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

* cherrypick

config/pmd/pmd.xml

@@ -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>
+

config/pmd/test_pmd.xml

@@ -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>

layoutparser-service/layoutparser-service-internal-api/build.gradle.kts

@@ -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"

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentData.java

@@ -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;

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentPage.java

@@ -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;

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentPositionData.java

@@ -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;

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentStructure.java

@@ -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;

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/DocumentTextData.java

@@ -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;

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/redaction/NodeType.java

@@ -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,

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

@@ -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"
@@ -10,12 +10,13 @@ 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:0.19.0")
+    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}")

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

@@ -51,7 +51,7 @@ 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;
@@ -84,7 +84,7 @@ public class LayoutParsingPipeline {
     TaasBlockificationService taasBlockificationService;
     DocuMineBlockificationService docuMineBlockificationService;
     RedactManagerBlockificationService redactManagerBlockificationService;
-    ViewerDocumentService viewerDocumentService;
+    LayoutGridService layoutGridService;
     ObservationRegistry observationRegistry;
 
 
@@ -94,16 +94,21 @@ public class LayoutParsingPipeline {
         log.info("Starting layout parsing for {}", layoutParsingRequest.identifier());
 
         File originFile = layoutParsingStorageService.getOriginFile(layoutParsingRequest.originFileStorageId());
-        File viewerDocumentFile = File.createTempFile("viewer_document", ".pdf");
+        File viewerDocumentFile = layoutParsingStorageService.getViewerDocFile(layoutParsingRequest.viewerDocumentStorageId())
+                .orElse(originFile);
 
         ImageServiceResponse imageServiceResponse = new ImageServiceResponse();
-        if (layoutParsingRequest.imagesFileStorageId().isPresent()) {
-            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId().get());
+        if (layoutParsingRequest.imagesFileStorageId()
+                .isPresent()) {
+            imageServiceResponse = layoutParsingStorageService.getImagesFile(layoutParsingRequest.imagesFileStorageId()
+                                                                                     .get());
         }
 
         TableServiceResponse tableServiceResponse = new TableServiceResponse();
-        if (layoutParsingRequest.tablesFileStorageId().isPresent()) {
-            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId().get());
+        if (layoutParsingRequest.tablesFileStorageId()
+                .isPresent()) {
+            tableServiceResponse = layoutParsingStorageService.getTablesFile(layoutParsingRequest.tablesFileStorageId()
+                                                                                     .get());
         }
 
         ClassificationDocument classificationDocument = parseLayout(layoutParsingRequest.layoutParsingType(),
@@ -111,50 +116,56 @@ public class LayoutParsingPipeline {
                 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));
-
-        log.info("Creating viewer document for {}", layoutParsingRequest.identifier());
-        viewerDocumentService.createViewerDocument(originFile, documentGraph, viewerDocumentFile, false);
         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();
-        viewerDocumentFile.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()))
+                                        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();
 
     }
@@ -164,9 +175,9 @@ public class LayoutParsingPipeline {
 
         AtomicReference<Document> documentReference = new AtomicReference<>();
 
-        Observation.createNotStarted("LayoutParsingPipeline", observationRegistry).contextualName("build-document-graph").observe(() -> {
-            documentReference.set(DocumentGraphFactory.buildDocumentGraph(classificationDocument));
-        });
+        Observation.createNotStarted("LayoutParsingPipeline", observationRegistry)
+                .contextualName("build-document-graph")
+                .observe(() -> documentReference.set(DocumentGraphFactory.buildDocumentGraph(classificationDocument)));
 
         return documentReference.get();
     }
@@ -175,14 +186,14 @@ public class LayoutParsingPipeline {
     private String buildSemanticNodeCountMessage(int numberOfPages, Map<NodeType, Long> semanticNodeCounts) {
 
         return String.format("%d pages with %d sections, %d headlines, %d paragraphs, %d tables with %d cells, %d headers, and %d footers parsed",
-                numberOfPages,
-                semanticNodeCounts.get(NodeType.SECTION) == null ? 0 : semanticNodeCounts.get(NodeType.SECTION),
-                semanticNodeCounts.get(NodeType.HEADLINE) == null ? 0 : semanticNodeCounts.get(NodeType.HEADLINE),
-                semanticNodeCounts.get(NodeType.PARAGRAPH) == null ? 0 : semanticNodeCounts.get(NodeType.PARAGRAPH),
-                semanticNodeCounts.get(NodeType.TABLE) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE),
-                semanticNodeCounts.get(NodeType.TABLE_CELL) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE_CELL),
-                semanticNodeCounts.get(NodeType.HEADER) == null ? 0 : semanticNodeCounts.get(NodeType.HEADER),
-                semanticNodeCounts.get(NodeType.FOOTER) == null ? 0 : semanticNodeCounts.get(NodeType.FOOTER));
+                             numberOfPages,
+                             semanticNodeCounts.get(NodeType.SECTION) == null ? 0 : semanticNodeCounts.get(NodeType.SECTION),
+                             semanticNodeCounts.get(NodeType.HEADLINE) == null ? 0 : semanticNodeCounts.get(NodeType.HEADLINE),
+                             semanticNodeCounts.get(NodeType.PARAGRAPH) == null ? 0 : semanticNodeCounts.get(NodeType.PARAGRAPH),
+                             semanticNodeCounts.get(NodeType.TABLE) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE),
+                             semanticNodeCounts.get(NodeType.TABLE_CELL) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE_CELL),
+                             semanticNodeCounts.get(NodeType.HEADER) == null ? 0 : semanticNodeCounts.get(NodeType.HEADER),
+                             semanticNodeCounts.get(NodeType.FOOTER) == null ? 0 : semanticNodeCounts.get(NodeType.FOOTER));
     }
 
 
@@ -313,9 +324,7 @@ public class LayoutParsingPipeline {
 
     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());

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

@@ -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);
+    }
+
 }

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

@@ -8,6 +8,7 @@ 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.pdfbox.Loader;
 import org.apache.pdfbox.pdmodel.PDDocument;
@@ -36,13 +37,6 @@ public class LayoutParsingStorageService {
     private final StorageService storageService;
     private final ObjectMapper objectMapper;
 
-
-    public PDDocument getOriginDocument(String storageId) throws IOException {
-
-        return Loader.loadPDF(getOriginFile(storageId));
-    }
-
-
     @Observed(name = "LayoutParsingStorageService", contextualName = "get-origin-file")
     public File getOriginFile(String storageId) throws IOException {
 
@@ -52,6 +46,18 @@ public class LayoutParsingStorageService {
     }
 
 
+    @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);
+    }
+
+
     public ImageServiceResponse getImagesFile(String storageId) throws IOException {
 
         try (InputStream inputStream = getObject(storageId)) {
@@ -137,7 +143,6 @@ public class LayoutParsingStorageService {
     public void storeViewerDocument(LayoutParsingRequest layoutParsingRequest, File out) {
 
         try (var in = new FileInputStream(out)) {
-
             storageService.storeObject(TenantContext.getTenantId(), layoutParsingRequest.viewerDocumentStorageId(), in);
         }
     }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/graph/Boundary.java

