDocument document) {
 
         Map<Integer, List<ClassificationSection>> sectionMap = new HashMap<>();
@@ -240,7 +255,7 @@ public class SectionsBuilderService {
     }
 
 
-    public ClassificationSection buildTextBlock(List<AbstractPageBlock> wordBlockList, String lastHeadline) {
+    private ClassificationSection buildTextBlock(List<AbstractPageBlock> wordBlockList, String lastHeadline) {
 
         ClassificationSection section = new ClassificationSection();
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/blockification/DocstrumBlockificationService.java

@@ -38,7 +38,7 @@ public class DocstrumBlockificationService {
     static final float THRESHOLD = 1f;
 
 
-    public ClassificationPage blockify(List<TextPositionSequence> textPositions, List<Cell> cells, boolean xyOder) {
+    public ClassificationPage blockify(List<TextPositionSequence> textPositions, List<Cell> cells, boolean xyOrder) {
 
         // Underlined or strikethrough are also in rulings but we dont want to split blocks with them so we use cells.
         List<Ruling> usedHorizonalRulings = new ArrayList<>();
@@ -52,7 +52,7 @@ public class DocstrumBlockificationService {
         });
 
         List<AbstractPageBlock> abstractPageBlocks = new ArrayList<>();
-        var zones = docstrumSegmentationService.segmentPage(textPositions, xyOder);
+        var zones = docstrumSegmentationService.segmentPage(textPositions, xyOrder);
         zones.forEach(zone -> {
 
             List<TextPositionSequence> textPositionSequences = new ArrayList<>();
@@ -63,7 +63,6 @@ public class DocstrumBlockificationService {
             });
 
             abstractPageBlocks.addAll(splitZonesAtRulings(textPositionSequences, usedHorizonalRulings, usedVerticalRulings));
-//            abstractPageBlocks.add(buildTextBlock(textPositionSequences, 0));
         });
 
         return new ClassificationPage(abstractPageBlocks);
@@ -147,6 +146,7 @@ public class DocstrumBlockificationService {
         }
 
         mergeZones(page.getTextBlocks());
+
     }
 
 
@@ -188,7 +188,7 @@ public class DocstrumBlockificationService {
     }
 
 
-    private void mergeZones(List<AbstractPageBlock> zones) {
+    public void mergeZones(List<AbstractPageBlock> zones) {
 
         ListIterator<AbstractPageBlock> itty = zones.listIterator();
         Set<AbstractPageBlock> toRemove = new HashSet<>();

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/blockification/TaasBlockificationService.java

