Merge branch 'main' into generic-iso2dcat

pom.xml

@@ -34,7 +34,7 @@
         <dependency>
             <groupId>ch.qos.logback</groupId>
             <artifactId>logback-classic</artifactId>
-            <version>1.4.14</version>
+            <version>1.5.6</version>
         </dependency>
 
         <!-- https://mvnrepository.com/artifact/jaxen/jaxen -->
@@ -50,7 +50,7 @@
         <dependency>
             <groupId>org.junit.jupiter</groupId>
             <artifactId>junit-jupiter-api</artifactId>
-            <version>5.10.1</version>
+            <version>5.10.2</version>
             <scope>test</scope>
         </dependency>
 
@@ -58,20 +58,20 @@
         <dependency>
             <groupId>org.json</groupId>
             <artifactId>json</artifactId>
-            <version>20231013</version>
+            <version>20240303</version>
         </dependency>
 
         <!-- https://mvnrepository.com/artifact/org.apache.httpcomponents.client5/httpclient5 -->
         <dependency>
             <groupId>org.apache.httpcomponents.client5</groupId>
             <artifactId>httpclient5</artifactId>
-            <version>5.3</version>
+            <version>5.3.1</version>
         </dependency>
 
         <dependency>
             <groupId>org.mockito</groupId>
             <artifactId>mockito-core</artifactId>
-            <version>5.8.0</version>
+            <version>5.11.0</version>
             <scope>test</scope>
         </dependency>
 
@@ -84,7 +84,7 @@
         <dependency>
             <groupId>uk.org.webcompere</groupId>
             <artifactId>system-stubs-jupiter</artifactId>
-            <version>2.1.5</version>
+            <version>2.1.6</version>
             <scope>test</scope>
         </dependency>
 
@@ -92,13 +92,13 @@
         <dependency>
             <groupId>io.undertow</groupId>
             <artifactId>undertow-core</artifactId>
-            <version>2.3.10.Final</version>
+            <version>2.3.12.Final</version>
         </dependency>
 
         <dependency>
             <groupId>org.junit.jupiter</groupId>
             <artifactId>junit-jupiter-engine</artifactId>
-            <version>5.10.1</version>
+            <version>5.10.2</version>
             <scope>test</scope>
         </dependency>
         

src/main/java/de/landsh/opendata/csw2dcat/CswInterface.java

 package de.landsh.opendata.csw2dcat;
 
+import org.apache.commons.lang3.StringEscapeUtils;
 import org.apache.commons.lang3.StringUtils;
 import org.apache.commons.lang3.math.NumberUtils;
 import org.apache.hc.client5.http.classic.methods.HttpPost;
@@ -16,7 +17,6 @@ import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
 import java.io.File;
-import java.io.FileOutputStream;
 import java.io.FileWriter;
 import java.io.IOException;
 import java.nio.charset.StandardCharsets;
@@ -45,11 +45,6 @@ public class CswInterface {
     private String[] typeInclude;
     private CloseableHttpClient httpClient = HttpClients.createMinimal();
     private String dumpCswResponse = null;
-
-    public void setVerboseLogging(String verboseLogging) {
-        this.verboseLogging = verboseLogging;
-    }
-
     /**
      * Name of a directory. If set all incoming messages will be logged into files in this directory.
      */
@@ -75,6 +70,10 @@ public class CswInterface {
         typeInclude = settings.typeInclude;
     }
 
+    public void setVerboseLogging(String verboseLogging) {
+        this.verboseLogging = verboseLogging;
+    }
+
     public void setTypeExclude(String[] typeExclude) {
         this.typeExclude = typeExclude;
     }
@@ -300,7 +299,7 @@ public class CswInterface {
         filter.append("        <ogc:PropertyIsEqualTo>" +
                         "          <ogc:PropertyName>operatesOn</ogc:PropertyName>" +
                         "          <ogc:Literal>")
-                .append(id)
+                .append(StringEscapeUtils.escapeXml11(id))
                 .append("</ogc:Literal>")
                 .append("        </ogc:PropertyIsEqualTo>\n");
 

src/main/java/de/landsh/opendata/csw2dcat/MDMetadata2Dataset.java

@@ -27,6 +27,8 @@ import org.slf4j.LoggerFactory;
 import java.io.IOException;
 import java.net.MalformedURLException;
 import java.net.URL;
+import java.net.URLEncoder;
+import java.nio.charset.StandardCharsets;
 import java.time.LocalDate;
 import java.util.*;
 import java.util.stream.Collectors;
@@ -37,12 +39,14 @@ import java.util.stream.Collectors;
 public class MDMetadata2Dataset {
 