@@ -12,6 +12,7 @@ import lombok.Setter;
 
 @Setter
 @EqualsAndHashCode
+@SuppressWarnings("PMD.AvoidFieldNameMatchingMethodName")
 public class Boundary implements Comparable<Boundary> {
 
     private int start;

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/table/Cell.java

@@ -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);
+    }
+
 }

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/table/TablePageBlock.java

@@ -1,14 +1,12 @@
 package com.knecon.fforesight.service.layoutparser.processor.model.table;
 
 import java.awt.geom.Point2D;
-import java.awt.geom.Rectangle2D;
 import java.util.ArrayList;
 import java.util.Collections;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
 import java.util.TreeMap;
-import java.util.stream.Collectors;
 
 import com.knecon.fforesight.service.layoutparser.processor.model.AbstractPageBlock;
 import com.knecon.fforesight.service.layoutparser.processor.model.PageBlockType;
@@ -21,7 +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
@@ -30,10 +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();
@@ -50,6 +52,7 @@ public class TablePageBlock extends AbstractPageBlock {
         return getColCount() == 0 || getRowCount() == 0;
     }
 
+
     public List<List<Cell>> getRows() {
 
         if (rows == null) {
@@ -80,7 +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);
 
     }
 
@@ -120,7 +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);
                     }
@@ -135,7 +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);
                 }
             }
@@ -151,7 +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);
                     }
@@ -162,7 +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);
                     }
@@ -173,7 +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);
                     }
@@ -187,17 +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()) {
@@ -206,11 +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);
             }
         }
 
@@ -221,29 +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<>();
@@ -254,42 +259,81 @@ 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 -> intersects(cell, c)).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);
 
-        return matrix;
-    }
-
-
-
-    public boolean intersects(Cell cell1, Cell cell2) {
-        if (cell1.getHeight() <= 0 || cell2.getHeight() <= 0) {
-            return false;
+        // 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<>());
         }
-        double x0 = cell1.getX() + 2;
-        double y0 = cell1.getY() + 2;
-        return (cell2.x + cell2.width > x0 &&
-                cell2.y + cell2.height > y0 &&
-                cell2.x < x0 + cell1.getWidth() -2 &&
-                cell2.y < y0 + cell1.getHeight() -2);
+
+        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);
+
+    }
+
 
     @Override
     public String getText() {
@@ -314,7 +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;
                         }
                     }
@@ -328,8 +372,6 @@ public class TablePageBlock extends AbstractPageBlock {
     }
 
 
-
-
     public String getTextAsHtml() {
 
         StringBuilder sb = new StringBuilder();

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/visualization/ColoredLine.java

@@ -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) {
-
-}

layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/visualization/PlacedText.java

@@ -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) {
-
-}

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

@@ -25,7 +25,8 @@ import lombok.extern.slf4j.Slf4j;
 @RequiredArgsConstructor
 public class RulingCleaningService {
 
-    private static final float THRESHOLD = 6;
+    private static final float THRESHOLD_Y = 6;
+    private static final float THRESHOLD_X = 2;
 
 
     public CleanRulings getCleanRulings(List<TableCells> tableCells, List<Ruling> rulings) {
@@ -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()) < THRESHOLD) {
+            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()) < THRESHOLD) {
+            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)));

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

@@ -5,7 +5,6 @@ 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;
@@ -25,55 +24,62 @@ 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) -> {
 
-    public boolean contains(Cell cell, double x, double y, double w, double h) {
+        Double cell1Size = cell1.getHeight() * cell1.getWidth();
+        Double cell2Size = cell2.getHeight() * cell2.getWidth();
+        return cell1Size.compareTo(cell2Size);
+    };
 
-        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 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);
+    };
 
 
     /**
@@ -89,22 +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) {
 
         List<Cell> cells = findCells(cleanRulings.getHorizontal(), cleanRulings.getVertical());
-
-        List<TextPageBlock> toBeRemoved = new ArrayList<>();
+        // 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() && contains(cell,
-                        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;
                 }
             }
@@ -114,39 +116,70 @@ public class TableExtractionService {
         DoubleComparisons.sort(cells, Rectangle.ILL_DEFINED_ORDER);
 
         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) {
+    private boolean doesCellContainTextBlock(Cell cell, TextPageBlock textBlock) {
+
+        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) {
@@ -160,7 +193,7 @@ public class TableExtractionService {
         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);
@@ -186,13 +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;
                     }
@@ -214,7 +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()) {
@@ -231,8 +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)) {
@@ -257,7 +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);
 
@@ -301,7 +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;

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

@@ -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)));
     }
 
 }

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

@@ -1,217 +0,0 @@
-package com.knecon.fforesight.service.layoutparser.processor.services.visualization;
-
-import java.awt.geom.AffineTransform;
-import java.awt.geom.Rectangle2D;
-import java.io.File;
-import java.io.IOException;
-import java.nio.file.Files;
-import java.nio.file.Path;
-import java.nio.file.StandardCopyOption;
-
-import org.apache.pdfbox.Loader;
-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 io.micrometer.observation.Observation;
-import io.micrometer.observation.ObservationRegistry;
-import io.micrometer.observation.annotation.Observed;
-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;
-    private final ObservationRegistry observationRegistry;
-
-
-    @SneakyThrows
-    @Observed(name = "ViewerDocumentService", contextualName = "create-viewer-document")
-    public void createViewerDocument(File originFile, Document document, File destinationFile, boolean layerVisibilityDefaultValue) {
-
-        Path tmpFile = Files.createTempFile("tmpViewerDocument", ".pdf");
-        PDDocument pdDocument = openPDDocument(originFile);
-        LayoutGrid layoutGrid = layoutGridService.createLayoutGrid(document);
-
-        PDOptionalContentGroup layer = addLayerToDocument(pdDocument, 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);
-            addLayerToPageResources(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();
-            }
-
-            if (pageNumber % 500 == 0 && pageNumber != 0) { // re-open document every once in a while to save on RAM
-                log.info("Incremental save after {} pages", pageNumber);
-                observedIncrementalSave(pdDocument, destinationFile);
-                pdDocument.close();
-                Files.copy(destinationFile.toPath(), tmpFile, StandardCopyOption.REPLACE_EXISTING);
-                pdDocument = openPDDocument(tmpFile.toFile());
-                layer = addLayerToDocument(pdDocument, layerVisibilityDefaultValue);
-            }
-
-        }
-        observedIncrementalSave(pdDocument, destinationFile);
-
-        tmpFile.toFile().delete();
-        pdDocument.close();
-    }
-
-
-    private static PDDocument openPDDocument(File tmpFile) throws IOException {
-
-        PDDocument pdDocument;
-        pdDocument = Loader.loadPDF(tmpFile);
-        pdDocument.setAllSecurityToBeRemoved(true);
-        return pdDocument;
-    }
-
-
-    @SneakyThrows
-    private void observedIncrementalSave(PDDocument pdDocument, File outputFile) {
-
-        Observation.createNotStarted("ViewerDocumentService", observationRegistry).contextualName("incremental-save").observe(() -> {
-            try {
-                pdDocument.save(outputFile);
-            } catch (IOException e) {
-                throw new RuntimeException(e);
-            }
-        });
-
-    }
-
-
-    private static void addLayerToPageResources(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, 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;
-        });
-    }
-
-}