@@ -1,330 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services.blockification;
-
-
-// TODO: figure out, why this fails the build
-// import static com.knecon.fforesight.service.layoutparser.processor.services.factory.SearchTextWithTextPositionFactory.HEIGHT_PADDING;
-
-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.Orientation;
-import com.knecon.fforesight.service.layoutparser.processor.model.table.Ruling;
-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.utils.RulingTextDirAdjustUtil;
-import org.springframework.stereotype.Service;
-
-import java.util.*;
-import java.util.regex.Matcher;
-import java.util.regex.Pattern;
-import java.util.stream.Stream;
-
-@Service
-@SuppressWarnings("all")
-public class TaasBlockificationService {
-
-    private static final float THRESHOLD = 1f;
-    private static final float Y_GAP_SPLIT_HEIGHT_MODIFIER = 1.25f; // multiplied with text height
-    private static final float INTERSECTS_Y_THRESHOLD = 4;// 2 * HEIGHT_PADDING // This is exactly 2 times our position height padding. This is required to find boxes that are visually intersecting.
-    private static final int X_GAP_SPLIT_CONSTANT = 50;
-    public static final int X_ALIGNMENT_THRESHOLD = 1;
-    public static final int NEGATIVE_X_GAP_THRESHOLD = -5;
-
-    private Pattern listIdentifier = Pattern.compile("^(?:(?:[1-9]|1\\d|20|[ivxlc]|[a-z])\\s*(?:[.)]))|\\uF0B7", Pattern.CASE_INSENSITIVE);
-
-
-    /**
-     * This method is building blocks by expanding the minX/maxX and minY/maxY value on each word that is not split by the conditions.
-     * This method must use text direction adjusted postions (DirAdj). Where {0,0} is on the upper left. Never try to change this!
-     * Rulings (Table lines) must be adjusted to the text directions as well, when checking if a block is split by a ruling.
-     *
-     * @param textPositions         The words of a page.
-     * @param horizontalRulingLines Horizontal table lines.
-     * @param verticalRulingLines   Vertical table lines.
-     * @return ClassificationPage object that contains the Textblock and text statistics.
-     */
-    public ClassificationPage blockify(List<TextPositionSequence> textPositions, List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines) {
-
-        List<TextPageBlock> classificationTextBlocks = constructFineGranularTextPageBlocks(textPositions, horizontalRulingLines, verticalRulingLines);
-        classificationTextBlocks = mergeTextPageBlocksAligningX(classificationTextBlocks);
-        classificationTextBlocks = mergeIntersectingTextBlocksUntilConvergence(classificationTextBlocks);
-
-        return new ClassificationPage(new ArrayList<>(classificationTextBlocks.stream().map(classificationTextBlock -> (AbstractPageBlock) classificationTextBlock).toList()));
-    }
-
-
-    private List<TextPageBlock> mergeIntersectingTextBlocksUntilConvergence(List<TextPageBlock> classificationTextBlocks) {
-
-        int currentSize = classificationTextBlocks.size();
-        while (true) {
-            classificationTextBlocks = mergeTextPageBlocksAlmostIntersecting(classificationTextBlocks);
-            if (classificationTextBlocks.size() == currentSize) {
-                break;
-            }
-            currentSize = classificationTextBlocks.size();
-        }
-        return classificationTextBlocks;
-    }
-
-
-    private List<TextPageBlock> mergeTextPageBlocksAligningX(List<TextPageBlock> classificationTextBlocks) {
-
-        if (classificationTextBlocks.isEmpty()) {
-            return new ArrayList<>();
-        }
-        List<List<TextPageBlock>> textBlocksToMerge = new LinkedList<>();
-        List<TextPageBlock> currentTextBlocksToMerge = new LinkedList<>();
-        textBlocksToMerge.add(currentTextBlocksToMerge);
-        TextPageBlock previousTextBlock = null;
-        Float lastLineGap = null;
-        for (TextPageBlock currentTextBlock : classificationTextBlocks) {
-            if (previousTextBlock == null) {
-                currentTextBlocksToMerge.add(currentTextBlock);
-                previousTextBlock = currentTextBlock;
-                continue;
-            }
-
-
-            Matcher listIdentifierPattern = listIdentifier.matcher(currentTextBlock.getText());
-            boolean isListIdentifier = listIdentifierPattern.find();
-
-            boolean yGap = Math.abs(currentTextBlock.getPdfMaxY() - previousTextBlock.getPdfMinY()) < previousTextBlock.getMostPopularWordHeight() * Y_GAP_SPLIT_HEIGHT_MODIFIER;
-
-            boolean sameFont = previousTextBlock.getMostPopularWordFont().equals(currentTextBlock.getMostPopularWordFont()) && previousTextBlock.getMostPopularWordFontSize() == currentTextBlock.getMostPopularWordFontSize();
-//            boolean yGap = previousTextBlock != null && currentTextBlock.getMinYDirAdj() - maxY > Math.min(word.getHeight(), prev.getHeight()) * Y_GAP_SPLIT_HEIGHT_MODIFIER;
-
-            boolean alignsXRight = Math.abs(currentTextBlock.getPdfMaxX() - previousTextBlock.getPdfMaxX()) < X_ALIGNMENT_THRESHOLD;
-            boolean alignsXLeft = Math.abs(currentTextBlock.getPdfMinX() - previousTextBlock.getPdfMinX()) < X_ALIGNMENT_THRESHOLD;
-//            boolean smallYGap = Math.abs(currentTextBlock.getPdfMaxY() - previousTextBlock.getPdfMinY()) < yGap;
-            if (yGap && sameFont && !isListIdentifier) {
-                currentTextBlocksToMerge.add(currentTextBlock);
-
-            } else {
-                currentTextBlocksToMerge = new LinkedList<>();
-                currentTextBlocksToMerge.add(currentTextBlock);
-                textBlocksToMerge.add(currentTextBlocksToMerge);
-            }
-            previousTextBlock = currentTextBlock;
-        }
-        return textBlocksToMerge.stream().map(TextPageBlock::merge).toList();
-    }
-
-
-    private List<TextPageBlock> mergeTextPageBlocksAlmostIntersecting(List<TextPageBlock> textPageBlocks) {
-
-        Set<TextPageBlock> alreadyMerged = new HashSet<>();
-        List<List<TextPageBlock>> textBlocksToMerge = new LinkedList<>();
-        for (TextPageBlock textPageBlock : textPageBlocks) {
-            if (alreadyMerged.contains(textPageBlock)) {
-                continue;
-            }
-            alreadyMerged.add(textPageBlock);
-            textBlocksToMerge.add(Stream.concat(Stream.of(textPageBlock),
-                            textPageBlocks.stream().filter(textPageBlock2 -> textPageBlock.almostIntersects(textPageBlock2, INTERSECTS_Y_THRESHOLD, 0) && !alreadyMerged.contains(textPageBlock2)).peek(alreadyMerged::add))
-                    .toList());
-        }
-        return textBlocksToMerge.stream().map(TextPageBlock::merge).toList();
-    }
-
-
-    private void assignOrientations(List<TextPageBlock> classificationTextBlocks) {
-
-        Iterator<TextPageBlock> itty = classificationTextBlocks.iterator();
-
-        TextPageBlock previousLeft = null;
-        TextPageBlock previousRight = null;
-        while (itty.hasNext()) {
-            TextPageBlock block = (TextPageBlock) itty.next();
-
-            if (previousLeft != null && block.getOrientation().equals(Orientation.LEFT)) {
-                if (previousLeft.getMinY() > block.getMinY() && block.getMaxY() + block.getMostPopularWordHeight() > previousLeft.getMinY()) {
-                    previousLeft.add(block);
-                    itty.remove();
-                    continue;
-                }
-            }
-
-            if (previousRight != null && block.getOrientation().equals(Orientation.RIGHT)) {
-                if (previousRight.getMinY() > block.getMinY() && block.getMaxY() + block.getMostPopularWordHeight() > previousRight.getMinY()) {
-                    previousRight.add(block);
-                    itty.remove();
-                    continue;
-                }
-            }
-
-            if (block.getOrientation().equals(Orientation.LEFT)) {
-                previousLeft = block;
-            } else if (block.getOrientation().equals(Orientation.RIGHT)) {
-                previousRight = block;
-            }
-        }
-
-        itty = classificationTextBlocks.iterator();
-        TextPageBlock previous = null;
-        while (itty.hasNext()) {
-            TextPageBlock block = (TextPageBlock) itty.next();
-
-            if (previous != null && previous.getOrientation().equals(Orientation.LEFT) && block.getOrientation().equals(Orientation.LEFT) && equalsWithThreshold(
-                    block.getMaxY(),
-                    previous.getMaxY()) || previous != null && previous.getOrientation().equals(Orientation.LEFT) && block.getOrientation()
-                    .equals(Orientation.RIGHT) && equalsWithThreshold(block.getMaxY(), previous.getMaxY())) {
-                previous.add(block);
-                itty.remove();
-                continue;
-            }
-
-            previous = block;
-        }
-    }
-
-
-    private List<TextPageBlock> constructFineGranularTextPageBlocks(List<TextPositionSequence> textPositions,
-                                                                    List<Ruling> horizontalRulingLines,
-                                                                    List<Ruling> verticalRulingLines) {
-
-        int indexOnPage = 0;
-        List<TextPositionSequence> wordClusterToCombine = new ArrayList<>();
-        List<TextPageBlock> classificationTextBlocks = new ArrayList<>();
-
-        float minX = 1000, maxX = 0, minY = 1000, maxY = 0;
-        TextPositionSequence prev = null;
-        // TODO: make static final constant
-
-
-        boolean wasSplitted = false;
-        Float splitX1 = null;
-        for (TextPositionSequence word : textPositions) {
-
-            Matcher listIdentifierPattern = listIdentifier.matcher(word.toString());
-
-            boolean yGap = prev != null && word.getMinYDirAdj() - maxY > Math.min(word.getHeight(), prev.getHeight()) * Y_GAP_SPLIT_HEIGHT_MODIFIER;
-            boolean sameLine = prev != null && equalsWithThreshold(prev.getMinYDirAdj(), word.getMinYDirAdj());
-            boolean positiveXGapInline = prev != null && maxX + X_GAP_SPLIT_CONSTANT < word.getMinXDirAdj() && sameLine;
-            boolean negativeXGap = prev != null && word.getMinXDirAdj() - minX < NEGATIVE_X_GAP_THRESHOLD;
-            boolean startFromTop = prev != null && word.getMinYDirAdj() < prev.getMinYDirAdj() - prev.getTextHeight();
-            boolean newLineAfterSplit = prev != null && word.getMinYDirAdj() != prev.getMinYDirAdj() && wasSplitted && splitX1 != word.getMinXDirAdj();
-            boolean splitByRuling = isSplitByRuling(minX, minY, maxX, maxY, word, horizontalRulingLines, verticalRulingLines);
-            boolean splitByDir = prev != null && !prev.getDir().equals(word.getDir());
-            boolean fontChange = prev != null && (!word.getFont().equals(prev.getFont()) || !word.getFontStyle()
-                    .equals(prev.getFontStyle()) || word.getFontSize() != prev.getFontSize());
-            boolean newline = prev != null && Math.abs(word.getMinYDirAdj() - prev.getMinYDirAdj()) > word.getHeight();
-            boolean isListIdentifier = listIdentifierPattern.matches();
-
-            if (prev != null && (prev.isParagraphStart() || negativeXGap || positiveXGapInline || yGap || startFromTop || splitByRuling || (newline && (fontChange || isListIdentifier)))) {
-//            if (prev != null && (lineSeparation || startFromTop || splitByX || splitByDir || isSplitByRuling)) {
-
-                Orientation prevOrientation = null;
-                if (!classificationTextBlocks.isEmpty()) {
-                    prevOrientation = classificationTextBlocks.get(classificationTextBlocks.size() - X_ALIGNMENT_THRESHOLD).getOrientation();
-                }
-
-                TextPageBlock classificationTextBlock = TextPageBlock.fromTextPositionSequences(wordClusterToCombine);
-
-                classificationTextBlocks.add(classificationTextBlock);
-                wordClusterToCombine = new ArrayList<>();
-
-                if (positiveXGapInline && !splitByRuling) {
-                    wasSplitted = true;
-                    classificationTextBlock.setOrientation(Orientation.LEFT);
-                    splitX1 = word.getMinXDirAdj();
-                } else if (newLineAfterSplit && !splitByRuling) {
-                    wasSplitted = false;
-                    classificationTextBlock.setOrientation(Orientation.RIGHT);
-                    splitX1 = null;
-                } else if (prevOrientation != null && prevOrientation.equals(Orientation.RIGHT) && (yGap || !startFromTop || !positiveXGapInline || !newLineAfterSplit || !splitByRuling)) {
-                    classificationTextBlock.setOrientation(Orientation.LEFT);
-                }
-
-                minX = 1000;
-                maxX = 0;
-                minY = 1000;
-                maxY = 0;
-                prev = null;
-            }
-
-            wordClusterToCombine.add(word);
-
-            prev = word;
-            if (word.getMinXDirAdj() < minX) {
-                minX = word.getMinXDirAdj();
-            }
-            if (word.getMaxXDirAdj() > maxX) {
-                maxX = word.getMaxXDirAdj();
-            }
-            if (word.getMinYDirAdj() < minY) {
-                minY = word.getMinYDirAdj();
-            }
-            if (word.getMaxYDirAdj() > maxY) {
-                maxY = word.getMaxYDirAdj();
-            }
-        }
-
-        TextPageBlock classificationTextBlock = TextPageBlock.fromTextPositionSequences(wordClusterToCombine);
-        if (classificationTextBlock != null) {
-            classificationTextBlocks.add(classificationTextBlock);
-        }
-        return classificationTextBlocks;
-    }
-
-
-    private boolean equalsWithThreshold(float f1, float f2) {
-
-        return Math.abs(f1 - f2) < THRESHOLD;
-    }
-
-
-    private boolean isSplitByRuling(float minX,
-                                    float minY,
-                                    float maxX,
-                                    float maxY,
-                                    TextPositionSequence word,
-                                    List<Ruling> horizontalRulingLines,
-                                    List<Ruling> verticalRulingLines) {
-
-        return isSplitByRuling(maxX,
-                minY,
-                word.getMinXDirAdj(),
-                word.getMinYDirAdj(),
-                verticalRulingLines,
-                word.getDir().getDegrees(),
-                word.getPageWidth(),
-                word.getPageHeight()) //
-                || isSplitByRuling(minX,
-                minY,
-                word.getMinXDirAdj(),
-                word.getMaxYDirAdj(),
-                horizontalRulingLines,
-                word.getDir().getDegrees(),
-                word.getPageWidth(),
-                word.getPageHeight()) //
-                || isSplitByRuling(maxX,
-                minY,
-                word.getMinXDirAdj(),
-                word.getMinYDirAdj(),
-                horizontalRulingLines,
-                word.getDir().getDegrees(),
-                word.getPageWidth(),
-                word.getPageHeight()) //
-                || isSplitByRuling(minX,
-                minY,
-                word.getMinXDirAdj(),
-                word.getMaxYDirAdj(),
-                verticalRulingLines,
-                word.getDir().getDegrees(),
-                word.getPageWidth(),
-                word.getPageHeight()); //
-    }
-
-
-    private boolean isSplitByRuling(float previousX2, float previousY1, float currentX1, float currentY1, List<Ruling> rulingLines, float dir, float pageWidth, float pageHeight) {
-
-        for (Ruling ruling : rulingLines) {
-            var line = RulingTextDirAdjustUtil.convertToDirAdj(ruling, dir, pageWidth, pageHeight);
-            if (line.intersectsLine(previousX2, previousY1, currentX1, currentY1)) {
-                return true;
-            }
-        }
-        return false;
-    }
-
-}
-

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/classification/TaasClassificationService.java

