Merge branch 'DM-589' into 'main'

DM-589: Filter wrong detected cells that borders from rotation at scanning See merge request fforesight/layout-parser!83
DM-589: Filter wrong detected cells that borders from rotation at scanning
2023-11-20 16:07:27 +01:00 · 2023-11-20 15:54:02 +01:00 · 2023-11-16 12:48:48 +01:00 · 2023-11-16 12:40:49 +01:00 · 2023-11-16 09:13:41 +01:00 · 2023-11-16 08:51:12 +01:00
43 changed files with 855 additions and 510 deletions
@@ -36,9 +36,6 @@ public record LayoutParsingRequest(
        @Schema(description = "Path where the Simplified Text File will be stored.")//
        @NonNull String simplifiedTextStorageId,//
        @Schema(description = "Path where the Viewer Document PDF will be stored.")//
-        @NonNull String viewerDocumentStorageId,//
-        @Deprecated//
-        @Schema(description = "Path where the Section Grid will be stored.")//
-        @NonNull String sectionGridStorageId) {
+        @NonNull String viewerDocumentStorageId) {

 }
@@ -10,13 +10,13 @@ import java.util.HashMap;
 import java.util.List;
 import java.util.Map;

-import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;
 import org.apache.pdfbox.pdmodel.PDDocument;
 import org.apache.pdfbox.pdmodel.PDPage;
 import org.apache.pdfbox.pdmodel.common.PDRectangle;
 import org.apache.pdfbox.pdmodel.documentinterchange.markedcontent.PDMarkedContent;
 import org.springframework.stereotype.Service;

+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingFinishedEvent;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
@@ -35,7 +35,6 @@ import com.knecon.fforesight.service.layoutparser.processor.python_api.model.tab
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
 import com.knecon.fforesight.service.layoutparser.processor.services.BodyTextFrameService;
 import com.knecon.fforesight.service.layoutparser.processor.services.RulingCleaningService;
-import com.knecon.fforesight.service.layoutparser.processor.services.SectionGridCreatorService;
 import com.knecon.fforesight.service.layoutparser.processor.services.SectionsBuilderService;
 import com.knecon.fforesight.service.layoutparser.processor.services.SimplifiedSectionTextService;
 import com.knecon.fforesight.service.layoutparser.processor.services.TableExtractionService;
@@ -50,6 +49,7 @@ import com.knecon.fforesight.service.layoutparser.processor.services.mapper.Docu
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.TaasDocumentDataMapper;
 import com.knecon.fforesight.service.layoutparser.processor.services.parsing.PDFLinesTextStripper;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService;
+import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;

 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
@@ -64,7 +64,6 @@ public class LayoutParsingPipeline {
    private final CvTableParsingAdapter cvTableParsingAdapter;
    private final LayoutParsingStorageService layoutParsingStorageService;
    private final SectionsBuilderService sectionsBuilderService;
-    private final SectionGridCreatorService sectionGridCreatorService;
    private final TaasClassificationService taasClassificationService;
    private final RedactManagerClassificationService redactManagerClassificationService;
    private final DocuMineClassificationService docuMineClassificationService;
@@ -98,7 +97,6 @@ public class LayoutParsingPipeline {

            int numberOfPages = originDocument.getNumberOfPages();

-            layoutParsingStorageService.storeSectionGrid(layoutParsingRequest, sectionGridCreatorService.createSectionGrid(documentGraph));
            layoutParsingStorageService.storeDocumentData(layoutParsingRequest, DocumentDataMapper.toDocumentData(documentGraph));
            layoutParsingStorageService.storeSimplifiedText(layoutParsingRequest, simplifiedSectionTextService.toSimplifiedText(documentGraph));

@@ -116,16 +114,45 @@ public class LayoutParsingPipeline {
                    .identifier(layoutParsingRequest.identifier())
                    .numberOfPages(numberOfPages)
                    .duration(System.currentTimeMillis() - start)
-                    .message(format("Layout parsing is finished and files have been saved with Ids:\n Structure: %s\nText: %s\nPositions: %s\nPageData: %s",
+                    .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(numberOfPages, documentGraph.buildSemanticNodeCounts()),
                            layoutParsingRequest.structureFileStorageId(),
                            layoutParsingRequest.textBlockFileStorageId(),
                            layoutParsingRequest.positionBlockFileStorageId(),
-                            layoutParsingRequest.pageFileStorageId()))
+                            layoutParsingRequest.pageFileStorageId(),
+                            layoutParsingRequest.simplifiedTextStorageId(),
+                            layoutParsingRequest.viewerDocumentStorageId()))
                    .build();
        }
    }


+    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));
+    }
+
+
    @SneakyThrows
    public ClassificationDocument parseLayout(LayoutParsingType layoutParsingType,
                                              PDDocument originDocument,
@@ -160,18 +187,12 @@ public class LayoutParsingPipeline {
            boolean isLandscape = pdr.getWidth() > pdr.getHeight() && (rotation == 0 || rotation == 180) || pdr.getHeight() > pdr.getWidth() && (rotation == 90 || rotation == 270);

            PDRectangle cropbox = pdPage.getCropBox();
-            CleanRulings cleanRulings = rulingCleaningService.getCleanRulings(pdfTableCells.get(pageNumber),
-                    stripper.getRulings(),
-                    stripper.getMinCharWidth(),
-                    stripper.getMaxCharHeight());
+            CleanRulings cleanRulings = rulingCleaningService.getCleanRulings(pdfTableCells.get(pageNumber), stripper.getRulings());

            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 DOCUMINE ->
-                        docuMineBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
+                case REDACT_MANAGER -> redactManagerBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
+                case TAAS -> taasBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
+                case DOCUMINE -> docuMineBlockificationService.blockify(stripper.getTextPositionSequences(), cleanRulings.getHorizontal(), cleanRulings.getVertical());
            };
            classificationPage.setCleanRulings(cleanRulings);
            classificationPage.setRotation(rotation);
@@ -189,7 +210,8 @@ public class LayoutParsingPipeline {
                imageServiceResponseAdapter.findOcr(classificationPage);
            }

-            tableExtractionService.extractTables(cleanRulings, classificationPage, layoutParsingType);
+            tableExtractionService.extractTables(cleanRulings, classificationPage);
+
            buildPageStatistics(classificationPage);
            increaseDocumentStatistics(classificationPage, classificationDocument);

@@ -211,18 +233,20 @@ public class LayoutParsingPipeline {
    }


-    private Map<String, Rectangle2D> convertMarkedContents(List<PDMarkedContent> pdMarkedContents) {
-        Map<String, Rectangle2D> markedContentBboxes = new HashMap<>();
+    private Map<String, List<Rectangle2D>> convertMarkedContents(List<PDMarkedContent> pdMarkedContents) {
+
+        Map<String, List<Rectangle2D>> markedContentBboxes = new HashMap<>();
        markedContentBboxes.put(MarkedContentUtils.HEADER, MarkedContentUtils.getMarkedContentBboxPerLine(pdMarkedContents, MarkedContentUtils.HEADER));
        markedContentBboxes.put(MarkedContentUtils.FOOTER, MarkedContentUtils.getMarkedContentBboxPerLine(pdMarkedContents, MarkedContentUtils.FOOTER));
        return markedContentBboxes;
    }

+
    private void increaseDocumentStatistics(ClassificationPage classificationPage, ClassificationDocument document) {

-        if (!classificationPage.isLandscape()) {
-            document.getFontSizeCounter().addAll(classificationPage.getFontSizeCounter().getCountPerValue());
-        }
+       if (!classificationPage.isLandscape()) {
+        document.getFontSizeCounter().addAll(classificationPage.getFontSizeCounter().getCountPerValue());
+      }
        document.getFontCounter().addAll(classificationPage.getFontCounter().getCountPerValue());
        document.getTextHeightCounter().addAll(classificationPage.getTextHeightCounter().getCountPerValue());
        document.getFontStyleCounter().addAll(classificationPage.getFontStyleCounter().getCountPerValue());
@@ -17,7 +17,6 @@ import org.apache.pdfbox.pdmodel.PDDocument;
 import org.springframework.stereotype.Service;

 import com.fasterxml.jackson.databind.ObjectMapper;
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.section.SectionGrid;
 import com.iqser.red.storage.commons.service.StorageService;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.SimplifiedText;
@@ -84,11 +83,6 @@ public class LayoutParsingStorageService {
    }


-    public void storeSectionGrid(LayoutParsingRequest layoutParsingRequest, SectionGrid sectionGrid) {
-
-        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.sectionGridStorageId(), sectionGrid);
-    }
-

    public void storeResearchDocumentData(LayoutParsingRequest layoutParsingRequest, ResearchDocumentData researchDocumentData) {

@@ -72,9 +72,27 @@ public abstract class AbstractPageBlock {
    }


-    public boolean intersectsY(AbstractPageBlock atc) {
+    public boolean intersectsY(AbstractPageBlock apb) {

-        return this.minY <= atc.getMaxY() && this.maxY >= atc.getMinY();
+        return this.minY <= apb.getMaxY() && this.maxY >= apb.getMinY();
+    }
+
+
+    public boolean almostIntersects(AbstractPageBlock apb, float yThreshold, float xThreshold) {
+
+        return this.almostIntersectsX(apb, xThreshold) && this.almostIntersectsY(apb, yThreshold);
+    }
+
+
+    private boolean almostIntersectsY(AbstractPageBlock apb, float threshold) {
+
+        return this.minY - threshold <= apb.getMaxY() && this.maxY + threshold >= apb.getMinY();
+    }
+
+
+    private boolean almostIntersectsX(AbstractPageBlock apb, float threshold) {
+
+        return this.minX - threshold <= apb.getMaxX() && this.maxX + threshold >= apb.getMinX();
    }


@@ -3,7 +3,6 @@ package com.knecon.fforesight.service.layoutparser.processor.model;
 import java.util.ArrayList;
 import java.util.List;

-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.section.SectionGrid;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.StringFrequencyCounter;
 import com.knecon.fforesight.service.layoutparser.processor.model.text.UnclassifiedText;

@@ -25,7 +24,6 @@ public class ClassificationDocument {
    private StringFrequencyCounter fontStyleCounter = new StringFrequencyCounter();
    private boolean headlines;

-    private SectionGrid sectionGrid = new SectionGrid();
    private long rulesVersion;

 }
@@ -42,6 +42,6 @@ public class ClassificationPage {

    CleanRulings cleanRulings;

-    private Map<String, Rectangle2D> markedContentBboxPerType = new HashMap<>();
+    private Map<String, List<Rectangle2D>> markedContentBboxPerType = new HashMap<>();

 }
@@ -100,6 +100,11 @@ public class Document implements GenericSemanticNode {
    }


+    public Map<NodeType, Long> buildSemanticNodeCounts() {
+
+        return streamAllSubNodes().collect(Collectors.groupingBy(SemanticNode::getType, Collectors.counting()));
+    }
+
    @Override
    public String toString() {

@@ -34,7 +34,6 @@ public class Table implements SemanticNode {

    int numberOfRows;
    int numberOfCols;
-
    TextBlock textBlock;

    @Builder.Default
@@ -208,7 +207,6 @@ public class Table implements SemanticNode {
        return IntStream.range(0, numberOfCols).boxed().map(col -> getCell(row, col));
    }