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

@@ -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();

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

@@ -6,7 +6,7 @@ plugins {
     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"
 }
 
@@ -30,8 +30,8 @@ dependencies {
     implementation(project(":layoutparser-service-internal-api"))
 
     implementation("com.iqser.red.commons:storage-commons:2.45.0")
-    implementation("com.knecon.fforesight:tenant-commons:0.19.0")
-    implementation("com.knecon.fforesight:tracing-commons:0.3.0")
+    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")
@@ -39,6 +39,7 @@ dependencies {
     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:${springBootStarterVersion}")

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

@@ -148,6 +148,7 @@ public class HeadlinesGoldStandardIntegrationTest {
     @EqualsAndHashCode
     @AllArgsConstructor
     @ToString
+    @SuppressWarnings("PMD.AvoidFieldNameMatchingTypeName")
     private class Headline {
 
         private int page;

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

@@ -3,14 +3,19 @@ package com.knecon.fforesight.service.layoutparser.server.graph;
 import java.io.File;
 import java.nio.file.Path;
 
+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;
 
@@ -22,11 +27,42 @@ public class ViewerDocumentTest extends BuildDocumentTest {
 
         String fileName = "files/S-Metolachlor_RAR_01_Volume_1_2018-09-06.pdf";
         String tmpFileName = "/tmp/" + Path.of(fileName).getFileName() + "_VIEWER.pdf";
-        LayoutGridService layoutGridService = new LayoutGridService();
-        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(layoutGridService, null);
-        Document document = buildGraph(fileName, LayoutParsingType.REDACT_MANAGER);
+
         var documentFile = new ClassPathResource(fileName).getFile();
-        viewerDocumentService.createViewerDocument(documentFile, document, new File(tmpFileName), true);
+        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);
     }
 
 }
+

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

@@ -29,8 +29,6 @@ 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;
@@ -50,12 +48,6 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     @Autowired
     private RedactManagerClassificationService redactManagerClassificationService;
 
-    @Autowired
-    private CvTableParsingAdapter cvTableParsingAdapter;
-
-    @Autowired
-    private ImageServiceResponseAdapter imageServiceResponseAdapter;
-
     @Autowired
     private SectionsBuilderService sectionsBuilderService;
 
@@ -64,10 +56,10 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     public ClassificationDocument buildClassificationDocument(File originDocument, TableServiceResponse tableServiceResponse) {
 
         ClassificationDocument classificationDocument = layoutParsingPipeline.parseLayout(LayoutParsingType.REDACT_MANAGER,
-                originDocument,
-                new ImageServiceResponse(),
-                tableServiceResponse,
-                "document");
+                                                                                          originDocument,
+                                                                                          new ImageServiceResponse(),
+                                                                                          tableServiceResponse,
+                                                                                          "document");
 
         redactManagerClassificationService.classifyDocument(classificationDocument);
 
@@ -87,11 +79,11 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     @Test
     public void tablesToHtmlDebugger() throws IOException {
 
-        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/A20622A izRMS (CZ) fRR Part B9_Page185.pdf");
+        ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/T5 VV-640252-Page16.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
 
-        toHtml(document, "/tmp/A20622A izRMS (CZ) fRR Part B9_Page185.html");
+        toHtml(document, "/tmp/T5.html");
 
     }
 
@@ -109,6 +101,7 @@ public class PdfSegmentationServiceTest extends AbstractTest {
     }
 
 
+    @Disabled
     @Test
     public void testScanRotationBorderIsIgnored() throws IOException {
 
@@ -117,8 +110,16 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         var tableServiceResponse = objectMapper.readValue(cvTablesResource.getInputStream(), TableServiceResponse.class);
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile(), tableServiceResponse);
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        var tables = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList();
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        var tables = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList();
 
         // Quality of the table parsing is not good, because the file is rotated at scanning.
         // We only asset that the table border is not the page border.
@@ -140,12 +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");
     }
@@ -157,11 +158,22 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Spanning Cells.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock table = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
         assertThat(table.getColCount()).isEqualTo(6);
         assertThat(table.getRowCount()).isEqualTo(13);
-        assertThat(table.getRows().stream().mapToInt(List::size).sum()).isEqualTo(6 * 13);
+        assertThat(table.getRows()
+                           .stream()
+                           .mapToInt(List::size).sum()).isEqualTo(6 * 13);
     }
 
 
@@ -171,15 +183,37 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Table.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
         assertThat(firstTable.getColCount()).isEqualTo(8);
         assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).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();
     }
 
 
@@ -189,15 +223,37 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Merge Multi Page Table.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
         assertThat(firstTable.getColCount()).isEqualTo(9);
         assertThat(firstTable.getRowCount()).isEqualTo(5);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).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();
     }
 
 
@@ -207,19 +263,41 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         ClassPathResource pdfFileResource = new ClassPathResource("files/Minimal Examples/Rotated Table Headers.pdf");
 
         ClassificationDocument document = buildClassificationDocument(pdfFileResource.getFile());
-        assertThat(document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList())).isNotEmpty();
-        TablePageBlock firstTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).collect(Collectors.toList()).get(0);
+        assertThat(document.getSections()
+                           .stream()
+                           .flatMap(paragraph -> paragraph.getTables()
+                                   .stream())
+                           .collect(Collectors.toList())).isNotEmpty();
+        TablePageBlock firstTable = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(0);
         assertThat(firstTable.getColCount()).isEqualTo(8);
         assertThat(firstTable.getRowCount()).isEqualTo(1);
-        TablePageBlock secondTable = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).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();
     }
 
 
-    @Test // Non-sense test
+    @Test
     public void testDoc56Page170() throws IOException {
 
         ClassPathResource pdfFileResource = new ClassPathResource("files/SinglePages/56 Fludioxonil_RAR_12_Volume_3CA_B-7_2018-02-21_Page170.pdf");
@@ -230,8 +308,8 @@ public class PdfSegmentationServiceTest extends AbstractTest {
 
         validateTable(document, 0, 1, 1, 0, 0);
         validateTable(document, 1, 2, 2, 0, 0);
-        validateTable(document, 2, 6, 20, 0, 0);
-        validateTable(document, 3, 7, 31, 0, 0);
+        validateTable(document, 2, 4, 19, 12, 0);
+        validateTable(document, 3, 2, 12, 0, 0);
 