@@ -1,114 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services.classification;
-
-import java.util.List;
-import java.util.regex.Pattern;
-
-import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;
-import org.springframework.stereotype.Service;
-
-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;
-import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
-import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPageBlock;
-import com.knecon.fforesight.service.layoutparser.processor.services.BodyTextFrameService;
-import com.knecon.fforesight.service.layoutparser.processor.utils.PositionUtils;
-
-import lombok.RequiredArgsConstructor;
-import lombok.extern.slf4j.Slf4j;
-
-@Slf4j
-@Service
-@RequiredArgsConstructor
-public class TaasClassificationService {
-
-    private final BodyTextFrameService bodyTextFrameService;
-
-
-    public void classifyDocument(ClassificationDocument document) {
-
-
-        List<Float> headlineFontSizes = document.getFontSizeCounter().getHighterThanMostPopular();
-
-        log.debug("Document FontSize counters are: {}", document.getFontSizeCounter().getCountPerValue());
-
-        for (ClassificationPage page : document.getPages()) {
-
-            classifyPage(page, document, headlineFontSizes);
-        }
-    }
-
-
-    public void classifyPage(ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
-
-        for (AbstractPageBlock textBlock : page.getTextBlocks()) {
-            if (textBlock instanceof TextPageBlock) {
-                classifyBlock((TextPageBlock) textBlock, page, document, headlineFontSizes);
-            }
-        }
-    }
-
-
-    public void classifyBlock(TextPageBlock textBlock, ClassificationPage page, ClassificationDocument document, List<Float> headlineFontSizes) {
-
-        var bodyTextFrame = page.getBodyTextFrame();
-
-        if (document.getFontSizeCounter().getMostPopular() == null) {
-            textBlock.setClassification(PageBlockType.OTHER);
-            return;
-        }
-        if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.HEADER)
-                || PositionUtils.isOverBodyTextFrame(bodyTextFrame, textBlock, page.getRotation())) {
-            textBlock.setClassification(PageBlockType.HEADER);
-        } else if (MarkedContentUtils.intersects(textBlock, page.getMarkedContentBboxPerType(), MarkedContentUtils.FOOTER)
-                || PositionUtils.isUnderBodyTextFrame(bodyTextFrame, textBlock, page.getRotation())) {
-            textBlock.setClassification(PageBlockType.FOOTER);
-        } else if (page.getPageNumber() == 1 && (PositionUtils.getHeightDifferenceBetweenChunkWordAndDocumentWord(textBlock,
-                document.getTextHeightCounter().getMostPopular()) > 2.5 && textBlock.getHighestFontSize() > document.getFontSizeCounter().getMostPopular() || page.getTextBlocks()
-                .size() == 1)) {
-            if (!Pattern.matches("[0-9]+", textBlock.toString())) {
-                textBlock.setClassification(PageBlockType.TITLE);
-            }
-        } else if (textBlock.getMostPopularWordFontSize() > document.getFontSizeCounter()
-                .getMostPopular() && PositionUtils.getApproxLineCount(textBlock) < 4.9 && (textBlock.getMostPopularWordStyle().equals("bold") || !document.getFontStyleCounter()
-                .getCountPerValue()
-                .containsKey("bold") && textBlock.getMostPopularWordFontSize() > document.getFontSizeCounter().getMostPopular() + 1) && textBlock.getSequences()
-                .get(0)
-                .getTextPositions()
-                .get(0)
-                .getFontSizeInPt() >= textBlock.getMostPopularWordFontSize()) {
-
-            for (int i = 1; i <= headlineFontSizes.size(); i++) {
-                if (textBlock.getMostPopularWordFontSize() == headlineFontSizes.get(i - 1)) {
-                    textBlock.setClassification(PageBlockType.getHeadlineType(i));
-                    document.setHeadlines(true);
-                }
-            }
-        } else if (!textBlock.getText().startsWith("Figure ") && PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getMostPopularWordStyle()
-                .equals("bold") && !document.getFontStyleCounter().getMostPopular().equals("bold") && PositionUtils.getApproxLineCount(textBlock) < 2.9 && textBlock.getSequences()
-                .get(0)
-                .getTextPositions()
-                .get(0)
-                .getFontSizeInPt() >= textBlock.getMostPopularWordFontSize()) {
-            textBlock.setClassification(PageBlockType.getHeadlineType(headlineFontSizes.size() + 1));
-            document.setHeadlines(true);
-        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getMostPopularWordFontSize() == document.getFontSizeCounter()
-                .getMostPopular() && textBlock.getMostPopularWordStyle().equals("bold") && !document.getFontStyleCounter().getMostPopular().equals("bold")) {
-            textBlock.setClassification(PageBlockType.PARAGRAPH_BOLD);
-        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getMostPopularWordFont()
-                .equals(document.getFontCounter().getMostPopular()) && textBlock.getMostPopularWordStyle()
-                .equals(document.getFontStyleCounter().getMostPopular()) && textBlock.getMostPopularWordFontSize() == document.getFontSizeCounter().getMostPopular()) {
-            textBlock.setClassification(PageBlockType.PARAGRAPH);
-        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock) && textBlock.getMostPopularWordFontSize() == document.getFontSizeCounter()
-                .getMostPopular() && textBlock.getMostPopularWordStyle().equals("italic") && !document.getFontStyleCounter()
-                .getMostPopular()
-                .equals("italic") && PositionUtils.getApproxLineCount(textBlock) < 2.9) {
-            textBlock.setClassification(PageBlockType.PARAGRAPH_ITALIC);
-        } else if (PositionUtils.isWithinBodyTextFrame(bodyTextFrame, textBlock)) {
-            textBlock.setClassification(PageBlockType.PARAGRAPH_UNKNOWN);
-        } else {
-            textBlock.setClassification(PageBlockType.OTHER);
-        }
-    }
-
-}

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/DocstrumSegmentationService.java

@@ -29,7 +29,7 @@ public class DocstrumSegmentationService {
     private final ReadingOrderService readingOrderService;
 
 
-    public List<Zone> segmentPage(List<TextPositionSequence> textPositions, boolean xyOder) {
+    public List<Zone> segmentPage(List<TextPositionSequence> textPositions, boolean xyOrder) {
 
         List<Zone> zones = new ArrayList<>();
         zones.addAll(computeZones(textPositions, TextDirection.ZERO));
@@ -37,7 +37,7 @@ public class DocstrumSegmentationService {
         zones.addAll(computeZones(textPositions, TextDirection.HALF_CIRCLE));
         zones.addAll(computeZones(textPositions, TextDirection.THREE_QUARTER_CIRCLE));
 
-        return readingOrderService.resolve(zones, xyOder);
+        return readingOrderService.resolve(zones, xyOrder);
     }
 
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/model/BoundingBox.java

@@ -42,7 +42,16 @@ public abstract class BoundingBox {
 
     public boolean contains(Rectangle2D contained, double tolerance) {
 
-        return bBox.getX() <= contained.getX() + tolerance && bBox.getY() <= contained.getY() + tolerance && bBox.getX() + bBox.getWidth() >= contained.getX() + contained.getWidth() - tolerance && bBox.getY() + bBox.getHeight() >= contained.getY() + contained.getHeight() - tolerance;
+        return bBox.getX() <= contained.getX() + tolerance
+               && bBox.getY() <= contained.getY() + tolerance
+               && bBox.getX() + bBox.getWidth() >= contained.getX() + contained.getWidth() - tolerance
+               && bBox.getY() + bBox.getHeight() >= contained.getY() + contained.getHeight() - tolerance;
+    }
+
+
+    public boolean intersectsY(BoundingBox other) {
+
+        return this.getBBox().getMinY() <= other.getBBox().getMaxY() && this.getBBox().getMaxY() >= other.getBBox().getMinY();
     }
 
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/model/DisjointSets.java

@@ -1,194 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model;
-
-import java.util.AbstractSet;
-import java.util.Collection;
-import java.util.HashMap;
-import java.util.Iterator;
-import java.util.Map;
-import java.util.NoSuchElementException;
-import java.util.Set;
-
-public class DisjointSets<E> implements Iterable<Set<E>> {
-
-    private final Map<E, Entry<E>> map = new HashMap<>();
-
-
-    public DisjointSets(Collection<? extends E> collection) {
-
-        for (E element : collection) {
-            map.put(element, new Entry<E>(element));
-        }
-    }
-
-
-    public boolean areTogether(E e1, E e2) {
-
-        return map.get(e1).findRepresentative().equals(map.get(e2).findRepresentative());
-    }
-
-
-    public void union(E e1, E e2) {
-
-        Entry<E> r1 = map.get(e1).findRepresentative();
-        Entry<E> r2 = map.get(e2).findRepresentative();
-        if (!r1.equals(r2)) {
-            if (r1.size <= r2.size) {
-                r2.mergeWith(r1);
-            } else {
-                r1.mergeWith(r2);
-            }
-        }
-    }
-
-
-    @Override
-    public Iterator<Set<E>> iterator() {
-
-        return new Iterator<>() {
-
-            private final Iterator<Entry<E>> iterator = map.values().iterator();
-            private Entry<E> nextRepresentative;
-
-            {
-                findNextRepresentative();
-            }
-
-            @Override
-            public boolean hasNext() {
-
-                return nextRepresentative != null;
-            }
-
-
-            @Override
-            public Set<E> next() {
-
-                if (nextRepresentative == null) {
-                    throw new NoSuchElementException();
-                }
-                Set<E> result = nextRepresentative.asSet();
-                findNextRepresentative();
-                return result;
-            }
-
-
-            private void findNextRepresentative() {
-
-                while (iterator.hasNext()) {
-                    Entry<E> candidate = iterator.next();
-                    if (candidate.isRepresentative()) {
-                        nextRepresentative = candidate;
-                        return;
-                    }
-                }
-                nextRepresentative = null;
-            }
-
-
-            @Override
-            public void remove() {
-
-                throw new UnsupportedOperationException();
-            }
-
-        };
-    }
-
-
-    private static class Entry<E> {
-
-        private int size = 1;
-        private final E value;
-        private Entry<E> parent = this;
-        private Entry<E> next;
-        private Entry<E> last = this;
-
-
-        Entry(E value) {
-
-            this.value = value;
-        }
-
-
-        void mergeWith(Entry<E> otherRepresentative) {
-
-            size += otherRepresentative.size;
-            last.next = otherRepresentative;
-            last = otherRepresentative.last;
-            otherRepresentative.parent = this;
-        }
-
-
-        Entry<E> findRepresentative() {
-
-            Entry<E> representative = parent;
-            while (!representative.parent.equals(representative)) {
-                representative = representative.parent;
-            }
-            for (Entry<E> entry = this; !entry.equals(representative); ) {
-                Entry<E> nextEntry = entry.parent;
-                entry.parent = representative;
-                entry = nextEntry;
-            }
-            return representative;
-        }
-
-
-        boolean isRepresentative() {
-
-            return parent.equals(this);
-        }
-
-
-        Set<E> asSet() {
-
-            return new AbstractSet<E>() {
-
-                @Override
-                public Iterator<E> iterator() {
-
-                    return new Iterator<E>() {
-
-                        private Entry<E> nextEntry = findRepresentative();
-
-
-                        @Override
-                        public boolean hasNext() {
-
-                            return nextEntry != null;
-                        }
-
-
-                        @Override
-                        public E next() {
-
-                            if (nextEntry == null) {
-                                throw new NoSuchElementException();
-                            }
-                            E result = nextEntry.value;
-                            nextEntry = nextEntry.next;
-                            return result;
-                        }
-
-
-                        @Override
-                        public void remove() {
-
-                            throw new UnsupportedOperationException();
-                        }
-
-                    };
-                }
-
-
-                @Override
-                public int size() {
-
-                    return findRepresentative().size;
-                }
-            };
-        }
-
-    }
-
-}