-
    /**
     * Streams all TableCells row-wise and filters them with header == true.
     *
@@ -1,12 +1,14 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.table;

 import java.awt.geom.Point2D;
+import java.awt.geom.Rectangle2D;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
 import java.util.TreeMap;
+import java.util.stream.Collectors;

 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
@@ -252,7 +254,8 @@ public class TablePageBlock extends AbstractPageBlock {
                if (prevY != null && prevX != null) {
                    var cell = new Cell(new Point2D.Float(prevX, prevY), new Point2D.Float(x, y));

-                    var intersectionCell = cells.stream().filter(c -> cell.intersects(c) && cell.overlapRatio(c) > 0.1f).findFirst();
+                    var intersectionCell = cells.stream().filter(c -> intersects(cell, c)).findFirst();
+
                    intersectionCell.ifPresent(value -> cell.getTextBlocks().addAll(value.getTextBlocks()));
                    if (cell.hasMinimumSize()) {
                        row.add(cell);
@@ -273,6 +276,21 @@ public class TablePageBlock extends AbstractPageBlock {
    }


+
+    public boolean intersects(Cell cell1, Cell cell2) {
+        if (cell1.getHeight() <= 0 || cell2.getHeight() <= 0) {
+            return false;
+        }
+        double x0 = cell1.getX() + 2;
+        double y0 = cell1.getY() + 2;
+        return (cell2.x + cell2.width > x0 &&
+                cell2.y + cell2.height > y0 &&
+                cell2.x < x0 + cell1.getWidth() -2 &&
+                cell2.y < y0 + cell1.getHeight() -2);
+    }
+
+
+
    @Override
    public String getText() {

@@ -17,7 +17,6 @@ import lombok.SneakyThrows;
@AllArgsConstructor
 public class RedTextPosition {

-    private String textMatrix;
    private float[] position;

    @JsonIgnore
@@ -56,8 +55,6 @@ public class RedTextPosition {

        pos.setFontSizeInPt(textPosition.getFontSizeInPt());

-        pos.setTextMatrix(textPosition.getTextMatrix().toString());
-
        var position = new float[4];

        position[0] = textPosition.getXDirAdj();
@@ -224,6 +224,9 @@ public class TextPositionSequence implements CharSequence {
    @JsonAttribute(ignore = true)
    public String getFont() {

+        if (textPositions.get(0).getFontName() == null) {
+            return "none";
+        }
        return textPositions.get(0).getFontName().toLowerCase(Locale.ROOT).replaceAll(",bold", "").replaceAll(",italic", "");
    }

@@ -231,7 +234,9 @@ public class TextPositionSequence implements CharSequence {
    @JsonIgnore
    @JsonAttribute(ignore = true)
    public String getFontStyle() {
-
+        if (textPositions.get(0).getFontName() == null) {
+            return "standard";
+        }
        String lowercaseFontName = textPositions.get(0).getFontName().toLowerCase(Locale.ROOT);

        if (lowercaseFontName.contains("bold") && lowercaseFontName.contains("italic")) {
@@ -243,7 +248,6 @@ public class TextPositionSequence implements CharSequence {
        } else {
            return "standard";
        }
-
    }


@@ -8,6 +8,7 @@ import java.util.Map;

 import org.springframework.stereotype.Service;

+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.PageInfo;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableCells;
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;

@@ -24,24 +25,26 @@ public class CvTableParsingAdapter {
        Map<Integer, List<TableCells>> tableCells = new HashMap<>();
        tableServiceResponse.getData()
                .forEach(tableData -> tableCells.computeIfAbsent(tableData.getPageInfo().getNumber(), tableCell -> new ArrayList<>())
-                        .addAll(convertTableCells(tableData.getTableCells())));
+                        .addAll(convertTableCells(tableData.getTableCells(), tableData.getPageInfo())));

        return tableCells;
    }


-    private Collection<TableCells> convertTableCells(List<TableCells> tableCells) {
+    private Collection<TableCells> convertTableCells(List<TableCells> tableCells, PageInfo pageInfo) {

        List<TableCells> cvParsedTableCells = new ArrayList<>();

-        tableCells.forEach(t -> cvParsedTableCells.add(TableCells.builder()
-                .y0(t.getY0())
-                .x1(t.getX1())
-                .y1(t.getY1())
-                .x0(t.getX0())
-                .width(t.getWidth())
-                .height(t.getHeight())
-                .build()));
+        tableCells.stream()
+                .filter(cell -> cell.getWidth() < pageInfo.getWidth() * 0.98 && cell.getHeight() < pageInfo.getHeight() * 0.98)
+                .forEach(t -> cvParsedTableCells.add(TableCells.builder()
+                        .y0(t.getY0())
+                        .x1(t.getX1())
+                        .y1(t.getY1())
+                        .x0(t.getX0())
+                        .width(t.getWidth())
+                        .height(t.getHeight())
+                        .build()));

        return cvParsedTableCells;
    }
@@ -12,9 +12,9 @@ import java.util.Map;

 import org.springframework.stereotype.Service;

-import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableCells;
 import com.knecon.fforesight.service.layoutparser.processor.model.table.CleanRulings;
 import com.knecon.fforesight.service.layoutparser.processor.model.table.Ruling;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableCells;
 import com.knecon.fforesight.service.layoutparser.processor.utils.DoubleComparisons;

 import lombok.RequiredArgsConstructor;
@@ -25,10 +25,13 @@ import lombok.extern.slf4j.Slf4j;
@RequiredArgsConstructor
 public class RulingCleaningService {

-    public CleanRulings getCleanRulings(List<TableCells> tableCells, List<Ruling> rulings, float minCharWidth, float maxCharHeight) {
+    private static final float THRESHOLD = 6;
+
+
+    public CleanRulings getCleanRulings(List<TableCells> tableCells, List<Ruling> rulings) {

        if (!rulings.isEmpty()) {
-            snapPoints(rulings, minCharWidth, maxCharHeight);
+            snapPoints(rulings);
        }

        List<Ruling> vrs = new ArrayList<>();
@@ -53,14 +56,11 @@ public class RulingCleaningService {
        }
        List<Ruling> horizontalRulingLines = collapseOrientedRulings(hrs);

-        return CleanRulings.builder()
-                .vertical(verticalRulingLines)
-                .horizontal(horizontalRulingLines)
-                .build();
+        return CleanRulings.builder().vertical(verticalRulingLines).horizontal(horizontalRulingLines).build();
    }


-    public void snapPoints(List<? extends Line2D.Float> rulings, float xThreshold, float yThreshold) {
+    public void snapPoints(List<? extends Line2D.Float> rulings) {

        // collect points and keep a Line -> p1,p2 map
        Map<Line2D.Float, Point2D[]> linesToPoints = new HashMap<>();
@@ -81,7 +81,7 @@ public class RulingCleaningService {

        for (Point2D p : points.subList(1, points.size() - 1)) {
            List<Point2D> last = groupedPoints.get(groupedPoints.size() - 1);
-            if (Math.abs(p.getX() - last.get(0).getX()) < xThreshold) {
+            if (Math.abs(p.getX() - last.get(0).getX()) < THRESHOLD) {
                groupedPoints.get(groupedPoints.size() - 1).add(p);
            } else {
                groupedPoints.add(new ArrayList<>(Collections.singletonList(p)));
@@ -108,7 +108,7 @@ public class RulingCleaningService {

        for (Point2D p : points.subList(1, points.size() - 1)) {
            List<Point2D> last = groupedPoints.get(groupedPoints.size() - 1);
-            if (Math.abs(p.getY() - last.get(0).getY()) < yThreshold) {
+            if (Math.abs(p.getY() - last.get(0).getY()) < THRESHOLD) {
                groupedPoints.get(groupedPoints.size() - 1).add(p);
            } else {
                groupedPoints.add(new ArrayList<>(Collections.singletonList(p)));
@@ -1,146 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services;
-
-import java.awt.geom.Rectangle2D;
-import java.util.HashMap;
-import java.util.List;
-import java.util.Map;
-import java.util.Objects;
-import java.util.Set;
-import java.util.function.BiConsumer;
-import java.util.function.BinaryOperator;
-import java.util.function.Function;
-import java.util.function.Supplier;
-import java.util.stream.Collector;
-import java.util.stream.Stream;
-
-import org.springframework.stereotype.Service;
-
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.Point;
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.section.CellRectangle;
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.section.SectionGrid;
-import com.iqser.red.service.persistence.service.v1.api.shared.model.redactionlog.section.SectionRectangle;
-import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.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 lombok.RequiredArgsConstructor;
-
-@Service
-@RequiredArgsConstructor
-public class SectionGridCreatorService {
-
-    public SectionGrid createSectionGrid(Document document) {
-
-        Map<Integer, List<SectionRectangle>> sectionBBox = document.streamAllSubNodesOfType(NodeType.SECTION).map(SemanticNode::getBBox).collect(new SectionGridCollector());
-        Map<Integer, List<SectionRectangle>> paragraphBBox = document.streamAllSubNodesOfType(NodeType.PARAGRAPH).map(SemanticNode::getBBox).collect(new SectionGridCollector());
-        Map<Integer, List<SectionRectangle>> headlineBBox = document.streamAllSubNodesOfType(NodeType.HEADLINE).map(SemanticNode::getBBox).collect(new SectionGridCollector());
-        Map<Integer, List<SectionRectangle>> tableBBox = document.streamAllSubNodesOfType(NodeType.TABLE).map(node -> (Table) node).collect(new TableGridCollector());
-        var sectionGrid = new SectionGrid();
-
-        sectionGrid.setRectanglesPerPage(mergeMapsByConcatenatingLists(//
-                mergeMapsByConcatenatingLists(paragraphBBox, headlineBBox), //
-                mergeMapsByConcatenatingLists(sectionBBox, tableBBox)));
-
-        return sectionGrid;
-    }
-
-
-    private static abstract class GridCollector<T> implements Collector<T, Map<Integer, List<SectionRectangle>>, Map<Integer, List<SectionRectangle>>> {
-
-        @Override
-        public Supplier<Map<Integer, List<SectionRectangle>>> supplier() {
-
-            return HashMap::new;
-        }
-
-
-        @Override
-        public Function<Map<Integer, List<SectionRectangle>>, Map<Integer, List<SectionRectangle>>> finisher() {
-
-            return Function.identity();
-        }
-
-
-        @Override
-        public BinaryOperator<Map<Integer, List<SectionRectangle>>> combiner() {
-
-            return SectionGridCreatorService::mergeMapsByConcatenatingLists;
-        }
-
-
-        @Override
-        public Set<Characteristics> characteristics() {
-
-            return Set.of(Characteristics.IDENTITY_FINISH, Characteristics.CONCURRENT, Characteristics.UNORDERED);
-        }
-
-    }
-
-    private static class TableGridCollector extends GridCollector<Table> {
-
-        @Override
-        public BiConsumer<Map<Integer, List<SectionRectangle>>, Table> accumulator() {
-
-            return (map, table) -> table.getPages()
-                    .forEach(page -> map.merge(page.getNumber(), List.of(toSectionRectangle(table, page, table.getPages().size())), SectionGridCreatorService::concatLists));
-        }
-
-
-        private static SectionRectangle toSectionRectangle(Table table, Page page, int numberOfParts) {
-
-            Rectangle2D rect = table.getBBox().get(page);
-            List<CellRectangle> tableCellRectangles = table.streamTableCells()
-                    .map(TableCell::getBBox)
-                    .map(map -> map.get(page))
-                    .filter(Objects::nonNull)
-                    .map(rectangle2D -> new CellRectangle(new Point((float) rectangle2D.getX(), (float) rectangle2D.getY()),
-                            (float) rectangle2D.getWidth(),
-                            (float) rectangle2D.getHeight()))
-                    .toList();
-            return new SectionRectangle(new Point((float) rect.getX(), (float) rect.getY()),
-                    (float) rect.getWidth(),
-                    (float) rect.getHeight(),
-                    1,
-                    numberOfParts,
-                    tableCellRectangles);
-        }
-
-    }
-
-    private static class SectionGridCollector extends GridCollector<Map<Page, Rectangle2D>> {
-
-        @Override
-        public BiConsumer<Map<Integer, List<SectionRectangle>>, Map<Page, Rectangle2D>> accumulator() {
-
-            return (mapToKeep, mapToMerge) -> mapToMerge.forEach((page, rectangle) -> mapToKeep.merge(page.getNumber(),
-                    List.of(toSectionRectangle(rectangle, mapToMerge.values().size())),
-                    SectionGridCreatorService::concatLists));
-
-        }
-
-
-        private static SectionRectangle toSectionRectangle(Rectangle2D rect, int numberOfParts) {
-
-            return new SectionRectangle(new Point((float) rect.getX(), (float) rect.getY()), (float) rect.getWidth(), (float) rect.getHeight(), 1, numberOfParts, null);
-        }
-
-    }
-
-
-    private static Map<Integer, List<SectionRectangle>> mergeMapsByConcatenatingLists(Map<Integer, List<SectionRectangle>> mapToKeep,
-                                                                                      Map<Integer, List<SectionRectangle>> mapToMerge) {
-
-        mapToMerge.forEach((page, rectangle) -> mapToKeep.merge(page, rectangle, SectionGridCreatorService::concatLists));
-        return mapToKeep;
-    }
-
-
-    private static List<SectionRectangle> concatLists(List<SectionRectangle> l1, List<SectionRectangle> l2) {
-
-        return Stream.concat(l1.stream(), l2.stream()).toList();
-    }
-
-}
@@ -12,7 +12,6 @@ import java.util.Set;

