RO-Crate exporter PoC

pom.xml

@@ -21,7 +21,7 @@
         <skipUnitTests>false</skipUnitTests>
         <skipIntegrationTests>false</skipIntegrationTests>
         <it.groups>integration</it.groups>
-        
+
         <!-- By default, this module will produce a WAR file. -->
         <!-- This will be switched within the container profile! -->
         <packaging.type>war</packaging.type>
@@ -225,7 +225,7 @@
             <artifactId>jakarta.json</artifactId>
             <scope>provided</scope>
         </dependency>
-        
+
         <!-- JSON-B -->
         <dependency>
             <groupId>jakarta.json.bind</groupId>
@@ -237,7 +237,7 @@
             <artifactId>yasson</artifactId>
             <scope>test</scope>
         </dependency>
-        
+
         <!-- Jakarta Faces & related -->
         <dependency>
             <groupId>org.glassfish</groupId>
@@ -450,7 +450,7 @@
             <artifactId>caffeine</artifactId>
             <version>3.1.8</version>
         </dependency>
-        
+
         <!-- New and Improved GDCC XOAI library! -->
         <dependency>
             <groupId>io.gdcc</groupId>
@@ -660,7 +660,7 @@
             <version>4.0.4</version>
             <scope>test</scope>
         </dependency>
-        
+
         <dependency>
             <groupId>org.mockito</groupId>
             <artifactId>mockito-core</artifactId>
@@ -690,6 +690,12 @@
             <artifactId>hazelcast</artifactId>
             <scope>test</scope>
         </dependency>
+        <!-- https://mvnrepository.com/artifact/edu.kit.datamanager/ro-crate-java -->
+        <dependency>
+            <groupId>edu.kit.datamanager</groupId>
+            <artifactId>ro-crate-java</artifactId>
+            <version>1.1.0-rc.1</version>
+        </dependency>
     </dependencies>
     <build>
         <testResources>

src/main/java/edu/harvard/iq/dataverse/export/RoCrateExporter.java

@@ -0,0 +1,98 @@
+package edu.harvard.iq.dataverse.export;
+
+import com.google.auto.service.AutoService;
+import edu.harvard.iq.dataverse.DatasetServiceBean;
+import edu.harvard.iq.dataverse.rocrate.RoCrateManager;
+import edu.harvard.iq.dataverse.util.BundleUtil;
+import io.gdcc.spi.export.ExportDataProvider;
+import io.gdcc.spi.export.ExportException;
+import io.gdcc.spi.export.Exporter;
+import jakarta.ws.rs.core.MediaType;
+
+import java.io.FileInputStream;
+import java.io.OutputStream;
+import java.nio.file.Files;
+import java.nio.file.Paths;
+import java.util.Locale;
+import java.util.Optional;
+import jakarta.enterprise.inject.spi.CDI;
+
+/**
+ * Exports the dataset metadata as an RO-Crate JSON using the metadatablocks of the dataset as the schema
+ * of the RO-Crate. It means, that all data that is present in the dataset in Dataverse is exported as is without
+ * mapping it to the Schema.org vocabulary, which is the default schema for RO-Crates.
+ *
+ * This class has been extracted from the ARP project (https://science-research-data.hu/en) in the frame of
+ * FAIR-IMPACT's 1st Open Call "Enabling FAIR Signposting and RO-Crate for content/metadata discovery and consumption".
+ *
+ * @author Balázs E. Pataki <balazs.pataki@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @author Norbert Finta <norbert.finta@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @version 1.0
+ */
+@AutoService(Exporter.class)
+public class RoCrateExporter implements Exporter {
+
+    @Override
+    public void exportDataset(ExportDataProvider exportDataProvider, OutputStream outputStream) throws ExportException {
+        // The exporter is not loaded as an EJB by the ExporterService but is initialized directly, which means
+        // we cannot inject EJB-s with the usual @EJB annotation. Instead, we look them up at runtime
+        var roCrateManager = CDI.current().select(RoCrateManager.class).get();
+        var datasetService = CDI.current().select(DatasetServiceBean.class).get();
+
+        // exportDataProvider doesn't provide the Dataset object only the JSON representation, so we get this to
+        // read the ID of the Dataset and then load it via the datasetService
+        var json = exportDataProvider.getDatasetJson();
+        var ds = datasetService.find(Long.valueOf(json.get("id").toString()));
+
+        // Get the version specified in the json. Note: it is actually always the latest version, one cannot
+        // export metadata of older versions.
+        var versionId = Long.valueOf(json.getJsonObject("datasetVersion").get("id").toString());
+        var version = ds.getVersionFromId(versionId);
+
+        // Now we can create the RO-Crate using the roCrateManager
+        String roCratePath = roCrateManager.getRoCratePath(version);
+        if (!Files.exists(Paths.get(roCratePath))) {
+            try {
+                roCrateManager.createOrUpdateRoCrate(version);
+            } catch (Exception e) {
+                throw new ExportException(e.getMessage());
+            }
+        }
+
+        try (FileInputStream fis = new FileInputStream(roCratePath)) {
+            byte[] buffer = new byte[1024];
+            int bytesRead;
+            while ((bytesRead = fis.read(buffer)) != -1) {
+                outputStream.write(buffer, 0, bytesRead);
+            }
+        } catch (Exception e) {
+            throw new ExportException(e.getMessage());
+        }
+    }
+
+    @Override
+    public String getFormatName() {
+        return "rocrate";
+    }
+
+    @Override
+    public String getDisplayName(Locale locale) {
+        String displayName = BundleUtil.getStringFromBundle("dataset.exportBtn.itemLabel.rocrate",locale);
+        return Optional.ofNullable(displayName).orElse("RO-Crate");
+    }
+
+    @Override
+    public Boolean isHarvestable() {
+        return true;
+    }
+
+    @Override
+    public Boolean isAvailableToUsers() {
+        return true;
+    }
+
+    @Override
+    public String getMediaType() {
+        return MediaType.APPLICATION_JSON;
+    }
+}

src/main/java/edu/harvard/iq/dataverse/rocrate/DefaultRoCrateConformsToIdProvider.java