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/model/Line.java

@@ -12,7 +12,7 @@ import lombok.Data;
 @Data
 public class Line extends BoundingBox {
 
-    private static final double WORD_DISTANCE_MULTIPLIER = 0.2;
+    private static final double WORD_DISTANCE_MULTIPLIER = 0.18;
 
     private final double x0;
     private final double y0;
@@ -78,11 +78,7 @@ public class Line extends BoundingBox {
 
     private double computeHeight() {
 
-        double sum = 0.0;
-        for (Character component : characters) {
-            sum += component.getHeight();
-        }
-        return sum / characters.size();
+        return characters.stream().map(Character::getHeight).reduce(0d, Double::sum) / characters.size();
     }
 
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/model/UnionFind.java

@@ -0,0 +1,31 @@
+package com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model;
+
+import java.util.Collection;
+import java.util.LinkedHashMap;
+import java.util.LinkedHashSet;
+import java.util.Map;
+import java.util.Set;
+
+public class UnionFind<T> extends org.jgrapht.alg.util.UnionFind<T> {
+
+    public UnionFind(Set<T> elements) {
+
+        super(elements);
+    }
+
+
+    public Collection<Set<T>> getGroups() {
+
+        Map<T, Set<T>> setRep = new LinkedHashMap<>();
+        for (T t : getParentMap().keySet()) {
+            T representative = find(t);
+            if (!setRep.containsKey(representative)) {
+                setRep.put(representative, new LinkedHashSet<>());
+            }
+            setRep.get(representative).add(t);
+        }
+
+        return setRep.values();
+    }
+
+}

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/service/LineBuilderService.java

@@ -2,14 +2,15 @@ package com.knecon.fforesight.service.layoutparser.processor.services.docstrum.s
 
 import java.util.ArrayList;
 import java.util.Comparator;
+import java.util.HashSet;
 import java.util.List;
 
 import org.springframework.stereotype.Service;
 
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.AngleFilter;
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.Character;
-import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.DisjointSets;
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.Line;
+import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.UnionFind;
 
 @Service
 public class LineBuilderService {
@@ -24,7 +25,8 @@ public class LineBuilderService {
         double maxHorizontalDistance = characterSpacing * CHARACTER_SPACING_DISTANCE_MULTIPLIER;
         double maxVerticalDistance = lineSpacing * MAX_VERTICAL_CHARACTER_DISTANCE;
 
-        DisjointSets<Character> sets = new DisjointSets<>(characters);
+        UnionFind<Character> unionFind = new UnionFind<>(new HashSet<>(characters));
+
         AngleFilter filter = new AngleFilter(-ANGLE_TOLERANCE, ANGLE_TOLERANCE);
 
         characters.forEach(character -> {
@@ -33,13 +35,13 @@ public class LineBuilderService {
                 double y = neighbor.getVerticalDistance() / maxVerticalDistance;
                 if (character.getTextPosition().getDir() == neighbor.getCharacter().getTextPosition().getDir() && filter.matches(neighbor) && Math.pow(x, 2) + Math.pow(y,
                         2) <= 1) {
-                    sets.union(character, neighbor.getCharacter());
+                    unionFind.union(character, neighbor.getCharacter());
                 }
             });
         });
 
         List<Line> lines = new ArrayList<>();
-        sets.forEach(group -> {
+        unionFind.getGroups().forEach(group -> {
             List<Character> lineCharacters = new ArrayList<>(group);
             lineCharacters.sort(Comparator.comparingDouble(Character::getX));
             lines.add(new Line(lineCharacters, characterSpacing));

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/service/NearestNeighbourService.java

@@ -18,7 +18,7 @@ public class NearestNeighbourService {
 
     public void findNearestNeighbors(List<Character> characters) {
 
-        if (characters.isEmpty()) {
+        if (characters.isEmpty() || characters.size() == 1) {
             return;
         }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/service/ReadingOrderService.java

@@ -2,8 +2,10 @@ package com.knecon.fforesight.service.layoutparser.processor.services.docstrum.s
 
 import java.util.ArrayList;
 import java.util.Comparator;
+import java.util.HashMap;
 import java.util.List;
 import java.util.ListIterator;
+import java.util.Map;
 
 import org.springframework.stereotype.Service;
 
@@ -15,22 +17,43 @@ import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.ut
 public class ReadingOrderService {
 
     private static final double THRESHOLD = 5;
+    public static final double MULTI_COLUMN_DETECTION_THRESHOLD = 1.5;
 
 
-    public List<Zone> resolve(List<Zone> zones, boolean xyOrder) {
+    public List<Zone> resolve(List<Zone> zones, boolean xyReadingOrder) {
 
         if (zones.isEmpty() || zones.size() == 1) {
             return zones;
         }
 
-        if (xyOrder) {
-//            QuickSort.sort(zones, new ZoneComparator());
-            zones.sort(Comparator.comparing(BoundingBox::getY, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, THRESHOLD))
-                    .thenComparing(BoundingBox::getX, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, 0)));
-            return zones;
+        if (xyReadingOrder) {
+            return resolveSingleColumnReadingOrder(zones);
         }
 
-        return resolveMultiColumnReadingOder(zones);
+        Map<Long, Integer> histogram = new HashMap<>();
+        for (Zone zone : zones) {
+            long minY = Math.round(zone.getBBox().getMinY());
+            long maxY = Math.round(zone.getBBox().getMaxY());
+            for (long i = minY; i <= maxY; i++) {
+                histogram.put(i, histogram.getOrDefault(i, 0) + 1);
+            }
+        }
+
+        if (histogram.values().stream().mapToInt(Integer::intValue).average().orElse(1) < MULTI_COLUMN_DETECTION_THRESHOLD) {
+            return resolveSingleColumnReadingOrder(zones);
+        } else {
+
+            return resolveMultiColumnReadingOder(zones);
+        }
+
+    }
+
+
+    private static List<Zone> resolveSingleColumnReadingOrder(List<Zone> zones) {
+
+        zones.sort(Comparator.comparing(BoundingBox::getY, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, THRESHOLD))
+                .thenComparing(BoundingBox::getX, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, THRESHOLD)));
+        return zones;
     }
 
 
@@ -75,6 +98,48 @@ public class ReadingOrderService {
         middle.sort(Comparator.comparing(BoundingBox::getY, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, THRESHOLD))
                 .thenComparing(BoundingBox::getX, (o1, o2) -> DoubleUtils.compareDouble(o1, o2, THRESHOLD)));
 
+        List<Zone> leftNotIntersecting = new ArrayList<>();
+        for (Zone leftZone : leftOf) {
+            boolean intersects = false;
+            for (Zone rightZone : rightOf) {
+                if (leftZone.intersectsY(rightZone)) {
+                    intersects = true;
+                    break;
+                }
+                // early stopping
+                if (rightZone.getBBox().getMinY() > leftZone.getBBox().getMaxY()) {
+                    break;
+                }
+            }
+            if (!intersects) {
+                leftNotIntersecting.add(leftZone);
+            }
+        }
+
+        List<Zone> rightNotIntersecting = new ArrayList<>();
+        for (Zone rightZone : rightOf) {
+            boolean intersects = false;
+            for (Zone leftZone : leftOf) {
+                if (rightZone.intersectsY(leftZone)) {
+                    intersects = true;
+                    break;
+                }
+                // early stopping
+                if (leftZone.getBBox().getMinY() > rightZone.getBBox().getMaxY()) {
+                    break;
+                }
+            }
+            if (!intersects) {
+                rightNotIntersecting.add(rightZone);
+            }
+        }
+
+        leftOf.removeAll(leftNotIntersecting);
+        rightOf.removeAll(rightNotIntersecting);
+
+        middle.addAll(leftNotIntersecting);
+        middle.addAll(rightNotIntersecting);
+
         List<Zone> sortedZones = new ArrayList<>();
         sortedZones.addAll(leftOf);
         sortedZones.addAll(rightOf);

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/docstrum/service/ZoneBuilderService.java

@@ -2,14 +2,15 @@ package com.knecon.fforesight.service.layoutparser.processor.services.docstrum.s
 
 import java.util.ArrayList;
 import java.util.Comparator;
+import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
 
 import org.springframework.stereotype.Service;
 
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.Character;
-import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.DisjointSets;
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.Line;
+import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.UnionFind;
 import com.knecon.fforesight.service.layoutparser.processor.services.docstrum.model.Zone;
 
 @Service
@@ -40,7 +41,7 @@ public class ZoneBuilderService {
         double minHorizontalMergeDistance = characterSpacing * MIN_HORIZONTAL_MERGE_DISTANCE_MULTIPLIER;
         double maxVerticalMergeDistance = lineSpacing * MAX_VERTICAL_MERGE_DISTANCE_MULTIPLIER;
 
