RED-8550: Faulty table recognition and text duplication leads to huge sections

* cherrypick
Merge branch 'RED-8156' into 'main'
2024-02-21 14:19:48 +01:00 · 2024-02-07 10:42:21 +01:00 · 2024-02-06 17:17:26 +01:00 · 2024-02-06 17:09:21 +01:00 · 2024-02-06 17:06:53 +01:00 · 2024-02-06 17:03:38 +01:00
109 changed files with 2692 additions and 1079 deletions
@@ -10,7 +10,7 @@ deploy:
  script:
    - echo "Building with gradle version ${BUILDVERSION}"
    - gradle -Pversion=${BUILDVERSION} publish
-    - gradle bootBuildImage --cleanCache --publishImage -PbuildbootDockerHostNetwork=true -Pversion=${BUILDVERSION}
+    - gradle bootBuildImage --publishImage -PbuildbootDockerHostNetwork=true -Pversion=${BUILDVERSION}
    - echo "BUILDVERSION=$BUILDVERSION" >> version.env
  artifacts:
    reports:
@@ -24,6 +24,8 @@ tasks.named<Test>("test") {
    reports {
        junitXml.outputLocation.set(layout.buildDirectory.dir("reports/junit"))
    }
+    minHeapSize = "512m"
+    maxHeapSize = "2048m"
 }

 tasks.test {
@@ -9,12 +9,13 @@
    </description>

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

-</ruleset>
+</ruleset>
+
@@ -10,14 +10,14 @@


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

-</ruleset>
+</ruleset>
@@ -1,6 +1,6 @@
 plugins {
    id("com.knecon.fforesight.java-conventions")
-    id("io.freefair.lombok") version "8.2.2"
+    id("io.freefair.lombok") version "8.4"
 }

 description = "layoutparser-service-internal-api"
@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

+import java.io.Serializable;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -12,7 +14,7 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
@Schema(description = "Object containing the complete document layout parsing information. It is split into 4 categories, structure, text, positions and pages: " + "The document tree structure of SemanticNodes such as Section, Paragraph, Headline, etc. " + "The text, which is stored as separate blocks of data. " + "The text positions, which are also stored as separate blocks. The Blocks are equal to the text blocks in length and order. " + "The page information.")
-public class DocumentData {
+public class DocumentData implements Serializable {

    @Schema(description = "Contains information about the document's pages.")
    DocumentPage[] documentPages;
@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

+import java.io.Serializable;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -14,7 +16,7 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
@Schema(description = "Object containing information about the document's pages.")
-public class DocumentPage {
+public class DocumentPage implements Serializable {

    @Schema(description = "The page number, starting with 1.")
    int number;
@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

+import java.io.Serializable;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -14,7 +16,7 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
@Schema(description = "Object containing text positional information of a specific text block. A document is split into multiple text blocks, which are supposed to be read in order. Every text block can only occur on a single page.")
-public class DocumentPositionData {
+public class DocumentPositionData implements Serializable {

    @Schema(description = "Identifier of the text block.")
    Long id;
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

 import java.awt.geom.Rectangle2D;
+import java.io.Serializable;
 import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
@@ -20,14 +21,13 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
@Schema(description = "Object containing information about the parsed tree structure of the SemanticNodes, such as Section, Paragraph, Headline etc inside of the document.")
-public class DocumentStructure {
+public class DocumentStructure implements Serializable {

    @Schema(description = "The root EntryData represents the Document.")
    EntryData root;

-
    @Schema(description = "Object containing the extra field names, a table has in its properties field.")
-    public static class TableProperties {
+    public static class TableProperties implements Serializable {

        public static final String NUMBER_OF_ROWS = "numberOfRows";
        public static final String NUMBER_OF_COLS = "numberOfCols";
@@ -35,7 +35,7 @@ public class DocumentStructure {
    }

    @Schema(description = "Object containing the extra field names, an Image has in its properties field.")
-    public static class ImageProperties {
+    public static class ImageProperties implements Serializable {

        public static final String TRANSPARENT = "transparent";
        public static final String IMAGE_TYPE = "imageType";
@@ -45,7 +45,7 @@ public class DocumentStructure {
    }

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

        public static final String B_BOX = "bBox";
        public static final String ROW = "row";
@@ -56,6 +56,7 @@ public class DocumentStructure {

    public static final String RECTANGLE_DELIMITER = ";";

+
    public static Rectangle2D parseRectangle2D(String bBox) {

        List<Float> floats = Arrays.stream(bBox.split(RECTANGLE_DELIMITER)).map(Float::parseFloat).toList();
@@ -100,7 +101,7 @@ public class DocumentStructure {
    @AllArgsConstructor
    @FieldDefaults(level = AccessLevel.PRIVATE)
    @Schema(description = "Object containing information of a SemanticNode and also structuring the layout with children.")
-    public static class EntryData {
+    public static class EntryData implements Serializable {

        @Schema(description = "Type of the semantic node.", allowableValues = {"DOCUMENT", "SECTION", "PARAGRAPH", "HEADLINE", "TABLE", "TABLE_CELL", "HEADER", "FOOTER", "IMAGE"})
        NodeType type;
@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

+import java.io.Serializable;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AccessLevel;
 import lombok.AllArgsConstructor;
@@ -14,7 +16,7 @@ import lombok.experimental.FieldDefaults;
@AllArgsConstructor
@FieldDefaults(level = AccessLevel.PRIVATE)
@Schema(description = "Object containing text information of a specific text block. A document is split into multiple text blocks, which are supposed to be read in order. Every text block can only occur on a single page.")
-public class DocumentTextData {
+public class DocumentTextData implements Serializable {

    @Schema(description = "Identifier of the text block.")
    Long id;
@@ -1,8 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.redaction;

+import java.io.Serializable;
 import java.util.Locale;

-public enum NodeType {
+public enum NodeType implements Serializable {
    DOCUMENT,
    SECTION,
    HEADLINE,
@@ -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) {

 }
@@ -1,6 +1,6 @@
 plugins {
    id("com.knecon.fforesight.java-conventions")
-    id("io.freefair.lombok") version "8.2.2"
+    id("io.freefair.lombok") version "8.4"
 }

 description = "layoutparser-service-processor"
@@ -10,13 +10,14 @@ val pdfBoxVersion = "3.0.0"

 dependencies {
    implementation(project(":layoutparser-service-internal-api"))
+    implementation(project(":viewer-doc-processor"))

    implementation("com.iqser.red.service:persistence-service-shared-api-v1:2.144.0") {
        exclude("org.springframework.boot", "spring-boot-starter-security")
        exclude("org.springframework.boot", "spring-boot-starter-validation")
    }
-    implementation("com.knecon.fforesight:tenant-commons:0.10.0")
-    implementation("com.iqser.red.commons:storage-commons:2.40.0")
+    implementation("com.knecon.fforesight:tenant-commons:0.21.0")
+    implementation("com.iqser.red.commons:storage-commons:2.45.0")

    implementation("org.apache.pdfbox:pdfbox:${pdfBoxVersion}")
    implementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
@@ -3,20 +3,23 @@ package com.knecon.fforesight.service.layoutparser.processor;
 import static java.lang.String.format;

 import java.awt.geom.Rectangle2D;
-import java.io.ByteArrayOutputStream;
+import java.io.File;
 import java.io.IOException;
+import java.nio.file.Files;
 import java.util.ArrayList;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
+import java.util.concurrent.atomic.AtomicReference;

-import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;
+import org.apache.pdfbox.Loader;
 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 +38,6 @@ 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;
@@ -49,99 +51,183 @@ import com.knecon.fforesight.service.layoutparser.processor.services.factory.Doc
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.DocumentDataMapper;
 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.services.visualization.LayoutGridService;
+import com.knecon.fforesight.service.layoutparser.processor.utils.MarkedContentUtils;

+import io.micrometer.observation.Observation;
+import io.micrometer.observation.ObservationRegistry;
+import io.micrometer.observation.annotation.Observed;
+import lombok.AccessLevel;
 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
+import lombok.experimental.FieldDefaults;
 import lombok.extern.slf4j.Slf4j;

+@SuppressWarnings("PMD.CloseResource")
@Slf4j
@Service
@RequiredArgsConstructor
+@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class LayoutParsingPipeline {

-    private final ImageServiceResponseAdapter imageServiceResponseAdapter;
-    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;
-    private final SimplifiedSectionTextService simplifiedSectionTextService;
-    private final BodyTextFrameService bodyTextFrameService;
-    private final RulingCleaningService rulingCleaningService;
-    private final TableExtractionService tableExtractionService;
-    private final TaasBlockificationService taasBlockificationService;
-    private final DocuMineBlockificationService docuMineBlockificationService;
-    private final RedactManagerBlockificationService redactManagerBlockificationService;
-    private final ViewerDocumentService viewerDocumentService;
+    ImageServiceResponseAdapter imageServiceResponseAdapter;
+    CvTableParsingAdapter cvTableParsingAdapter;
+    LayoutParsingStorageService layoutParsingStorageService;
+    SectionsBuilderService sectionsBuilderService;
+    TaasClassificationService taasClassificationService;
+    RedactManagerClassificationService redactManagerClassificationService;
+    DocuMineClassificationService docuMineClassificationService;
+    SimplifiedSectionTextService simplifiedSectionTextService;
+    BodyTextFrameService bodyTextFrameService;
+    RulingCleaningService rulingCleaningService;
+    TableExtractionService tableExtractionService;
+    TaasBlockificationService taasBlockificationService;
+    DocuMineBlockificationService docuMineBlockificationService;
+    RedactManagerBlockificationService redactManagerBlockificationService;
+    LayoutGridService layoutGridService;
+    ObservationRegistry observationRegistry;


    public LayoutParsingFinishedEvent parseLayoutAndSaveFilesToStorage(LayoutParsingRequest layoutParsingRequest) throws IOException {

        long start = System.currentTimeMillis();
+        log.info("Starting layout parsing for {}", layoutParsingRequest.identifier());

-        try (PDDocument originDocument = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId())) {
-            ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
-            if (layoutParsingRequest.imagesFileStorageId().isPresent()) {
-                imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId().get());
-            }
+        File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId());
+        File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId())
+                .orElse(originFile);

-            TableServiceResponse tableServiceResponse = new TableServiceResponse();
-            if (layoutParsingRequest.tablesFileStorageId().isPresent()) {
-                tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId().get());
-            }
-
-            ClassificationDocument classificationDocument = parseLayout(layoutParsingRequest.layoutParsingType(), originDocument, imageServiceResponse, tableServiceResponse);
-            Document documentGraph = DocumentGraphFactory.buildDocumentGraph(classificationDocument);
-
-            int numberOfPages = originDocument.getNumberOfPages();
-
-            layoutParsingStorageService.storeSectionGrid(layoutParsingRequest, sectionGridCreatorService.createSectionGrid(documentGraph));
-            layoutParsingStorageService.storeDocumentData(layoutParsingRequest, DocumentDataMapper.toDocumentData(documentGraph));
-            layoutParsingStorageService.storeSimplifiedText(layoutParsingRequest, simplifiedSectionTextService.toSimplifiedText(documentGraph));
-
-            try (var out = new ByteArrayOutputStream()) {
-                viewerDocumentService.createViewerDocument(originDocument, documentGraph, out, false);
-                layoutParsingStorageService.storeViewerDocument(layoutParsingRequest, out);
-            }
-
-            if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.TAAS)) {
-                var researchDocumentData = TaasDocumentDataMapper.fromDocument(documentGraph);
-                layoutParsingStorageService.storeResearchDocumentData(layoutParsingRequest, researchDocumentData);
-            }
-
-            return LayoutParsingFinishedEvent.builder()
-                    .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",
-                            layoutParsingRequest.structureFileStorageId(),
-                            layoutParsingRequest.textBlockFileStorageId(),
-                            layoutParsingRequest.positionBlockFileStorageId(),
-                            layoutParsingRequest.pageFileStorageId()))
-                    .build();
+        ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
+        if (layoutParsingRequest.imagesFileStorageId()
+                .isPresent()) {
+            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId()
+                                                                                     .get());
        }
+
+        TableServiceResponse tableServiceResponse = new TableServiceResponse();
+        if (layoutParsingRequest.tablesFileStorageId()
+                .isPresent()) {
+            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId()
+                                                                                     .get());
+        }
+
+        ClassificationDocument classificationDocument = parseLayout(layoutParsingRequest.layoutParsingType(),
+                originFile,
+                imageServiceResponse,
+                tableServiceResponse,
+                layoutParsingRequest.identifier().toString());
+
+        log.info("Building document graph for {}", layoutParsingRequest.identifier());
+
+        Document documentGraph = observeBuildDocumentGraph(classificationDocument);
+
+        log.info("Creating viewer document for {}", layoutParsingRequest.identifier());
+
+        layoutGridService.addLayoutGrid(viewerDocumentFile, documentGraph, viewerDocumentFile, false);
+
+        log.info("Storing resulting files for {}", layoutParsingRequest.identifier());
+
+        layoutParsingStorageService.storeDocumentData(layoutParsingRequest, DocumentDataMapper.toDocumentData(documentGraph));
+        layoutParsingStorageService.storeSimplifiedText(layoutParsingRequest, simplifiedSectionTextService.toSimplifiedText(documentGraph));
+        layoutParsingStorageService.storeViewerDocument(layoutParsingRequest, viewerDocumentFile);
+
+        if (layoutParsingRequest.layoutParsingType().equals(LayoutParsingType.TAAS)) {
+            log.info("Building research document data for {}", layoutParsingRequest.identifier());
+            var researchDocumentData = TaasDocumentDataMapper.fromDocument(documentGraph);
+            layoutParsingStorageService.storeResearchDocumentData(layoutParsingRequest, researchDocumentData);
+        }
+
+        if (!viewerDocumentFile.equals(originFile)) {
+            viewerDocumentFile.delete();
+        }
+        originFile.delete();
+
+        return LayoutParsingFinishedEvent.builder()
+                .identifier(layoutParsingRequest.identifier())
+                .numberOfPages(documentGraph.getNumberOfPages())
+                .duration(System.currentTimeMillis() - start)
+                .message(format("""
+                                        Layout parsing has finished in %.02f s.
+                                        identifiers: %s
+                                        %s
+                                        Files have been saved with Ids:
+                                        Structure: %s
+                                        Text: %s
+                                        Positions: %s
+                                        PageData: %s
+                                        Simplified Text: %s
+                                        Viewer Doc: %s""",
+                                ((float) (System.currentTimeMillis() - start)) / 1000,
+                                layoutParsingRequest.identifier(),
+                                buildSemanticNodeCountMessage(documentGraph.getNumberOfPages(), documentGraph.buildSemanticNodeCounts()),
+                                layoutParsingRequest.structureFileStorageId(),
+                                layoutParsingRequest.textBlockFileStorageId(),
+                                layoutParsingRequest.positionBlockFileStorageId(),
+                                layoutParsingRequest.pageFileStorageId(),
+                                layoutParsingRequest.simplifiedTextStorageId(),
+                                layoutParsingRequest.viewerDocumentStorageId()))
+                .build();
+
+    }
+
+
+    private Document observeBuildDocumentGraph(ClassificationDocument classificationDocument) {
+
+        AtomicReference<Document> documentReference = new AtomicReference<>();
+
+        Observation.createNotStarted("LayoutParsingPipeline", observationRegistry)
+                .contextualName("build-document-graph")
+                .observe(() -> documentReference.set(DocumentGraphFactory.buildDocumentGraph(classificationDocument)));
+
+        return documentReference.get();
+    }
+
+
+    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
+    @Observed(name = "LayoutParsingPipeline", contextualName = "parse-layout")
    public ClassificationDocument parseLayout(LayoutParsingType layoutParsingType,
-                                              PDDocument originDocument,
+                                              File originFile,
                                              ImageServiceResponse imageServiceResponse,
-                                              TableServiceResponse tableServiceResponse) {
+                                              TableServiceResponse tableServiceResponse,
+                                              String identifier) {