 import org.springframework.stereotype.Service;

-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.ClassificationPage;
 import com.knecon.fforesight.service.layoutparser.processor.model.table.Cell;
@@ -66,6 +65,17 @@ public class TableExtractionService {
    };


+    public boolean contains(Cell cell, double x, double y, double w, double h) {
+
+        if (cell.isEmpty() || w <= 0 || h <= 0) {
+            return false;
+        }
+        double x0 = cell.getX();
+        double y0 = cell.getY();
+        return (x >= x0 - 2 && y >= y0 - 2 && (x + w) <= x0 + cell.getWidth() + 2 && (y + h) <= y0 + cell.getHeight() + 2);
+    }
+
+
    /**
     * Finds tables on a page and moves textblocks into cells of the found tables.
     * Note: This algorithm uses Pdf Coordinate System where {0,0} rotated with the page rotation.
@@ -79,16 +89,17 @@ public class TableExtractionService {
     * @param cleanRulings The lines used to build the table.
     * @param page         Page object that contains textblocks and statistics.
     */
-    public void extractTables(CleanRulings cleanRulings, ClassificationPage page, LayoutParsingType layoutParsingType) {
+    public void extractTables(CleanRulings cleanRulings, ClassificationPage page) {

-        List<Cell> cells = findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical(), layoutParsingType);
+        List<Cell> cells = findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical());

        List<TextPageBlock> toBeRemoved = new ArrayList<>();

        for (AbstractPageBlock abstractPageBlock : page.getTextBlocks()) {
            TextPageBlock textBlock = (TextPageBlock) abstractPageBlock;
            for (Cell cell : cells) {
-                if (cell.hasMinimumSize() && cell.intersects(textBlock.getPdfMinX(),
+                if (cell.hasMinimumSize() && contains(cell,
+                        textBlock.getPdfMinX(),
                        textBlock.getPdfMinY(),
                        textBlock.getPdfMaxX() - textBlock.getPdfMinX(),
                        textBlock.getPdfMaxY() - textBlock.getPdfMinY())) {
@@ -102,7 +113,7 @@ public class TableExtractionService {
        cells = new ArrayList<>(new HashSet<>(cells));
        DoubleComparisons.sort(cells, Rectangle.ILL_DEFINED_ORDER);

-        List<Rectangle> spreadsheetAreas = findSpreadsheetsFromCells(cells).stream().filter(r -> r.getWidth() > 0f && r.getHeight() > 0f).toList();
+        List<Rectangle> spreadsheetAreas = findSpreadsheetsFromCells(cells);

        List<TablePageBlock> tables = new ArrayList<>();
        for (Rectangle area : spreadsheetAreas) {
@@ -135,16 +146,14 @@ public class TableExtractionService {
    }


-    public List<Cell> findCells(List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines, LayoutParsingType layoutParsingType) {
+    public List<Cell> findCells(List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines) {

-        if (layoutParsingType.equals(LayoutParsingType.TAAS)) {
-            // TODO: breaks some tables, for example "1 Abamectin Prr.pdf" try to fix this upstream in RulingCleaningService
-            for (Ruling r : horizontalRulingLines) {
-                if (r.getX2() < r.getX1()) {
-                    double a = r.getX2();
-                    r.x2 = (float) r.getX1();
-                    r.x1 = (float) a;
-                }
+        // Fix for 211.pdf
+        for (Ruling r : horizontalRulingLines) {
+            if (r.getX2() < r.getX1()) {
+                double a = r.getX2();
+                r.x2 = (float) r.getX1();
+                r.x1 = (float) a;
            }
        }

@@ -1,13 +1,8 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.blockification;

-import java.util.ArrayList;
-import java.util.Iterator;
-import java.util.LinkedList;
-import java.util.List;
-import java.util.regex.Matcher;
-import java.util.regex.Pattern;

-import org.springframework.stereotype.Service;
+// 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;
@@ -16,14 +11,25 @@ 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;
+    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);


    /**
@@ -39,14 +45,28 @@ public class TaasBlockificationService {
    public ClassificationPage blockify(List<TextPositionSequence> textPositions, List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines) {

        List<TextPageBlock> classificationTextBlocks = constructFineGranularTextPageBlocks(textPositions, horizontalRulingLines, verticalRulingLines);
-
-        classificationTextBlocks = mergeFineGranularTextPageBlocks(classificationTextBlocks);
+        classificationTextBlocks = mergeTextPageBlocksAligningX(classificationTextBlocks);
+        classificationTextBlocks = mergeIntersectingTextBlocksUntilConvergence(classificationTextBlocks);

        return new ClassificationPage(new ArrayList<>(classificationTextBlocks.stream().map(classificationTextBlock -> (AbstractPageBlock) classificationTextBlock).toList()));
    }


-    private List<TextPageBlock> mergeFineGranularTextPageBlocks(List<TextPageBlock> classificationTextBlocks) {
+    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<>();
@@ -55,16 +75,29 @@ public class TaasBlockificationService {
        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;
            }
-            boolean alignsXRight = Math.abs(currentTextBlock.getPdfMaxX() - previousTextBlock.getPdfMaxX()) < 1;
-            boolean smallYGap = Math.abs(currentTextBlock.getPdfMaxY() - previousTextBlock.getPdfMinY()) < 5;
-            if (alignsXRight && smallYGap) {
+
+
+            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);
@@ -76,6 +109,23 @@ public class TaasBlockificationService {
    }


+    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();
@@ -128,8 +178,8 @@ public class TaasBlockificationService {


    private List<TextPageBlock> constructFineGranularTextPageBlocks(List<TextPositionSequence> textPositions,
-                                                                                        List<Ruling> horizontalRulingLines,
-                                                                                        List<Ruling> verticalRulingLines) {
+                                                                    List<Ruling> horizontalRulingLines,
+                                                                    List<Ruling> verticalRulingLines) {

        int indexOnPage = 0;
        List<TextPositionSequence> wordClusterToCombine = new ArrayList<>();
@@ -138,18 +188,18 @@ public class TaasBlockificationService {
        float minX = 1000, maxX = 0, minY = 1000, maxY = 0;
        TextPositionSequence prev = null;
        // TODO: make static final constant
-        var listIdentitifier = Pattern.compile("\\b(?:[1-9]|1\\d|20|[ivxlc]|[a-z])\\s*(?:[.)])", Pattern.CASE_INSENSITIVE);
+

        boolean wasSplitted = false;
        Float splitX1 = null;
        for (TextPositionSequence word : textPositions) {

-            Matcher listIdentifierPattern = listIdentitifier.matcher(word.toString());
+            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 < -5;
+            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);
@@ -164,7 +214,7 @@ public class TaasBlockificationService {

                Orientation prevOrientation = null;
                if (!classificationTextBlocks.isEmpty()) {
-                    prevOrientation = classificationTextBlocks.get(classificationTextBlocks.size() - 1).getOrientation();
+                    prevOrientation = classificationTextBlocks.get(classificationTextBlocks.size() - X_ALIGNMENT_THRESHOLD).getOrientation();
                }

                TextPageBlock classificationTextBlock = TextPageBlock.fromTextPositionSequences(wordClusterToCombine);
@@ -133,7 +133,7 @@ public class SearchTextWithTextPositionFactory {

    private static void addTextPositionWithFontType(RedTextPosition currentTextPosition, String fontType, List<Integer> fontTypePositions, int stringIdx) {

-        if (currentTextPosition.getFontName().toLowerCase(Locale.ROOT).contains(fontType)) {
+        if (currentTextPosition.getFontName() != null && currentTextPosition.getFontName().toLowerCase(Locale.ROOT).contains(fontType)) {
            fontTypePositions.add(stringIdx);
        }
    }
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.mapper;

 import java.awt.geom.Rectangle2D;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.Locale;
 import java.util.Map;
@@ -8,6 +9,7 @@ import java.util.Map;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Image;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.ImageType;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Page;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;

@@ -1,18 +1,34 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.parsing;

-import com.knecon.fforesight.service.layoutparser.processor.model.table.Ruling;
-import com.knecon.fforesight.service.layoutparser.processor.model.text.RedTextPosition;
-import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
-import lombok.Getter;
-import lombok.Setter;
-import lombok.SneakyThrows;
-import lombok.extern.slf4j.Slf4j;
+import java.awt.color.CMMException;
+import java.awt.geom.Point2D;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.List;
+
 import org.apache.pdfbox.contentstream.operator.Operator;
 import org.apache.pdfbox.contentstream.operator.OperatorName;
-import org.apache.pdfbox.contentstream.operator.color.*;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColor;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColorN;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingColorSpace;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceCMYKColor;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceGrayColor;
+import org.apache.pdfbox.contentstream.operator.color.SetNonStrokingDeviceRGBColor;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingColor;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingColorN;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingColorSpace;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceCMYKColor;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceGrayColor;
+import org.apache.pdfbox.contentstream.operator.color.SetStrokingDeviceRGBColor;
 import org.apache.pdfbox.contentstream.operator.markedcontent.BeginMarkedContentSequenceWithProperties;
 import org.apache.pdfbox.contentstream.operator.markedcontent.EndMarkedContentSequence;
-import org.apache.pdfbox.contentstream.operator.state.*;
+import org.apache.pdfbox.contentstream.operator.state.SetFlatness;
+import org.apache.pdfbox.contentstream.operator.state.SetLineCapStyle;
+import org.apache.pdfbox.contentstream.operator.state.SetLineDashPattern;
+import org.apache.pdfbox.contentstream.operator.state.SetLineJoinStyle;
+import org.apache.pdfbox.contentstream.operator.state.SetLineMiterLimit;
+import org.apache.pdfbox.contentstream.operator.state.SetLineWidth;
+import org.apache.pdfbox.contentstream.operator.state.SetRenderingIntent;
 import org.apache.pdfbox.contentstream.operator.text.SetFontAndSize;
 import org.apache.pdfbox.cos.COSBase;
 import org.apache.pdfbox.cos.COSNumber;
@@ -21,11 +37,14 @@ import org.apache.pdfbox.pdmodel.PDPage;
 import org.apache.pdfbox.pdmodel.graphics.color.PDColor;
 import org.apache.pdfbox.text.TextPosition;