-        DisjointSets<Line> sets = new DisjointSets<>(lines);
+        UnionFind<Line> unionFind = new UnionFind<>(new HashSet<>(lines));
 
         double meanHeight = calculateMeanHeight(lines);
 
@@ -50,20 +51,20 @@ public class ZoneBuilderService {
                     double scale = Math.min(outerLine.getHeight(), innerLine.getHeight()) / meanHeight;
                     scale = Math.max(MIN_LINE_SIZE_SCALE, Math.min(scale, MAX_LINE_SIZE_SCALE));
 
-                    if (!sets.areTogether(outerLine, innerLine) && outerLine.angularDifference(innerLine) <= ANGLE_TOLERANCE) {
+                    if (!unionFind.inSameSet(outerLine, innerLine) && outerLine.angularDifference(innerLine) <= ANGLE_TOLERANCE) {
 
                         double horizontalDistance = outerLine.horizontalDistance(innerLine) / scale;
                         double verticalDistance = outerLine.verticalDistance(innerLine) / scale;
 
                         if (minHorizontalDistance <= horizontalDistance && verticalDistance <= maxVerticalDistance //
                                 || minHorizontalMergeDistance <= horizontalDistance && verticalDistance <= maxVerticalMergeDistance) {
-                            sets.union(outerLine, innerLine);
+                            unionFind.union(outerLine, innerLine);
                         }
                     }
                 }));
 
         List<Zone> zones = new ArrayList<>();
-        sets.forEach(group -> {
+        unionFind.getGroups().forEach(group -> {
             zones.add(new Zone(new ArrayList<>(group)));
         });
 
@@ -99,7 +100,7 @@ public class ZoneBuilderService {
         double minVerticalDistance = 0;
         double maxVerticalDistance = lineSpacing * MAX_VERTICAL_MERGE_DISTANCE;
 
-        DisjointSets<Line> sets = new DisjointSets<>(lines);
+        UnionFind<Line> unionFind = new UnionFind<>(new HashSet<>(lines));
 
         lines.forEach(outer -> {
 
@@ -110,7 +111,7 @@ public class ZoneBuilderService {
                     double verticalDistance = outer.verticalDistance(inner);
 
                     if (horizontalDistance <= maxHorizontalDistance && minVerticalDistance <= verticalDistance && verticalDistance <= maxVerticalDistance) {
-                        sets.union(outer, inner);
+                        unionFind.union(outer, inner);
                     } else if (minVerticalDistance <= verticalDistance && verticalDistance <= maxVerticalDistance && Math.abs(horizontalDistance - Math.min(outer.getLength(),
                             inner.getLength())) < 0.1) {
                         boolean characterOverlap = false;
@@ -127,7 +128,7 @@ public class ZoneBuilderService {
                             }
                         }
                         if (!characterOverlap && overlappingCount <= 2) {
-                            sets.union(outer, inner);
+                            unionFind.union(outer, inner);
                         }
                     }
                 }
@@ -135,15 +136,16 @@ public class ZoneBuilderService {
         });
 
         List<Line> outputZone = new ArrayList<>();
-        for (Set<Line> group : sets) {
-            List<Character> components = new ArrayList<>();
+        for (Set<Line> group : unionFind.getGroups()) {
+            List<Character> characters = new ArrayList<>();
             for (Line line : group) {
-                components.addAll(line.getCharacters());
+                characters.addAll(line.getCharacters());
             }
-            components.sort(Comparator.comparingDouble(Character::getX));
+            characters.sort(Comparator.comparingDouble(Character::getX));
 
-            outputZone.add(new Line(components, characterSpacing));
+            outputZone.add(new Line(characters, characterSpacing));
         }
+
         return new Zone(outputZone);
     }
 

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/DocumentGraphFactory.java

@@ -16,6 +16,7 @@ import java.util.Set;
 import java.util.stream.Collectors;
 
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.LayoutEngine;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 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.ClassificationFooter;
@@ -48,14 +49,14 @@ import lombok.experimental.UtilityClass;
 @UtilityClass
 public class DocumentGraphFactory {
 
-    public Document buildDocumentGraph(ClassificationDocument document) {
+    public Document buildDocumentGraph(LayoutParsingType layoutParsingType, ClassificationDocument document) {
 
         Document documentGraph = new Document();
         Context context = new Context(documentGraph);
 
         document.getPages().forEach(context::buildAndAddPageWithCounter);
         document.getSections().stream().flatMap(section -> section.getImages().stream()).forEach(image -> context.getImages().add(image));
-        addSections(document, context);
+        addSections(layoutParsingType, document, context);
         addHeaderAndFooterToEachPage(document, context);
 
         documentGraph.setNumberOfPages(context.pages.size());
@@ -66,9 +67,9 @@ public class DocumentGraphFactory {
     }
 
 
-    private void addSections(ClassificationDocument document, Context context) {
+    private void addSections(LayoutParsingType layoutParsingType, ClassificationDocument document, Context context) {
 
-        document.getSections().forEach(section -> SectionNodeFactory.addSection(null, section.getNonEmptyPageBlocks(), section.getImages(), context));
+        document.getSections().forEach(section -> SectionNodeFactory.addSection(layoutParsingType, null, section.getNonEmptyPageBlocks(), section.getImages(), context));
     }
 
 
@@ -90,6 +91,7 @@ public class DocumentGraphFactory {
         List<TextPageBlock> textBlocks = new ArrayList<>();
         textBlocks.add(originalTextBlock);
         textBlocks.addAll(textBlocksToMerge);
+
         AtomicTextBlock textBlock = context.textBlockFactory.buildAtomicTextBlock(TextPositionOperations.mergeAndSortTextPositionSequenceByYThenX(textBlocks), node, context, page);
 
         if (node instanceof DuplicatedParagraph duplicatedParagraph) {

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/SectionNodeFactory.java

@@ -4,12 +4,14 @@ import static java.lang.String.format;
 import static java.util.Collections.emptyList;
 import static java.util.stream.Collectors.groupingBy;
 
+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 com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.GenericSemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
@@ -24,7 +26,11 @@ import lombok.experimental.UtilityClass;
 @UtilityClass
 public class SectionNodeFactory {
 
-    public void addSection(GenericSemanticNode parentNode, List<AbstractPageBlock> pageBlocks, List<ClassifiedImage> images, DocumentGraphFactory.Context context) {
+    public void addSection(LayoutParsingType layoutParsingType,
+                           GenericSemanticNode parentNode,
+                           List<AbstractPageBlock> pageBlocks,
+                           List<ClassifiedImage> images,
+                           DocumentGraphFactory.Context context) {
 
         if (pageBlocks.isEmpty()) {
             return;
@@ -37,11 +43,11 @@ public class SectionNodeFactory {
 
         section.setTreeId(getTreeId(parentNode, context, section));
 
-        addFirstHeadlineDirectlyToSection(pageBlocks, context, section);
+        addFirstHeadlineDirectlyToSection(layoutParsingType, pageBlocks, context, section);
         if (containsTablesAndTextBlocks(pageBlocks)) {
-            splitPageBlocksIntoSubSections(pageBlocks).forEach(subSectionPageBlocks -> addSection(section, subSectionPageBlocks, emptyList(), context));
+            splitPageBlocksIntoSubSections(pageBlocks).forEach(subSectionPageBlocks -> addSection(layoutParsingType, section, subSectionPageBlocks, emptyList(), context));
         } else {
-            addTablesAndParagraphsAndHeadlinesToSection(pageBlocks, context, section);
+            addTablesAndParagraphsAndHeadlinesToSection(layoutParsingType, pageBlocks, context, section);
         }
 
         images.stream().distinct().forEach(image -> DocumentGraphFactory.addImage(section, image, context));
@@ -58,16 +64,19 @@ public class SectionNodeFactory {
     }
 
 
-    private void addFirstHeadlineDirectlyToSection(List<AbstractPageBlock> pageBlocks, DocumentGraphFactory.Context context, Section section) {
+    private void addFirstHeadlineDirectlyToSection(LayoutParsingType layoutParsingType, List<AbstractPageBlock> pageBlocks, DocumentGraphFactory.Context context, Section section) {
 
         if (pageBlocks.get(0).isHeadline()) {
-            addTablesAndParagraphsAndHeadlinesToSection(List.of(pageBlocks.get(0)), context, section);
+            addTablesAndParagraphsAndHeadlinesToSection(layoutParsingType, List.of(pageBlocks.get(0)), context, section);
             pageBlocks.remove(0);
         }
     }
 
 
-    private void addTablesAndParagraphsAndHeadlinesToSection(List<AbstractPageBlock> pageBlocks, DocumentGraphFactory.Context context, Section section) {
+    private void addTablesAndParagraphsAndHeadlinesToSection(LayoutParsingType layoutParsingType,
+                                                             List<AbstractPageBlock> pageBlocks,
+                                                             DocumentGraphFactory.Context context,
+                                                             Section section) {
 