@@ -0,0 +1,45 @@
+package edu.harvard.iq.dataverse.rocrate;
+
+import edu.harvard.iq.dataverse.Dataset;
+import edu.harvard.iq.dataverse.MetadataBlock;
+import edu.harvard.iq.dataverse.MetadataBlockServiceBean;
+import edu.kit.datamanager.ro_crate.entities.data.RootDataEntity;
+import jakarta.ejb.EJB;
+import jakarta.ejb.Stateless;
+import jakarta.inject.Named;
+
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * Manages RO-Crate conformsTo values based on Dataverse metadatablocks.
+ *
+ * This class has been extracted from the ARP project (https://science-research-data.hu/en) in the frame of
+ * FAIR-IMPACT's 1st Open Call "Enabling FAIR Signposting and RO-Crate for content/metadata discovery and consumption".
+ *
+ * @author Balázs E. Pataki <balazs.pataki@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @author Norbert Finta <norbert.finta@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @version 1.0
+ */
+@Stateless
+@Named
+public class DefaultRoCrateConformsToIdProvider implements RoCrateConformsToIdProvider {
+    @EJB
+    protected MetadataBlockServiceBean metadataBlockSvc;
+
+    private static String ID_PREFIX = "https://dataverse.org/metadatablocks/";
+
+    @Override
+    public List<String> generateConformsToIds(Dataset dataset, RootDataEntity rootDataEntity) {
+        return dataset.getOwner().getMetadataBlocks().stream()
+                .map(metadataBlock -> ID_PREFIX+metadataBlock.getName())
+                .collect(Collectors.toList());
+    }
+
+    @Override
+    public List<MetadataBlock> findMetadataBlockForConformsToIds(List<String> ids) {
+        return ids.stream()
+                .map(id -> metadataBlockSvc.findByName(id.substring(ID_PREFIX.length())))
+                .collect(Collectors.toList());
+    }
+}

src/main/java/edu/harvard/iq/dataverse/rocrate/DefaultRoCrateNameProvider.java

@@ -0,0 +1,33 @@
+package edu.harvard.iq.dataverse.rocrate;
+
+import edu.harvard.iq.dataverse.DatasetFieldCompoundValue;
+import jakarta.ejb.Stateless;
+import jakarta.inject.Named;
+
+import java.util.stream.Collectors;
+
+/**
+ * Helps to generate a value for the "name" field of RO-Crate objects for a Dataverse compound field. This
+ * RoCrateNameProvider implementation genetates a string value similar to the Metadata tab of the Dataverse dataset
+ * UI.
+ *
+ * This class has been extracted from the ARP project (https://science-research-data.hu/en) in the frame of
+ * FAIR-IMPACT's 1st Open Call "Enabling FAIR Signposting and RO-Crate for content/metadata discovery and consumption".
+ *
+ * @author Balázs E. Pataki <balazs.pataki@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @author Norbert Finta <norbert.finta@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @version 1.0
+ */
+@Stateless
+@Named
+public class DefaultRoCrateNameProvider implements RoCrateNameProvider {
+    @Override
+    public String generateRoCrateName(DatasetFieldCompoundValue compoundValue) {
+        var nameFieldValue = compoundValue.getDisplayValueMap().entrySet().stream()
+                .map(o -> o.getValue())
+                .collect(Collectors.joining(" "));
+        nameFieldValue = (nameFieldValue.length() > 80) ? nameFieldValue.substring(0, 77) + "..." : nameFieldValue;
+
+        return nameFieldValue;
+    }
+}

src/main/java/edu/harvard/iq/dataverse/rocrate/RoCrateConformsToIdProvider.java

@@ -0,0 +1,24 @@
+package edu.harvard.iq.dataverse.rocrate;
+
+import edu.harvard.iq.dataverse.Dataset;
+import edu.harvard.iq.dataverse.MetadataBlock;
+import edu.kit.datamanager.ro_crate.entities.data.RootDataEntity;
+
+import java.util.List;
+
+/**
+ * Manages RO-Crate conformsTo values.
+ *
+ * This class has been extracted from the ARP project (https://science-research-data.hu/en) in the frame of
+ * FAIR-IMPACT's 1st Open Call "Enabling FAIR Signposting and RO-Crate for content/metadata discovery and consumption".
+ *
+ * @author Balázs E. Pataki <balazs.pataki@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @author Norbert Finta <norbert.finta@sztaki.hu>, SZTAKI, Department of Distributed Systems, https://dsd.sztaki.hu
+ * @version 1.0
+ */
+
+public interface RoCrateConformsToIdProvider {
+    List<String> generateConformsToIds(Dataset dataset, RootDataEntity rootDataEntity);
+
+    List<MetadataBlock> findMetadataBlockForConformsToIds(List<String> ids);
+}