+        PDDocument originDocument = openDocument(originFile);
+        addNumberOfPagesToTrace(originDocument.getNumberOfPages(), Files.size(originFile.toPath()));
        Map<Integer, List<TableCells>> pdfTableCells = cvTableParsingAdapter.buildCvParsedTablesPerPage(tableServiceResponse);
        Map<Integer, List<ClassifiedImage>> pdfImages = imageServiceResponseAdapter.buildClassifiedImagesPerPage(imageServiceResponse);

        ClassificationDocument classificationDocument = new ClassificationDocument();
        List<ClassificationPage> classificationPages = new ArrayList<>();

-        originDocument.setAllSecurityToBeRemoved(true);
        long pageCount = originDocument.getNumberOfPages();

        for (int pageNumber = 1; pageNumber <= pageCount; pageNumber++) {
+
+            if (pageNumber % 100 == 0) {
+                // re-open document every once in a while to save on RAM. This has no significant performance impact.
+                // This is due to PDFBox caching all images and some other stuff with Soft References. This dereferences them and forces the freeing of memory.
+                originDocument.close();
+                originDocument = openDocument(originFile);
+            }
+
+            if (pageNumber % 100 == 0 || pageNumber == pageCount || pageNumber == 1) {
+                log.info("Extracting text on Page {} for {}", pageNumber, identifier);
+            }
+
            classificationDocument.setPages(classificationPages);
            PDFLinesTextStripper stripper = new PDFLinesTextStripper();
            PDPage pdPage = originDocument.getPage(pageNumber - 1);
@@ -160,18 +246,12 @@ 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);
@@ -189,40 +269,62 @@ public class LayoutParsingPipeline {
                imageServiceResponseAdapter.findOcr(classificationPage);
            }

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

            classificationPages.add(classificationPage);
        }

-        bodyTextFrameService.setBodyTextFrames(classificationDocument, layoutParsingType);
+        originDocument.close();

+        log.info("Calculating BodyTextFrame for {}", identifier);
+        bodyTextFrameService.setBodyTextFrames(classificationDocument, layoutParsingType);
+        log.info("Classify TextBlocks for {}", identifier);
        switch (layoutParsingType) {
            case TAAS -> taasClassificationService.classifyDocument(classificationDocument);
            case DOCUMINE -> docuMineClassificationService.classifyDocument(classificationDocument);
            case REDACT_MANAGER -> redactManagerClassificationService.classifyDocument(classificationDocument);
        }

+        log.info("Building Sections for {}", identifier);
        sectionsBuilderService.buildSections(classificationDocument);
        sectionsBuilderService.addImagesToSections(classificationDocument);
-
        return classificationDocument;
    }


-    private Map<String, Rectangle2D> convertMarkedContents(List<PDMarkedContent> pdMarkedContents) {
-        Map<String, Rectangle2D> markedContentBboxes = new HashMap<>();
+    private void addNumberOfPagesToTrace(int numberOfPages, long size) {
+
+        if (observationRegistry.getCurrentObservation() != null) {
+            observationRegistry.getCurrentObservation().highCardinalityKeyValue("numberOfPages", String.valueOf(numberOfPages));
+            observationRegistry.getCurrentObservation().highCardinalityKeyValue("fileSize", String.valueOf(size));
+        }
+    }
+
+
+    @SneakyThrows
+    private PDDocument openDocument(File originFile) {
+
+        PDDocument document = Loader.loadPDF(originFile);
+        document.setAllSecurityToBeRemoved(true);
+        return document;
+    }
+
+
+    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());
-        }
+        document.getFontSizeCounter().addAll(classificationPage.getFontSizeCounter().getCountPerValue());
        document.getFontCounter().addAll(classificationPage.getFontCounter().getCountPerValue());
        document.getTextHeightCounter().addAll(classificationPage.getTextHeightCounter().getCountPerValue());
        document.getFontStyleCounter().addAll(classificationPage.getFontStyleCounter().getCountPerValue());
@@ -1,10 +1,24 @@
 package com.knecon.fforesight.service.layoutparser.processor;

+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.ComponentScan;
 import org.springframework.context.annotation.Configuration;

+import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+
+import io.micrometer.observation.ObservationRegistry;
+
@Configuration
@ComponentScan
 public class LayoutParsingServiceProcessorConfiguration {

+
+    @Bean
+    @Autowired
+    public ViewerDocumentService viewerDocumentService(ObservationRegistry registry) {
+
+        return new ViewerDocumentService(registry);
+    }
+
 }
@@ -1,23 +1,20 @@
 package com.knecon.fforesight.service.layoutparser.processor;

-import java.io.ByteArrayInputStream;
-import java.io.ByteArrayOutputStream;
 import java.io.File;
-import java.io.FileOutputStream;
+import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.nio.file.Files;
 import java.nio.file.Path;
 import java.nio.file.Paths;
 import java.nio.file.StandardOpenOption;
+import java.util.Optional;

-import org.apache.commons.io.IOUtils;
 import org.apache.pdfbox.Loader;
 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;
@@ -27,6 +24,7 @@ import com.knecon.fforesight.service.layoutparser.processor.python_api.model.ima
 import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
 import com.knecon.fforesight.tenantcommons.TenantContext;

+import io.micrometer.observation.annotation.Observed;
 import lombok.RequiredArgsConstructor;
 import lombok.SneakyThrows;
 import lombok.extern.slf4j.Slf4j;
@@ -39,17 +37,24 @@ public class LayoutParsingStorageService {
    private final StorageService storageService;
    private final ObjectMapper objectMapper;

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

-    public PDDocument getOriginFile(String storageId) throws IOException {
+        File tempFile = createTempFile("document", ".pdf");
+        storageService.downloadTo(TenantContext.getTenantId(), storageId, tempFile);
+        return tempFile;
+    }

-        try (var originDocumentInputStream = getObject(storageId)) {
-            File tempFile = createTempFile("document", ".pdf");
-            try (var tempFileOutputStream = new FileOutputStream(tempFile)) {
-                IOUtils.copy(originDocumentInputStream, tempFileOutputStream);
-                originDocumentInputStream.close();
-            }
-            return Loader.loadPDF(tempFile);
+
+    @Observed(name = "LayoutParsingStorageService", contextualName = "get-viewer-doc-file")
+    public Optional<File> getViewerDocFile(String storageId) throws IOException {
+
+        if (!storageService.objectExists(TenantContext.getTenantId(), storageId)) {
+            return Optional.empty();
        }
+        File tempFile = createTempFile("viewerDocument", ".pdf");
+        storageService.downloadTo(TenantContext.getTenantId(), storageId, tempFile);
+        return Optional.of(tempFile);
    }


@@ -75,6 +80,7 @@ public class LayoutParsingStorageService {
    }


+    @Observed(name = "LayoutParsingStorageService", contextualName = "store-document-data")
    public void storeDocumentData(LayoutParsingRequest layoutParsingRequest, DocumentData documentData) {

        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.structureFileStorageId(), documentData.getDocumentStructure());
@@ -84,12 +90,6 @@ public class LayoutParsingStorageService {
    }


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

        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.researchDocumentStorageId(), researchDocumentData);
@@ -121,6 +121,7 @@ public class LayoutParsingStorageService {
    }


+    @Observed(name = "LayoutParsingStorageService", contextualName = "store-simplified-text")
    public void storeSimplifiedText(LayoutParsingRequest layoutParsingRequest, SimplifiedText simplifiedText) {

        storageService.storeJSONObject(TenantContext.getTenantId(), layoutParsingRequest.simplifiedTextStorageId(), simplifiedText);
@@ -138,10 +139,10 @@ public class LayoutParsingStorageService {


    @SneakyThrows
-    public void storeViewerDocument(LayoutParsingRequest layoutParsingRequest, ByteArrayOutputStream out) {
-
-        try (var in = new ByteArrayInputStream(out.toByteArray())) {
+    @Observed(name = "LayoutParsingStorageService", contextualName = "store-viewer-document")
+    public void storeViewerDocument(LayoutParsingRequest layoutParsingRequest, File out) {

+        try (var in = new FileInputStream(out)) {
            storageService.storeObject(TenantContext.getTenantId(), layoutParsingRequest.viewerDocumentStorageId(), in);
        }
    }
@@ -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<>();

 }
@@ -12,6 +12,7 @@ import lombok.Setter;

@Setter
@EqualsAndHashCode
+@SuppressWarnings("PMD.AvoidFieldNameMatchingMethodName")
 public class Boundary implements Comparable<Boundary> {

    private int start;
@@ -10,7 +10,6 @@ import java.util.Set;
 import java.util.stream.Collectors;
 import java.util.stream.Stream;

-import com.amazonaws.services.kms.model.NotFoundException;
 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.DocumentTree;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.entity.RedactionEntity;
@@ -84,7 +83,7 @@ public class Document implements GenericSemanticNode {
    @Override
    public Headline getHeadline() {

-        return streamAllSubNodesOfType(NodeType.HEADLINE).map(node -> (Headline) node).findFirst().orElseThrow(() -> new NotFoundException("No Headlines found in this document!"));
+        return streamAllSubNodesOfType(NodeType.HEADLINE).map(node -> (Headline) node).findFirst().orElse(Headline.builder().build());
    }


@@ -100,6 +99,12 @@ public class Document implements GenericSemanticNode {
    }


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

@@ -1,7 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

+import java.awt.geom.Rectangle2D;
 import java.util.HashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
@@ -34,6 +36,9 @@ public class Footer implements GenericSemanticNode {
    @EqualsAndHashCode.Exclude
    Set<RedactionEntity> entities = new HashSet<>();

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;
+

    @Override
    public NodeType getType() {
@@ -62,4 +67,14 @@ public class Footer implements GenericSemanticNode {
        return treeId + ": " + NodeType.FOOTER + ": " + leafTextBlock.buildSummary();
    }

+
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {
+
+        if (bBoxCache == null) {
+            bBoxCache = GenericSemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
+
 }
@@ -1,7 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

+import java.awt.geom.Rectangle2D;
 import java.util.HashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
@@ -34,6 +36,9 @@ public class Header implements GenericSemanticNode {
    @EqualsAndHashCode.Exclude
    Set<RedactionEntity> entities = new HashSet<>();

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;
+

    @Override
    public boolean isLeaf() {
@@ -62,4 +67,14 @@ public class Header implements GenericSemanticNode {
        return treeId + ": " + NodeType.HEADER + ": " + leafTextBlock.buildSummary();
    }

+
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {
+
+        if (bBoxCache == null) {
+            bBoxCache = GenericSemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
+
 }
@@ -1,7 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

+import java.awt.geom.Rectangle2D;
 import java.util.HashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
@@ -34,6 +36,9 @@ public class Headline implements GenericSemanticNode {
    @EqualsAndHashCode.Exclude
    Set<RedactionEntity> entities = new HashSet<>();

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;
+

    @Override
    public NodeType getType() {
@@ -69,4 +74,14 @@ public class Headline implements GenericSemanticNode {
        return this;
    }

+
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {
+
+        if (bBoxCache == null) {
+            bBoxCache = GenericSemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
+
 }
@@ -1,7 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

+import java.awt.geom.Rectangle2D;
 import java.util.HashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
@@ -32,6 +34,9 @@ public class Paragraph implements GenericSemanticNode {
    @EqualsAndHashCode.Exclude
    Set<RedactionEntity> entities = new HashSet<>();

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;
+

    @Override
    public NodeType getType() {
@@ -60,4 +65,14 @@ public class Paragraph implements GenericSemanticNode {
        return treeId + ": " + NodeType.PARAGRAPH + ": " + leafTextBlock.buildSummary();
    }

+
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {
+
+        if (bBoxCache == null) {
+            bBoxCache = GenericSemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
+
 }
@@ -1,7 +1,9 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

+import java.awt.geom.Rectangle2D;
 import java.util.HashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Set;

 import com.knecon.fforesight.service.layoutparser.internal.api.data.redaction.NodeType;
@@ -35,6 +37,9 @@ public class Section implements GenericSemanticNode {
    @EqualsAndHashCode.Exclude
    Set<RedactionEntity> entities = new HashSet<>();

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;
+

    @Override
    public NodeType getType() {
@@ -74,4 +79,14 @@ public class Section implements GenericSemanticNode {
                .orElseGet(() -> getParent().getHeadline());
    }

+
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {
+
+        if (bBoxCache == null) {
+            bBoxCache = GenericSemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
+
 }
@@ -398,12 +398,11 @@ public interface SemanticNode {
     */
    default Map<Page, Rectangle2D> getBBox() {

-        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
        if (isLeaf()) {
-            return getBBoxFromLeafTextBlock(bBoxPerPage);
+            return getBBoxFromLeafTextBlock();
        }

-        return getBBoxFromChildren(bBoxPerPage);
+        return getBBoxFromChildren();
    }


@@ -426,25 +425,31 @@ public interface SemanticNode {

    /**
     * TODO: this produces unwanted results for sections spanning multiple columns.
-     *
-     * @param bBoxPerPage initial empty BoundingBox
+     * Computes the Union of the bounding boxes of all children recursively.
     * @return The union of the BoundingBoxes of all children
     */
-    private Map<Page, Rectangle2D> getBBoxFromChildren(Map<Page, Rectangle2D> bBoxPerPage) {
+    private Map<Page, Rectangle2D> getBBoxFromChildren() {

-        return streamChildren().map(SemanticNode::getBBox).reduce((map1, map2) -> {
-            map1.forEach((page, rectangle) -> map2.merge(page, rectangle, (rect1, rect2) -> rect1.createUnion(rect2).getBounds2D()));
-            return map2;
-        }).orElse(bBoxPerPage);
+        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
+        List<Map<Page, Rectangle2D>> childrenBBoxes = streamChildren().map(SemanticNode::getBBox).toList();
+        Set<Page> pages = childrenBBoxes.stream().flatMap(map -> map.keySet().stream()).collect(Collectors.toSet());
+        for (Page page : pages) {
+            Rectangle2D bBoxOnPage = childrenBBoxes.stream()
+                    .filter(childBboxPerPage -> childBboxPerPage.containsKey(page))
+                    .map(childBboxPerPage -> childBboxPerPage.get(page))
+                    .collect(RectangleTransformations.collectBBox());
+            bBoxPerPage.put(page, bBoxOnPage);
+        }
+        return bBoxPerPage;
    }


    /**
-     * @param bBoxPerPage initial empty BoundingBox
     * @return The union of all BoundingBoxes of the TextBlock of this node
     */
-    private Map<Page, Rectangle2D> getBBoxFromLeafTextBlock(Map<Page, Rectangle2D> bBoxPerPage) {
+    private Map<Page, Rectangle2D> getBBoxFromLeafTextBlock() {

+        Map<Page, Rectangle2D> bBoxPerPage = new HashMap<>();
        Map<Page, List<AtomicTextBlock>> atomicTextBlockPerPage = getTextBlock().getAtomicTextBlocks().stream().collect(Collectors.groupingBy(AtomicTextBlock::getPage));
        atomicTextBlockPerPage.forEach((page, atbs) -> bBoxPerPage.put(page, RectangleTransformations.bBoxUnionAtomicTextBlock(atbs)));
        return bBoxPerPage;
@@ -2,10 +2,12 @@ package com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes;

 import static java.lang.String.format;