-import java.awt.color.CMMException;
-import java.awt.geom.Point2D;
-import java.io.IOException;
-import java.util.ArrayList;
-import java.util.List;
+import com.knecon.fforesight.service.layoutparser.processor.model.table.Ruling;
+import com.knecon.fforesight.service.layoutparser.processor.model.text.RedTextPosition;
+import com.knecon.fforesight.service.layoutparser.processor.model.text.TextPositionSequence;
+
+import lombok.Getter;
+import lombok.Setter;
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;

@Getter
@Slf4j
@@ -36,11 +55,6 @@ public class PDFLinesTextStripper extends PDFTextStripper {
    private final List<Ruling> graphicsPath = new ArrayList<>();
    @Setter
    protected PDPage pdpage;
-    private int minCharWidth;
-    private int maxCharWidth;
-    private int minCharHeight;
-    private int maxCharHeight;
-

    private float path_x;
    private float path_y;
@@ -73,7 +87,6 @@ public class PDFLinesTextStripper extends PDFTextStripper {
        this.addOperator(new SetFontAndSize(this));
        this.addOperator(new SetLineWidth(this));

-
        addOperator(new BeginMarkedContentSequenceWithProperties(this));
 //        addOperator(new BeginMarkedContentSequence(this));
        addOperator(new EndMarkedContentSequence(this));
@@ -232,33 +245,15 @@ public class PDFLinesTextStripper extends PDFTextStripper {
                        .get(textPositionSequences.get(textPositionSequences.size() - 1).getTextPositions().size() - 1);
            }

-            int charWidth = (int) textPositions.get(i).getWidthDirAdj();
-            if (charWidth < minCharWidth) {
-                minCharWidth = charWidth;
-            }
-            if (charWidth > maxCharWidth) {
-                maxCharWidth = charWidth;
-            }
-
-            int charHeight = (int) textPositions.get(i).getHeightDir();
-            if (charHeight < minCharHeight) {
-                minCharHeight = charHeight;
-            }
-            if (charWidth > maxCharHeight) {
-                maxCharHeight = charHeight;
-            }
-
            if (i == 0 && (textPositions.get(i).getUnicode().equals(" ") || textPositions.get(i).getUnicode().equals("\u00A0") || textPositions.get(i).getUnicode().equals("\t"))) {
                startIndex++;
                continue;
            }

            // Strange but sometimes this is happening, for example: Metolachlor2.pdf
-            if (i > 0 && textPositions.get(i).getXDirAdj() < textPositions.get(i - 1).getXDirAdj()) {
+            if (checkIfCurrentPositionIsToTheRightOfPreviousPosition(i, textPositions)) {
                List<TextPosition> sublist = textPositions.subList(startIndex, i);
-                if (!(sublist.isEmpty() || sublist.size() == 1 && (sublist.get(0).getUnicode().equals(" ") || sublist.get(0).getUnicode().equals("\u00A0") || sublist.get(0)
-                        .getUnicode()
-                        .equals("\t")))) {
+                if (checkIfSequenceContainsOnlyWhitespaces(sublist)) {
                    textPositionSequences.add(new TextPositionSequence(sublist, pageNumber, i == textPositions.size() - 1 && isParagraphStart));
                }
                startIndex = i;
@@ -266,9 +261,7 @@ public class PDFLinesTextStripper extends PDFTextStripper {

            if (textPositions.get(i).getRotation() == 0 && i > 0 && textPositions.get(i).getX() > textPositions.get(i - 1).getEndX() + 1) {
                List<TextPosition> sublist = textPositions.subList(startIndex, i);
-                if (!(sublist.isEmpty() || sublist.size() == 1 && (sublist.get(0).getUnicode().equals(" ") || sublist.get(0).getUnicode().equals("\u00A0") || sublist.get(0)
-                        .getUnicode()
-                        .equals("\t")))) {
+                if (checkIfSequenceContainsOnlyWhitespaces(sublist)) {
                    textPositionSequences.add(new TextPositionSequence(sublist, pageNumber, i == textPositions.size() - 1 && isParagraphStart));
                }
                startIndex = i;
@@ -278,13 +271,10 @@ public class PDFLinesTextStripper extends PDFTextStripper {
                    .getUnicode()
                    .equals("\t")) && i <= textPositions.size() - 2) {
                List<TextPosition> sublist = textPositions.subList(startIndex, i);
-                if (!(sublist.isEmpty() || sublist.size() == 1 && (sublist.get(0).getUnicode().equals(" ") || sublist.get(0).getUnicode().equals("\u00A0") || sublist.get(0)
-                        .getUnicode()
-                        .equals("\t")))) {
+                if (checkIfSequenceContainsOnlyWhitespaces(sublist)) {

                    // Remove false sequence ends (whitespaces)
-                    if (previous != null && sublist.get(0).getYDirAdj() == previous.getYDirAdj() && sublist.get(0)
-                            .getXDirAdj() - (previous.getXDirAdj() + previous.getWidthDirAdj()) < 0.01) {
+                    if (checkIfGapSizeBetweenCharactersSmallerThanMaximum(previous, sublist, 0.01f)) {
                        for (TextPosition t : sublist) {
                            textPositionSequences.get(textPositionSequences.size() - 1).add(t);
                        }
@@ -319,13 +309,29 @@ public class PDFLinesTextStripper extends PDFTextStripper {
    }


+    public boolean checkIfCurrentPositionIsToTheRightOfPreviousPosition(int i, List<TextPosition> textPositions) {
+
+        return i > 0 && textPositions.get(i).getXDirAdj() < textPositions.get(i - 1).getXDirAdj();
+    }
+
+
+    public boolean checkIfSequenceContainsOnlyWhitespaces(List<TextPosition> sublist) {
+
+        return !(sublist.isEmpty() || sublist.size() == 1 && (sublist.get(0).getUnicode().equals(" ") || sublist.get(0).getUnicode().equals("\u00A0") || sublist.get(0)
+                .getUnicode()
+                .equals("\t")));
+    }
+
+
+    public boolean checkIfGapSizeBetweenCharactersSmallerThanMaximum(RedTextPosition previous, List<TextPosition> sublist, float maximumGapSize) {
+
+        return previous != null && sublist.get(0).getYDirAdj() == previous.getYDirAdj() && sublist.get(0)
+                .getXDirAdj() - (previous.getXDirAdj() + previous.getWidthDirAdj()) < maximumGapSize;
+    }
+
    @Override
    public String getText(PDDocument doc) throws IOException {

-        minCharWidth = Integer.MAX_VALUE;
-        maxCharWidth = 0;
-        minCharHeight = Integer.MAX_VALUE;
-        maxCharHeight = 0;
        textPositionSequences.clear();
        rulings.clear();
        graphicsPath.clear();
@@ -1711,7 +1711,7 @@ public class PDFTextStripper extends LegacyPDFStreamEngine {
        int numberOfStrings = line.size();
        for (int i = 0; i < numberOfStrings; i++) {
            WordWithTextPositions word = line.get(i);
-            word.getTextPositions().sort(Comparator.comparing(TextPosition::getX));
+            word.getTextPositions().sort(Comparator.comparing(TextPosition::getXDirAdj));
            writeString(word.getText(), word.getTextPositions(), isParagraphEnd && i == numberOfStrings - 1);
            if (i < numberOfStrings - 1) {
                writeWordSeparator();
@@ -51,7 +51,6 @@ public class ViewerDocumentService {
    @SneakyThrows
    public void createViewerDocument(PDDocument pdDocument, Document document, OutputStream outputStream, boolean layerVisibilityDefaultValue) {

-        log.info("Start Viewer Document Creation");
        LayoutGrid layoutGrid = layoutGridService.createLayoutGrid(document);
        // PDDocument.save() is very slow, since it actually traverses the entire pdf and writes a new one.
        // If we collect all COSDictionaries we changed and tell it explicitly to only add the changed ones by using saveIncremental it's very fast.
@@ -122,7 +121,6 @@ public class ViewerDocumentService {
        dictionariesToUpdate.add(pdDocument.getDocumentInformation().getCOSObject());
 //        dictionariesToUpdate.add(pdDocument.getDocument().getTrailer());
        pdDocument.saveIncremental(outputStream, dictionariesToUpdate);
-        log.info("Saved Viewer Document");
    }


@@ -9,6 +9,7 @@ import org.apache.pdfbox.text.TextPosition;

 import java.awt.geom.Rectangle2D;
 import java.util.Collection;
+import java.util.Collections;
 import java.util.List;
 import java.util.Map;
 import java.util.stream.Collectors;
@@ -19,10 +20,10 @@ public class MarkedContentUtils {
    public static final String HEADER = "Header";
    public static final String FOOTER = "Footer";

-    public Rectangle2D getMarkedContentBboxPerLine(List<PDMarkedContent> markedContents, String subtype) {
+    public List<Rectangle2D> getMarkedContentBboxPerLine(List<PDMarkedContent> markedContents, String subtype) {

        if (markedContents == null) {
-            return null;
+            return Collections.emptyList();
        }

        var markedContentByYPosition = markedContents.stream()
@@ -37,19 +38,19 @@ public class MarkedContentUtils {
                .collect(Collectors.groupingBy(TextPosition::getY));

        if (markedContentByYPosition.isEmpty()) {
-            return null;
+            return Collections.emptyList();
        }

-        return RectangleTransformations.rectangle2DBBox(markedContentByYPosition.values().stream()
+        return markedContentByYPosition.values().stream()
                .map(textPositions -> new TextPositionSequence(textPositions.stream()
                        .toList(), 0, true)
                        .getRectangle())
-                .map(t -> new Rectangle2D.Float(t.getTopLeft().getX(), t.getTopLeft().getY() - Math.abs(t.getHeight()), t.getWidth(), Math.abs(t.getHeight()))).collect(Collectors.toList()));
+                .map(t -> new Rectangle2D.Float(t.getTopLeft().getX(), t.getTopLeft().getY() - Math.abs(t.getHeight()), t.getWidth(), Math.abs(t.getHeight()))).collect(Collectors.toList());
    }