     }
 
@@ -265,29 +343,30 @@ public class PdfSegmentationServiceTest extends AbstractTest {
         validateTable(document, 0, 8, 8, 0, 0);
 
         List<List<String>> values = Arrays.asList(Arrays.asList("Annex point Reference within DAR/RAR",
-                        "Author, date",
-                        "Study title",
-                        "Analytical method Author, date, No.",
-                        "Technique, LOQ of the method, validated working range",
-                        "Method meets analytical validation criteria",
-                        "Remarks (in case validation criteria are not met)",
-                        "Acceptability of the method"),
-                Arrays.asList("Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
-                        "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
-                Arrays.asList("CA 7.1.2.1.1 DAR (2009)",
-                        "Evans P.G. 2001 TMJ4569B, VV-323245",
-                        "Azoxystrobin Laboratory Degradation Study in Three Soil Types, Sampled from Holland and the United Kingdom",
-                        "Method: RAM 269 Johnson R.I., Tummon O.J., Earl M. 1995 RJ1864B, VV-377731 Johnson R.I., Tummon O.J., Earl M. 1998 RAM 269/02, VV-124072 Johnson R.I., Tummon O.J., Earl M. 2000 RAM 269/03, VV-123986 Validation: Robinson N.J. 2001 TMJ4617B, VV-895845",
-                        "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
-                        "Y",
-                        "N/A",
-                        "Y"));
+                                                                "Author, date",
+                                                                "Study title",
+                                                                "Analytical method Author, date, No.",
+                                                                "Technique, LOQ of the method, validated working range",
+                                                                "Method meets analytical validation criteria",
+                                                                "Remarks (in case validation criteria are not met)",
+                                                                "Acceptability of the method"),
+                                                  Arrays.asList(
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies",
+                                                          "Part (a) Methods in soil, water, sediment, air and any additional matrices used in support of environmental fate studies"),
+                                                  Arrays.asList("CA 7.1.2.1.1 DAR (2009)",
+                                                                "Evans P.G. 2001 TMJ4569B, VV-323245",
+                                                                "Azoxystrobin Laboratory Degradation Study in Three Soil Types, Sampled from Holland and the United Kingdom",
+                                                                "Method: RAM 269 Johnson R.I., Tummon O.J., Earl M. 1995 RJ1864B, VV-377731 Johnson R.I., Tummon O.J., Earl M. 1998 RAM 269/02, VV-124072 Johnson R.I., Tummon O.J., Earl M. 2000 RAM 269/03, VV-123986 Validation: Robinson N.J. 2001 TMJ4617B, VV-895845",
+                                                                "LC-MS/MS LOQ: 0.01 mg/kg (R401553 (SYN50165 7), R402173 (SYN501114 )) or 0.02 mg/kg (azoxystrobin, R230310, R234886) Working range: 0.02-1.0 or 0.01-0.5 mg/kg (depending on analyte) Other supporting quantificati on methods: HPLC-UV GC-MSD",
+                                                                "Y",
+                                                                "N/A",
+                                                                "Y"));
 
         validateTable(document, 0, values);
 
@@ -579,10 +658,109 @@ 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();
+        var tables = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList();
         StringBuilder sb = new StringBuilder();
 
         int currentPage = 1;
@@ -603,9 +781,19 @@ public class PdfSegmentationServiceTest extends AbstractTest {
 
     private void validateTable(ClassificationDocument document, int tableIndex, int colCount, int rowCount, int emptyCellsCountCorrect, int emptyCellsCountIncorrect) {
 
-        TablePageBlock table = document.getSections().stream().flatMap(paragraph -> paragraph.getTables().stream()).toList().get(tableIndex);
+        TablePageBlock table = document.getSections()
+                .stream()
+                .flatMap(paragraph -> paragraph.getTables()
+                        .stream())
+                .toList()
+                .get(tableIndex);
         List<List<Cell>> rows = table.getRows();
-        int emptyCellsFoundFound = rows.stream().flatMap(List::stream).toList().stream().filter(f -> f.toString().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()));
@@ -620,11 +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);
@@ -637,7 +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);
 
     }
 

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

@@ -26,8 +26,6 @@ import com.knecon.fforesight.service.layoutparser.processor.services.RulingClean
 import com.knecon.fforesight.service.layoutparser.processor.services.factory.DocumentGraphFactory;
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.DocumentDataMapper;
 import com.knecon.fforesight.service.layoutparser.processor.services.mapper.PropertiesMapper;
-import com.knecon.fforesight.service.layoutparser.processor.services.visualization.LayoutGridService;
-import com.knecon.fforesight.service.layoutparser.processor.services.visualization.ViewerDocumentService;
 import com.knecon.fforesight.service.layoutparser.server.utils.BuildDocumentTest;
 import com.knecon.fforesight.service.layoutparser.server.utils.visualizations.PdfDraw;
 
@@ -57,9 +55,6 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
     @SneakyThrows
     public void testTableExtraction() {
 
-        LayoutGridService layoutGridService = new LayoutGridService();
-        ViewerDocumentService viewerDocumentService = new ViewerDocumentService(layoutGridService, null);
-
         ClassPathResource resource = new ClassPathResource("files");
         List<String> pdfFileNames = Files.walk(resource.getFile().toPath())
                 .filter(path -> path.getFileName().toString().endsWith(".pdf"))
@@ -67,8 +62,8 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
                 .map(Path::toString)
                 .toList();
 
-        for (int i = 0; i < pdfFileNames.size(); i++) {
-            writeJsons(Path.of(pdfFileNames.get(i)));
+        for (String pdfFileName : pdfFileNames) {
+            writeJsons(Path.of(pdfFileName));
         }
     }
 
@@ -88,13 +83,13 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
                 filename.toFile().toString()));
         DocumentData documentDataBefore = DocumentDataMapper.toDocumentData(documentGraphBefore);
         DocumentData documentDataAfter = DocumentDataMapper.toDocumentData(documentGraphAfter);
-        if (!compareStructures(documentDataBefore.getDocumentStructure(), documentDataAfter.getDocumentStructure(), filename.getFileName().toString())) {
-            String tmpFileNameBefore = "C:/Users/YANNIK~1/AppData/Local/Temp/before." + filename.getFileName().toString();
+        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 = "C:/Users/YANNIK~1/AppData/Local/Temp/after." + filename.getFileName().toString();
+            String tmpFileNameAfter = "/tmp/after." + filename.getFileName().toString();
             try (PDDocument pdDocument = Loader.loadPDF(filename.toFile())) {
                 PdfDraw.drawDocumentGraph(pdDocument, documentGraphAfter);
                 pdDocument.save(tmpFileNameAfter);
@@ -105,9 +100,9 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
 
 
     @SneakyThrows
-    private boolean compareStructures(DocumentStructure structure1, DocumentStructure structure2, String pdfName) {
+    private boolean compareStructures(DocumentStructure structure1, DocumentStructure structure2) {
 