+import java.awt.geom.Rectangle2D;
 import java.util.Collection;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Locale;
+import java.util.Map;
 import java.util.Set;
 import java.util.stream.IntStream;
 import java.util.stream.Stream;
@@ -34,13 +36,14 @@ public class Table implements SemanticNode {

    int numberOfRows;
    int numberOfCols;
-
    TextBlock textBlock;

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

+    @EqualsAndHashCode.Exclude
+    Map<Page, Rectangle2D> bBoxCache;

    /**
     * Streams all entities in this table, that appear in a row, which contains any of the provided strings.
@@ -208,7 +211,6 @@ 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.
     *
@@ -313,5 +315,12 @@ public class Table implements SemanticNode {

        return treeId.toString() + ": " + NodeType.TABLE + ": #cols: " + numberOfCols + ", #rows: " + numberOfRows + ", " + this.getTextBlock().buildSummary();
    }
+    @Override
+    public Map<Page, Rectangle2D> getBBox() {

+        if (bBoxCache == null) {
+            bBoxCache = SemanticNode.super.getBBox();
+        }
+        return bBoxCache;
+    }
 }
@@ -76,4 +76,14 @@ public class Cell extends Rectangle {
        return this.getHeight() >= MIN_SIZE && this.getWidth() >= MIN_SIZE;
    }

+    public boolean nearlyIntersects(Cell other) {
+
+        if (this.getHeight() <= 0 || other.getHeight() <= 0) {
+            return false;
+        }
+        double x0 = this.getX() + 2;
+        double y0 = this.getY() + 2;
+        return (other.x + other.width > x0 && other.y + other.height > y0 && other.x < x0 + this.getWidth() - 2 && other.y < y0 + this.getHeight() - 2);
+    }
+
 }
@@ -19,7 +19,7 @@ import lombok.extern.slf4j.Slf4j;
@Slf4j
 public class TablePageBlock extends AbstractPageBlock {

-    private final TreeMap<CellPosition, Cell> cells = new TreeMap<>();
+    private final TreeMap<CellPosition, Cell> cellTreeMap = new TreeMap<>();

    private final int rotation;
    @Getter
@@ -28,10 +28,14 @@ public class TablePageBlock extends AbstractPageBlock {
    private int unrotatedRowCount;
    private int unrotatedColCount;
    private List<List<Cell>> rows;
+    @Getter
+    @Setter
+    private List<Cell> cells;


    public TablePageBlock(List<Cell> cells, Rectangle area, int rotation) {

+        this.cells = cells;
        addCells(cells);
        minX = area.getLeft();
        minY = area.getBottom();
@@ -48,6 +52,7 @@ public class TablePageBlock extends AbstractPageBlock {
        return getColCount() == 0 || getRowCount() == 0;
    }

+
    public List<List<Cell>> getRows() {

        if (rows == null) {
@@ -78,7 +83,10 @@ public class TablePageBlock extends AbstractPageBlock {

    public int getColCount() {

-        return getRows().stream().mapToInt(List::size).max().orElse(0);
+        return getRows().stream()
+                .mapToInt(List::size)
+                .max()
+                .orElse(0);

    }

@@ -118,7 +126,8 @@ public class TablePageBlock extends AbstractPageBlock {
                List<Cell> cellsToTheTop = new ArrayList<>();
                for (int i = 0; i < rowIndex; i++) {
                    try {
-                        cellsToTheTop.add(rows.get(i).get(colIndex));
+                        cellsToTheTop.add(rows.get(i)
+                                                  .get(colIndex));
                    } catch (IndexOutOfBoundsException e) {
                        log.debug("No cell {} in row {}, ignoring.", colIndex, rowIndex);
                    }
@@ -133,7 +142,8 @@ public class TablePageBlock extends AbstractPageBlock {
                if (lastHeaderCell != null) {
                    cell.getHeaderCells().add(lastHeaderCell);
                }
-                if (!cell.getTextBlocks().isEmpty() && cell.getTextBlocks().get(0).getMostPopularWordStyle().equals("bold")) {
+                if (!cell.getTextBlocks().isEmpty() && cell.getTextBlocks()
+                        .get(0).getMostPopularWordStyle().equals("bold")) {
                    cell.setHeaderCell(true);
                }
            }
@@ -149,7 +159,7 @@ public class TablePageBlock extends AbstractPageBlock {
            for (int i = 0; i < unrotatedColCount; i++) { // rows
                List<Cell> lastRow = new ArrayList<>();
                for (int j = unrotatedRowCount - 1; j >= 0; j--) { // cols
-                    Cell cell = cells.get(new CellPosition(j, i));
+                    Cell cell = cellTreeMap.get(new CellPosition(j, i));
                    if (cell != null) {
                        lastRow.add(cell);
                    }
@@ -160,7 +170,7 @@ public class TablePageBlock extends AbstractPageBlock {
            for (int i = unrotatedColCount - 1; i >= 0; i--) { // rows
                List<Cell> lastRow = new ArrayList<>();
                for (int j = 0; j < unrotatedRowCount; j++) { // cols
-                    Cell cell = cells.get(new CellPosition(j, i));
+                    Cell cell = cellTreeMap.get(new CellPosition(j, i));
                    if (cell != null) {
                        lastRow.add(cell);
                    }
@@ -171,7 +181,7 @@ public class TablePageBlock extends AbstractPageBlock {
            for (int i = 0; i < unrotatedRowCount; i++) {
                List<Cell> lastRow = new ArrayList<>();
                for (int j = 0; j < unrotatedColCount; j++) {
-                    Cell cell = cells.get(new CellPosition(i, j)); // JAVA_8 use getOrDefault()
+                    Cell cell = cellTreeMap.get(new CellPosition(i, j)); // JAVA_8 use getOrDefault()
                    if (cell != null) {
                        lastRow.add(cell);
                    }
@@ -185,17 +195,6 @@ public class TablePageBlock extends AbstractPageBlock {
    }


-    private void add(Cell chunk, int row, int col) {
-
-        unrotatedRowCount = Math.max(unrotatedRowCount, row + 1);
-        unrotatedColCount = Math.max(unrotatedColCount, col + 1);
-
-        CellPosition cp = new CellPosition(row, col);
-        cells.put(cp, chunk);
-
-    }
-
-
    private void addCells(List<Cell> cells) {

        if (cells.isEmpty()) {
@@ -204,11 +203,12 @@ public class TablePageBlock extends AbstractPageBlock {

        cells.removeIf(cell -> cell.getWidth() < 1.1 || cell.getHeight() < 1.1);

-        List<List<Cell>> rowsOfCells = calculateStructure(cells);
+        List<List<Cell>> rowsOfCellsMatrix = calculateTableStructure(cells);

-        for (int i = 0; i < rowsOfCells.size(); i++) {
-            for (int j = 0; j < rowsOfCells.get(i).size(); j++) {
-                add(rowsOfCells.get(i).get(j), i, j);
+        for (int i = 0; i < rowsOfCellsMatrix.size(); i++) {
+            for (int j = 0; j < rowsOfCellsMatrix.get(i).size(); j++) {
+                addCellToRowAndCol(rowsOfCellsMatrix.get(i)
+                                           .get(j), i, j);
            }
        }

@@ -219,29 +219,36 @@ public class TablePageBlock extends AbstractPageBlock {
     * Calculates the structure of the table. For spanning rows and columns multiple cells with the same values will be inserted.
     *
     * @param cells The found cells
-     * @return TablePageBlock Structure
+     * @return TablePageBlock Structure as a rows of cells matrix
     */
-    private List<List<Cell>> calculateStructure(List<Cell> cells) {
-
-        List<List<Cell>> matrix = new ArrayList<>();
+    private List<List<Cell>> calculateTableStructure(List<Cell> cells) {

        if (cells.isEmpty()) {
-            return matrix;
+            return new ArrayList<>();
        }

        Set<Float> uniqueX = new HashSet<>();
        Set<Float> uniqueY = new HashSet<>();
-        cells.stream().filter(c -> !c.getTextBlocks().isEmpty() || c.getHeight() > 3 && c.getWidth() > 3).forEach(c -> {
-            uniqueX.add(c.getLeft());
-            uniqueX.add(c.getRight());
-            uniqueY.add(c.getBottom());
-            uniqueY.add(c.getTop());
-        });
+        cells.stream()
+                .filter(c -> !c.getTextBlocks().isEmpty() || c.getHeight() > 3 && c.getWidth() > 3)
+                .forEach(c -> {
+                    uniqueX.add(c.getLeft());
+                    uniqueX.add(c.getRight());
+                    uniqueY.add(c.getBottom());
+                    uniqueY.add(c.getTop());
+                });

-        var sortedUniqueX = uniqueX.stream().sorted().toList();
-        var sortedUniqueY = uniqueY.stream().sorted().toList();
+        var sortedUniqueX = uniqueX.stream()
+                .sorted()
+                .toList();
+        var sortedUniqueY = uniqueY.stream()
+                .sorted()
+                .toList();
+
+        List<List<Cell>> rowsOfCells = new ArrayList<>();

        Float prevY = null;
+
        for (Float y : sortedUniqueY) {

            List<Cell> row = new ArrayList<>();
@@ -252,24 +259,79 @@ 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();
-                    intersectionCell.ifPresent(value -> cell.getTextBlocks().addAll(value.getTextBlocks()));
                    if (cell.hasMinimumSize()) {
+
+                        cells.stream()
+                                .filter(cell::nearlyIntersects)
+                                .forEach(intersectingCell -> cell.getTextBlocks().addAll(intersectingCell.getTextBlocks()));
+
                        row.add(cell);
                    }
                }
                prevX = x;
            }

-            if (prevY != null && prevX != null && !row.isEmpty()) {
-                matrix.add(row);
+            // exclude empty rows and rows where all text blocks are empty
+            if (prevY != null && prevX != null && !row.isEmpty() && !row.stream()
+                    .allMatch(cell -> cell.getTextBlocks().isEmpty())) {
+
+                rowsOfCells.add(row);
            }
            prevY = y;
        }

-        Collections.reverse(matrix);
+        Collections.reverse(rowsOfCells);
+
+        // now cells are removed which are part of a column without any text blocks
+        // this is done by first computing the inverse matrix which contains call columns of cells
+        // then the column indices that have to be removed are determined
+        List<List<Cell>> columnsOfCells = new ArrayList<>();
+        int maxRowLength = rowsOfCells.stream()
+                .map(List::size)
+                .max(java.util.Comparator.naturalOrder())
+                .orElse(0);
+        for (int i = 0; i < maxRowLength; i++) {
+            columnsOfCells.add(new ArrayList<>());
+        }
+
+        for (List<Cell> row : rowsOfCells) {
+            for (int j = 0; j < row.size(); j++) {
+                columnsOfCells.get(j).add(row.get(j));
+            }
+        }
+
+        List<Integer> columnIndicesToRemove = new ArrayList<>();
+        int columnIndex = 0;
+        for (List<Cell> col : columnsOfCells) {
+            if (col.stream()
+                    .allMatch(cell -> cell.getTextBlocks().isEmpty())) {
+                columnIndicesToRemove.add(columnIndex);
+            }
+            columnIndex++;
+        }
+        columnIndicesToRemove.sort(Collections.reverseOrder());
+
+        // update all rows so that the values of the empty columns get removed
+        var rowsOfCellsBefore = new ArrayList<>(rowsOfCells);
+        rowsOfCells = new ArrayList<>();
+        for (List<Cell> row : rowsOfCellsBefore) {
+            var updatedRow = new ArrayList<>(row);
+            columnIndicesToRemove.forEach(idxToRemove -> updatedRow.remove(updatedRow.get(idxToRemove)));
+            rowsOfCells.add(updatedRow);
+        }
+
+        return rowsOfCells;
+    }
+
+
+    private void addCellToRowAndCol(Cell cell, int row, int col) {
+
+        unrotatedRowCount = Math.max(unrotatedRowCount, row + 1);
+        unrotatedColCount = Math.max(unrotatedColCount, col + 1);
+
+        CellPosition cp = new CellPosition(row, col);
+        cellTreeMap.put(cp, cell);

-        return matrix;
    }


@@ -296,7 +358,7 @@ public class TablePageBlock extends AbstractPageBlock {
                            if (!first) {
                                sb.append("\n");
                            }
-                            sb.append('\"').append(textBlock.getText().replaceAll("\"", "\\\"")).append('\"');
+                            sb.append('\"').append(textBlock.getText().replaceAll("\"", "\"")).append('\"');
                            first = false;
                        }
                    }
@@ -310,8 +372,6 @@ public class TablePageBlock extends AbstractPageBlock {
    }


-
-
    public String getTextAsHtml() {

        StringBuilder sb = new StringBuilder();
@@ -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();
@@ -9,8 +9,6 @@ import java.util.stream.Collectors;

 import org.apache.pdfbox.text.TextPosition;

-import com.dslplatform.json.JsonAttribute;
-import com.fasterxml.jackson.annotation.JsonIgnore;
 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.Rectangle;

@@ -142,8 +140,7 @@ public class TextPositionSequence implements CharSequence {
     *
     * @return the text direction adjusted minX value
     */
-    @JsonIgnore
-    @JsonAttribute(ignore = true)
+
    public float getMinXDirAdj() {

        return textPositions.get(0).getXDirAdj();
@@ -157,8 +154,7 @@ public class TextPositionSequence implements CharSequence {
     *
     * @return the text direction adjusted maxX value
     */
-    @JsonIgnore
-    @JsonAttribute(ignore = true)
+
    public float getMaxXDirAdj() {

        return textPositions.get(textPositions.size() - 1).getXDirAdj() + textPositions.get(textPositions.size() - 1).getWidthDirAdj() + HEIGHT_PADDING;
@@ -172,8 +168,7 @@ public class TextPositionSequence implements CharSequence {
     *
     * @return the text direction adjusted minY value. The upper border of the bounding box of the word.
     */
-    @JsonIgnore
-    @JsonAttribute(ignore = true)
+
    public float getMinYDirAdj() {

        return textPositions.get(0).getYDirAdj() - getTextHeight();
@@ -187,8 +182,7 @@ public class TextPositionSequence implements CharSequence {
     *
     * @return the text direction adjusted maxY value. The lower border of the bounding box of the word.
     */
-    @JsonIgnore
-    @JsonAttribute(ignore = true)
+
    public float getMaxYDirAdj() {

        return textPositions.get(0).getYDirAdj();
@@ -196,42 +190,38 @@ public class TextPositionSequence implements CharSequence {
    }


-    @JsonIgnore
-    @JsonAttribute(ignore = true)
    public float getTextHeight() {

        return textPositions.get(0).getHeightDir() + HEIGHT_PADDING;
    }


-    @JsonIgnore
-    @JsonAttribute(ignore = true)
    public float getHeight() {

        return getMaxYDirAdj() - getMinYDirAdj();
    }


-    @JsonIgnore
-    @JsonAttribute(ignore = true)
    public float getWidth() {

        return getMaxXDirAdj() - getMinXDirAdj();
    }


-    @JsonIgnore
-    @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", "");
    }


-    @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")) {
@@ -243,20 +233,15 @@ public class TextPositionSequence implements CharSequence {
        } else {
            return "standard";
        }
-
    }


-    @JsonIgnore
-    @JsonAttribute(ignore = true)
    public float getFontSize() {

        return textPositions.get(0).getFontSizeInPt();
    }


-    @JsonIgnore
-    @JsonAttribute(ignore = true)
    public float getSpaceWidth() {

        return textPositions.get(0).getWidthOfSpace();
@@ -272,8 +257,7 @@ public class TextPositionSequence implements CharSequence {
     *
     * @return bounding box of the word in Pdf Coordinate System
     */
-    @JsonIgnore
-    @JsonAttribute(ignore = true)
+
    @SneakyThrows
    public Rectangle getRectangle() {