-    public boolean intersects(TextPageBlock textBlock, Map<String, Rectangle2D> markedContentBboxPerType, String type) {
-        return markedContentBboxPerType.get(type) != null && markedContentBboxPerType.get(type).intersects(textBlock.getPdfMinX(), textBlock.getPdfMinY(), textBlock.getWidth(), textBlock.getHeight());
+    public boolean intersects(TextPageBlock textBlock, Map<String, List<Rectangle2D>> markedContentBboxPerType, String type) {
+        return markedContentBboxPerType.get(type) != null && markedContentBboxPerType.get(type).stream().anyMatch(rectangle -> rectangle.intersects(textBlock.getPdfMinX(), textBlock.getPdfMinY(), textBlock.getWidth(), textBlock.getHeight()));
    }

 }
@@ -19,9 +19,10 @@ public final class PositionUtils {

        double threshold = textBlock.getMostPopularWordHeight() * 3;

-        if (textBlock.getPdfMinX() + threshold > btf.getTopLeft().getX() && textBlock.getPdfMaxX() - threshold < btf.getTopLeft()
-                .getX() + btf.getWidth() && textBlock.getPdfMinY() + threshold > btf.getTopLeft().getY() && textBlock.getPdfMaxY() - threshold < btf.getTopLeft()
-                .getY() + btf.getHeight()) {
+        if (textBlock.getPdfMinX() + threshold > btf.getTopLeft().getX()
+                && textBlock.getPdfMaxX() - threshold < btf.getTopLeft().getX() + btf.getWidth()
+                && textBlock.getPdfMinY() + threshold > btf.getTopLeft().getY()
+                && textBlock.getPdfMaxY() - threshold < btf.getTopLeft().getY() + btf.getHeight()) {
            return true;
        } else {
            return false;
@@ -19,8 +19,8 @@ dependencies {
    implementation(project(":layoutparser-service-processor"))
    implementation(project(":layoutparser-service-internal-api"))

-    implementation("com.iqser.red.commons:storage-commons:2.40.0")
-    implementation("com.knecon.fforesight:tenant-commons:0.10.0")
+    implementation("com.iqser.red.commons:storage-commons:2.45.0")
+    implementation("com.knecon.fforesight:tenant-commons:0.14.0")
    implementation("org.springframework.boot:spring-boot-starter-actuator:3.1.3")
    implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
    implementation("org.springframework.boot:spring-boot-starter-amqp:3.1.3")
@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.server.queue;

+import java.util.Arrays;
+
 import org.springframework.amqp.AmqpRejectAndDontRequeueException;
 import org.springframework.amqp.core.Message;
 import org.springframework.amqp.rabbit.annotation.RabbitHandler;
@@ -45,12 +47,12 @@ public class MessageHandler {
        }
        LayoutParsingFinishedEvent layoutParsingFinishedEvent = layoutParsingPipeline.parseLayoutAndSaveFilesToStorage(layoutParsingRequest);
        sendLayoutParsingFinishedEvent(layoutParsingFinishedEvent);
-        log.info("Layout parsing finished {} in {} ms", layoutParsingRequest.identifier(), layoutParsingFinishedEvent.duration());
    }


    public void sendLayoutParsingFinishedEvent(LayoutParsingFinishedEvent layoutParsingFinishedEvent) {

+        Arrays.stream(layoutParsingFinishedEvent.message().split("\n")).forEach(log::info);
        rabbitTemplate.convertAndSend(LayoutParsingQueueNames.LAYOUT_PARSING_FINISHED_EVENT_QUEUE, layoutParsingFinishedEvent);
    }

@@ -0,0 +1,34 @@
+package com.knecon.fforesight.service.layoutparser.server;
+
+import java.util.Arrays;
+
+import org.junit.jupiter.api.Test;
+import org.springframework.beans.factory.annotation.Autowired;
+
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingFinishedEvent;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
+import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingPipeline;
+import com.knecon.fforesight.service.layoutparser.server.utils.AbstractTest;
+
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;
+
+@Slf4j
+public class LayoutparserEnd2EndTest extends AbstractTest {
+
+    @Autowired
+    private LayoutParsingPipeline layoutParsingPipeline;
+
+
+    @Test
+    @SneakyThrows
+    public void testLayoutParserEndToEnd() {
+
+        prepareStorage("files/bdr/btd_19_053_1905391.pdf");
+        LayoutParsingRequest layoutParsingRequest = buildDefaultLayoutParsingRequest(LayoutParsingType.TAAS);
+        LayoutParsingFinishedEvent finishedEvent = layoutParsingPipeline.parseLayoutAndSaveFilesToStorage(layoutParsingRequest);
+        Arrays.stream(finishedEvent.message().split("\n")).forEach(log::info);
+    }
+
+}
@@ -47,7 +47,7 @@ public class DocumentGraphVisualizationTest extends BuildDocumentTest {
    @Disabled
    public void visualizeCraftedDocument() {

-        String filename = "files/crafted document.pdf";
+        String filename = "files/1 Abamectin_prr.pdf";
        visualizePdf(filename);
    }

@@ -0,0 +1,203 @@
+package com.knecon.fforesight.service.layoutparser.server.graph;
+
+import java.io.File;
+import java.io.IOException;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+import lombok.SneakyThrows;
+import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.cos.COSArray;
+import org.apache.pdfbox.cos.COSBase;
+import org.apache.pdfbox.cos.COSDictionary;
+import org.apache.pdfbox.cos.COSNumber;
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.PDPage;
+import org.apache.pdfbox.pdmodel.documentinterchange.logicalstructure.PDStructureElement;
+import org.apache.pdfbox.pdmodel.documentinterchange.logicalstructure.PDStructureNode;
+import org.apache.pdfbox.pdmodel.documentinterchange.markedcontent.PDMarkedContent;
+import org.apache.pdfbox.text.PDFMarkedContentExtractor;
+import org.apache.pdfbox.text.TextPosition;
+import org.junit.jupiter.api.BeforeEach;
+import org.junit.jupiter.api.Disabled;
+import org.junit.jupiter.api.Test;
+import org.springframework.core.io.ClassPathResource;
+
+/**
+ * @author mkl
+ */
+@Disabled
+public class ExtractMarkedContentTest {
+    final static File RESULT_FOLDER = new File("target/test-outputs", "extract");
+
+    @BeforeEach
+    public void setUpBeforeClass() throws Exception {
+        RESULT_FOLDER.mkdirs();
+    }
+
+    /**
+     * <a href="https://stackoverflow.com/questions/54956720/how-to-replace-a-space-with-a-word-while-extract-the-data-from-pdf-using-pdfbox">
+     * How to replace a space with a word while extract the data from PDF using PDFBox
+     * </a>
+     * <br/>
+     * <a href="https://drive.google.com/open?id=10ZkdPlGWzMJeahwnQPzE6V7s09d1nvwq">
+     * test.pdf
+     * </a> as "testWPhromma.pdf"
+     * <p>
+     * This test shows how to, in principle, extract tagged text.
+     * </p>
+     */
+    @Test
+    @SneakyThrows
+    public void testExtractTestWPhromma() throws IOException {
+        System.out.printf("\n\n===\n%s\n===\n", "testWPhromma.pdf");
+        try (PDDocument document = Loader.loadPDF(new ClassPathResource("files/bdr/Drucksache_19_9865.pdf").getFile())) {
+
+            Map<PDPage, Map<Integer, PDMarkedContent>> markedContents = new HashMap<>();
+
+            for (PDPage page : document.getPages()) {
+                PDFMarkedContentExtractor extractor = new PDFMarkedContentExtractor();
+                extractor.processPage(page);
+
+                Map<Integer, PDMarkedContent> theseMarkedContents = new HashMap<>();
+                markedContents.put(page, theseMarkedContents);
+                for (PDMarkedContent markedContent : extractor.getMarkedContents()) {
+                    theseMarkedContents.put(markedContent.getMCID(), markedContent);
+                }
+            }
+
+            PDStructureNode root = document.getDocumentCatalog().getStructureTreeRoot();
+            showStructure(root, markedContents);
+            document.close();
+        }
+    }
+
+    /**
+     * <a href="https://stackoverflow.com/questions/59192443/get-tags-related-bboxs-even-though-there-is-no-attributes-a-in-document-cata">
+     * Get tag's related BBox's even though there is no attributes (/A in document catalog structure) related to Layout in PDFBox?
+     * </a>
+     * <br/>
+     * <a href="https://drive.google.com/file/d/1_-tuWuReaTvrDsqQwldTnPYrMHSpXIWp/view?usp=sharing">
+     * res_multipage.pdf
+     * </a>
+     * <p>
+     * This test shows how to, in principle, extract tagged text from this document.
+     * </p>
+     */
+    @Test
+    public void testExtractResMultipage() throws IOException {
+        System.out.printf("\n\n===\n%s\n===\n", "res_multipage.pdf");
+
+        try(PDDocument document = Loader.loadPDF(new ClassPathResource("files/bdr/Drucksache_19_9865.pdf").getFile())) {
+
+            Map<PDPage, Map<Integer, PDMarkedContent>> markedContents = new HashMap<>();
+
+            for (PDPage page : document.getPages()) {
+                PDFMarkedContentExtractor extractor = new PDFMarkedContentExtractor();
+                extractor.processPage(page);
+
+                Map<Integer, PDMarkedContent> theseMarkedContents = new HashMap<>();
+                markedContents.put(page, theseMarkedContents);
+                for (PDMarkedContent markedContent : extractor.getMarkedContents()) {
+                    theseMarkedContents.put(markedContent.getMCID(), markedContent);
+                }
+            }
+
+            PDStructureNode root = document.getDocumentCatalog().getStructureTreeRoot();
+            showStructure(root, markedContents);
+            document.close();
+        }
+    }
+
+    /**
+     * <a href="https://issues.apache.org/jira/browse/PDFBOX-5613">
+     * PDFBOX-5613 - uncorrent paragraph split
+     * </a>
+     * <br/>
+     * <a href="https://issues.apache.org/jira/secure/attachment/13058595/Daily%20Report.pdf">
+     * Daily Report.pdf
+     * </a>
+     * <p>
+     * Making use of the marked content, the contents of the file can better be extracted.
+     * </p>
+     */
+    @Test
+    public void testExtractDailyReport() throws IOException {
+        System.out.printf("\n\n===\n%s\n===\n", "Daily Report.pdf");
+        try (PDDocument document = Loader.loadPDF(new ClassPathResource("files/bdr/Drucksache_19_9865.pdf").getFile())) {
+
+            Map<PDPage, Map<Integer, PDMarkedContent>> markedContents = new HashMap<>();
+
+            for (PDPage page : document.getPages()) {
+                PDFMarkedContentExtractor extractor = new PDFMarkedContentExtractor();
+                extractor.processPage(page);
+
+                Map<Integer, PDMarkedContent> theseMarkedContents = new HashMap<>();
+                markedContents.put(page, theseMarkedContents);
+                for (PDMarkedContent markedContent : extractor.getMarkedContents()) {
+                    theseMarkedContents.put(markedContent.getMCID(), markedContent);
+                }
+            }
+
+            PDStructureNode root = document.getDocumentCatalog().getStructureTreeRoot();
+            showStructure(root, markedContents);
+            document.close();
+        }
+    }
+
+    /**
+     * @see #testExtractTestWPhromma()
+     */
+    void showStructure(PDStructureNode node, Map<PDPage, Map<Integer, PDMarkedContent>> markedContents) {
+        String structType = null;
+        PDPage page = null;
+        if (node instanceof PDStructureElement) {
+            PDStructureElement element = (PDStructureElement) node;
+            structType = element.getStructureType();
+            page = element.getPage();
+        }
+        Map<Integer, PDMarkedContent> theseMarkedContents = markedContents.get(page);
+        System.out.printf("<%s>\n", structType);
+        for (Object object : node.getKids()) {
+            if (object instanceof COSArray) {
+                for (COSBase base : (COSArray) object) {
+                    if (base instanceof COSDictionary) {
+                        showStructure(PDStructureNode.create((COSDictionary) base), markedContents);
+                    } else if (base instanceof COSNumber) {
+                        showContent(((COSNumber)base).intValue(), theseMarkedContents);
+                    } else {
+                        System.out.printf("?%s\n", base);
+                    }
+                }
+            } else if (object instanceof PDStructureNode) {
+                showStructure((PDStructureNode) object, markedContents);
+            } else if (object instanceof Integer) {
+                showContent((Integer)object, theseMarkedContents);
+            } else {
+                System.out.printf("?%s\n", object);
+            }
+
+        }
+        System.out.printf("</%s>\n", structType);
+    }
+
+    /**
+     * @see #showStructure(PDStructureNode, Map)
+     * @see #testExtractTestWPhromma()
+     */
+    void showContent(int mcid, Map<Integer, PDMarkedContent> theseMarkedContents) {
+        PDMarkedContent markedContent = theseMarkedContents != null ? theseMarkedContents.get(mcid) : null;
+        List<Object> contents = markedContent != null ? markedContent.getContents() : Collections.emptyList();
+        StringBuilder textContent =  new StringBuilder();
+        for (Object object : contents) {
+            if (object instanceof TextPosition) {
+                textContent.append(((TextPosition)object).getUnicode());
+            } else {
+                textContent.append("?" + object);
+            }
+        }
+        System.out.printf("%s\n", textContent);
+    }
+}
@@ -2,14 +2,31 @@ package com.knecon.fforesight.service.layoutparser.server.graph;