-        List listStructure1 = structure1.streamAllEntries()
+        List<Table> listStructure1 = structure1.streamAllEntries()
                 .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
                 .map(DocumentStructure.EntryData::getProperties)
                 .map(properties -> {
@@ -117,7 +112,7 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
                 })
                 .toList();
 
-        List listStructure2 = structure2.streamAllEntries()
+        List<Table> listStructure2 = structure2.streamAllEntries()
                 .filter(entryData -> entryData.getType().equals(NodeType.TABLE))
                 .map(DocumentStructure.EntryData::getProperties)
                 .map(properties -> {
@@ -128,8 +123,8 @@ public class RulingCleaningServiceTest extends BuildDocumentTest {
                 .toList();
 
         for (int i = 0; i < listStructure1.size(); i++) {
-            Table tableNode1 = (Table) listStructure1.get(i);
-            Table tableNode2 = (Table) listStructure2.get(i);
+            Table tableNode1 = listStructure1.get(i);
+            Table tableNode2 = listStructure2.get(i);
             if (tableNode1.getNumberOfRows() != tableNode2.getNumberOfRows() || tableNode1.getNumberOfCols() != tableNode2.getNumberOfCols()) {
                 return false;
             }

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

@@ -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')
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/ContentStreams.java

@@ -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) {
+
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/ColoredLine.java

@@ -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) {
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/ColoredRectangle.java

@@ -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) {
 
 }

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/EmbeddableFont.java

@@ -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);
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/FilledRectangle.java

@@ -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;

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/LayoutGrid.java

@@ -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());
         }
     }
 

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/OperatorWithArguments.java

@@ -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) {
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/PlacedText.java

@@ -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());
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/Standard14EmbeddableFont.java

@@ -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;
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/Visualizations.java

@@ -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;
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/model/VisualizationsOnPage.java

@@ -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

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/pdf/ClassifiedContentStream.java

@@ -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) {
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/pdf/SinglePDContentStream.java

@@ -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;
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/service/ContentStreamClassifier.java

@@ -0,0 +1,121 @@
+package com.knecon.fforesight.service.viewerdoc.service;
+
+import java.util.LinkedList;
+import java.util.List;
+import java.util.Optional;
+
+import org.apache.pdfbox.contentstream.PDContentStream;
+import org.apache.pdfbox.contentstream.operator.OperatorName;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.pdmodel.PDPage;
+
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.model.OperatorWithArguments;
+import com.knecon.fforesight.service.viewerdoc.pdf.ClassifiedContentStream;
+import com.knecon.fforesight.service.viewerdoc.pdf.SinglePDContentStream;
+
+import lombok.SneakyThrows;
+import lombok.experimental.UtilityClass;
+
+@UtilityClass
+public class ContentStreamClassifier {
+
+    public List<ClassifiedContentStream> getClassifiedContentStreams(PDPage page) {
+
+        List<SinglePDContentStream> streams = new LinkedList<>();
+        page.getContentStreams().forEachRemaining(stream -> streams.add(new SinglePDContentStream(stream)));
+        return ContentStreamClassifier.classifySingleContentStreams(page, streams);
+    }
+
+
+    public List<ClassifiedContentStream> classifySingleContentStreams(PDPage page, List<SinglePDContentStream> streams) {
+
+        return streams.stream().map(singlePDContentStream -> classifySingleContentStream(page, singlePDContentStream)).toList();
+    }
+
+
+    private ClassifiedContentStream classifySingleContentStream(PDPage page, SinglePDContentStream singlePDContentStream) {
+
+        ContentStreams.Identifier classification = classifyContentStream(singlePDContentStream, page);
+        return new ClassifiedContentStream(singlePDContentStream, classification);
+    }
+
+
+    /**
+     * We assume all of our layers are written escaped, so only unknown content streams need to be escaped.
+     *
+     * @param classifiers List of all content streams of a page with their classification
+     * @return false, if any content stream with classification other is not prefixed with an ESCAPE_START and suffixed with an ESCAPE_END
+     */
+    public boolean areAllContentStreamsEscaped(List<ClassifiedContentStream> classifiers) {
+
+        int escapeDepth = 0;
+        for (ClassifiedContentStream classifier : classifiers) {
+            if (classifier.classification().equals(ContentStreams.OTHER) && escapeDepth == 0) {
+                return false;
+            }
+            if (classifier.classification().equals(ContentStreams.ESCAPE_START)) {
+                escapeDepth++;
+            }
+            if (classifier.classification().equals(ContentStreams.ESCAPE_END)) {
+                escapeDepth--;
+            }
+        }
+        return escapeDepth == 0;
+    }
+
+
+    @SneakyThrows
+    public ContentStreams.Identifier classifyContentStream(PDContentStream contentStream, PDPage page) {
+
+        List<OperatorWithArguments> operatorsWithArguments = ContentStreamUtility.parseLeadingOperators(contentStream, 2);
+        if (operatorsWithArguments.isEmpty()) {
+            return ContentStreams.OTHER;
+        }
+        OperatorWithArguments firstOperator = operatorsWithArguments.get(0);
+
+        // If we wrap the content streams we append and prepend a content stream with exactly one operator "q" or "Q".
+        if (operatorsWithArguments.size() == 1) {
+            if (firstOperator.operator().getName().equals(OperatorName.SAVE)) {
+                return ContentStreams.ESCAPE_START;
+            }
+            if (firstOperator.operator().getName().equals(OperatorName.RESTORE)) {
+                return ContentStreams.ESCAPE_END;
+            }
+        }
+
+        // In previous versions we did not set a marked content with an explicit name. Instead, we wrote an optional content group (OCG) with the name "Layout grid".
+        // This OCG is then assigned a COSName by PDFBox. Usually its "oc1".
+        // Thus, in order to find this name we need to look in the page resources to find the COSName assigned to the OCG.
+        // This COSName can then be found as an argument for the first operator in the content stream.
+        if (firstOperator.operator().getName().equals(OperatorName.BEGIN_MARKED_CONTENT_SEQ)) {
+            Optional<COSName> layoutGridOCGName = ContentStreamUtility.findLayoutGridOCGName(page);
+            if (layoutGridOCGName.isPresent()) {
+                if (arumentsContainLayoutGridOCG(firstOperator, layoutGridOCGName.get())) {
+                    return ContentStreams.KNECON_LAYOUT;
+                }
+            }
+        }
+
+        if (!firstOperator.operator().getName().equals(OperatorName.BEGIN_MARKED_CONTENT)) {
+            return ContentStreams.OTHER;
+        }
+
+        Optional<COSName> firstCOSNameFromArguments = firstOperator.arguments().stream().filter(c -> c instanceof COSName).map(c -> (COSName) c).findFirst();
+
+        if (firstCOSNameFromArguments.isEmpty()) {
+            return ContentStreams.OTHER;
+        }
+
+        var cosName = firstCOSNameFromArguments.get();
+
+        return ContentStreams.allContentStreams.stream().filter(identifier -> identifier.cosName().equals(cosName)).findAny().orElse(ContentStreams.OTHER);
+    }
+
+
+    private static boolean arumentsContainLayoutGridOCG(OperatorWithArguments operator, COSName layoutGridOCGName) {
+
+        return operator.arguments().stream().filter(c -> c instanceof COSName).map(c -> (COSName) c).anyMatch(cosName -> cosName.equals(layoutGridOCGName));
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/service/ContentStreamUtility.java