@@ -1,8 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
-
-import java.awt.Color;
-import java.awt.geom.Line2D;
-
-public record ColoredLine(Line2D line, Color color) {
-
-}
@@ -1,7 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
-
-import java.awt.geom.Point2D;
-
-public record PlacedText(String text, Point2D lineStart) {
-
-}
@@ -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 +25,14 @@ 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_Y = 6;
+    private static final float THRESHOLD_X = 2;
+
+
+    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 +57,11 @@ 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 +82,7 @@ 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_X) {
                groupedPoints.get(groupedPoints.size() - 1).add(p);
            } else {
                groupedPoints.add(new ArrayList<>(Collections.singletonList(p)));
@@ -108,7 +109,7 @@ 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_Y) {
                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();
-    }
-
-}
@@ -5,14 +5,12 @@ import java.util.ArrayList;
 import java.util.Comparator;
 import java.util.HashMap;
 import java.util.HashSet;
-import java.util.Iterator;
 import java.util.List;
 import java.util.Map;
 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;
@@ -26,45 +24,63 @@ import com.knecon.fforesight.service.layoutparser.processor.utils.DoubleComparis
@Service
 public class TableExtractionService {

-    private static final Comparator<Point2D> X_FIRST_POINT_COMPARATOR = (arg0, arg1) -> {
+    private static final int MAX_TABLE_OUTER_POINT_TOLERANCE = 10;
+    private static final int MAX_TABLE_CONTAINED_CELLS_WITH_TEXT = 1;
+    private static final float SPREADSHEET_AREA_TOLERANCE = 0.001f;
+
+    private static final Comparator<Point2D> X_FIRST_POINT_COMPARATOR = (point1, point2) -> {

        int rv = 0;
-        float arg0X = DoubleComparisons.round(arg0.getX(), 2);
-        float arg0Y = DoubleComparisons.round(arg0.getY(), 2);
-        float arg1X = DoubleComparisons.round(arg1.getX(), 2);
-        float arg1Y = DoubleComparisons.round(arg1.getY(), 2);
+        float point1X = DoubleComparisons.round(point1.getX(), 2);
+        float point1Y = DoubleComparisons.round(point1.getY(), 2);
+        float point2X = DoubleComparisons.round(point2.getX(), 2);
+        float point2Y = DoubleComparisons.round(point2.getY(), 2);

-        if (arg0X > arg1X) {
+        if (point1X > point2X) {
            rv = 1;
-        } else if (arg0X < arg1X) {
+        } else if (point1X < point2X) {
            rv = -1;
-        } else if (arg0Y > arg1Y) {
+        } else if (point1Y > point2Y) {
            rv = 1;
-        } else if (arg0Y < arg1Y) {
+        } else if (point1Y < point2Y) {
            rv = -1;
        }
        return rv;
    };
-    private static final Comparator<Point2D> POINT_COMPARATOR = (arg0, arg1) -> {
+    private static final Comparator<Point2D> Y_FIRST_POINT_COMPARATOR = (point1, point2) -> {

        int rv = 0;
-        float arg0X = DoubleComparisons.round(arg0.getX(), 2);
-        float arg0Y = DoubleComparisons.round(arg0.getY(), 2);
-        float arg1X = DoubleComparisons.round(arg1.getX(), 2);
-        float arg1Y = DoubleComparisons.round(arg1.getY(), 2);
+        float point1X = DoubleComparisons.round(point1.getX(), 2);
+        float point1Y = DoubleComparisons.round(point1.getY(), 2);
+        float point2X = DoubleComparisons.round(point2.getX(), 2);
+        float point2Y = DoubleComparisons.round(point2.getY(), 2);

-        if (arg0Y > arg1Y) {
+        if (point1Y > point2Y) {
            rv = 1;
-        } else if (arg0Y < arg1Y) {
+        } else if (point1Y < point2Y) {
            rv = -1;
-        } else if (arg0X > arg1X) {
+        } else if (point1X > point2X) {
            rv = 1;
-        } else if (arg0X < arg1X) {
+        } else if (point1X < point2X) {
            rv = -1;
        }
        return rv;
    };

+    private static final Comparator<Cell> CELL_SIZE_COMPARATOR = (cell1, cell2) -> {
+
+        Double cell1Size = cell1.getHeight() * cell1.getWidth();
+        Double cell2Size = cell2.getHeight() * cell2.getWidth();
+        return cell1Size.compareTo(cell2Size);
+    };
+
+    private static final Comparator<Rectangle> RECTANGLE_SIZE_COMPARATOR = (rect1, rect2) -> {
+
+        Double rect1Size = rect1.getHeight() * rect1.getWidth();
+        Double rect2Size = rect2.getHeight() * rect2.getWidth();
+        return rect1Size.compareTo(rect2Size);
+    };
+

    /**
     * Finds tables on a page and moves textblocks into cells of the found tables.
@@ -79,21 +95,18 @@ 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) {

-        List<Cell> cells = findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical(), layoutParsingType);
+    public void extractTables(CleanRulings cleanRulings, ClassificationPage page) {

-        List<TextPageBlock> toBeRemoved = new ArrayList<>();
+        List<Cell> cells = findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical());
+        // sort cells by size (height * width) ascending so that textBlocks are always assigned to the smallest cells that contain them
+        cells.sort(CELL_SIZE_COMPARATOR);

        for (AbstractPageBlock abstractPageBlock : page.getTextBlocks()) {
            TextPageBlock textBlock = (TextPageBlock) abstractPageBlock;
            for (Cell cell : cells) {
-                if (cell.hasMinimumSize() && cell.intersects(textBlock.getPdfMinX(),
-                        textBlock.getPdfMinY(),
-                        textBlock.getPdfMaxX() - textBlock.getPdfMinX(),
-                        textBlock.getPdfMaxY() - textBlock.getPdfMinY())) {
+                if (cell.hasMinimumSize() && doesCellContainTextBlock(cell, textBlock)) {
                    cell.addTextBlock(textBlock);
-                    toBeRemoved.add(textBlock);
                    break;
                }
            }
@@ -102,56 +115,85 @@ 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);
+        // sort spreadsheetAreas by size (height * width) ascending so that cells are placed in the smallest tables first
+        // this way no cell duplication occurs when tables are contained in other tables and only the most inner table contains the cells
+        spreadsheetAreas.sort(RECTANGLE_SIZE_COMPARATOR);

        List<TablePageBlock> tables = new ArrayList<>();
        for (Rectangle area : spreadsheetAreas) {

-            List<Cell> overlappingCells = new ArrayList<>();
+            List<Cell> containedCells = new ArrayList<>();
            for (Cell c : cells) {
-                if (c.hasMinimumSize() && c.intersects(area)) {
-                    overlappingCells.add(c);
+                if (c.hasMinimumSize() && area.contains(c)) {
+                    containedCells.add(c);
                }
            }
-            tables.add(new TablePageBlock(overlappingCells, area, page.getRotation()));
+
+            var containedCellsWithText = containedCells.stream()
+                    .filter(cell -> !cell.getTextBlocks().isEmpty())
+                    .count();
+
+            // verify if table would contain fewer cells with text than the threshold allows
+            if (containedCellsWithText >= MAX_TABLE_CONTAINED_CELLS_WITH_TEXT) {
+                tables.add(new TablePageBlock(containedCells, area, page.getRotation()));
+                cells.removeAll(containedCells);
+            }
        }

        for (TablePageBlock table : tables) {
            int position = -1;

-            Iterator<AbstractPageBlock> itty = page.getTextBlocks().iterator();
-            while (itty.hasNext()) {
-                AbstractPageBlock textBlock = itty.next();
-                if (textBlock instanceof TextPageBlock ? table.containsBlock((TextPageBlock) textBlock) : table.contains(textBlock) && position == -1) {
-                    position = page.getTextBlocks().indexOf(textBlock);
+            for (AbstractPageBlock pageBlock : page.getTextBlocks()) {
+                if (pageBlock instanceof TextPageBlock ? table.containsBlock((TextPageBlock) pageBlock) : table.contains(pageBlock) && position == -1) {
+                    position = page.getTextBlocks().indexOf(pageBlock);
                }
            }
            if (position != -1) {
                page.getTextBlocks().add(position, table);
+
+                var toBeRemoved = table.getCells()
+                        .stream()
+                        .map(Cell::getTextBlocks)
+                        .flatMap(List::stream)
+                        .toList();
+                // remove text blocks from the page that were also added with the table (from its contained cells)
+                page.getTextBlocks().removeAll(toBeRemoved);
            }
        }
-
-        page.getTextBlocks().removeAll(toBeRemoved);
    }


-    public List<Cell> findCells(List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines, LayoutParsingType layoutParsingType) {
+    private boolean doesCellContainTextBlock(Cell cell, TextPageBlock textBlock) {

-        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;
-                }
+        double x = textBlock.getPdfMinX();
+        double y = textBlock.getPdfMinY();
+        double w = textBlock.getPdfMaxX() - textBlock.getPdfMinX();
+        double h = textBlock.getPdfMaxY() - textBlock.getPdfMinY();
+        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);
+    }
+
+
+    private List<Cell> findCells(List<Ruling> horizontalRulingLines, List<Ruling> verticalRulingLines) {
+
+        // 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;
            }
        }

        List<Cell> cellsFound = new ArrayList<>();
        Map<Point2D, Ruling[]> intersectionPoints = Ruling.findIntersections(horizontalRulingLines, verticalRulingLines);
        List<Point2D> intersectionPointsList = new ArrayList<>(intersectionPoints.keySet());
-        intersectionPointsList.sort(POINT_COMPARATOR);
+        intersectionPointsList.sort(Y_FIRST_POINT_COMPARATOR);

        for (int i = 0; i < intersectionPointsList.size(); i++) {
            Point2D topLeft = intersectionPointsList.get(i);
@@ -177,13 +219,14 @@ public class TableExtractionService {
                    continue;
                }
                for (Point2D yPoint : yPoints) {
-                    // is there an horizontal edge b/w topLeft and yPoint ?
+                    // is there a horizontal edge b/w topLeft and yPoint ?
                    if (!hv[0].equals(intersectionPoints.get(yPoint)[0])) {
                        continue;
                    }
                    Point2D btmRight = new Point2D.Float((float) yPoint.getX(), (float) xPoint.getY());
-                    if (intersectionPoints.containsKey(btmRight) && intersectionPoints.get(btmRight)[0].equals(intersectionPoints.get(xPoint)[0]) && intersectionPoints.get(btmRight)[1].equals(
-                            intersectionPoints.get(yPoint)[1])) {
+                    if (intersectionPoints.containsKey(btmRight)
+                        && intersectionPoints.get(btmRight)[0].equals(intersectionPoints.get(xPoint)[0])
+                        && intersectionPoints.get(btmRight)[1].equals(intersectionPoints.get(yPoint)[1])) {
                        cellsFound.add(new Cell(topLeft, btmRight));
                        break outer;
                    }
@@ -205,7 +248,6 @@ public class TableExtractionService {
        Set<Point2D> pointSet = new HashSet<>();
        Map<Point2D, Point2D> edgesH = new HashMap<>();
        Map<Point2D, Point2D> edgesV = new HashMap<>();
-        int i = 0;

        for (Rectangle cell : cells) {
            for (Point2D pt : cell.getPoints()) {
@@ -222,8 +264,9 @@ public class TableExtractionService {
        pointsSortX.sort(X_FIRST_POINT_COMPARATOR);
        // Y first sort
        List<Point2D> pointsSortY = new ArrayList<>(pointSet);
-        pointsSortY.sort(POINT_COMPARATOR);
+        pointsSortY.sort(Y_FIRST_POINT_COMPARATOR);

+        int i = 0;
        while (i < pointSet.size()) {
            float currY = (float) pointsSortY.get(i).getY();
            while (i < pointSet.size() && DoubleComparisons.feq(pointsSortY.get(i).getY(), currY)) {
@@ -248,7 +291,8 @@ public class TableExtractionService {
        Point2D nextVertex;
        while (!edgesH.isEmpty()) {
            ArrayList<PolygonVertex> polygon = new ArrayList<>();
-            Point2D first = edgesH.keySet().iterator().next();
+            Point2D first = edgesH.keySet()
+                    .iterator().next();
            polygon.add(new PolygonVertex(first, Direction.HORIZONTAL));
            edgesH.remove(first);

@@ -292,7 +336,14 @@ public class TableExtractionService {
                bottom = (float) Math.max(bottom, pt.point.getY());
                right = (float) Math.max(right, pt.point.getX());
            }
-            rectangles.add(new Rectangle(top, left, right - left, bottom - top));
+
+            // do not add polygons with too many outer points as they are unlikely to be tables
+            if (poly.size() <= MAX_TABLE_OUTER_POINT_TOLERANCE) {
+                rectangles.add(new Rectangle(top - SPREADSHEET_AREA_TOLERANCE,
+                                             left - SPREADSHEET_AREA_TOLERANCE,
+                                             right - left + 2 * SPREADSHEET_AREA_TOLERANCE,
+                                             bottom - top + 2 * SPREADSHEET_AREA_TOLERANCE));
+            }
        }

        return rectangles;
@@ -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();
@@ -1,12 +1,11 @@
 package com.knecon.fforesight.service.layoutparser.processor.services.visualization;

-import static com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService.LINE_WIDTH;
-
 import java.awt.Color;
 import java.awt.geom.Line2D;
 import java.awt.geom.Point2D;
 import java.awt.geom.Rectangle2D;
 import java.awt.geom.RectangularShape;
+import java.io.File;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.Comparator;
@@ -24,25 +23,59 @@ import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Pa
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.SemanticNode;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Table;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.TableCell;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.ColoredLine;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.ColoredRectangle;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.LayoutGrid;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.PlacedText;
 import com.knecon.fforesight.service.layoutparser.processor.utils.RectangleTransformations;
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.model.ColoredLine;
+import com.knecon.fforesight.service.viewerdoc.model.ColoredRectangle;
+import com.knecon.fforesight.service.viewerdoc.model.LayoutGrid;
+import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
+import com.knecon.fforesight.service.viewerdoc.model.Standard14EmbeddableFont;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;
+
+import io.micrometer.observation.annotation.Observed;
+import lombok.AccessLevel;
+import lombok.RequiredArgsConstructor;
+import lombok.SneakyThrows;
+import lombok.experimental.FieldDefaults;

@Service
+@RequiredArgsConstructor
+@FieldDefaults(level = AccessLevel.PRIVATE, makeFinal = true)
 public class LayoutGridService {

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


-    public LayoutGrid createLayoutGrid(Document document) {
+    @SneakyThrows
+    @Observed(name = "ViewerDocumentService", contextualName = "create-viewer-document")
+    public void addLayoutGrid(File originFile, Document document, File destinationFile, boolean layerVisibilityDefaultValue) {
+
+        LayoutGrid layoutGrid = createLayoutGrid(document);
+
+        viewerDocumentService.addVisualizationsOnPage(originFile,
+                destinationFile,
+                Visualizations.builder()
+                        .layer(ContentStreams.KNECON_LAYOUT)
+                        .visualizationsOnPages(layoutGrid.getVisualizationsPerPages())
+                        .layerVisibilityDefaultValue(layerVisibilityDefaultValue)
+                        .build());
+    }
+
+
+    private LayoutGrid createLayoutGrid(Document document) {