 import java.io.FileOutputStream;
 import java.nio.file.Path;
+import java.util.List;

 import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.pdmodel.PDDocument;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
+import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.core.io.ClassPathResource;

+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
+import com.knecon.fforesight.service.layoutparser.processor.model.ClassificationDocument;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
+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.python_api.model.image.ImageServiceResponse;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
+import com.knecon.fforesight.service.layoutparser.processor.services.SectionsBuilderService;
+import com.knecon.fforesight.service.layoutparser.processor.services.classification.RedactManagerClassificationService;
+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.PropertiesMapper;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService;
 import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest;
@@ -18,19 +35,38 @@ import lombok.SneakyThrows;

 public class ViewerDocumentTest extends BuildDocumentTest {

+    @Autowired
+    private SectionsBuilderService sectionsBuilderService;
+
+    @Autowired
+    private RedactManagerClassificationService redactManagerClassificationService;
+
    @Test
-    @Disabled
    @SneakyThrows
    public void testViewerDocument() {

+        String fileName = "files/bdr/notMergedParagraphs.pdf";
+        String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
        LayoutGridService layoutGridService = new LayoutGridService();
        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(layoutGridService);
-        String fileName = "files/new/VV-511309_OCR.pdf";
-        Document document = buildGraph(fileName, LayoutParsingType.DOCUMINE);
-        String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
+        Document document = buildGraph(fileName, LayoutParsingType.TAAS);
        try (var pdDocument = Loader.loadPDF(new ClassPathResource(fileName).getFile()); var out = new FileOutputStream(tmpFileName)) {
            viewerDocumentService.createViewerDocument(pdDocument, document, out, true);
        }
    }

+    public ClassificationDocument buildClassificationDocument(PDDocument originDocument) {
+
+        ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
+                originDocument,
+                new ImageServiceResponse(),
+                new TableServiceResponse());
+
+        redactManagerClassificationService.classifyDocument(classificationDocument);
+
+        sectionsBuilderService.buildSections(classificationDocument);
+
+        return classificationDocument;
+    }
+
 }
@@ -27,14 +27,14 @@ public class TextPositionSequenceTest {
            + "  \"pageWidth\": 600\n"  //
            + "}";

-    private final ObjectSerializer objectSerializer = new ObjectSerializer(new ObjectMapper(), new StorageProperties());
+    private final ObjectSerializer objectSerializer = new ObjectSerializer(new ObjectMapper());


    @Test
    @SneakyThrows
    public void testDeserializationWithJackson() {

-        TextPositionSequence textPositionSequence = objectSerializer.deserializeWithJackson(new ByteArrayInputStream(TEXT_POSITION_SEQUENCE_AS_JSON.getBytes(StandardCharsets.UTF_8)),
+        TextPositionSequence textPositionSequence = objectSerializer.deserialize(new ByteArrayInputStream(TEXT_POSITION_SEQUENCE_AS_JSON.getBytes(StandardCharsets.UTF_8)),
                TextPositionSequence.class);

        assertPropertiesAfterJsonDeserialization(textPositionSequence);
@@ -1,5 +1,27 @@
 package com.knecon.fforesight.service.layoutparser.server.segmentation;

+import static org.assertj.core.api.Assertions.assertThat;
+
+import java.awt.geom.Rectangle2D;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.nio.file.Path;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Locale;
+import java.util.Map;
+import java.util.stream.Collectors;
+
+import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.junit.jupiter.api.Disabled;
+import org.junit.jupiter.api.Test;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.core.io.ClassPathResource;
+
 import com.fasterxml.jackson.databind.ObjectMapper;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingPipeline;
@@ -15,19 +37,8 @@ import com.knecon.fforesight.service.layoutparser.processor.python_api.model.tab
 import com.knecon.fforesight.service.layoutparser.processor.services.SectionsBuilderService;
 import com.knecon.fforesight.service.layoutparser.processor.services.classification.RedactManagerClassificationService;
 import com.knecon.fforesight.service.layoutparser.server.utils.AbstractTest;
+
 import lombok.SneakyThrows;
-import org.apache.pdfbox.Loader;
-import org.apache.pdfbox.pdmodel.PDDocument;
-import org.junit.jupiter.api.Test;
-import org.springframework.beans.factory.annotation.Autowired;
-import org.springframework.core.io.ClassPathResource;
-
-import java.awt.geom.Rectangle2D;
-import java.io.IOException;
-import java.util.*;
-import java.util.stream.Collectors;
-
-import static org.assertj.core.api.Assertions.assertThat;

 public class PdfSegmentationServiceTest extends AbstractTest {

@@ -50,12 +61,13 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    private SectionsBuilderService sectionsBuilderService;


-    public ClassificationDocument buildClassificationDocument(PDDocument originDocument) {
+    @SneakyThrows
+    public ClassificationDocument buildClassificationDocument(PDDocument originDocument, TableServiceResponse tableServiceResponse) {

-        ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.DOCUMINE,
+        ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
                originDocument,
                new ImageServiceResponse(),
-                new TableServiceResponse());
+                tableServiceResponse);

        redactManagerClassificationService.classifyDocument(classificationDocument);

@@ -65,6 +77,58 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    }


+    @SneakyThrows
+    public ClassificationDocument buildClassificationDocument(PDDocument originDocument) {
+
+        return buildClassificationDocument(originDocument, new TableServiceResponse());
+    }
+
+
+    @Test
+    public void tablesToHtmlDebugger() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/A20622A izRMS (CZ) fRR Part B9_Page185.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
+
+        toHtml(document, "/tmp/A20622A izRMS (CZ) fRR Part B9_Page185.html");
+
+    }
+
+
+    @Test
+    public void tablesToHtmlDebuggerWithCVResponse() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.pdf");
+        ClassPathResource cvTablesResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.TABLES.json");
+        var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()), tableServiceResponse);
+
+        toHtml(document, "/tmp/ScanRotationBorder.html");
+    }
+
+
+    @Test
+    public void testScanRotationBorderIsIgnored() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.pdf");
+        ClassPathResource cvTablesResource = new ClassPathResource("files/cv_tables/ScanRotationBorder.TABLES.json");
+        var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(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();
+
+        // 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.
+        tables.forEach(table -> {
+            assertThat(table.getMinX()).isGreaterThan(75);
+            assertThat(table.getMaxX()).isLessThan(512);
+        });
+    }
+
+
    @Test
    @SneakyThrows
    public void testMapping() {
@@ -155,7 +219,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    }


-    @Test
+    @Test // Non-sense test
    public void testDoc56Page170() throws IOException {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/56 Fludioxonil_RAR_12_Volume_3CA_B-7_2018-02-21_Page170.pdf");
@@ -166,8 +230,25 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        validateTable(document, 0, 1, 1, 0, 0);
        validateTable(document, 1, 2, 2, 0, 0);
-        validateTable(document, 2, 7, 20, 0, 140);
-        validateTable(document, 3, 8, 31, 0, 170);
+        validateTable(document, 2, 6, 20, 0, 0);
+        validateTable(document, 3, 7, 31, 0, 0);
+
+    }
+
+
+    @Test
+    public void testDoc211() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/211.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
+
+        validateTableSize(document, 4);
+
+        validateTable(document, 0, 5, 4, 0, 0);
+        validateTable(document, 1, 5, 15, 14, 0);
+        validateTable(document, 2, 5, 14, 11, 0);
+        validateTable(document, 3, 5, 3, 0, 0);

    }

@@ -181,7 +262,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        validateTableSize(document, 1);

-        validateTable(document, 0, 8, 8, 0, 2);
+        validateTable(document, 0, 8, 8, 0, 0);

        List<List<String>> values = Arrays.asList(Arrays.asList("Annex point Reference within DAR/RAR",
                        "Author, date",
@@ -191,18 +272,18 @@ public class PdfSegmentationServiceTest extends AbstractTest {
                        "Method meets analytical validation criteria",
                        "Remarks (in case validation criteria are not met)",
                        "Acceptability of the method"),
-                Arrays.asList("",
+                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 in a Trial Carried",
+                        "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",
@@ -220,6 +301,8 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));

+        toHtml(document, "/tmp/html.html");
+
        validateTableSize(document, 4);

        validateTable(document, 0, 3, 2, 0, 0);
@@ -231,17 +314,29 @@ public class PdfSegmentationServiceTest extends AbstractTest {


    @Test
+    @Disabled // FIXME Fake Redactions leads to more cells, no solution for this currently
    public void testDocA20622APartB9Page185() throws IOException {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/A20622A izRMS (CZ) fRR Part B9_Page185.pdf");

        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));

-        validateTableSize(document, 2);
+        validateTableSize(document, 1);

-        validateTable(document, 0, 5, 5, 0, 23);
-        validateTable(document, 1, 11, 9, 0, 36);
+        validateTable(document, 0, 7, 4, 0, 0);
+    }

+
+    @Test
+    public void testDocA20622APartB9Page185FixedDoc() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/A20622A izRMS (CZ) fRR Part B9_Page185_fixed.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
+
+        validateTableSize(document, 1);
+
+        validateTable(document, 0, 7, 4, 0, 0);
    }


@@ -328,7 +423,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));

        validateTableSize(document, 1);
-        validateTable(document, 0, 10, 6, 0, 1);
+        validateTable(document, 0, 10, 6, 0, 0);

    }

@@ -450,8 +545,8 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        validateTableSize(document, 2);

