Merge branch 'CLARI-003' into 'main'

CLARI-003: add treeId to StructureObject

See merge request fforesight/layout-parser!176

layoutparser-service/layoutparser-service-internal-api/src/main/java/com/knecon/fforesight/service/layoutparser/internal/api/data/taas/StructureObject.java

@@ -1,5 +1,7 @@
 package com.knecon.fforesight.service.layoutparser.internal.api.data.taas;
 
+import java.util.List;
+
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.AllArgsConstructor;
 import lombok.Builder;
@@ -13,6 +15,8 @@ public class StructureObject {
 
     @Schema(description = "The ID of this StructureObject.")
     Integer structureObjectNumber;
+    @Schema(description = "The Tree ID of this StructureObject.")
+    List<Integer> treeId;
     @Schema(description = "This value indicates the start of the string offsets in this Object, with respect to the reading order.")
     int page;
     @Schema(description = "This stringOffset indicates the start of the string offsets in this Object, with respect to the reading order of the entire document. It is equal to the previous' StructureObject stringOffset + its length.")

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

@@ -21,7 +21,7 @@ public class LayoutParsingServiceProcessorConfiguration {
     public IViewerDocumentService viewerDocumentService(ObservationRegistry registry, LayoutparserSettings settings) {
 
         if (!Strings.isNullOrEmpty(settings.getPdftronLicense())) {
-            return new PDFTronViewerDocumentService(registry, settings.getPdftronLicense());
+            return new PDFTronViewerDocumentService(registry);
         } else {
             return new ViewerDocumentService(registry);
         }

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

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

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

@@ -37,6 +37,7 @@ dependencies {
     implementation("com.amazonaws:aws-java-sdk-s3:1.12.536")
     implementation("org.springframework.cloud:spring-cloud-starter-openfeign:4.0.4")
     implementation("net.logstash.logback:logstash-logback-encoder:7.4")
+    implementation("com.pdftron:PDFNet:10.5.0")
 
     // for integration testing only
     testImplementation(project(":viewer-doc-processor"))

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

@@ -0,0 +1,45 @@
+package com.knecon.fforesight.service.layoutparser.server;
+
+import org.springframework.stereotype.Component;
+
+import com.google.common.base.Strings;
+import com.knecon.fforesight.service.layoutparser.processor.LayoutparserSettings;
+import com.pdftron.pdf.PDFNet;
+
+import jakarta.annotation.PostConstruct;
+import jakarta.annotation.PreDestroy;
+import lombok.RequiredArgsConstructor;
+import lombok.SneakyThrows;
+import lombok.extern.slf4j.Slf4j;
+
+@Slf4j
+@Component
+@RequiredArgsConstructor
+public class PDFNetInitializer {
+
+    private final LayoutparserSettings settings;
+
+
+    @SneakyThrows
+    @PostConstruct
+    // Do not change back to application runner, if it is application runner it takes messages from the queue before PDFNet is initialized, that leads to UnsatisfiedLinkError.
+    public void init() {
+
+        if (Strings.isNullOrEmpty(settings.getPdftronLicense())) {
+            return;
+        }
+        log.info("Initializing Native Libraries");
+        log.info("Setting pdftron license: {}", settings.getPdftronLicense());
+        PDFNet.setTempPath("/tmp/pdftron");
+        PDFNet.initialize(settings.getPdftronLicense());
+
+    }
+
+
+    @PreDestroy
+    public void terminate() {
+
+        PDFNet.terminate();
+    }
+
+}

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

@@ -26,7 +26,6 @@ import com.pdftron.pdf.ElementReader;
 import com.pdftron.pdf.ElementWriter;
 import com.pdftron.pdf.Font;
 import com.pdftron.pdf.PDFDoc;
-import com.pdftron.pdf.PDFNet;
 import com.pdftron.pdf.Page;
 import com.pdftron.pdf.PageIterator;
 import com.pdftron.pdf.ocg.Group;
@@ -43,7 +42,6 @@ import lombok.extern.slf4j.Slf4j;
 public class PDFTronViewerDocumentService implements IViewerDocumentService {
 
     private final ObservationRegistry registry;
-    private final String pdftronLicense;
 
 
     @Override
@@ -51,7 +49,6 @@ public class PDFTronViewerDocumentService implements IViewerDocumentService {
     @SneakyThrows
     public synchronized void addVisualizationsOnPage(File originFile, File destinationFile, List<Visualizations> visualizations) {
 
-        PDFNet.initialize(pdftronLicense);
 
         // originFile and destinationFile might be the same, so we use a temp file.
         // Otherwise, saving the document might corrupt the file
@@ -108,7 +105,6 @@ public class PDFTronViewerDocumentService implements IViewerDocumentService {
             saveDocument(pdfDoc, destinationFile);
         }
 
-        PDFNet.terminate();
     }