        LayoutGrid layoutGrid = new LayoutGrid(document.getNumberOfPages());
        document.streamAllSubNodes().forEach(semanticNode -> {
@@ -103,11 +136,11 @@ public class LayoutGridService {
            List<ColoredLine> coloredLines = layoutGrid.getVisualizationsPerPages().get(page.getNumber() - 1).getColoredLines();
            xs.forEach(x -> {
                Line2D line = new Line2D.Double(new Point2D.Double(x, tableBBox.getMaxY()), new Point2D.Double(x, tableBBox.getMinY()));
-                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR));
+                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
            });
            ys.forEach(y -> {
                Line2D line = new Line2D.Double(new Point2D.Double(tableBBox.getMinX(), y), new Point2D.Double(tableBBox.getMaxX(), y));
-                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR));
+                coloredLines.add(new ColoredLine(line, INNER_LINES_COLOR, LINE_WIDTH));
            });
        }
    }
@@ -135,9 +168,9 @@ public class LayoutGridService {
            List<Line2D> lines = createLinesFromRectangle(r, firstPage.getRotation());
            // add string to top line
            var firstLine = lines.remove(0);
-            coloredLines.add(new ColoredLine(firstLine, color));
+            coloredLines.add(new ColoredLine(firstLine, color, LINE_WIDTH));
            for (Line2D line : lines) {
-                coloredLines.add(new ColoredLine(line, color));
+                coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
            }
            return;
        }
@@ -152,6 +185,7 @@ public class LayoutGridService {
    }


+    @SneakyThrows
    private void addPlacedText(Page page, Rectangle2D textBBox, String s, LayoutGrid layoutGrid) {

        Point2D.Float upperLeftCorner = switch (page.getRotation()) {
@@ -161,7 +195,8 @@ public class LayoutGridService {
            default -> new Point2D.Float((float) (textBBox.getMinX()), (float) textBBox.getMaxY());
        };
        var placedTexts = layoutGrid.getVisualizationsPerPages().get(page.getNumber() - 1).getPlacedTexts();
-        placedTexts.add(new PlacedText(s, upperLeftCorner));
+        upperLeftCorner.setLocation(upperLeftCorner.getX() - ((FONT.getStringWidth(s) / 1000) * FONT_SIZE + (2 * LINE_WIDTH) + 4), upperLeftCorner.getY() - FONT_SIZE);
+        placedTexts.add(PlacedText.textFacingUp(s, upperLeftCorner, FONT_SIZE, Color.BLACK, FONT));
    }


@@ -176,9 +211,9 @@ public class LayoutGridService {
        midPageLines.remove(1);
        // add string to left line
        var leftLine = midPageLines.remove(1);
-        coloredLines.add(new ColoredLine(leftLine, color));
+        coloredLines.add(new ColoredLine(leftLine, color, LINE_WIDTH));
        for (Line2D line : midPageLines) {
-            coloredLines.add(new ColoredLine(line, color));
+            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
        }
    }

@@ -192,9 +227,9 @@ public class LayoutGridService {
        lastPageLines.remove(0);
        // add string to left line
        var leftLine = lastPageLines.remove(2);
-        coloredLines.add(new ColoredLine(leftLine, color));
+        coloredLines.add(new ColoredLine(leftLine, color, LINE_WIDTH));
        for (Line2D line : lastPageLines) {
-            coloredLines.add(new ColoredLine(line, color));
+            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
        }
    }

@@ -208,9 +243,9 @@ public class LayoutGridService {
        firstPageLines.remove(2);
        // add string to top line
        var firstLine = firstPageLines.remove(0);
-        coloredLines.add(new ColoredLine(firstLine, color));
+        coloredLines.add(new ColoredLine(firstLine, color, LINE_WIDTH));
        for (Line2D line : firstPageLines) {
-            coloredLines.add(new ColoredLine(line, color));
+            coloredLines.add(new ColoredLine(line, color, LINE_WIDTH));
        }
    }

@@ -276,7 +311,10 @@ public class LayoutGridService {
    private void addAsRectangle(SemanticNode semanticNode, LayoutGrid layoutGrid, Color color) {

        semanticNode.getBBox()
-                .forEach((page, textBBox) -> layoutGrid.getVisualizationsPerPages().get(page.getNumber() - 1).getColoredRectangles().add(new ColoredRectangle(textBBox, color)));
+                .forEach((page, textBBox) -> layoutGrid.getVisualizationsPerPages()
+                        .get(page.getNumber() - 1)
+                        .getColoredRectangles()
+                        .add(new ColoredRectangle(textBBox, color, LINE_WIDTH)));
    }

 }
@@ -1,181 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services.visualization;
-
-import java.awt.geom.AffineTransform;
-import java.awt.geom.Rectangle2D;
-import java.io.IOException;
-import java.io.OutputStream;
-import java.util.HashSet;
-import java.util.Set;
-
-import org.apache.pdfbox.cos.COSDictionary;
-import org.apache.pdfbox.cos.COSName;
-import org.apache.pdfbox.pdmodel.PDDocument;
-import org.apache.pdfbox.pdmodel.PDDocumentCatalog;
-import org.apache.pdfbox.pdmodel.PDPage;
-import org.apache.pdfbox.pdmodel.PDPageContentStream;
-import org.apache.pdfbox.pdmodel.PDResources;
-import org.apache.pdfbox.pdmodel.font.PDFont;
-import org.apache.pdfbox.pdmodel.font.PDType1Font;
-import org.apache.pdfbox.pdmodel.font.Standard14Fonts;
-import org.apache.pdfbox.pdmodel.graphics.optionalcontent.PDOptionalContentGroup;
-import org.apache.pdfbox.pdmodel.graphics.optionalcontent.PDOptionalContentProperties;
-import org.apache.pdfbox.pdmodel.graphics.state.PDExtendedGraphicsState;
-import org.apache.pdfbox.util.Matrix;
-import org.springframework.stereotype.Service;
-
-import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.ColoredLine;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.ColoredRectangle;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.FilledRectangle;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.LayoutGrid;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.PlacedText;
-import com.knecon.fforesight.service.layoutparser.processor.model.visualization.VisualizationsOnPage;
-
-import lombok.RequiredArgsConstructor;
-import lombok.SneakyThrows;
-import lombok.extern.slf4j.Slf4j;
-
-@Slf4j
-@Service
-@RequiredArgsConstructor
-public class ViewerDocumentService {
-
-
-    private static final String LAYER_NAME = "Layout grid";
-    private static final int FONT_SIZE = 10;
-    public static final float LINE_WIDTH = 1f;
-
-    private final LayoutGridService layoutGridService;
-
-
-    @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.
-        Set<COSDictionary> dictionariesToUpdate = new HashSet<>();
-        PDOptionalContentGroup layer = addLayerToDocument(pdDocument, dictionariesToUpdate, layerVisibilityDefaultValue);
-        PDFont font = new PDType1Font(Standard14Fonts.FontName.HELVETICA);
-
-        for (int pageNumber = 0; pageNumber < pdDocument.getNumberOfPages(); pageNumber++) {
-            PDPage pdPage = pdDocument.getPage(pageNumber);
-
-            AffineTransform textDeRotationMatrix = getTextDeRotationTransform(pdPage);
-            addLayerToPageRessources(pdPage);
-
-            // We need to save the graphics state before, such that our appended content cannot be affected by previous content streams with side effects,
-            // e.g. not escaped matrix transformations.
-            escapePreviousContents(pdDocument, pdPage);
-
-            VisualizationsOnPage visualizationsOnPage = layoutGrid.getVisualizationsPerPages().get(pageNumber);
-            assert pageNumber == visualizationsOnPage.getPageNumber();
-            // We need to append to the content stream, otherwise the content could be overlapped by following content.
-            try (var contentStream = new PDPageContentStream(pdDocument, pdPage, PDPageContentStream.AppendMode.APPEND, true)) {
-
-                contentStream.beginMarkedContent(COSName.OC, layer);
-                contentStream.saveGraphicsState();
-
-                contentStream.setLineWidth(LINE_WIDTH);
-                for (ColoredLine coloredLine : visualizationsOnPage.getColoredLines()) {
-                    contentStream.setStrokingColor(coloredLine.color());
-                    contentStream.moveTo((float) coloredLine.line().getX1(), (float) coloredLine.line().getY1());
-                    contentStream.lineTo((float) coloredLine.line().getX2(), (float) coloredLine.line().getY2());
-                    contentStream.stroke();
-                }
-                for (ColoredRectangle coloredRectangle : visualizationsOnPage.getColoredRectangles()) {
-                    contentStream.setStrokingColor(coloredRectangle.color());
-                    Rectangle2D r = coloredRectangle.rectangle2D();
-                    contentStream.addRect((float) r.getX(), (float) r.getY(), (float) r.getWidth(), (float) r.getHeight());
-                    contentStream.stroke();
-                }
-                for (FilledRectangle filledRectangle : visualizationsOnPage.getFilledRectangles()) {
-                    contentStream.setNonStrokingColor(filledRectangle.color());
-                    PDExtendedGraphicsState graphicsState = new PDExtendedGraphicsState();
-                    graphicsState.setNonStrokingAlphaConstant(filledRectangle.alpha());
-                    contentStream.setGraphicsStateParameters(graphicsState);
-                    Rectangle2D r = filledRectangle.rectangle2D();
-                    contentStream.addRect((float) r.getX(), (float) r.getY(), (float) r.getWidth(), (float) r.getHeight());
-                    contentStream.fill();
-                }
-                for (PlacedText placedText : visualizationsOnPage.getPlacedTexts()) {
-                    contentStream.setFont(font, FONT_SIZE);
-                    contentStream.beginText();
-                    Matrix textMatrix = new Matrix((float) textDeRotationMatrix.getScaleX(),
-                            (float) textDeRotationMatrix.getShearX(),
-                            (float) textDeRotationMatrix.getShearY(),
-                            (float) textDeRotationMatrix.getScaleY(),
-                            (float) placedText.lineStart().getX(),
-                            (float) placedText.lineStart().getY());
-                    textMatrix.translate(-((font.getStringWidth(placedText.text()) / 1000) * FONT_SIZE + (2 * LINE_WIDTH) + 4), -FONT_SIZE);
-                    contentStream.setTextMatrix(textMatrix);
-                    contentStream.showText(placedText.text());
-                    contentStream.endText();
-                }
-                contentStream.restoreGraphicsState();
-                contentStream.endMarkedContent();
-            }
-            dictionariesToUpdate.add(pdPage.getCOSObject());
-            dictionariesToUpdate.add(pdPage.getResources().getCOSObject());
-        }
-        dictionariesToUpdate.add(pdDocument.getDocumentInformation().getCOSObject());
-//        dictionariesToUpdate.add(pdDocument.getDocument().getTrailer());
-        pdDocument.saveIncremental(outputStream, dictionariesToUpdate);
-        log.info("Saved Viewer Document");
-    }
-
-
-    private static void addLayerToPageRessources(PDPage pdPage) {
-
-        PDResources resources = pdPage.getResources();
-        if (resources == null) {
-            resources = new PDResources();
-            pdPage.setResources(resources);
-        }
-    }
-
-
-    private static void escapePreviousContents(PDDocument pdDocument, PDPage pdPage) throws IOException {
-
-        try (var contentStream = new PDPageContentStream(pdDocument, pdPage, PDPageContentStream.AppendMode.PREPEND, false)) {
-            contentStream.saveGraphicsState();
-        }
-        try (var contentStream = new PDPageContentStream(pdDocument, pdPage, PDPageContentStream.AppendMode.APPEND, false)) {
-            contentStream.restoreGraphicsState();
-        }
-    }
-
-
-    private static PDOptionalContentGroup addLayerToDocument(PDDocument pdDocument, Set<COSDictionary> dictionariesToUpdate, boolean layerVisibilityDefaultValue) {
-
-        PDDocumentCatalog catalog = pdDocument.getDocumentCatalog();
-        PDOptionalContentProperties ocprops = catalog.getOCProperties();
-        if (ocprops == null) {
-            ocprops = new PDOptionalContentProperties();
-            catalog.setOCProperties(ocprops);
-        }
-        PDOptionalContentGroup layer = null;
-        if (ocprops.hasGroup(LAYER_NAME)) {
-            layer = ocprops.getGroup(LAYER_NAME);
-        } else {
-            layer = new PDOptionalContentGroup(LAYER_NAME);
-            ocprops.addGroup(layer);
-        }
-        ocprops.setGroupEnabled(layer, layerVisibilityDefaultValue);
-        dictionariesToUpdate.add(catalog.getCOSObject());
-        return layer;
-    }
-
-
-    private static AffineTransform getTextDeRotationTransform(PDPage page) {
-
-        return AffineTransform.getQuadrantRotateInstance(switch (page.getRotation()) {
-            case 90 -> 3;
-            case 180 -> 2;
-            case 270 -> 1;
-            default -> 0;
-        });
-    }
-
-}
@@ -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;
@@ -30,8 +30,6 @@ public class TableMergingUtility {
            if (consecutiveTable.getColCount() == originalTablePageBlock.getColCount() && !hasTableHeader(consecutiveTable) && outerBoundaryAlignsX(originalTablePageBlock,
                    consecutiveTable)) {
                consecutiveTablesWithSameColCountAndHeaders.add(consecutiveTable);
-            } else {
-                break;
            }
        }
        return Stream.concat(Stream.of(originalTablePageBlock), consecutiveTablesWithSameColCountAndHeaders.stream()).toList();
@@ -3,40 +3,53 @@ import org.springframework.boot.gradle.tasks.bundling.BootBuildImage
 plugins {
    application
    id("com.knecon.fforesight.java-conventions")
-    id("org.springframework.boot") version "3.1.3"
+    id("org.springframework.boot") version "3.1.5"
    id("io.spring.dependency-management") version "1.1.3"
    id("org.sonarqube") version "4.3.0.3225"
-    id("io.freefair.lombok") version "8.2.2"
+    id("io.freefair.lombok") version "8.4"
 //    id("org.graalvm.buildtools.native") version "0.9.23"
 }

 description = "layoutparser-service-server"

+configurations {
+    all {
+        exclude(group = "commons-logging", module = "commons-logging")
+        exclude(group = "org.springframework.boot", module = "spring-boot-starter-log4j2")
+        exclude(group = "com.iqser.red.commons", module = "logging-commons")
+    }
+}
+
+
 val jacksonVersion = "2.15.2"
 val pdfBoxVersion = "3.0.0"
+val springBootStarterVersion = "3.1.5"

 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("org.springframework.boot:spring-boot-starter-actuator:3.1.3")
+    implementation("com.iqser.red.commons:storage-commons:2.45.0")
+    implementation("com.knecon.fforesight:tenant-commons:0.21.0")
+    implementation("com.knecon.fforesight:tracing-commons:0.5.0")
+    implementation("org.springframework.boot:spring-boot-starter-actuator:${springBootStarterVersion}")
+    implementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
    implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
-    implementation("org.springframework.boot:spring-boot-starter-amqp:3.1.3")
    implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
+    implementation("net.logstash.logback:logstash-logback-encoder:7.4")

    // for integration testing only
+    testImplementation(project(":viewer-doc-processor"))
    testImplementation(project(":layoutparser-service-internal-api"))

-    testImplementation("org.springframework.boot:spring-boot-starter-amqp:3.1.3")
+    testImplementation("org.springframework.boot:spring-boot-starter-amqp:${springBootStarterVersion}")
+    testImplementation("org.springframework.boot:spring-boot-starter-test:${springBootStarterVersion}")
    testImplementation("com.iqser.red.service:persistence-service-shared-api-v1:2.144.0")
    testImplementation("com.iqser.red.commons:jackson-commons:1.3.0")
    testImplementation("com.fasterxml.jackson.module:jackson-module-afterburner:${jacksonVersion}")
    testImplementation("com.fasterxml.jackson.datatype:jackson-datatype-jsr310:${jacksonVersion}")
    testImplementation("org.apache.pdfbox:pdfbox:${pdfBoxVersion}")
    testImplementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
-    testImplementation("org.springframework.boot:spring-boot-starter-test:3.1.3")
 }