@@ -0,0 +1,78 @@
+package com.knecon.fforesight.service.viewerdoc.service;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.Optional;
+import java.util.Set;
+
+import org.apache.pdfbox.contentstream.PDContentStream;
+import org.apache.pdfbox.contentstream.operator.Operator;
+import org.apache.pdfbox.cos.COSBase;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.cos.COSString;
+import org.apache.pdfbox.pdfparser.PDFStreamParser;
+import org.apache.pdfbox.pdmodel.PDPage;
+import org.apache.pdfbox.pdmodel.common.PDStream;
+
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.model.OperatorWithArguments;
+import com.knecon.fforesight.service.viewerdoc.pdf.ClassifiedContentStream;
+import com.knecon.fforesight.service.viewerdoc.pdf.SinglePDContentStream;
+
+import lombok.experimental.UtilityClass;
+
+@UtilityClass
+public class ContentStreamUtility {
+
+    public static List<OperatorWithArguments> parseLeadingOperators(PDContentStream contentStream,
+                                                                    int numberOfOperatorsToRead) throws IOException {
+
+        List<COSBase> arguments = new ArrayList<>();
+        PDFStreamParser parser = new PDFStreamParser(contentStream);
+        List<OperatorWithArguments> operatorsWithArguments = new LinkedList<>();
+        for (int i = 0; i < numberOfOperatorsToRead; ) {
+            Object token = parser.parseNextToken();
+            if (token == null) {
+                break;
+            }
+            if (token instanceof Operator operator) {
+                operatorsWithArguments.add(new OperatorWithArguments(operator, arguments));
+                arguments = new ArrayList<>();
+                i++;
+            } else {
+                arguments.add((COSBase) token);
+            }
+
+        }
+        return operatorsWithArguments;
+    }
+
+
+    public static Optional<COSName> findLayoutGridOCGName(PDPage page) {
+
+        Optional<COSName> layoutGridOCGName = Optional.empty();
+        var resourceIterator = page.getResources().getPropertiesNames();
+        for (COSName cosName : resourceIterator) {
+            COSBase cosBase = page.getResources().getProperties(cosName).getCOSObject().getDictionaryObject(COSName.NAME);
+            if (cosBase instanceof COSString string) {
+                if (ContentStreams.KNECON_LAYOUT.name().equals(string.getString())) {
+                    layoutGridOCGName = Optional.of(cosName);
+                }
+            }
+        }
+        return layoutGridOCGName;
+    }
+
+
+    public static List<PDStream> removeLayerFromContentStreams(Set<ContentStreams.Identifier> layers, List<ClassifiedContentStream> classifiers) {
+
+        return classifiers.stream()
+                .filter(classifiedContentStream -> !layers.contains(classifiedContentStream.classification()))
+                .map(ClassifiedContentStream::contentStream)
+                .map(SinglePDContentStream::getPdStream)
+                .toList();
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/main/java/com/knecon/fforesight/service/viewerdoc/service/ViewerDocumentService.java