-        validateTable(document, 0, 6, 8, 0, 2);
-        validateTable(document, 1, 6, 8, 0, 1);
+        validateTable(document, 0, 6, 8, 0, 0);
+        validateTable(document, 1, 6, 8, 0, 0);

    }

@@ -484,12 +579,37 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    }


+    @SneakyThrows
+    private void toHtml(ClassificationDocument document, String filename) {
+
+        var tables = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList();
+        StringBuilder sb = new StringBuilder();
+
+        int currentPage = 1;
+        for (var table : tables) {
+            if (currentPage != table.getPage()) {
+                currentPage = table.getPage();
+                sb.append("---------------------- Page ").append(currentPage).append("--------------\n");
+            }
+            sb.append("\n\n");
+            sb.append(table.getTextAsHtml());
+        }
+
+        try (FileOutputStream fileOutputStream = new FileOutputStream(Path.of(filename).toFile())) {
+            fileOutputStream.write(sb.toString().getBytes());
+        }
+    }
+
+
    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);
        List<List<Cell>> rows = table.getRows();
        int emptyCellsFoundFound = rows.stream().flatMap(List::stream).toList().stream().filter(f -> f.toString().equals("")).toList().size();

+        for (List<Cell> row : table.getRows()) {
+            row.forEach(r -> System.out.println(r.toString()));
+        }
        assertThat(emptyCellsFoundFound).isEqualTo(emptyCellsCountCorrect + emptyCellsCountIncorrect);

        assertThat(table.getColCount()).isEqualTo(colCount);
@@ -1,21 +1,39 @@
 package com.knecon.fforesight.service.layoutparser.server.services;

+import java.nio.file.Files;
 import java.nio.file.Path;
 import java.util.Collections;
 import java.util.LinkedList;
 import java.util.List;

+import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.pdmodel.PDDocument;
 import org.junit.jupiter.api.Test;
+import org.springframework.core.io.ClassPathResource;

+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentData;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.DocumentStructure;
+import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageContents;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
+import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
 import com.knecon.fforesight.service.layoutparser.processor.model.table.CleanRulings;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.image.ImageServiceResponse;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
 import com.knecon.fforesight.service.layoutparser.processor.services.PageContentExtractor;
 import com.knecon.fforesight.service.layoutparser.processor.services.RulingCleaningService;
+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.PropertiesMapper;
+import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
+import com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService;
+import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest;
 import com.knecon.fforesight.service.layoutparser.server.utils.visualizations.PdfDraw;

 import lombok.SneakyThrows;

-public class RulingCleaningServiceTest {
+public class RulingCleaningServiceTest extends BuildDocumentTest {

    @Test
 //    @Disabled
@@ -25,13 +43,96 @@ public class RulingCleaningServiceTest {
        String fileName = "files/211.pdf";
        String lineFileName = "/tmp/" + Path.of(fileName).getFileName().toString() + "_LINES.pdf";
        List<PageContents> pageContents = PageContentExtractor.getSortedPageContents(fileName);
-        PdfDraw.drawLinesPerPage(fileName, pageContents.stream().map(PageContents::getRulings).toList(), lineFileName);
-
        RulingCleaningService rulingCleaningService = new RulingCleaningService();
+        PdfDraw.drawLinesPerPage(fileName, pageContents.stream().map(PageContents::getRulings).toList(), lineFileName);
        List<CleanRulings> cleanRulingsPerPage = new LinkedList<>();
        for (PageContents pageContent : pageContents) {
-            cleanRulingsPerPage.add(rulingCleaningService.getCleanRulings(Collections.emptyList(), pageContent.getRulings(), 8, 20));
+            cleanRulingsPerPage.add(rulingCleaningService.getCleanRulings(Collections.emptyList(), pageContent.getRulings()));
+        }
+
+    }
+
+
+    @Test
+    @SneakyThrows
+    public void testTableExtraction() {
+
+        LayoutGridService layoutGridService = new LayoutGridService();
+        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(layoutGridService);
+
+        ClassPathResource resource = new ClassPathResource("files");
+        List<String> pdfFileNames = Files.walk(resource.getFile().toPath())
+                .filter(path -> path.getFileName().toString().endsWith(".pdf"))
+                .map(Path::toAbsolutePath)
+                .map(Path::toString)
+                .toList();
+
+        for (int i = 0; i < pdfFileNames.size(); i++) {
+            writeJsons(Path.of(pdfFileNames.get(i)));
        }
    }

+
+    @SneakyThrows
+    private void writeJsons(Path filename) {
+
+        Document documentGraphBefore = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
+                Loader.loadPDF(filename.toFile()),
+                new ImageServiceResponse(),
+                new TableServiceResponse()));
+        Document documentGraphAfter = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
+                Loader.loadPDF(filename.toFile()),
+                new ImageServiceResponse(),
+                new TableServiceResponse()));
+        DocumentData documentDataBefore = DocumentDataMapper.toDocumentData(documentGraphBefore);
+        DocumentData documentDataAfter = DocumentDataMapper.toDocumentData(documentGraphAfter);
+        if (!compareStructures(documentDataBefore.getDocumentStructure(), documentDataAfter.getDocumentStructure(), filename.getFileName().toString())) {
+            String tmpFileNameBefore = "C:/Users/YANNIK~1/AppData/Local/Temp/before." + filename.getFileName().toString();
+            try (PDDocument pdDocument = Loader.loadPDF(filename.toFile())) {
+                PdfDraw.drawDocumentGraph(pdDocument, documentGraphBefore);
+                pdDocument.save(tmpFileNameBefore);
+            }
+            String tmpFileNameAfter = "C:/Users/YANNIK~1/AppData/Local/Temp/after." + filename.getFileName().toString();
+            try (PDDocument pdDocument = Loader.loadPDF(filename.toFile())) {
+                PdfDraw.drawDocumentGraph(pdDocument, documentGraphAfter);
+                pdDocument.save(tmpFileNameAfter);
+
+            }
+        }
+    }
+
+
+    @SneakyThrows
+    private boolean compareStructures(DocumentStructure structure1, DocumentStructure structure2, String pdfName) {
+
+        List listStructure1 = structure1.streamAllEntries()
+                .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
+                .map(DocumentStructure.EntryData::getProperties)
+                .map(properties -> {
+                    var builder = Table.builder();
+                    PropertiesMapper.parseTableProperties(properties, builder);
+                    return builder.build();
+                })
+                .toList();
+
+        List listStructure2 = structure2.streamAllEntries()
+                .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
+                .map(DocumentStructure.EntryData::getProperties)
+                .map(properties -> {
+                    var builder = Table.builder();
+                    PropertiesMapper.parseTableProperties(properties, builder);
+                    return builder.build();
+                })
+                .toList();
+
+        for (int i = 0; i < listStructure1.size(); i++) {
+            Table tableNode1 = (Table) listStructure1.get(i);
+            Table tableNode2 = (Table) listStructure2.get(i);
+            if (tableNode1.getNumberOfRows() != tableNode2.getNumberOfRows() || tableNode1.getNumberOfCols() != tableNode2.getNumberOfCols()) {
+                return false;
+            }
+        }
+        return true;
+    }
+
 }
@@ -1,8 +1,15 @@
 package com.knecon.fforesight.service.layoutparser.server.utils;

-import java.io.InputStream;
-import java.util.Optional;
-
+import com.iqser.red.commons.jackson.ObjectMapperFactory;
+import com.iqser.red.storage.commons.service.StorageService;
+import com.iqser.red.storage.commons.utils.FileSystemBackedStorageService;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
+import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
+import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingStorageService;
+import com.knecon.fforesight.service.layoutparser.server.Application;
+import com.knecon.fforesight.tenantcommons.TenantContext;
+import com.knecon.fforesight.tenantcommons.TenantsClient;
+import lombok.SneakyThrows;
 import org.junit.jupiter.api.AfterEach;
 import org.junit.jupiter.api.BeforeEach;
 import org.junit.jupiter.api.extension.ExtendWith;
@@ -12,21 +19,12 @@ import org.springframework.boot.autoconfigure.EnableAutoConfiguration;
 import org.springframework.boot.autoconfigure.amqp.RabbitAutoConfiguration;
 import org.springframework.boot.test.context.SpringBootTest;
 import org.springframework.boot.test.mock.mockito.MockBean;
-import org.springframework.context.annotation.ComponentScan;
-import org.springframework.context.annotation.Configuration;
-import org.springframework.context.annotation.Import;
+import org.springframework.context.annotation.*;
 import org.springframework.core.io.ClassPathResource;
 import org.springframework.test.context.junit.jupiter.SpringExtension;

-import com.iqser.red.storage.commons.service.StorageService;
-import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingRequest;
-import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
-import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingStorageService;
-import com.knecon.fforesight.service.layoutparser.server.Application;
-import com.knecon.fforesight.tenantcommons.TenantContext;
-import com.knecon.fforesight.tenantcommons.TenantsClient;
-
-import lombok.SneakyThrows;
+import java.io.InputStream;
+import java.util.Optional;

@ExtendWith(SpringExtension.class)
@SpringBootTest(classes = Application.class, webEnvironment = SpringBootTest.WebEnvironment.RANDOM_PORT)
@@ -53,7 +51,6 @@ public abstract class AbstractTest {
    protected final static String POSITION_FILE_ID = "positions";
    protected final static String PAGES_FILE_ID = "pages";
    protected final static String TENANT_ID = "tenant";
-    protected final static String SECTION_GRID_ID = "section";
    protected final static String VIEWER_DOCUMENT_ID = "viewer";
    protected final static String SIMPLIFIED_ID = "simplified";

@@ -70,7 +67,6 @@ public abstract class AbstractTest {
                .positionBlockFileStorageId(POSITION_FILE_ID)
                .pageFileStorageId(PAGES_FILE_ID)
                .simplifiedTextStorageId(SIMPLIFIED_ID)
-                .sectionGridStorageId(SECTION_GRID_ID)
                .viewerDocumentStorageId(VIEWER_DOCUMENT_ID)
                .build();
    }
@@ -101,8 +97,13 @@ public abstract class AbstractTest {
    protected LayoutParsingRequest prepareStorage(InputStream fileInputStream) {

        storageService.storeObject(TenantContext.getTenantId(), ORIGIN_FILE_ID, fileInputStream);
+        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+    }
+
+    protected LayoutParsingRequest buildDefaultLayoutParsingRequest(LayoutParsingType layoutParsingType) {
+
        return LayoutParsingRequest.builder()
-                .layoutParsingType(LayoutParsingType.REDACT_MANAGER)
+                .layoutParsingType(layoutParsingType)
                .originFileStorageId(ORIGIN_FILE_ID)
                .tablesFileStorageId(Optional.of(TABLE_FILE_ID))
                .imagesFileStorageId(Optional.of(IMAGE_FILE_ID))
@@ -111,12 +112,10 @@ public abstract class AbstractTest {
                .positionBlockFileStorageId(POSITION_FILE_ID)
                .pageFileStorageId(PAGES_FILE_ID)
                .simplifiedTextStorageId(SIMPLIFIED_ID)
-                .sectionGridStorageId(SECTION_GRID_ID)
                .viewerDocumentStorageId(VIEWER_DOCUMENT_ID)
                .build();
    }

-
    @SneakyThrows
    protected LayoutParsingRequest prepareStorage(String file, String cvServiceResponseFile, String imageInfoFile) {