 /*
@@ -5,6 +5,7 @@ import org.springframework.boot.actuate.autoconfigure.security.servlet.Managemen
 import org.springframework.boot.autoconfigure.ImportAutoConfiguration;
 import org.springframework.boot.autoconfigure.SpringBootApplication;
 import org.springframework.boot.autoconfigure.security.servlet.SecurityAutoConfiguration;
+import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.Import;

 import com.amazonaws.services.s3.model.metrics.MetricsConfiguration;
@@ -13,8 +14,11 @@ import com.knecon.fforesight.service.layoutparser.processor.LayoutParsingService
 import com.knecon.fforesight.service.layoutparser.server.queue.MessagingConfiguration;
 import com.knecon.fforesight.tenantcommons.MultiTenancyAutoConfiguration;

+import io.micrometer.observation.ObservationRegistry;
+import io.micrometer.observation.aop.ObservedAspect;
+
@ImportAutoConfiguration({MultiTenancyAutoConfiguration.class})
-@Import({MetricsConfiguration.class, StorageAutoConfiguration.class,  LayoutParsingServiceProcessorConfiguration.class, MessagingConfiguration.class})
+@Import({MetricsConfiguration.class, StorageAutoConfiguration.class, LayoutParsingServiceProcessorConfiguration.class, MessagingConfiguration.class})
@SpringBootApplication(exclude = {SecurityAutoConfiguration.class, ManagementWebSecurityAutoConfiguration.class})
 public class Application {

@@ -23,4 +27,11 @@ public class Application {
        SpringApplication.run(Application.class, args);
    }

+
+    @Bean
+    public ObservedAspect observedAspect(ObservationRegistry observationRegistry) {
+
+        return new ObservedAspect(observationRegistry);
+    }
+
 }
@@ -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);
    }

@@ -24,8 +24,7 @@ public class MessagingConfiguration {
    @Bean
    public Queue layoutparsingResponseQueue() {

-        return QueueBuilder.durable(LayoutParsingQueueNames.LAYOUT_PARSING_FINISHED_EVENT_QUEUE)//
-                .withArgument("x-dead-letter-exchange", "").withArgument("x-dead-letter-routing-key", LayoutParsingQueueNames.LAYOUT_PARSING_DLQ).build();
+        return QueueBuilder.durable(LayoutParsingQueueNames.LAYOUT_PARSING_FINISHED_EVENT_QUEUE).build();
    }


@@ -5,10 +5,19 @@ tenant-user-management-service.url: "http://tenant-user-management-service:8080/

 fforesight.tenants.remote: true

+logging.pattern.level: "%5p [${spring.application.name},%X{traceId:-},%X{spanId:-}]"
+
+logging.type: ${LOGGING_TYPE:CONSOLE}
+kubernetes.namespace: ${NAMESPACE:default}
+project.version: 1.0-SNAPSHOT
+
+
 server:
  port: 8080

 spring:
+  application:
+    name: layoutparser-service
  main:
    allow-circular-references: true # FIXME
  rabbitmq:
@@ -32,6 +41,13 @@ management:
    prometheus.enabled: ${monitoring.enabled:false}
    health.enabled: true
  endpoints.web.exposure.include: prometheus, health
+  tracing:
+    enabled: ${TRACING_ENABLED:false}
+    sampling:
+      probability: ${TRACING_PROBABILITY:1.0}
+  otlp:
+    tracing:
+      endpoint: ${OTLP_ENDPOINT:http://otel-collector-opentelemetry-collector.otel-collector:4318/v1/traces}


 storage:
@@ -0,0 +1,17 @@
+<configuration>
+
+    <springProperty scope="configuration" name="logType" source="logging.type"/>
+    <springProperty scope="context" name="application.name" source="spring.application.name"/>
+    <springProperty scope="context" name="version" source="project.version"/>
+    <include resource="org/springframework/boot/logging/logback/defaults.xml"/>
+    <include resource="org/springframework/boot/logging/logback/console-appender.xml"/>
+
+    <appender name="JSON" class="ch.qos.logback.core.ConsoleAppender">
+        <encoder class="net.logstash.logback.encoder.LogstashEncoder"/>
+    </appender>
+
+    <root level="INFO">
+        <appender-ref ref="${logType}"/>
+    </root>
+
+</configuration>
@@ -46,12 +46,11 @@ public class BdrJsonBuildTest extends AbstractTest {
    @SneakyThrows
    protected Document buildGraph(File file) {

-        try (PDDocument pdDocument = Loader.loadPDF(file)) {
-            return DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.TAAS,
-                    pdDocument,
-                    new ImageServiceResponse(),
-                    new TableServiceResponse()));
-        }
+        return DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.TAAS,
+                file,
+                new ImageServiceResponse(),
+                new TableServiceResponse(),
+                file.toString()));
    }


@@ -6,7 +6,6 @@ import java.util.List;
 import java.util.Set;
 import java.util.stream.Collectors;

-import org.apache.pdfbox.Loader;
 import org.assertj.core.api.Assertions;
 import org.junit.jupiter.api.Test;
 import org.junit.jupiter.api.extension.ExtendWith;
@@ -96,9 +95,10 @@ public class HeadlinesGoldStandardIntegrationTest {
        goldStandardLog.getRedactionLogEntry().forEach(e -> goldStandardHeadlines.add(new Headline(e.getPositions().get(0).getPage(), e.getValue())));

        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                Loader.loadPDF(pdfFileResource.getFile()),
+                pdfFileResource.getFile(),
                new ImageServiceResponse(),
-                new TableServiceResponse()));
+                new TableServiceResponse(),
+                filePath));

        var foundHeadlines = documentGraph.streamAllSubNodes()
                .map(SemanticNode::getHeadline)
@@ -148,6 +148,7 @@ public class HeadlinesGoldStandardIntegrationTest {
    @EqualsAndHashCode
    @AllArgsConstructor
    @ToString
+    @SuppressWarnings("PMD.AvoidFieldNameMatchingTypeName")
    private class Headline {

        private int page;
@@ -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/Wie weiter bei Kristeneinrichtungen.pdf");
+        LayoutParsingRequest layoutParsingRequest = buildDefaultLayoutParsingRequest(LayoutParsingType.REDACT_MANAGER);
+        LayoutParsingFinishedEvent finishedEvent = layoutParsingPipeline.parseLayoutAndSaveFilesToStorage(layoutParsingRequest);
+        Arrays.stream(finishedEvent.message().split("\n")).forEach(log::info);
+    }
+
+}
@@ -6,6 +6,7 @@ import java.nio.file.Files;
 import java.nio.file.Path;
 import java.util.List;

+import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.springframework.core.io.ClassPathResource;

@@ -17,7 +18,9 @@ import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest
 import lombok.SneakyThrows;

 public class DocumentDataTests extends BuildDocumentTest {
+
    @Test
+    @Disabled // This test takes waaaaaay too long, it's ridiculous
    @SneakyThrows
    public void createDocumentDataForAllFiles() {

@@ -36,11 +39,12 @@ public class DocumentDataTests extends BuildDocumentTest {
        for (String pdfFileName : pdfFileNames) {
            System.out.println(pdfFileName);
            DocumentData documentData = DocumentDataMapper.toDocumentData(buildGraph(resource.getFile().toPath().getParent().relativize(Path.of(pdfFileName)).toString()));
-                        File outputFile = Path.of(outPath).resolve(resource.getFile().toPath().relativize(Path.of(pdfFileName))).toFile();
+            File outputFile = Path.of(outPath).resolve(resource.getFile().toPath().relativize(Path.of(pdfFileName))).toFile();
            outputFile.toPath().getParent().toFile().mkdirs();
            try (var out = new FileOutputStream(outputFile.toString().replace(".pdf", ".json"))) {
                ObjectMapperFactory.create().writeValue(out, documentData);
            }
        }
    }
+
 }
@@ -5,7 +5,6 @@ import java.io.FileOutputStream;
 import java.nio.file.Files;
 import java.nio.file.Path;

-import org.apache.pdfbox.Loader;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.springframework.core.io.ClassPathResource;
@@ -56,9 +55,10 @@ public class DocumentGraphJsonWritingTest extends BuildDocumentTest {
    private void writeJsons(Path filename) {

        Document documentGraph = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                Loader.loadPDF(filename.toFile()),
+                filename.toFile(),
                new ImageServiceResponse(),
-                new TableServiceResponse()));
+                new TableServiceResponse(),
+                filename.toFile().toString()));

        DocumentData documentData = DocumentDataMapper.toDocumentData(documentGraph);
        ObjectMapper mapper = ObjectMapperFactory.create();
@@ -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);
+    }
+}
@@ -1,36 +1,68 @@
 package com.knecon.fforesight.service.layoutparser.server.graph;

-import java.io.FileOutputStream;
+import java.io.File;
 import java.nio.file.Path;

-import org.apache.pdfbox.Loader;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.springframework.core.io.ClassPathResource;

+import com.iqser.red.commons.jackson.ObjectMapperFactory;
 import com.knecon.fforesight.service.layoutparser.internal.api.queue.LayoutParsingType;
 import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Document;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.image.ImageServiceResponse;
+import com.knecon.fforesight.service.layoutparser.processor.python_api.model.table.TableServiceResponse;
+import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
-import com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService;
 import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest;
+import com.knecon.fforesight.service.viewerdoc.service.ViewerDocumentService;

 import lombok.SneakyThrows;

 public class ViewerDocumentTest extends BuildDocumentTest {

    @Test
-    @Disabled
    @SneakyThrows
    public void testViewerDocument() {

-        LayoutGridService layoutGridService = new LayoutGridService();
-        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(layoutGridService);
-        String fileName = "files/new/VV-511309_OCR.pdf";
-        Document document = buildGraph(fileName, LayoutParsingType.DOCUMINE);
+        String fileName = "files/S-Metolachlor_RAR_01_Volume_1_2018-09-06.pdf";
        String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
-        try (var pdDocument = Loader.loadPDF(new ClassPathResource(fileName).getFile()); var out = new FileOutputStream(tmpFileName)) {
-            viewerDocumentService.createViewerDocument(pdDocument, document, out, true);
-        }
+
+        var documentFile = new ClassPathResource(fileName).getFile();
+        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
+        LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);
+
+        Document document = buildGraph(fileName, LayoutParsingType.REDACT_MANAGER);
+        long start = System.currentTimeMillis();
+        layoutGridService.addLayoutGrid(documentFile, document, new File(tmpFileName), true);
+        System.out.printf("Total time: %.2fs%n", ((float) (System.currentTimeMillis() - start)) / 1000);
+    }
+
+
+    @Test
+    @Disabled
+    @SneakyThrows
+    public void testViewerDocumentWithTables() {
+
+        String fileName = "files/cv_tables/brokenTablesOnOcr_ocred.pdf";
+        String tableFileName = "files/cv_tables/brokenTablesOnOcr_ocred.TABLES.json";
+        String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
+
+        var mapper = ObjectMapperFactory.create();
+        var tableResponse = mapper.readValue(new ClassPathResource(tableFileName).getInputStream(), TableServiceResponse.class);
+        var documentFile = new ClassPathResource(fileName).getFile();
+
+        var classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.DOCUMINE,
+                                                                       documentFile,
+                                                                       new ImageServiceResponse(),
+                                                                       tableResponse,
+                                                                       Path.of(fileName).getFileName().toFile().toString());
+        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(null);
+        LayoutGridService layoutGridService = new LayoutGridService(viewerDocumentService);
+        Document document = DocumentGraphFactory.buildDocumentGraph(classificationDocument);
+
+        layoutGridService.addLayoutGrid(documentFile, document, new File(tmpFileName), true);
    }

 }
+
@@ -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 +1,26 @@
 package com.knecon.fforesight.service.layoutparser.server.segmentation;

+import static org.assertj.core.api.Assertions.assertThat;
+
+import java.awt.geom.Rectangle2D;
+import java.io.File;
+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.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;
@@ -8,26 +29,13 @@ import com.knecon.fforesight.service.layoutparser.processor.model.graph.nodes.Im
 import com.knecon.fforesight.service.layoutparser.processor.model.image.ClassifiedImage;
 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.adapter.CvTableParsingAdapter;
-import com.knecon.fforesight.service.layoutparser.processor.python_api.adapter.ImageServiceResponseAdapter;
 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.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 {

@@ -40,22 +48,18 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    @Autowired
    private RedactManagerClassificationService redactManagerClassificationService;

-    @Autowired
-    private CvTableParsingAdapter cvTableParsingAdapter;
-
-    @Autowired
-    private ImageServiceResponseAdapter imageServiceResponseAdapter;
-
    @Autowired
    private SectionsBuilderService sectionsBuilderService;


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

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

        redactManagerClassificationService.classifyDocument(classificationDocument);

@@ -65,6 +69,67 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    }


+    @SneakyThrows
+    public ClassificationDocument buildClassificationDocument(File originDocument) {
+
+        return buildClassificationDocument(originDocument, new TableServiceResponse());
+    }
+
+
+    @Test
+    public void tablesToHtmlDebugger() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T5 VV-640252-Page16.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        toHtml(document, "/tmp/T5.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(pdfFileResource.getFile(), tableServiceResponse);
+
+        toHtml(document, "/tmp/ScanRotationBorder.html");
+    }
+
+
+    @Disabled
+    @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(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() {
@@ -76,12 +141,12 @@ public class PdfSegmentationServiceTest extends AbstractTest {
        imageServiceResponse.getData()
                .forEach(imageMetadata -> images.computeIfAbsent(imageMetadata.getPosition().getPageNumber(), x -> new ArrayList<>())
                        .add(new ClassifiedImage(new Rectangle2D.Double(imageMetadata.getPosition().getX1(),
-                                imageMetadata.getPosition().getY1(),
-                                imageMetadata.getGeometry().getWidth(),
-                                imageMetadata.getGeometry().getHeight()),
-                                ImageType.valueOf(imageMetadata.getClassification().getLabel().toUpperCase(Locale.ROOT)),
-                                imageMetadata.isAlpha(),
-                                imageMetadata.getPosition().getPageNumber())));
+                                                                        imageMetadata.getPosition().getY1(),
+                                                                        imageMetadata.getGeometry().getWidth(),
+                                                                        imageMetadata.getGeometry().getHeight()),
+                                                 ImageType.valueOf(imageMetadata.getClassification().getLabel().toUpperCase(Locale.ROOT)),
+                                                 imageMetadata.isAlpha(),
+                                                 imageMetadata.getPosition().getPageNumber())));

        System.out.println("object");
    }
@@ -92,12 +157,23 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Spanning Cells.pdf");

-        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock table = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
        assertThat(table.getColCount()).isEqualTo(6);
        assertThat(table.getRowCount()).isEqualTo(13);
-        assertThat(table.getRows().stream().mapToInt(List::size).sum()).isEqualTo(6 * 13);
+        assertThat(table.getRows()
+                           .stream()
+                           .mapToInt(List::size).sum()).isEqualTo(6 * 13);
    }


@@ -106,16 +182,38 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Table.pdf");

-        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
        assertThat(firstTable.getColCount()).isEqualTo(8);
        assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(1);
+        TablePageBlock secondTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(1);
        assertThat(secondTable.getColCount()).isEqualTo(8);
        assertThat(secondTable.getRowCount()).isEqualTo(2);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(0).stream().map(Collections::singletonList).collect(Collectors.toList());
-        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
+                .get(0)
+                .stream()
+                .map(Collections::singletonList)
+                .collect(Collectors.toList());
+        assertThat(secondTable.getRows()
+                           .stream()
+                           .allMatch(row -> row.stream()
+                                   .map(Cell::getHeaderCells)
+                                   .toList().equals(firstTableHeaderCells))).isTrue();
    }