@@ -0,0 +1,316 @@
+package com.knecon.fforesight.service.viewerdoc.service;
+
+import java.awt.geom.AffineTransform;
+import java.awt.geom.Rectangle2D;
+import java.io.File;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.nio.file.Files;
+import java.nio.file.Path;
+import java.nio.file.StandardCopyOption;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Optional;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+import org.apache.pdfbox.Loader;
+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.graphics.optionalcontent.PDOptionalContentGroup;
+import org.apache.pdfbox.pdmodel.graphics.optionalcontent.PDOptionalContentProperties;
+import org.apache.pdfbox.pdmodel.graphics.state.PDExtendedGraphicsState;
+import org.apache.pdfbox.pdmodel.graphics.state.RenderingMode;
+import org.apache.pdfbox.util.Matrix;
+
+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.FilledRectangle;
+import com.knecon.fforesight.service.viewerdoc.model.PlacedText;
+import com.knecon.fforesight.service.viewerdoc.model.Visualizations;
+import com.knecon.fforesight.service.viewerdoc.model.VisualizationsOnPage;
+import com.knecon.fforesight.service.viewerdoc.pdf.ClassifiedContentStream;
+
+import io.micrometer.observation.Observation;
+import io.micrometer.observation.ObservationRegistry;
+import io.micrometer.observation.annotation.Observed;
+import lombok.RequiredArgsConstructor;
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;
+
+@Slf4j
+@RequiredArgsConstructor
+public class ViewerDocumentService {
+
+    private final ObservationRegistry registry;
+
+
+    public void addVisualizationsOnPage(File originFile, File destinationFile, Visualizations visualizations) {
+
+        addVisualizationsOnPage(originFile, destinationFile, List.of(visualizations));
+    }
+
+
+    @Observed(name = "ViewerDocumentService", contextualName = "add-visualizations")
+    @SneakyThrows
+    public void addVisualizationsOnPage(File originFile, File destinationFile, List<Visualizations> visualizations) {
+
+        // originFile and destinationFile might be the same, so we use a temp file.
+        // Otherwise, saving the document might corrupt the file
+        Path tmpFile = Files.createTempFile("tmpViewerDocument", ".pdf");
+        Files.copy(originFile.toPath(), tmpFile, StandardCopyOption.REPLACE_EXISTING);
+
+        PDDocument pdDocument = openPDDocument(tmpFile.toFile());
+
+        enrichObservation(pdDocument, visualizations.stream().map(Visualizations::getLayer).toList());
+
+        Set<ContentStreams.Identifier> allLayers = visualizations.stream().map(Visualizations::getLayer).collect(Collectors.toUnmodifiableSet());
+
+        Map<ContentStreams.Identifier, PDOptionalContentGroup> optionalContentGroupMap = addLayersToDocument(visualizations, pdDocument);
+
+        for (int pageNumber = 0; pageNumber < pdDocument.getNumberOfPages(); pageNumber++) {
+            PDPage pdPage = pdDocument.getPage(pageNumber);
+            createPageResourcesIfNotPresent(pdPage); // needed for optionalContentGroups
+
+            List<ClassifiedContentStream> classifiers = ContentStreamClassifier.getClassifiedContentStreams(pdPage);
+
+            pdPage.setContents(ContentStreamUtility.removeLayerFromContentStreams(allLayers, classifiers));
+
+            AffineTransform textDeRotationMatrix = getTextDeRotationTransform(pdPage);
+
+            if (!ContentStreamClassifier.areAllContentStreamsEscaped(classifiers)) {
+                // 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.
+                wrapContentStreams(pdDocument, pdPage);
+            }
+
+            for (Visualizations visualization : visualizations) {
+                if (!visualization.getVisualizationsOnPages().containsKey(pageNumber)) {
+                    continue;
+                }
+                // 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(visualization.getLayer().cosName());
+
+                    if (optionalContentGroupMap.containsKey(visualization.getLayer())) {
+                        contentStream.beginMarkedContent(COSName.OC, optionalContentGroupMap.get(visualization.getLayer()));
+                    }
+
+                    contentStream.saveGraphicsState();
+
+                    drawVisualizationsToContentStream(pdDocument, visualization.getVisualizationsOnPages().get(pageNumber), contentStream, textDeRotationMatrix);
+
+                    contentStream.restoreGraphicsState();
+
+                    if (optionalContentGroupMap.containsKey(visualization.getLayer())) {
+                        contentStream.endMarkedContent();
+                    }
+
+                    contentStream.endMarkedContent();
+                }
+
+            }
+            if (pageNumber % 500 == 0 && pageNumber != 0) { // re-open document every once in a while to save on RAM
+                log.info("Incremental save after {}/{} pages", pageNumber, pdDocument.getNumberOfPages());
+                observedIncrementalSave(pdDocument, destinationFile);
+                pdDocument.close();
+                Files.copy(destinationFile.toPath(), tmpFile, StandardCopyOption.REPLACE_EXISTING);
+                pdDocument = openPDDocument(tmpFile.toFile());
+            }
+        }
+        observedIncrementalSave(pdDocument, destinationFile);
+
+        pdDocument.close();
+        assert tmpFile.toFile().delete();
+    }
+
+
+    private static Map<ContentStreams.Identifier, PDOptionalContentGroup> addLayersToDocument(List<Visualizations> visualizations, PDDocument pdDocument) {
+
+        Map<ContentStreams.Identifier, PDOptionalContentGroup> optionalContentGroupMap = new HashMap<>();
+        for (Visualizations visualization : visualizations) {
+            addLayerToDocument(visualization.getLayer(), pdDocument, visualization.isLayerVisibilityDefaultValue())//
+                    .ifPresent(ocg -> optionalContentGroupMap.put(visualization.getLayer(), ocg));
+        }
+        return optionalContentGroupMap;
+    }
+
+
+    private static void drawVisualizationsToContentStream(PDDocument pdDocument,
+                                                          VisualizationsOnPage visualizationsOnPage,
+                                                          PDPageContentStream contentStream,
+                                                          AffineTransform textDeRotationMatrix) throws IOException {
+
+        if (visualizationsOnPage.isMakePathsInvisible()) {
+            contentStream.addRect(0, 0, 1, 1);
+            contentStream.clip();
+        }
+
+        for (ColoredLine coloredLine : visualizationsOnPage.getColoredLines()) {
+            contentStream.setLineWidth(coloredLine.lineWidth());
+            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.setLineWidth(coloredRectangle.lineWidth());
+            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()) {
+            PDFont font = placedText.font().embed(pdDocument);
+            contentStream.setFont(font, placedText.fontSize());
+            contentStream.beginText();
+            contentStream.setNonStrokingColor(placedText.color());
+            if (placedText.renderingMode().isPresent()) {
+                contentStream.setRenderingMode(placedText.renderingMode().get());
+            } else {
+                contentStream.setRenderingMode(RenderingMode.FILL);
+            }
+            Matrix textMatrix = getTextMatrix(placedText, textDeRotationMatrix);
+            contentStream.setTextMatrix(textMatrix);
+            contentStream.showText(placedText.text());
+            contentStream.endText();
+        }
+    }
+
+
+    private void enrichObservation(PDDocument pdDocument, List<ContentStreams.Identifier> layers) {
+
+        if (registry == null || registry.getCurrentObservation() == null || registry.isNoop()) {
+            return;
+        }
+        registry.getCurrentObservation().highCardinalityKeyValue("numberOfPages", String.valueOf(pdDocument.getNumberOfPages()));
+        for (int i = 0; i < layers.size(); i++) {
+            ContentStreams.Identifier layer = layers.get(i);
+
+            registry.getCurrentObservation().highCardinalityKeyValue("layer_" + i, String.valueOf(layer.name()));
+        }
+    }
+
+
+    private static void wrapContentStreams(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 Matrix getTextMatrix(PlacedText placedText, AffineTransform textDeRotationMatrix) {
+
+        Matrix textMatrix;
+        if (placedText.textMatrix().isEmpty()) {
+            textMatrix = new Matrix((float) textDeRotationMatrix.getScaleX(),
+                    (float) textDeRotationMatrix.getShearX(),
+                    (float) textDeRotationMatrix.getShearY(),
+                    (float) textDeRotationMatrix.getScaleY(),
+                    (float) placedText.lineStart().getX(),
+                    (float) placedText.lineStart().getY());
+        } else {
+            textMatrix = placedText.textMatrix().get();
+        }
+        return textMatrix;
+    }
+
+
+    private static Optional<PDOptionalContentGroup> addLayerToDocument(ContentStreams.Identifier layer, PDDocument pdDocument, boolean layerVisibilityDefaultValue) {
+
+        if (layer.optionalContent()) {
+            return Optional.of(addLayerToDocument(pdDocument, layer.name(), layerVisibilityDefaultValue));
+
+        }
+        return Optional.empty();
+
+    }
+
+
+    private static PDOptionalContentGroup addLayerToDocument(PDDocument pdDocument, String layerName, 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(layerName)) {
+            layer = ocprops.getGroup(layerName);
+        } else {
+            layer = new PDOptionalContentGroup(layerName);
+            ocprops.addGroup(layer);
+        }
+        ocprops.setGroupEnabled(layer, layerVisibilityDefaultValue);
+        return layer;
+    }
+
+
+    private static PDDocument openPDDocument(File tmpFile) throws IOException {
+
+        PDDocument pdDocument;
+        pdDocument = Loader.loadPDF(tmpFile);
+        pdDocument.setAllSecurityToBeRemoved(true);
+        return pdDocument;
+    }
+
+
+    @SneakyThrows
+    private void observedIncrementalSave(PDDocument pdDocument, File outputFile) {
+
+        Observation.createNotStarted("ViewerDocumentService", registry).contextualName("incremental-save").observe(() -> {
+            try (var out = new FileOutputStream(outputFile)) {
+                pdDocument.save(out);
+            } catch (IOException e) {
+                throw new RuntimeException(e);
+            }
+        });
+    }
+
+
+    private static void createPageResourcesIfNotPresent(PDPage pdPage) {
+
+        PDResources resources = pdPage.getResources();
+        if (resources == null) {
+            resources = new PDResources();
+            pdPage.setResources(resources);
+        }
+    }
+
+
+    private static AffineTransform getTextDeRotationTransform(PDPage page) {
+
+        return AffineTransform.getQuadrantRotateInstance(switch (page.getRotation()) {
+            case 90 -> 3;
+            case 180 -> 2;
+            case 270 -> 1;
+            default -> 0;
+        });
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/test/java/com/knecon/fforesight/service/viewerdoc/service/ContentStreamClassifierTest.java