         Set<AbstractPageBlock> alreadyMerged = new HashSet<>();
         List<AbstractPageBlock> remainingBlocks = new LinkedList<>(pageBlocks);
@@ -80,13 +89,23 @@ public class SectionNodeFactory {
             remainingBlocks.removeAll(alreadyMerged);
 
             if (abstractPageBlock instanceof TextPageBlock) {
-                List<TextPageBlock> textBlocks = findTextBlocksWithSameClassificationAndAlignsY((TextPageBlock) abstractPageBlock, remainingBlocks);
-                alreadyMerged.addAll(textBlocks);
-                DocumentGraphFactory.addParagraphOrHeadline(section, (TextPageBlock) abstractPageBlock, context, textBlocks);
+
+                switch (layoutParsingType) {
+                    case CLARIFYND, CLARIFYND_PARAGRAPH_DEBUG -> {
+                        alreadyMerged.add(abstractPageBlock);
+                        remainingBlocks.remove(abstractPageBlock);
+                        DocumentGraphFactory.addParagraphOrHeadline(section, (TextPageBlock) abstractPageBlock, context, new ArrayList<>());
+                    }
+                    default -> {
+                        List<TextPageBlock> textBlocks = findTextBlocksWithSameClassificationAndAlignsY((TextPageBlock) abstractPageBlock, remainingBlocks);
+                        alreadyMerged.addAll(textBlocks);
+                        DocumentGraphFactory.addParagraphOrHeadline(section, (TextPageBlock) abstractPageBlock, context, textBlocks);
+                    }
+                }
             } else if (abstractPageBlock instanceof TablePageBlock tablePageBlock) {
                 List<TablePageBlock> tablesToMerge = TableMergingUtility.findConsecutiveTablesWithSameColCountAndSameHeaders(tablePageBlock, remainingBlocks);
                 alreadyMerged.addAll(tablesToMerge);
-                TableNodeFactory.addTable(section, tablesToMerge, context);
+                TableNodeFactory.addTable(layoutParsingType, section, tablesToMerge, context);
             } else {
                 throw new RuntimeException(format("Unhandled AbstractPageBlockType %s!", abstractPageBlock.getClass()));
             }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/TableNodeFactory.java

@@ -7,16 +7,17 @@ import java.util.List;
 import java.util.Set;
 import java.util.stream.Collectors;
 
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
-import com.knecon.fforesight.service.layoutparser.processor.model.table.Cell;
-import com.knecon.fforesight.service.layoutparser.processor.model.table.TablePageBlock;
-import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.GenericSemanticNode;
 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.Table;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.textblock.TextBlock;
+import com.knecon.fforesight.service.layoutparser.processor.model.table.Cell;
+import com.knecon.fforesight.service.layoutparser.processor.model.table.TablePageBlock;
+import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
 import com.knecon.fforesight.service.layoutparser.processor.utils.TextPositionOperations;
 
 import lombok.experimental.UtilityClass;
@@ -27,7 +28,7 @@ public class TableNodeFactory {
     public final double TABLE_CELL_MERGE_CONTENTS_SIZE_THRESHOLD = 0.05;
 
 
-    public void addTable(GenericSemanticNode parentNode, List<TablePageBlock> tablesToMerge, DocumentGraphFactory.Context context) {
+    public void addTable(LayoutParsingType layoutParsingType, GenericSemanticNode parentNode, List<TablePageBlock> tablesToMerge, DocumentGraphFactory.Context context) {
 
         setPageNumberInCells(tablesToMerge);
         Set<Page> pages = tablesToMerge.stream().map(AbstractPageBlock::getPage).map(context::getPage).collect(Collectors.toSet());
@@ -43,7 +44,7 @@ public class TableNodeFactory {
 
         List<Integer> treeId = context.getDocumentTree().createNewChildEntryAndReturnId(parentNode, table);
         table.setTreeId(treeId);
-        addTableCells(mergedRows, table, context);
+        addTableCells(layoutParsingType, mergedRows, table, context);
 
         ifTableHasNoHeadersSetFirstRowAsHeaders(table);
     }
@@ -88,18 +89,18 @@ public class TableNodeFactory {
     }
 
 
-    private void addTableCells(List<List<Cell>> rows, Table table, DocumentGraphFactory.Context context) {
+    private void addTableCells(LayoutParsingType layoutParsingType, List<List<Cell>> rows, Table table, DocumentGraphFactory.Context context) {
 
         for (int rowIndex = 0; rowIndex < rows.size(); rowIndex++) {
             for (int colIndex = 0; colIndex < rows.get(rowIndex).size(); colIndex++) {
-                addTableCell(rows.get(rowIndex).get(colIndex), rowIndex, colIndex, table, context);
+                addTableCell(layoutParsingType, rows.get(rowIndex).get(colIndex), rowIndex, colIndex, table, context);
             }
         }
     }
 
 
     @SuppressWarnings("PMD.UnusedPrivateMethod") // PMD actually flags this wrong
-    private void addTableCell(Cell cell, int rowIndex, int colIndex, Table tableNode, DocumentGraphFactory.Context context) {
+    private void addTableCell(LayoutParsingType layoutParsingType, Cell cell, int rowIndex, int colIndex, Table tableNode, DocumentGraphFactory.Context context) {
 
         Page page = context.getPage(cell.getPageNumber());
 
@@ -116,7 +117,7 @@ public class TableNodeFactory {
             textBlock = context.getTextBlockFactory().buildAtomicTextBlock(cell.getTextBlocks().get(0).getSequences(), tableCell, context, page);
             tableCell.setLeafTextBlock(textBlock);
         } else if (firstTextBlockIsHeadline(cell)) {
-            SectionNodeFactory.addSection(tableCell, cell.getTextBlocks().stream().map(tb -> (AbstractPageBlock) tb).toList(), emptyList(), context);
+            SectionNodeFactory.addSection(layoutParsingType, tableCell, cell.getTextBlocks().stream().map(tb -> (AbstractPageBlock) tb).toList(), emptyList(), context);
         } else if (cellAreaIsSmallerThanPageAreaTimesThreshold(cell, page)) {
             List<TextPositionSequence> sequences = TextPositionOperations.mergeAndSortTextPositionSequenceByYThenX(cell.getTextBlocks());
             textBlock = context.getTextBlockFactory().buildAtomicTextBlock(sequences, tableCell, context, page);

layoutparser-service/layoutparser-service-server/src/main/java/com/knecon/fforesight/service/layoutparser/server/queue/MessageHandler.java

@@ -37,7 +37,7 @@ public class MessageHandler {
 
         LayoutParsingRequest layoutParsingRequest = objectMapper.readValue(message.getBody(), LayoutParsingRequest.class);
 
-        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.TAAS) && layoutParsingRequest.researchDocumentStorageId() == null) {
+        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.CLARIFYND) && layoutParsingRequest.researchDocumentStorageId() == null) {
             throw new IllegalArgumentException("ResearchDocumentDataStorageId is null!");
         }
         log.info("Layout parsing request received {}", layoutParsingRequest.identifier());

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/BdrJsonBuildTest.java

@@ -48,12 +48,13 @@ public class BdrJsonBuildTest extends AbstractTest {
     @SneakyThrows
     protected Document buildGraph(File file) {
 
-        return DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.TAAS,
-                file,
-                new ImageServiceResponse(),
-                new TableServiceResponse(),
-                new VisualLayoutParsingResponse(),
-                file.toString()));
+        return DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.CLARIFYND,
+                layoutParsingPipeline.parseLayout(LayoutParsingType.CLARIFYND,
+                        file,
+                        new ImageServiceResponse(),
+                        new TableServiceResponse(),
+                        new VisualLayoutParsingResponse(),
+                        file.toString()));
     }
 
 

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/HeadlinesGoldStandardIntegrationTest.java

@@ -95,12 +95,13 @@ public class HeadlinesGoldStandardIntegrationTest {
         goldStandardLog.getRedactionLogEntry().removeIf(r -> !r.isRedacted() || r.getChanges().get(r.getChanges().size() - 1).getType().equals(ChangeType.REMOVED));
         goldStandardLog.getRedactionLogEntry().forEach(e -> goldStandardHeadlines.add(new Headline(e.getPositions().get(0).getPage(), e.getValue())));
 
-        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                pdfFileResource.getFile(),
-                new ImageServiceResponse(),
-                new TableServiceResponse(),
-                new VisualLayoutParsingResponse(),
-                filePath));
+        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.REDACT_MANAGER_OLD,
+                layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_OLD,
+                        pdfFileResource.getFile(),
+                        new ImageServiceResponse(),
+                        new TableServiceResponse(),
+                        new VisualLayoutParsingResponse(),
+                        filePath));
 
         var foundHeadlines = documentGraph.streamAllSubNodes()
                 .map(SemanticNode::getHeadline)

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/LayoutparserEnd2EndTest.java

@@ -26,7 +26,7 @@ public class LayoutparserEnd2EndTest extends AbstractTest {
     public void testLayoutParserEndToEnd() {
 
         prepareStorage("files/bdr/Wie weiter bei Kristeneinrichtungen.pdf");
-        LayoutParsingRequest layoutParsingRequest = buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+        LayoutParsingRequest layoutParsingRequest = buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER_OLD);
         LayoutParsingFinishedEvent finishedEvent = layoutParsingPipeline.parseLayoutAndSaveFilesToStorage(layoutParsingRequest);
         Arrays.stream(finishedEvent.message().split("\n")).forEach(log::info);
     }

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/graph/DocumentGraphJsonWritingTest.java

@@ -55,12 +55,13 @@ public class DocumentGraphJsonWritingTest extends BuildDocumentTest {
     @SneakyThrows
     private void writeJsons(Path filename) {
 
-        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                filename.toFile(),
-                new ImageServiceResponse(),
-                new TableServiceResponse(),
-                new VisualLayoutParsingResponse(),
-                filename.toFile().toString()));
+        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.REDACT_MANAGER_OLD,
+                layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_OLD,
+                        filename.toFile(),
+                        new ImageServiceResponse(),
+                        new TableServiceResponse(),
+                        new VisualLayoutParsingResponse(),
+                        filename.toFile().toString()));
 
         DocumentData documentData = DocumentDataMapper.toDocumentData(documentGraph);
         ObjectMapper mapper = ObjectMapperFactory.create();

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/graph/ViewerDocumentTest.java

@@ -26,14 +26,14 @@ public class ViewerDocumentTest extends BuildDocumentTest {
     @SneakyThrows
     public void testViewerDocument() {
 