@@ -124,16 +222,38 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Multi Page Table.pdf");

-        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
        assertThat(firstTable.getColCount()).isEqualTo(9);
        assertThat(firstTable.getRowCount()).isEqualTo(5);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(1);
+        TablePageBlock secondTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(1);
        assertThat(secondTable.getColCount()).isEqualTo(9);
        assertThat(secondTable.getRowCount()).isEqualTo(6);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(firstTable.getRowCount() - 1).stream().map(Cell::getHeaderCells).collect(Collectors.toList());
-        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
+                .get(firstTable.getRowCount() - 1)
+                .stream()
+                .map(Cell::getHeaderCells)
+                .collect(Collectors.toList());
+        assertThat(secondTable.getRows()
+                           .stream()
+                           .allMatch(row -> row.stream()
+                                   .map(Cell::getHeaderCells)
+                                   .toList().equals(firstTableHeaderCells))).isTrue();
    }


@@ -142,16 +262,38 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Rotated Table Headers.pdf");

-        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
        assertThat(firstTable.getColCount()).isEqualTo(8);
        assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(1);
+        TablePageBlock secondTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(1);
        assertThat(secondTable.getColCount()).isEqualTo(8);
        assertThat(secondTable.getRowCount()).isEqualTo(6);
-        List<List<Cell>> firstTableHeaderCells = firstTable.getRows().get(0).stream().map(Collections::singletonList).collect(Collectors.toList());
-        assertThat(secondTable.getRows().stream().allMatch(row -> row.stream().map(Cell::getHeaderCells).toList().equals(firstTableHeaderCells))).isTrue();
+        List<List<Cell>> firstTableHeaderCells = firstTable.getRows()
+                .get(0)
+                .stream()
+                .map(Collections::singletonList)
+                .collect(Collectors.toList());
+        assertThat(secondTable.getRows()
+                           .stream()
+                           .allMatch(row -> row.stream()
+                                   .map(Cell::getHeaderCells)
+                                   .toList().equals(firstTableHeaderCells))).isTrue();
    }


@@ -160,14 +302,31 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/56 Fludioxonil_RAR_12_Volume_3CA_B-7_2018-02-21_Page170.pdf");

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

        validateTableSize(document, 4);

        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, 4, 19, 12, 0);
+        validateTable(document, 3, 2, 12, 0, 0);
+
+    }
+
+
+    @Test
+    public void testDoc211() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/211.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(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);

    }

@@ -177,36 +336,37 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/VV-931175_Page1.pdf");

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

        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",
-                        "Study title",
-                        "Analytical method Author, date, No.",
-                        "Technique, LOQ of the method, validated working range",
-                        "Method meets analytical validation criteria",
-                        "Remarks (in case validation criteria are not met)",
-                        "Acceptability of the method"),
-                Arrays.asList("",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
-                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",
-                        "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
-                        "Y",
-                        "N/A",
-                        "Y"));
+                                                                "Author, date",
+                                                                "Study title",
+                                                                "Analytical method Author, date, No.",
+                                                                "Technique, LOQ of the method, validated working range",
+                                                                "Method meets analytical validation criteria",
+                                                                "Remarks (in case validation criteria are not met)",
+                                                                "Acceptability of the method"),
+                                                  Arrays.asList(
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
+                                                  Arrays.asList("CA 7.1.2.1.1 DAR (2009)",
+                                                                "Evans P.G. 2001 TMJ4569B, VV-323245",
+                                                                "Azoxystrobin Laboratory Degradation Study in Three Soil Types, Sampled from Holland and the United Kingdom",
+                                                                "Method: RAM 269 Johnson R.I., Tummon O.J., Earl M. 1995 RJ1864B, VV-377731 Johnson R.I., Tummon O.J., Earl M. 1998 RAM 269/02, VV-124072 Johnson R.I., Tummon O.J., Earl M. 2000 RAM 269/03, VV-123986 Validation: Robinson N.J. 2001 TMJ4617B, VV-895845",
+                                                                "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
+                                                                "Y",
+                                                                "N/A",
+                                                                "Y"));

        validateTable(document, 0, values);

@@ -218,7 +378,9 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/27 A8637C - EU AIR3 - MCP Section 1 - Identity of the plant protection product_Page6.pdf");

-        ClassificationDocument document = buildClassificationDocument(Loader.loadPDF(pdfFileResource.getFile()));
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        toHtml(document, "/tmp/html.html");

        validateTableSize(document, 4);

@@ -231,17 +393,29 @@ 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()));
+        ClassificationDocument document = buildClassificationDocument(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(pdfFileResource.getFile());
+
+        validateTableSize(document, 1);
+
+        validateTable(document, 0, 7, 4, 0, 0);
    }


@@ -250,7 +424,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/A20622A izZRMS (CZ) fRR Part B7_Page123.pdf");

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

        validateTableSize(document, 6);

@@ -269,7 +443,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/77 Pirimicarb_RAR_08_Volume_3CA_B-6_2017-12-04_Page11.pdf");

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

        validateTableSize(document, 3);

@@ -285,7 +459,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/95 Trinexapac-ethyl_RAR_08_Volume_3CA_B-6_2018-01-10_Page532.pdf");

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

        validateTableSize(document, 1);
        validateTable(document, 0, 9, 9, 0, 0);
@@ -298,7 +472,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/52 Fludioxonil_RAR_07_Volume_3CA_B-5_2018-02-21_Page175.pdf");

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

        validateTableSize(document, 1);

@@ -312,7 +486,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/52 Fludioxonil_RAR_07_Volume_3CA_B-5_2018-02-21_Page174.pdf");

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

        validateTableSize(document, 1);
        validateTable(document, 0, 9, 6, 7, 0);
@@ -325,10 +499,10 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/19 Chlorothalonil RAR 08 Volume 3CA B 6b metabolites Oct 2017_Page35.pdf");

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

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

    }

@@ -338,7 +512,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/19 Chlorothalonil RAR 08 Volume 3CA B 6b metabolites Oct 2017_Page161.pdf");

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

        validateTableSize(document, 2);
        validateTable(document, 0, 2, 2, 0, 0);
@@ -353,7 +527,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
        ClassPathResource pdfFileResource = new ClassPathResource(
                "files/SinglePages/47 Cyprodinil - EU AIR3 - MCA Section 5 Supplement - Toxicological and metabolism studies on the active substance_Page30.pdf");

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

        validateTableSize(document, 2);

@@ -369,7 +543,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
        ClassPathResource pdfFileResource = new ClassPathResource(
                "files/SinglePages/49 Cyprodinil - EU AIR3 - MCA Section 8 Supplement - Ecotoxicological studies on the active substance_Page61.pdf");

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

        validateTableSize(document, 2);

@@ -384,7 +558,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/81 Pirimicarb_RAR_20_Volume_3CP_A10788A (_Pirimor_)_B-9_2017-12-04_Page54.pdf");

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

        validateTableSize(document, 2);

@@ -399,7 +573,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/85 Pydiflumetofen_DAR_08_Volume_3CA_B-6_2017-07-26_Page134.pdf");

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

        validateTableSize(document, 2);

@@ -414,7 +588,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/Thiabendazole DAR Addendum for ED_April_2020_Page18.pdf");

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

        validateTableSize(document, 4);

@@ -431,7 +605,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/15 - Pretilachlor - Acute Oral Toxicity (Up and Down Procedure) - Rat_Page18.pdf");

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

        validateTableSize(document, 1);

@@ -446,12 +620,12 @@ public class PdfSegmentationServiceTest extends AbstractTest {
        ClassPathResource pdfFileResource = new ClassPathResource(
                "files/SinglePages/28 A8637C - EU AIR3 - MCP Section 10 - Ecotoxicological studies on the plant protection product_Page23.pdf");

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

        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);

    }

@@ -461,7 +635,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/24 - SYN549522 - Acute Oral Toxicity - Rats_Page17.pdf");

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

        validateTableSize(document, 1);

@@ -475,7 +649,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {

        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/30 - Dicamba - Acute Oral Toxicity - Rats_Page5.pdf");

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

        validateTableSize(document, 1);

@@ -484,12 +658,146 @@ public class PdfSegmentationServiceTest extends AbstractTest {
    }


+    @Test
+    public void testT0() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T0 TableWithMergedCells.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 1);
+
+        validateTable(document, 0, 6, 8, 0, 0);
+    }
+
+
+    @Test
+    public void testT1() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T1 MultipleNestedTable.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 4);
+
+        validateTable(document, 0, 3, 3, 0, 0);
+        validateTable(document, 1, 3, 5, 2, 0);
+        validateTable(document, 2, 3, 3, 1, 0);
+        validateTable(document, 3, 3, 3, 0, 0);
+
+    }
+
+
+    @Test
+    public void testT2() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T2 MultipleTables.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 6);
+
+        validateTable(document, 0, 5, 5, 0, 0);
+        validateTable(document, 1, 5, 6, 0, 0);
+        validateTable(document, 2, 5, 5, 0, 0);
+        validateTable(document, 3, 5, 5, 0, 0);
+        validateTable(document, 4, 5, 5, 0, 0);
+        validateTable(document, 5, 5, 5, 0, 0);
+
+    }
+
+
+    @Test
+    public void testT3() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T3 S-Meto_Page29.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 1);
+
+        validateTable(document, 0, 6, 5, 0, 0);
+
+    }
+
+
+    @Test
+    public void testT4() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T4 138 IDD0000261736_Page16.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 1);
+
+        validateTable(document, 0, 5, 8, 1, 0);
+
+    }
+
+
+    @Test
+    public void testT5() throws IOException {
+
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T5 VV-640252-Page16.pdf");
+
+        ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
+
+        validateTableSize(document, 6);
+        validateTable(document, 0, 1, 1, 0, 0);
+        validateTable(document, 1, 1, 1, 0, 0);
+        validateTable(document, 2, 1, 1, 0, 0);
+        validateTable(document, 3, 1, 1, 0, 0);
+        validateTable(document, 4, 1, 1, 0, 0);
+        validateTable(document, 5, 1, 1, 0, 0);
+
+    }
+
+
+    @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);
+        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();
+        int emptyCellsFoundFound = rows.stream()
+                .flatMap(List::stream)
+                .toList()
+                .stream()
+                .filter(f -> f.toString().isEmpty())
+                .toList().size();

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

        assertThat(table.getColCount()).isEqualTo(colCount);
@@ -500,11 +808,20 @@ public class PdfSegmentationServiceTest extends AbstractTest {

    private void validateTable(ClassificationDocument document, int tableIndex, List<List<String>> values) {

-        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(tableIndex);
+        TablePageBlock table = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(tableIndex);
        List<List<Cell>> rows = table.getRows();

-        List<Cell> rowsFlattened = rows.stream().flatMap(List::stream).toList();
-        List<String> valuesFlattened = values.stream().flatMap(List::stream).toList();
+        List<Cell> rowsFlattened = rows.stream()
+                .flatMap(List::stream)
+                .toList();
+        List<String> valuesFlattened = values.stream()
+                .flatMap(List::stream)
+                .toList();

        for (int i = 0; i < valuesFlattened.size(); i++) {
            Cell cell = rowsFlattened.get(i);
@@ -517,7 +834,11 @@ public class PdfSegmentationServiceTest extends AbstractTest {

    private void validateTableSize(ClassificationDocument document, int tableSize) {

-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().size()).isEqualTo(tableSize);
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .toList().size()).isEqualTo(tableSize);

    }

@@ -1,21 +1,37 @@
 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.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 +41,95 @@ 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() {
+
+        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 (String pdfFileName : pdfFileNames) {
+            writeJsons(Path.of(pdfFileName));
        }
    }

+
+    @SneakyThrows
+    private void writeJsons(Path filename) {
+
+        Document documentGraphBefore = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
+                filename.toFile(),
+                new ImageServiceResponse(),
+                new TableServiceResponse(),
+                filename.toFile().toString()));
+        Document documentGraphAfter = DocumentGraphFactory.buildDocumentGraph(layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
+                filename.toFile(),
+                new ImageServiceResponse(),
+                new TableServiceResponse(),
+                filename.toFile().toString()));
+        DocumentData documentDataBefore = DocumentDataMapper.toDocumentData(documentGraphBefore);
+        DocumentData documentDataAfter = DocumentDataMapper.toDocumentData(documentGraphAfter);
+        if (!compareStructures(documentDataBefore.getDocumentStructure(), documentDataAfter.getDocumentStructure())) {
+            String tmpFileNameBefore = "/tmp/before." + filename.getFileName().toString();
+            try (PDDocument pdDocument = Loader.loadPDF(filename.toFile())) {
+                PdfDraw.drawDocumentGraph(pdDocument, documentGraphBefore);
+                pdDocument.save(tmpFileNameBefore);
+            }
+            String tmpFileNameAfter = "/tmp/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) {
+
+        List<Table> 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<Table> 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 = listStructure1.get(i);
+            Table tableNode2 = listStructure2.get(i);
+            if (tableNode1.getNumberOfRows() != tableNode2.getNumberOfRows() || tableNode1.getNumberOfCols() != tableNode2.getNumberOfCols()) {
+                return false;
+            }
+        }
+        return true;
+    }
+
 }
@@ -1,6 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.server.utils;

 import java.io.InputStream;
+import java.util.Map;
 import java.util.Optional;

 import org.junit.jupiter.api.AfterEach;
@@ -12,13 +13,17 @@ 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.Bean;
 import org.springframework.context.annotation.ComponentScan;
 import org.springframework.context.annotation.Configuration;
 import org.springframework.context.annotation.Import;
+import org.springframework.context.annotation.Primary;
 import org.springframework.core.io.ClassPathResource;
 import org.springframework.test.context.junit.jupiter.SpringExtension;

+import 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;
@@ -53,7 +58,6 @@ 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 +74,6 @@ 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 +104,15 @@ 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)
+                .identifier(Map.of("fileId", "1337"))
+                .layoutParsingType(layoutParsingType)
                .originFileStorageId(ORIGIN_FILE_ID)
                .tablesFileStorageId(Optional.of(TABLE_FILE_ID))
                .imagesFileStorageId(Optional.of(IMAGE_FILE_ID))
@@ -111,7 +121,6 @@ public abstract class AbstractTest {
                .positionBlockFileStorageId(POSITION_FILE_ID)
                .pageFileStorageId(PAGES_FILE_ID)
                .simplifiedTextStorageId(SIMPLIFIED_ID)
-                .sectionGridStorageId(SECTION_GRID_ID)
                .viewerDocumentStorageId(VIEWER_DOCUMENT_ID)
                .build();
    }
@@ -135,19 +144,7 @@ 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 +162,13 @@ public abstract class AbstractTest {
    @ComponentScan("com.knecon.fforesight.service.layoutparser")
    public static class TestConfiguration {

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

 }
@@ -1,9 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.server.utils;

-import java.io.InputStream;
+import java.io.File;

-import org.apache.pdfbox.Loader;
-import org.apache.pdfbox.pdmodel.PDDocument;
 import org.springframework.beans.factory.annotation.Autowired;
 import org.springframework.core.io.ClassPathResource;

@@ -25,11 +23,9 @@ public abstract class BuildDocumentTest extends AbstractTest {
    @SneakyThrows
    protected ClassificationDocument parseLayout(String filename, LayoutParsingType layoutParsingType) {

-        ClassPathResource fileResource = new ClassPathResource(filename);
+        File fileResource = new ClassPathResource(filename).getFile();
        prepareStorage(filename);
-        try (InputStream inputStream = fileResource.getInputStream(); PDDocument pdDocument = Loader.loadPDF(inputStream.readAllBytes())) {
-            return layoutParsingPipeline.parseLayout(layoutParsingType, pdDocument, layoutParsingStorageService.getImagesFile(IMAGE_FILE_ID), new TableServiceResponse());
-        }
+        return layoutParsingPipeline.parseLayout(layoutParsingType, fileResource, layoutParsingStorageService.getImagesFile(IMAGE_FILE_ID), new TableServiceResponse(), filename);
    }