@@ -0,0 +1,124 @@
+package com.knecon.fforesight.service.viewerdoc.service;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+
+import java.io.File;
+import java.nio.file.Files;
+import java.util.List;
+import java.util.Set;
+import java.util.stream.Collectors;
+
+import org.apache.pdfbox.Loader;
+import org.apache.pdfbox.cos.COSName;
+import org.apache.pdfbox.pdmodel.PDDocument;
+import org.apache.pdfbox.pdmodel.PDPage;
+import org.junit.jupiter.api.Test;
+
+import com.knecon.fforesight.service.viewerdoc.ContentStreams;
+import com.knecon.fforesight.service.viewerdoc.pdf.ClassifiedContentStream;
+
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;
+
+@Slf4j
+public class ContentStreamClassifierTest {
+
+    @Test
+    @SneakyThrows
+    public void testClassification() {
+
+        File pdfFile = new File(Thread.currentThread().getContextClassLoader().getResource("viewerDocLayers.pdf").getFile());
+
+        try (PDDocument document = Loader.loadPDF(pdfFile)) {
+
+            PDPage page = document.getPage(0);
+
+            List<ClassifiedContentStream> classifieds = ContentStreamClassifier.getClassifiedContentStreams(page);
+
+            logContentStreamClassifications(classifieds);
+
+            assertEquals(11, classifieds.size());
+            assertEquals(ContentStreams.ESCAPE_START, classifieds.get(0).classification());
+            for (int i = 1; i < 9; i++) {
+                assertEquals(ContentStreams.OTHER, classifieds.get(i).classification());
+            }
+            assertEquals(ContentStreams.ESCAPE_END, classifieds.get(9).classification());
+            assertEquals(ContentStreams.KNECON_LAYOUT, classifieds.get(10).classification());
+            assertTrue(ContentStreamClassifier.areAllContentStreamsEscaped(classifieds));
+        }
+    }
+
+
+    @Test
+    @SneakyThrows
+    public void testRemoveLayoutLayer() {
+
+        File pdfFile = new File(Thread.currentThread().getContextClassLoader().getResource("viewerDocLayers.pdf").getFile());
+        File tmpFile = Files.createTempFile("removedLayout", ".pdf").toFile();
+
+        try (PDDocument document = Loader.loadPDF(pdfFile)) {
+
+            PDPage page = document.getPage(0);
+
+            List<ClassifiedContentStream> classifieds = ContentStreamClassifier.getClassifiedContentStreams(page);
+            page.setContents(ContentStreamUtility.removeLayerFromContentStreams(Set.of(ContentStreams.KNECON_LAYOUT), classifieds));
+
+            document.save(tmpFile);
+        }
+        try (PDDocument document2 = Loader.loadPDF(tmpFile)) {
+
+            PDPage page2 = document2.getPage(0);
+
+            List<ClassifiedContentStream> classifieds2 = ContentStreamClassifier.getClassifiedContentStreams(page2);
+
+            logContentStreamClassifications(classifieds2);
+
+            assertEquals(10, classifieds2.size());
+            assertEquals(ContentStreams.ESCAPE_START, classifieds2.get(0).classification());
+            for (int i = 1; i < 9; i++) {
+                assertEquals(ContentStreams.OTHER, classifieds2.get(i).classification());
+            }
+            assertEquals(ContentStreams.ESCAPE_END, classifieds2.get(9).classification());
+            assertTrue(ContentStreamClassifier.areAllContentStreamsEscaped(classifieds2));
+        }
+        assert tmpFile.delete();
+    }
+
+
+    @Test
+    @SneakyThrows
+    public void testClassificationForOldLayers() {
+
+        File pdfFile = new File(Thread.currentThread().getContextClassLoader().getResource("oldViewerDocLayers.pdf").getFile());
+        try (PDDocument document = Loader.loadPDF(pdfFile)) {
+
+            PDPage page = document.getPage(0);
+
+            List<ClassifiedContentStream> classifieds = ContentStreamClassifier.getClassifiedContentStreams(page);
+
+            logContentStreamClassifications(classifieds);
+
+            assertEquals(11, classifieds.size());
+            assertEquals(ContentStreams.ESCAPE_START, classifieds.get(0).classification());
+            for (int i = 1; i < 9; i++) {
+                assertEquals(ContentStreams.OTHER, classifieds.get(i).classification());
+            }
+            assertEquals(ContentStreams.ESCAPE_END, classifieds.get(9).classification());
+            assertEquals(ContentStreams.KNECON_LAYOUT, classifieds.get(10).classification());
+            assertTrue(ContentStreamClassifier.areAllContentStreamsEscaped(classifieds));
+        }
+    }
+
+
+    private static void logContentStreamClassifications(List<ClassifiedContentStream> classifieds) {
+
+        log.info("number of content streams: {}", classifieds.size());
+        log.info("Classifications: {}", classifieds.stream()//
+                .map(ClassifiedContentStream::classification)//
+                .map(ContentStreams.Identifier::cosName)//
+                .map(COSName::getName)//
+                .collect(Collectors.joining(", ")));
+    }
+
+}

layoutparser-service/viewer-doc-processor/src/test/resources/log4j2-test.xml

@@ -0,0 +1,16 @@
+<Configuration>
+
+    <Appenders>
+        <Console name="CONSOLE" target="SYSTEM_OUT">
+            <PatternLayout pattern="%d{HH:mm:ss.SSS} [%t] %-5level %logger{36} - %msg%n"/>
+        </Console>
+    </Appenders>
+
+    <Loggers>
+        <Root level="warn">
+            <AppenderRef ref="CONSOLE"/>
+        </Root>
+        <Logger name="com.knecon" level="info"/>
+    </Loggers>
+
+</Configuration>

settings.gradle.kts

@@ -1,13 +1,9 @@
-/*
- * This file was generated by the Gradle 'init' task.
- *
- * This project uses @Incubating APIs which are subject to change.
- */
-
 rootProject.name = "layoutparser"
 include(":layoutparser-service-server")
 include(":layoutparser-service-processor")
 include(":layoutparser-service-internal-api")
+include("viewer-doc-processor")
 project(":layoutparser-service-server").projectDir = file("layoutparser-service/layoutparser-service-server")
 project(":layoutparser-service-processor").projectDir = file("layoutparser-service/layoutparser-service-processor")
 project(":layoutparser-service-internal-api").projectDir = file("layoutparser-service/layoutparser-service-internal-api")
+project(":viewer-doc-processor").projectDir = file("layoutparser-service/viewer-doc-processor")