-        String fileName = "files/new/270 rotated text on non rotated pages.pdf";
+        String fileName = "files/S-Metolachlor_RAR_01_Volume_1_2018-09-06.pdf";
         String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
 
         var documentFile = new ClassPathResource(fileName).getFile();
         ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
         LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);
 
-        Document document = buildGraph(fileName, LayoutParsingType.DOCSTRUM_XY);
+        Document document = buildGraph(fileName, LayoutParsingType.REDACT_MANAGER);
         long start = System.currentTimeMillis();
         layoutGridService.addLayoutGrid(documentFile, document, new File(tmpFileName), true);
         System.out.printf("Total time: %.2fs%n", ((float) (System.currentTimeMillis() - start)) / 1000);
@@ -61,7 +61,7 @@ public class ViewerDocumentTest extends BuildDocumentTest {
                 Path.of(fileName).getFileName().toFile().toString());
         ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
         LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);
-        Document document = DocumentGraphFactory.buildDocumentGraph(classificationDocument);
+        Document document = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.DOCUMINE, classificationDocument);
 
         layoutGridService.addLayoutGrid(documentFile, document, new File(tmpFileName), true);
     }

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/segmentation/PdfSegmentationServiceTest.java

@@ -56,12 +56,12 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     @SneakyThrows
     public ClassificationDocument buildClassificationDocument(File originDocument, TableServiceResponse tableServiceResponse) {
 
-        ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                                                                                          originDocument,
-                                                                                          new ImageServiceResponse(),
-                                                                                          tableServiceResponse,
-                                                                                          new VisualLayoutParsingResponse(),
-                                                                                          "document");
+        ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_OLD,
+                originDocument,
+                new ImageServiceResponse(),
+                tableServiceResponse,
+                new VisualLayoutParsingResponse(),
+                "document");
 
         redactManagerClassificationService.classifyDocument(classificationDocument);
 
@@ -112,16 +112,8 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile(), tableServiceResponse);
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .collect(Collectors.toList())).isNotEmpty();
-        var tables = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList();
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        var tables = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList();
 
         // Quality of the table parsing is not good, because the file is rotated at scanning.
         // We only asset that the table border is not the page border.
@@ -143,12 +135,12 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         imageServiceResponse.getData()
                 .forEach(imageMetadata -> images.computeIfAbsent(imageMetadata.getPosition().getPageNumber(), x -> new ArrayList<>())
                         .add(new ClassifiedImage(new Rectangle2D.Double(imageMetadata.getPosition().getX1(),
-                                                                        imageMetadata.getPosition().getY1(),
-                                                                        imageMetadata.getGeometry().getWidth(),
-                                                                        imageMetadata.getGeometry().getHeight()),
-                                                 ImageType.valueOf(imageMetadata.getClassification().getLabel().toUpperCase(Locale.ROOT)),
-                                                 imageMetadata.isAlpha(),
-                                                 imageMetadata.getPosition().getPageNumber())));
+                                imageMetadata.getPosition().getY1(),
+                                imageMetadata.getGeometry().getWidth(),
+                                imageMetadata.getGeometry().getHeight()),
+                                ImageType.valueOf(imageMetadata.getClassification().getLabel().toUpperCase(Locale.ROOT)),
+                                imageMetadata.isAlpha(),
+                                imageMetadata.getPosition().getPageNumber())));
 
         System.out.println("object");
     }
@@ -160,22 +152,11 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Spanning Cells.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock table = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(0);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(0);
         assertThat(table.getColCount()).isEqualTo(6);
         assertThat(table.getRowCount()).isEqualTo(13);
-        assertThat(table.getRows()
-                           .stream()
-                           .mapToInt(List::size).sum()).isEqualTo(6 * 13);
+        assertThat(table.getRows().stream().mapToInt(List::size).sum()).isEqualTo(6 * 13);
     }
 
 
@@ -185,37 +166,15 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Table.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(0);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(0);
         assertThat(firstTable.getColCount()).isEqualTo(8);
         assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(1);
+        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(1);
         assertThat(secondTable.getColCount()).isEqualTo(8);
         assertThat(secondTable.getRowCount()).isEqualTo(2);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
-                .get(0)
-                .stream()
-                .map(Collections::singletonList)
-                .collect(Collectors.toList());
-        assertThat(secondTable.getRows()
-                           .stream()
-                           .allMatch(row -> row.stream()
-                                   .map(Cell::getHeaderCells)
-                                   .toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(0).stream().map(Collections::singletonList).collect(Collectors.toList());
+        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
     }
 
 
@@ -225,37 +184,15 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Multi Page Table.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(0);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(0);
         assertThat(firstTable.getColCount()).isEqualTo(9);
         assertThat(firstTable.getRowCount()).isEqualTo(5);
-        TablePageBlock secondTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(1);
+        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(1);
         assertThat(secondTable.getColCount()).isEqualTo(9);
         assertThat(secondTable.getRowCount()).isEqualTo(6);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
-                .get(firstTable.getRowCount() - 1)
-                .stream()
-                .map(Cell::getHeaderCells)
-                .collect(Collectors.toList());
-        assertThat(secondTable.getRows()
-                           .stream()
-                           .allMatch(row -> row.stream()
-                                   .map(Cell::getHeaderCells)
-                                   .toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(firstTable.getRowCount() - 1).stream().map(Cell::getHeaderCells).collect(Collectors.toList());
+        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
     }
 
 
@@ -265,37 +202,15 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Rotated Table Headers.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(0);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(0);
         assertThat(firstTable.getColCount()).isEqualTo(8);
         assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(1);
+        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(1);
         assertThat(secondTable.getColCount()).isEqualTo(8);
         assertThat(secondTable.getRowCount()).isEqualTo(6);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
-                .get(0)
-                .stream()
-                .map(Collections::singletonList)
-                .collect(Collectors.toList());
-        assertThat(secondTable.getRows()
-                           .stream()
-                           .allMatch(row -> row.stream()
-                                   .map(Cell::getHeaderCells)
-                                   .toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(0).stream().map(Collections::singletonList).collect(Collectors.toList());
+        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
     }
 
 
@@ -345,30 +260,29 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         validateTable(document, 0, 8, 8, 0, 0);
 
         List<List<String>> values = Arrays.asList(Arrays.asList("Annex point Reference within DAR/RAR",
-                                                                "Author, date",
-                                                                "Study title",
-                                                                "Analytical method Author, date, No.",
-                                                                "Technique, LOQ of the method, validated working range",
-                                                                "Method meets analytical validation criteria",
-                                                                "Remarks (in case validation criteria are not met)",
-                                                                "Acceptability of the method"),
-                                                  Arrays.asList(
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
-                                                  Arrays.asList("CA 7.1.2.1.1 DAR (2009)",
-                                                                "Evans P.G. 2001 TMJ4569B, VV-323245",
-                                                                "Azoxystrobin Laboratory Degradation Study in Three Soil Types, Sampled from Holland and the United Kingdom",
-                                                                "Method: RAM 269 Johnson R.I., Tummon O.J., Earl M. 1995 RJ1864B, VV-377731 Johnson R.I., Tummon O.J., Earl M. 1998 RAM 269/02, VV-124072 Johnson R.I., Tummon O.J., Earl M. 2000 RAM 269/03, VV-123986 Validation: Robinson N.J. 2001 TMJ4617B, VV-895845",
-                                                                "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
-                                                                "Y",
-                                                                "N/A",
-                                                                "Y"));
+                        "Author, date",
+                        "Study title",
+                        "Analytical method Author, date, No.",
+                        "Technique, LOQ of the method, validated working range",
+                        "Method meets analytical validation criteria",
+                        "Remarks (in case validation criteria are not met)",
+                        "Acceptability of the method"),
+                Arrays.asList("Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
+                Arrays.asList("CA 7.1.2.1.1 DAR (2009)",
+                        "Evans P.G. 2001 TMJ4569B, VV-323245",
+                        "Azoxystrobin Laboratory Degradation Study in Three Soil Types, Sampled from Holland and the United Kingdom",
+                        "Method: RAM 269 Johnson R.I., Tummon O.J., Earl M. 1995 RJ1864B, VV-377731 Johnson R.I., Tummon O.J., Earl M. 1998 RAM 269/02, VV-124072 Johnson R.I., Tummon O.J., Earl M. 2000 RAM 269/03, VV-123986 Validation: Robinson N.J. 2001 TMJ4617B, VV-895845",
+                        "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
+                        "Y",
+                        "N/A",
+                        "Y"));
 
         validateTable(document, 0, values);
 
@@ -757,11 +671,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     @SneakyThrows
     private void toHtml(ClassificationDocument document, String filename) {
 
-        var tables = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList();
+        var tables = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList();
         StringBuilder sb = new StringBuilder();
 
         int currentPage = 1;
@@ -782,19 +692,9 @@ public class PdfSegmentationServiceTest extends AbstractTest {
 
     private void validateTable(ClassificationDocument document, int tableIndex, int colCount, int rowCount, int emptyCellsCountCorrect, int emptyCellsCountIncorrect) {
 
-        TablePageBlock table = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(tableIndex);
+        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(tableIndex);
         List<List<Cell>> rows = table.getRows();
-        int emptyCellsFoundFound = rows.stream()
-                .flatMap(List::stream)
-                .toList()
-                .stream()
-                .filter(f -> f.toString().isEmpty())
-                .toList().size();
+        int emptyCellsFoundFound = rows.stream().flatMap(List::stream).toList().stream().filter(f -> f.toString().isEmpty()).toList().size();
 
         for (List<Cell> row : table.getRows()) {
             row.forEach(r -> System.out.println(r.toString()));
@@ -809,20 +709,11 @@ public class PdfSegmentationServiceTest extends AbstractTest {
 
     private void validateTable(ClassificationDocument document, int tableIndex, List<List<String>> values) {
 