@@ -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();
-    }
-
-}
@@ -7,6 +7,9 @@ fforesight.tenants.remote: true
 server:
  port: 8080

+logging.pattern.level: "%5p [${spring.application.name},%X{traceId:-},%X{spanId:-}]"
+logging.type: ${LOGGING_TYPE:CONSOLE}
+
 spring:
  main:
    allow-circular-references: true # FIXME
@@ -0,0 +1,17 @@
+<configuration>
+
+    <springProperty scope="configuration" name="logType" source="logging.type"/>
+    <springProperty scope="context" name="application.name" source="spring.application.name"/>
+    <springProperty scope="context" name="version" source="project.version"/>
+    <include resource="org/springframework/boot/logging/logback/defaults.xml"/>
+    <include resource="org/springframework/boot/logging/logback/console-appender.xml"/>
+
+    <appender name="JSON" class="ch.qos.logback.core.ConsoleAppender">
+        <encoder class="net.logstash.logback.encoder.LogstashEncoder"/>
+    </appender>
+
+    <root level="INFO">
+        <appender-ref ref="${logType}"/>
+    </root>
+
+</configuration>
@@ -0,0 +1,19 @@
+plugins {
+    id("com.knecon.fforesight.java-conventions")
+    id("io.freefair.lombok") version "8.2.2"
+}
+
+description = "Library for adding/removing layers in the viewer document"
+
+var pdfBoxVersion = "3.0.0"
+
+dependencies {
+    implementation("org.apache.pdfbox:pdfbox:${pdfBoxVersion}")
+    implementation("org.apache.pdfbox:pdfbox-tools:${pdfBoxVersion}")
+    implementation("org.slf4j:slf4j-api:1.7.25")
+    implementation("com.knecon.fforesight:tracing-commons:0.5.0")
+
+    testImplementation("org.apache.logging.log4j:log4j-slf4j-impl:2.22.1")
+    testImplementation("org.junit.jupiter:junit-jupiter")
+    testImplementation platform('org.junit:junit-bom:5.10.0')
+}
@@ -0,0 +1,33 @@
+package com.knecon.fforesight.service.viewerdoc;
+
+import java.util.List;
+
+import org.apache.pdfbox.cos.COSName;
+
+import lombok.AccessLevel;
+import lombok.experimental.FieldDefaults;
+
+@FieldDefaults(makeFinal = true, level = AccessLevel.PUBLIC)
+public class ContentStreams {
+
+    public static Identifier KNECON_LAYOUT = new Identifier("Layout grid", COSName.getPDFName("KNECON_LAYOUT"), true);
+
+    public static Identifier KNECON_OCR = new Identifier("OCR", COSName.getPDFName("KNECON_OCR"), false);
+
+    public static Identifier KNECON_OCR_TEXT_DEBUG = new Identifier("OCR Text", COSName.getPDFName("KNECON_OCR_TEXT_DEBUG"), true);
+
+    public static Identifier KNECON_OCR_BBOX_DEBUG = new Identifier("OCR Boxes", COSName.getPDFName("KNECON_OCR_BBOX_DEBUG"), true);
+
+    public static Identifier OTHER = new Identifier("other", COSName.getPDFName("OTHER"), false);
+
+    public static Identifier ESCAPE_START = new Identifier("escape start", COSName.getPDFName("ESCAPE_START"), false);
+
+    public static Identifier ESCAPE_END = new Identifier("escape start", COSName.getPDFName("ESCAPE_END"), false);
+
+    public static List<Identifier> allContentStreams = List.of(KNECON_LAYOUT, KNECON_OCR, KNECON_OCR_BBOX_DEBUG, KNECON_OCR_TEXT_DEBUG, OTHER, ESCAPE_START, ESCAPE_END);
+
+    public record Identifier(String name, COSName cosName, boolean optionalContent) {
+
+    }
+
+}
@@ -0,0 +1,8 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import java.awt.Color;
+import java.awt.geom.Line2D;
+
+public record ColoredLine(Line2D line, Color color, float lineWidth) {
+
+}
@@ -1,8 +1,8 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
+package com.knecon.fforesight.service.viewerdoc.model;

 import java.awt.Color;
 import java.awt.geom.Rectangle2D;

-public record ColoredRectangle(Rectangle2D rectangle2D, Color color) {
+public record ColoredRectangle(Rectangle2D rectangle2D, Color color, float lineWidth) {

 }
@@ -0,0 +1,10 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.font.PDFont;
+
+public interface EmbeddableFont {
+
+    PDFont embed(PDDocument document);
+
+}
@@ -1,4 +1,4 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
+package com.knecon.fforesight.service.viewerdoc.model;

 import java.awt.Color;
 import java.awt.geom.Rectangle2D;
@@ -1,7 +1,7 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
+package com.knecon.fforesight.service.viewerdoc.model;

-import java.util.ArrayList;
-import java.util.List;
+import java.util.HashMap;
+import java.util.Map;

 import lombok.AccessLevel;
 import lombok.Getter;
@@ -12,15 +12,15 @@ import lombok.experimental.FieldDefaults;
 public class LayoutGrid {

    int numberOfPages;
-    List<VisualizationsOnPage> visualizationsPerPages;
+    Map<Integer, VisualizationsOnPage> visualizationsPerPages;


    public LayoutGrid(int numberOfPages) {

        this.numberOfPages = numberOfPages;
-        this.visualizationsPerPages = new ArrayList<>(numberOfPages);
+        this.visualizationsPerPages = new HashMap<>();
        for (int i = 0; i < numberOfPages; i++) {
-            this.visualizationsPerPages.add(VisualizationsOnPage.builder().pageNumber(i).build());
+            this.visualizationsPerPages.put(i, VisualizationsOnPage.builder().build());
        }
    }

@@ -0,0 +1,10 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import java.util.List;
+
+import org.apache.pdfbox.contentstream.operator.Operator;
+import org.apache.pdfbox.cos.COSBase;
+
+public record OperatorWithArguments(Operator operator, List<COSBase> arguments) {
+
+}
@@ -0,0 +1,17 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import java.awt.Color;
+import java.awt.geom.Point2D;
+import java.util.Optional;
+
+import org.apache.pdfbox.pdmodel.graphics.state.RenderingMode;
+import org.apache.pdfbox.util.Matrix;
+
+public record PlacedText(String text, Point2D lineStart, Color color, float fontSize, EmbeddableFont font, Optional<Matrix> textMatrix, Optional<RenderingMode> renderingMode) {
+
+    public static PlacedText textFacingUp(String text, Point2D lineStart, float fontSize, Color color, EmbeddableFont font) {
+
+        return new PlacedText(text, lineStart, color, fontSize, font, Optional.empty(), Optional.empty());
+    }
+
+}
@@ -0,0 +1,37 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.font.PDFont;
+import org.apache.pdfbox.pdmodel.font.PDType1Font;
+import org.apache.pdfbox.pdmodel.font.Standard14Fonts;
+
+import lombok.RequiredArgsConstructor;
+import lombok.SneakyThrows;
+
+@RequiredArgsConstructor
+public class Standard14EmbeddableFont implements EmbeddableFont {
+
+    private final PDType1Font font;
+
+
+    public static Standard14EmbeddableFont helvetica() {
+
+        return new Standard14EmbeddableFont(new PDType1Font(Standard14Fonts.FontName.HELVETICA));
+    }
+
+
+    @SneakyThrows
+    public float getStringWidth(String text) {
+
+        return font.getStringWidth(text);
+    }
+
+
+    @Override
+    public PDFont embed(PDDocument document) {
+
+        // no need to embed anything
+        return font;
+    }
+
+}
@@ -0,0 +1,23 @@
+package com.knecon.fforesight.service.viewerdoc.model;
+
+import java.util.Map;
+
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+
+import lombok.AccessLevel;
+import lombok.AllArgsConstructor;
+import lombok.Builder;
+import lombok.Getter;
+import lombok.experimental.FieldDefaults;
+
+@Builder
+@Getter
+@AllArgsConstructor
+@FieldDefaults(level = AccessLevel.PRIVATE)
+public class Visualizations {
+
+    ContentStreams.Identifier layer;
+    Map<Integer, VisualizationsOnPage> visualizationsOnPages;
+    boolean layerVisibilityDefaultValue;
+
+}
@@ -1,4 +1,4 @@
-package com.knecon.fforesight.service.layoutparser.processor.model.visualization;
+package com.knecon.fforesight.service.viewerdoc.model;

 import java.util.LinkedList;
 import java.util.List;
@@ -13,7 +13,7 @@ import lombok.experimental.FieldDefaults;
@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
 public class VisualizationsOnPage {

-    int pageNumber;
+    boolean makePathsInvisible;
    @Builder.Default
    List<PlacedText> placedTexts = new LinkedList<>();
    @Builder.Default
@@ -0,0 +1,7 @@
+package com.knecon.fforesight.service.viewerdoc.pdf;
+
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+
+public record ClassifiedContentStream(SinglePDContentStream contentStream, ContentStreams.Identifier classification) {
+
+}
@@ -0,0 +1,61 @@
+package com.knecon.fforesight.service.viewerdoc.pdf;
+
+import java.io.IOException;
+import java.io.InputStream;
+
+import org.apache.pdfbox.contentstream.PDContentStream;
+import org.apache.pdfbox.io.RandomAccessInputStream;
+import org.apache.pdfbox.io.RandomAccessRead;
+import org.apache.pdfbox.pdmodel.PDResources;
+import org.apache.pdfbox.pdmodel.common.PDRectangle;
+import org.apache.pdfbox.pdmodel.common.PDStream;
+import org.apache.pdfbox.util.Matrix;
+
+import lombok.AccessLevel;
+import lombok.Getter;
+import lombok.RequiredArgsConstructor;
+import lombok.experimental.FieldDefaults;
+
+@Getter
+@RequiredArgsConstructor
+@FieldDefaults(makeFinal = true, level = AccessLevel.PRIVATE)
+public class SinglePDContentStream implements PDContentStream {
+
+    PDStream pdStream;
+
+
+    @Override
+    public InputStream getContents() throws IOException {
+
+        return new RandomAccessInputStream(getContentsForRandomAccess());
+    }
+
+
+    @Override
+    public RandomAccessRead getContentsForRandomAccess() throws IOException {
+
+        return pdStream.getCOSObject().createView();
+    }
+
+
+    @Override
+    public PDResources getResources() {
+
+        return null;
+    }
+
+
+    @Override
+    public PDRectangle getBBox() {
+
+        return null;
+    }
+
+
+    @Override
+    public Matrix getMatrix() {
+
+        return null;
+    }
+
+}
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Maverick Studer	18a28e82d0	RED-8550: Faulty table recognition and text duplication leads to huge sections * cherrypick	2024-02-21 14:19:48 +01:00
Kilian Schüttler	3c9049dc8a	Merge branch 'RED-8156' into 'main' RED-8156: refactor ViewerDocumentService as a dependency for ocr-service See merge request fforesight/layout-parser!99	2024-02-07 10:42:21 +01:00
Kilian Schuettler	015984891f	RED-8156: refactor ViewerDocumentService as a dependency for ocr-service * fix pmd	2024-02-06 17:17:26 +01:00
Kilian Schuettler	66fcb62833	RED-8156: refactor ViewerDocumentService as a dependency for ocr-service * fix pmd	2024-02-06 17:09:21 +01:00
Kilian Schuettler	48824f56a8	RED-8156: refactor ViewerDocumentService as a dependency for ocr-service * fix pmd	2024-02-06 17:06:53 +01:00
Kilian Schuettler	785628537f	RED-8156: refactor ViewerDocumentService as a dependency for ocr-service * various improvements to experimental parsing steps * added embed fonts functionality to viewer doc * fix checkstyle	2024-02-06 17:03:38 +01:00
Kilian Schuettler	23eb0c40a3	RED-8156: refactor ViewerDocumentService as a dependency for ocr-service * various improvements to experimental parsing steps * added embed fonts functionality to viewer doc	2024-02-06 16:59:51 +01:00
Dominique Eifländer	1b4aaf4454	Merge branch 'RED-8171' into 'main' RED-8171: Traces do not stop at @Async See merge request fforesight/layout-parser!98	2024-02-02 13:34:44 +01:00
Dominique Eifländer	e4f3557b36	RED-8171: Traces do not stop at @Async	2024-02-02 13:22:57 +01:00
Timo Bejan	9be3c86297	Merge branch 'RED-8085' into 'main' Red 8085 See merge request fforesight/layout-parser!96	2024-01-29 10:31:36 +01:00
Timo Bejan	88855de2da	Red 8085	2024-01-29 10:31:36 +01:00
Dominique Eifländer	368a75e985	Merge branch 'RED-8106' into 'main' RED-8106: Make documentdata serializable See merge request fforesight/layout-parser!95	2023-12-22 13:33:02 +01:00
Dominique Eifländer	12344d57b2	RED-8106: Make documentdata serializable	2023-12-21 13:42:25 +01:00
Dominique Eifländer	9e854379e7	Merge branch 'RED-1137' into 'main' RED-1137: Do not observe actuator endpoints See merge request fforesight/layout-parser!92	2023-12-20 14:11:31 +01:00
Dominique Eifländer	b779c72041	RED-1137: Do not observe actuator endpoints	2023-12-20 14:05:00 +01:00
Dominique Eifländer	760a809900	Merge branch 'RED-7384' into 'main' RED-7384: fixes for migration See merge request fforesight/layout-parser!91	2023-12-20 12:40:00 +01:00
Kilian Schüttler	ba1c7c07ab	RED-7384: fixes for migration	2023-12-20 12:40:00 +01:00
Dominique Eifländer	ca0cbbcb49	Merge branch 'RED-5223' into 'main' RED-5223: Use tracing-commons from fforesight See merge request fforesight/layout-parser!90	2023-12-13 16:05:01 +01:00
Dominique Eifländer	da2cdc288e	RED-5223: Use tracing-commons from fforesight	2023-12-13 15:31:26 +01:00
Dominique Eifländer	68da328889	Merge branch 'queueHotfix' into 'main' hotfix: removed dlq from response queue to be equal to persistence-service See merge request fforesight/layout-parser!89	2023-12-13 09:54:43 +01:00
Dominique Eifländer	711548d1a7	hotfix: removed dlq from response queue to be equal to persistence-service	2023-12-13 09:47:27 +01:00
Dominique Eifländer	2bddcdafee	Merge branch 'RED-5223' into 'main' RED-5223: Enabled tracing, upgrade spring, use logstash-logback-encoder for json logs See merge request fforesight/layout-parser!88	2023-12-11 15:13:22 +01:00
Dominique Eifländer	750ccf4ce2	RED-5223: Enabled tracing, upgrade spring, use logstash-logback-encoder for json logs	2023-12-11 15:06:23 +01:00
Ali Oezyetimoglu	57b5d3f48e	Merge branch 'RED-7714' into 'main' RED-7715 - Add log4j config to enable switching between json/line logs See merge request fforesight/layout-parser!87	2023-12-06 13:00:28 +01:00
Andrei Isvoran	d8c9659469	RED-7715 - Add log4j config to enable switching between json/line logs	2023-12-06 11:59:42 +02:00
Kilian Schüttler	30f060e36c	Merge branch 'enable-caching' into 'main' enable caching in build See merge request fforesight/layout-parser!84	2023-11-24 10:36:55 +01:00
Kilian Schuettler	53a5824e6c	enable caching in build	2023-11-24 10:24:50 +01:00
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