@@ -135,19 +134,7 @@ public abstract class AbstractTest {
        storageService.storeObject(TenantContext.getTenantId(), TABLE_FILE_ID, cvServiceResponseFileStream);
        storageService.storeObject(TenantContext.getTenantId(), ORIGIN_FILE_ID, fileStream);

-        return LayoutParsingRequest.builder()
-                .layoutParsingType(LayoutParsingType.REDACT_MANAGER)
-                .originFileStorageId(ORIGIN_FILE_ID)
-                .tablesFileStorageId(Optional.of(TABLE_FILE_ID))
-                .imagesFileStorageId(Optional.of(IMAGE_FILE_ID))
-                .structureFileStorageId(STRUCTURE_FILE_ID)
-                .textBlockFileStorageId(TEXT_FILE_ID)
-                .positionBlockFileStorageId(POSITION_FILE_ID)
-                .pageFileStorageId(PAGES_FILE_ID)
-                .simplifiedTextStorageId(SIMPLIFIED_ID)
-                .sectionGridStorageId(SECTION_GRID_ID)
-                .viewerDocumentStorageId(VIEWER_DOCUMENT_ID)
-                .build();
+        return buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
    }


@@ -165,6 +152,14 @@ public abstract class AbstractTest {
    @ComponentScan("com.knecon.fforesight.service.layoutparser")
    public static class TestConfiguration {

+
+        @Bean
+        @Primary
+        public StorageService inmemoryStorage() {
+
+            return new FileSystemBackedStorageService(ObjectMapperFactory.create());
+        }
+
    }

 }
@@ -1,124 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.server.utils;
-
-import static java.io.File.createTempFile;
-
-import java.io.File;
-import java.io.FileInputStream;
-import java.io.FileOutputStream;
-import java.io.InputStream;
-import java.util.ArrayList;
-import java.util.HashMap;
-import java.util.List;
-import java.util.Map;
-import java.util.stream.Collectors;
-
-import org.apache.commons.io.IOUtils;
-import org.springframework.stereotype.Service;
-
-import com.fasterxml.jackson.databind.ObjectMapper;
-import com.iqser.red.commons.jackson.ObjectMapperFactory;
-import com.iqser.red.storage.commons.exception.StorageObjectDoesNotExist;
-import com.iqser.red.storage.commons.service.StorageService;
-
-import lombok.SneakyThrows;
-
-@Service
-public class FileSystemBackedStorageService implements StorageService {
-
-    private final Map<String, File> dataMap = new HashMap<>();
-
-
-    public FileSystemBackedStorageService() {
-
-    }
-
-
-    @SneakyThrows
-    @Override
-    public void downloadTo(String tenantId, String objectId, File destinationFile) {
-
-        var res = dataMap.get(objectId);
-        if (res == null) {
-            throw new StorageObjectDoesNotExist(new RuntimeException());
-        }
-
-        IOUtils.copy(new FileInputStream(res), new FileOutputStream(destinationFile));
-    }
-
-
-    @Override
-    public void deleteObject(String tenantId, String objectId) {
-
-        dataMap.remove(objectId);
-    }
-
-
-    @Override
-    public boolean objectExists(String tenantId, String objectId) {
-
-        return dataMap.containsKey(objectId);
-    }
-
-
-    @Override
-    @SneakyThrows
-    public <T> void storeJSONObject(String tenantId, String objectId, T any) {
-
-        File tempFile = createTempFile("test", ".tmp");
-        getMapper().writeValue(new FileOutputStream(tempFile), any);
-        dataMap.put(objectId, tempFile);
-    }
-
-
-    private ObjectMapper getMapper() {
-
-        return ObjectMapperFactory.create();
-    }
-
-
-    @Override
-    @SneakyThrows
-    public <T> T readJSONObject(String tenantId, String objectId, Class<T> clazz) {
-
-        if (dataMap.get(objectId) == null || !dataMap.get(objectId).exists()) {
-            throw new StorageObjectDoesNotExist("Stored object not found");
-        }
-        return getMapper().readValue(new FileInputStream(dataMap.get(objectId)), clazz);
-    }
-
-
-    public List<String> listPaths() {
-
-        return new ArrayList<>(dataMap.keySet());
-    }
-
-
-    public List<String> listFilePaths() {
-
-        return dataMap.values().stream().map(File::getAbsolutePath).collect(Collectors.toList());
-    }
-
-
-    @Override
-    @SneakyThrows
-    public void storeObject(String tenantId, String objectId, InputStream stream) {
-
-        File tempFile = createTempFile("test", ".tmp");
-
-        try (var fileOutputStream = new FileOutputStream(tempFile)) {
-            IOUtils.copy(stream, fileOutputStream);
-        }
-
-        dataMap.put(objectId, tempFile);
-    }
-
-
-    public void clearStorage() {
-
-        this.dataMap.forEach((k, v) -> {
-            v.delete();
-        });
-        this.dataMap.clear();
-    }
-
-}
Author	SHA1	Message	Date
Dominique Eifländer	e2bcf971c9	Merge branch 'DM-589' into 'main' DM-589: Filter wrong detected cells that borders from rotation at scanning See merge request fforesight/layout-parser!83	2023-11-20 16:07:27 +01:00
Dominique Eifländer	dacc2f7f43	DM-589: Filter wrong detected cells that borders from rotation at scanning	2023-11-20 15:54:02 +01:00
Dominique Eifländer	144a9591a2	Merge branch 'TAAS-103-hotfix' into 'main' * added back in if statement See merge request fforesight/layout-parser!82	2023-11-16 12:48:48 +01:00
yhampe	207d9dec97	* added back in if statement * removed not needed commentar	2023-11-16 12:40:49 +01:00
Yannik Hampe	09ee90222e	Merge branch 'TAAS-103' into 'main' TAAS-103: Table Detection and rotated text See merge request fforesight/layout-parser!81	2023-11-16 09:13:41 +01:00
yhampe	1316a067fe	* removed double chechking for height of cell	2023-11-16 08:51:12 +01:00
yhampe	e203210ade	* removed not needed properties	2023-11-16 08:23:58 +01:00
yhampe	b25d46291a	* checkstyle	2023-11-16 08:12:47 +01:00
yhampe	84148d3b6e	* fixed tests	2023-11-16 07:51:08 +01:00
Dominique Eifländer	a6ba66b1aa	TAAS-103: Fixed values in wrong cells	2023-11-15 13:36:46 +01:00
yhampe	c3e69b2cdf	* fixed bug with incorrect empty cell count by adding threshhold to cell.contains	2023-11-15 10:44:47 +01:00
yhampe	f69331e7d8	*renamed page to firstPage in DocumentStructure and Table	2023-11-07 10:21:19 +01:00
yhampe	01493dc033	TAAS-103: Table Detection and rotated text * added page property to DocumentStructure to be able to get page of found tables * added a method to TableExtractionService to get the table area * added calculateMinCharWidthAndMaxCharHeightInsideTable to LayoutParsingPipeline to calculate the values based upon table area * refactored PDFLinesTextStripper for better readability *removed textMatrix from RedTextPosition as it is no longer needed	2023-11-07 08:47:28 +01:00
yhampe	459e0c8be7	TAAS-103:	2023-11-07 08:39:15 +01:00
Kilian Schüttler	1b1f777706	Merge branch 'RED-7806' into 'main' RED-7806 - Specific customer document cannot be processed See merge request fforesight/layout-parser!79	2023-10-25 10:50:34 +02:00
Corina Olariu	0e0a811f9d	RED-7806 - Specific customer document cannot be processed - add brackets	2023-10-25 11:36:54 +03:00
Corina Olariu	efa3d75479	RED-7806 - Specific customer document cannot be processed - check for font name null before using to avoid the NPE	2023-10-25 09:16:47 +03:00
Kilian Schüttler	9abdc6d44d	Merge branch 'RED-7434' into 'main' RED-7434 - Remove Section Grid entirely See merge request fforesight/layout-parser!78	2023-10-20 10:07:01 +02:00
Corina Olariu	3bab61c446	RED-7434 - Remove Section Grid entirely - remove sectionGrid relation (including SectionGridCreatorService) - update junit tests	2023-10-20 09:09:22 +03:00
Dominique Eifländer	d17517d3c3	Merge branch 'hotfix-bdr-doc' into 'main' hotfix: Fixed parsing for specific taas document See merge request fforesight/layout-parser!77	2023-10-18 16:12:15 +02:00
Dominique Eifländer	567cbc178b	hotfix: Fixed parsing for specific taas document	2023-10-17 15:52:19 +02:00
Kilian Schüttler	3c53772765	Merge branch 'RED-7759' into 'main' RED-7759: Upgraded storage-commons to newest windwos compatible version See merge request fforesight/layout-parser!76	2023-10-13 12:21:04 +02:00
Dominique Eifländer	8647cf5a18	RED-7759: Upgraded storage-commons to newest windwos compatible version	2023-10-13 12:15:22 +02:00
Kilian Schüttler	310c07b200	Merge branch 'RED-7607-WIP' into 'main' RED-7607 - Rotating pages leads to lost annotations (RM & DM) See merge request fforesight/layout-parser!75	2023-10-05 13:34:12 +02:00
Corina Olariu	daba0bf8a6	RED-7607 - Rotating pages leads to lost annotations (RM & DM) - remove finally clause	2023-10-04 17:46:46 +03:00
Corina Olariu	3839de215c	RED-7607 - Rotating pages leads to lost annotations (RM & DM) - rollback to getDir().getDegrees()	2023-10-04 15:27:13 +03:00
Corina Olariu	b4d68594f1	RED-7607 - Rotating pages leads to lost annotations (RM & DM) - use rotation instead of getDir().getDegrees()	2023-10-04 14:22:15 +03:00
Corina Olariu	99ed331a1e	RED-7607 - Rotating pages leads to lost annotations (RM & DM) - use getXDirAdj instead of getX - add fontSizeCounter for landscape pages also	2023-10-04 14:13:38 +03:00
Corina Olariu	f2c0991987	RED-7607 - Rotating pages leads to lost annotations (RM & DM) - fix PMD findings	2023-10-04 14:09:46 +03:00
Timo Bejan	b8ef55e6e2	Merge branch 'TAAS-104' into 'main' TAAS-104: merge visually intersecting Paragraphs See merge request fforesight/layout-parser!73	2023-09-05 17:08:40 +02:00
Kilian Schuettler	5792ff4a93	TAAS-104: merge visually intersecting Paragraphs * fix build	2023-09-05 16:54:23 +02:00
Kilian Schuettler	621c3f269d	TAAS-104: merge visually intersecting Paragraphs	2023-09-05 16:09:05 +02:00
Dominique Eifländer	8dba392904	Merge branch 'RED-7461' into 'main' RED-7461: Fixed wrong textblock classifation if footer is marked as header See merge request fforesight/layout-parser!72	2023-09-01 12:14:38 +02:00
deiflaender	306a53ea79	RED-7461: Fixed wrong textblock classifation if footer is marked as header	2023-09-01 12:07:47 +02:00
Kilian Schüttler	754fd8f933	Merge branch 'TAAS-89' into 'main' TAAS-89: added log entry and an end2end test See merge request fforesight/layout-parser!71	2023-08-31 14:40:48 +02:00
Kilian Schuettler	28ec4c9ccb	TAAS-89: added log entry and an end2end test	2023-08-31 14:28:18 +02:00