-        TablePageBlock table = document.getSections()
-                .stream()
-                .flatMap(paragraph -> paragraph.getTables()
-                        .stream())
-                .toList()
-                .get(tableIndex);
+        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(tableIndex);
         List<List<Cell>> rows = table.getRows();
 
-        List<Cell> rowsFlattened = rows.stream()
-                .flatMap(List::stream)
-                .toList();
-        List<String> valuesFlattened = values.stream()
-                .flatMap(List::stream)
-                .toList();
+        List<Cell> rowsFlattened = rows.stream().flatMap(List::stream).toList();
+        List<String> valuesFlattened = values.stream().flatMap(List::stream).toList();
 
         for (int i = 0; i < valuesFlattened.size(); i++) {
             Cell cell = rowsFlattened.get(i);
@@ -835,11 +726,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
 
     private void validateTableSize(ClassificationDocument document, int tableSize) {
 
-        assertThat(document.getSections()
-                           .stream()
-                           .flatMap(paragraph -> paragraph.getTables()
-                                   .stream())
-                           .toList().size()).isEqualTo(tableSize);
+        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().size()).isEqualTo(tableSize);
 
     }
 

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/services/BodyTextFrameServiceTest.java

@@ -21,7 +21,7 @@ class BodyTextFrameServiceTest extends BuildDocumentTest {
 
         String filename = "files/211.pdf";
         String outputFilename = "/tmp/" + Path.of(filename).getFileName() + "_MAINBODY.pdf";
-        ClassificationDocument document = parseLayout(filename, LayoutParsingType.TAAS);
+        ClassificationDocument document = parseLayout(filename, LayoutParsingType.CLARIFYND);
         PdfDraw.drawRectanglesPerPage(filename,
                 document.getPages().stream().map(page -> List.of(RectangleTransformations.toRectangle2D(page.getBodyTextFrame()))).toList(),
                 outputFilename);

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/services/RulingCleaningServiceTest.java

@@ -74,7 +74,7 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
             cleanRulingsPerPage.add(rulingCleaningService.getCleanRulings(Collections.emptyList(), pageContent.getRulings()));
         }
         var cleanRulings = cleanRulingsPerPage.stream().map(CleanRulings::getVertical).collect(Collectors.toList());
-        PdfDraw.drawLinesPerPage(fileName,  cleanRulings, lineFileName);
+        PdfDraw.drawLinesPerPage(fileName, cleanRulings, lineFileName);
 
     }
 
@@ -99,18 +99,20 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
     @SneakyThrows
     private void writeJsons(Path filename) {
 
-        Document documentGraphBefore = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                filename.toFile(),
-                new ImageServiceResponse(),
-                new TableServiceResponse(),
-                new VisualLayoutParsingResponse(),
-                filename.toFile().toString()));
-        Document documentGraphAfter = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                filename.toFile(),
-                new ImageServiceResponse(),
-                new TableServiceResponse(),
-                new VisualLayoutParsingResponse(),
-                filename.toFile().toString()));
+        Document documentGraphBefore = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.REDACT_MANAGER_OLD,
+                layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_OLD,
+                        filename.toFile(),
+                        new ImageServiceResponse(),
+                        new TableServiceResponse(),
+                        new VisualLayoutParsingResponse(),
+                        filename.toFile().toString()));
+        Document documentGraphAfter = DocumentGraphFactory.buildDocumentGraph(LayoutParsingType.REDACT_MANAGER_OLD,
+                layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER_OLD,
+                        filename.toFile(),
+                        new ImageServiceResponse(),
+                        new TableServiceResponse(),
+                        new VisualLayoutParsingResponse(),
+                        filename.toFile().toString()));
         DocumentData documentDataBefore = DocumentDataMapper.toDocumentData(documentGraphBefore);
         DocumentData documentDataAfter = DocumentDataMapper.toDocumentData(documentGraphAfter);
         if (!compareStructures(documentDataBefore.getDocumentStructure(), documentDataAfter.getDocumentStructure())) {

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/utils/AbstractTest.java

@@ -20,7 +20,6 @@ import org.springframework.context.annotation.Import;
 import org.springframework.context.annotation.Primary;
 import org.springframework.core.io.ClassPathResource;
 import org.springframework.test.context.junit.jupiter.SpringExtension;
-import org.xmlunit.builder.Input;
 
 import com.iqser.red.commons.jackson.ObjectMapperFactory;
 import com.iqser.red.storage.commons.service.StorageService;
@@ -68,7 +67,7 @@ public abstract class AbstractTest {
     protected LayoutParsingRequest buildStandardLayoutParsingRequest() {
 
         return LayoutParsingRequest.builder()
-                .layoutParsingType(LayoutParsingType.REDACT_MANAGER)
+                .layoutParsingType(LayoutParsingType.REDACT_MANAGER_OLD)
                 .originFileStorageId(ORIGIN_FILE_ID)
                 .tablesFileStorageId(Optional.of(TABLE_FILE_ID))
                 .imagesFileStorageId(Optional.of(IMAGE_FILE_ID))
@@ -99,7 +98,7 @@ public abstract class AbstractTest {
     @SneakyThrows
     protected LayoutParsingRequest prepareStorage(String file) {
 
-        return prepareStorage(file, "cv_table_parsing_response/empty.json", "image_service_response/empty.json","visual_layout_parsing_response/empty.json");
+        return prepareStorage(file, "cv_table_parsing_response/empty.json", "image_service_response/empty.json", "visual_layout_parsing_response/empty.json");
     }
 
 
@@ -107,7 +106,7 @@ public abstract class AbstractTest {
     protected LayoutParsingRequest prepareStorage(InputStream fileInputStream) {
 
         storageService.storeObject(TenantContext.getTenantId(), ORIGIN_FILE_ID, fileInputStream);
-        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER_OLD);
     }
 
 
@@ -140,6 +139,7 @@ public abstract class AbstractTest {
         return prepareStorage(pdfFileResource.getInputStream(), cvServiceResponseFileResource.getInputStream(), imageInfoFileResource.getInputStream());
     }
 
+
     @SneakyThrows
     protected LayoutParsingRequest prepareStorage(String file, String cvServiceResponseFile, String imageInfoFile, String visualLayoutParsingResponseFile) {
 
@@ -148,9 +148,13 @@ public abstract class AbstractTest {
         ClassPathResource imageInfoFileResource = new ClassPathResource(imageInfoFile);
         ClassPathResource visualLayoutParsingResponseResource = new ClassPathResource(visualLayoutParsingResponseFile);
 
-        return prepareStorage(pdfFileResource.getInputStream(), cvServiceResponseFileResource.getInputStream(), imageInfoFileResource.getInputStream(), visualLayoutParsingResponseResource.getInputStream());
+        return prepareStorage(pdfFileResource.getInputStream(),
+                cvServiceResponseFileResource.getInputStream(),
+                imageInfoFileResource.getInputStream(),
+                visualLayoutParsingResponseResource.getInputStream());
     }
 
+
     @SneakyThrows
     protected LayoutParsingRequest prepareStorage(InputStream fileStream, InputStream cvServiceResponseFileStream, InputStream imageInfoStream) {
 
@@ -158,18 +162,22 @@ public abstract class AbstractTest {
         storageService.storeObject(TenantContext.getTenantId(), TABLE_FILE_ID, cvServiceResponseFileStream);
         storageService.storeObject(TenantContext.getTenantId(), ORIGIN_FILE_ID, fileStream);
 
-        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER_OLD);
     }
 
+
     @SneakyThrows
-    protected LayoutParsingRequest prepareStorage(InputStream fileStream, InputStream cvServiceResponseFileStream, InputStream imageInfoStream, InputStream visualLayoutParsingResponseFileStream) {
+    protected LayoutParsingRequest prepareStorage(InputStream fileStream,
+                                                  InputStream cvServiceResponseFileStream,
+                                                  InputStream imageInfoStream,
+                                                  InputStream visualLayoutParsingResponseFileStream) {
 
         storageService.storeObject(TenantContext.getTenantId(), IMAGE_FILE_ID, imageInfoStream);
         storageService.storeObject(TenantContext.getTenantId(), TABLE_FILE_ID, cvServiceResponseFileStream);
         storageService.storeObject(TenantContext.getTenantId(), ORIGIN_FILE_ID, fileStream);
-        storageService.storeObject(TenantContext.getTenantId(),VISUAL_LAYOUT_FILE,visualLayoutParsingResponseFileStream );
+        storageService.storeObject(TenantContext.getTenantId(), VISUAL_LAYOUT_FILE, visualLayoutParsingResponseFileStream);
 
-        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER_OLD);
     }
 
 

layoutparser-service/layoutparser-service-server/src/test/java/com/knecon/fforesight/service/layoutparser/server/utils/BuildDocumentTest.java

@@ -38,7 +38,7 @@ public abstract class BuildDocumentTest extends AbstractTest {
     @SneakyThrows
     protected Document buildGraph(String filename) {
 
-        return buildGraph(filename, LayoutParsingType.REDACT_MANAGER);
+        return buildGraph(filename, LayoutParsingType.REDACT_MANAGER_OLD);
     }
 
 
@@ -51,7 +51,7 @@ public abstract class BuildDocumentTest extends AbstractTest {
             prepareStorage(filename);
         }
 
-        return DocumentGraphFactory.buildDocumentGraph(parseLayout(filename, layoutParsingType));
+        return DocumentGraphFactory.buildDocumentGraph(layoutParsingType, parseLayout(filename, layoutParsingType));
     }
 
 }

layoutparser-service/viewer-doc-processor/build.gradle

@@ -1,6 +1,6 @@
 plugins {
     id("com.knecon.fforesight.java-conventions")
-    id("io.freefair.lombok") version "8.2.2"
+    id("io.freefair.lombok") version "8.4"
 }
 
 description = "Library for adding/removing layers in the viewer document"