     private static final String LANGUAGE_DE = "de";
+    private static final String LANGUAGE_EN = "en";
+
     private static final Logger log = LoggerFactory.getLogger(MDMetadata2Dataset.class);
     private static final String XPATH_OPERATION_GET_CAPABILITIES = "gmd:identificationInfo/*/srv:containsOperations/srv:SV_OperationMetadata[srv:operationName/gco:CharacterString/text()='GetCapabilities']";
-    private static final Property PROPERTY_applicable_Legislation = ResourceFactory.createProperty("http://data.europa.eu/r5r/", "applicableLegislation");
     static IRIFactory factory = IRIFactory.iriImplementation();
     private final Model model;
     private final Map<String, String> keywordMapping = new HashMap<>();
+    private final Map<String, JSONObject> url2license = new HashMap<>();
     Map<String, String[]> themeMapping = new HashMap<>();
     boolean ignoreInvalidMapping = true;
     private CswInterface cswInterface;
@@ -77,6 +81,12 @@ public class MDMetadata2Dataset {
             final String value = entry.getString(key);
             keywordMapping.put(key, value);
         }
+
+        JSONArray licenses = new JSONArray(new JSONTokener(Objects.requireNonNull(getClass().getResourceAsStream("/licenses.json"))));
+        for (Object l : licenses) {
+            JSONObject license = (JSONObject) l;
+            url2license.put(license.getString("uri"), license);
+        }
     }
 
     static String getTextOrNull(Node node) {
@@ -501,10 +511,14 @@ public class MDMetadata2Dataset {
 
         final String applicationProfile = getTextOrNull(onlineResource.selectSingleNode("gmd:applicationProfile/*"));
 
-        Resource format;
+        Resource format = null;
         if (StringUtils.startsWith(applicationProfile, "http://publications.europa.eu/resource/authority/file-type/")) {
             format = ResourceFactory.createResource(applicationProfile);
-        } else {
+        } else if (StringUtils.isNotEmpty(applicationProfile)) {
+            format = Mapping.mapFormat(applicationProfile);
+        }
+
+        if (format == null) {
             // Guess the format.
             format = guessFormat(url);
         }
@@ -520,6 +534,7 @@ public class MDMetadata2Dataset {
             }  else if( url.contains("feed") && url.toLowerCase().contains("atom") && url.endsWith(".xml")) {
                 format = ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/ATOM");
             }
+
         }
 
 
@@ -533,18 +548,39 @@ public class MDMetadata2Dataset {
     }
 
     private void addLicenseInformation(Resource resource, List<Node> otherLegalConstraints) {
-        final JSONObject licenceInformation = findLicenseInformation(otherLegalConstraints);
-        if (licenceInformation != null) {
-            resource.addLiteral(DCATAPde.licenseAttributionByText, licenceInformation.getString("quelle"));
-            if (licenceInformation.has("url") && licenceInformation.getString("url").startsWith("http://dcat-ap.de/def/licenses/")) {
-                // Some publishers specify the licence as the URL and not the id.
-                resource.addProperty(DCTerms.license, model.createResource(licenceInformation.getString("url")));
+        final JSONObject licenseInformation = findLicenseInformation(otherLegalConstraints);
+
+        if (licenseInformation != null) {
+            resource.addLiteral(DCATAPde.licenseAttributionByText, licenseInformation.getString("quelle"));
+
+            if (licenseInformation.has("url")) {
+                String licenseURL = licenseInformation.getString("url");
+
+                if (licenseURL.startsWith("http://dcat-ap.de/def/licenses/")) {
+                    resource.addProperty(DCTerms.license, model.createResource(licenseURL));
                 } else {
-                resource.addProperty(DCTerms.license, model.createResource("http://dcat-ap.de/def/licenses/" + licenceInformation.getString("id")));
+                    final JSONObject license = url2license.get(licenseURL);
+
+                    if (license != null && license.has("uri")) {
+                        resource.addProperty(DCTerms.license, model.createResource(license.getString("uri")));
+                    } else {
+                        log.debug("Unknown license: {}", licenseURL);
+                        handleUnknownLicense(resource, licenseInformation.getString("id"));
+                    }
+                }
+            } else {
+                handleUnknownLicense(resource, licenseInformation.getString("id"));
             }
         }
     }
 
+    private void handleUnknownLicense(Resource resource, String licenseId) {
+        licenseId = URLEncoder.encode(licenseId, StandardCharsets.UTF_8)
+                .replaceAll("%2F", "/"); // Replace the encoded slash with the original slash
+        resource.addProperty(DCTerms.license, model.createResource("http://dcat-ap.de/def/licenses/" + licenseId));
+        log.debug("Falling back to: {}", licenseId);
+    }
+
     private Resource guessFormat(String url) {
         final String lowerCaseURL = url.toLowerCase();
         if (lowerCaseURL.contains("service=wms")) {
@@ -554,9 +590,12 @@ public class MDMetadata2Dataset {
                 log.info("WMS with output format: {}", url);
                 return null;
             }
+        } else if (lowerCaseURL.contains("service=wfs") && lowerCaseURL.contains("getcapabilities")) {
+            return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/WFS_SRVC");
         } else if (lowerCaseURL.contains("service=wfs")) {
             if (!lowerCaseURL.contains("outputformat=")) {
-                return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/WFS_SRVC");
+                // assume that the WFS returns GML als standard output format
+                return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/GML");
             } else if (lowerCaseURL.contains("outputformat=shape-zip")) {
                 return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/SHP");
             } else if (lowerCaseURL.contains("outputformat=csv")) {
@@ -565,13 +604,15 @@ public class MDMetadata2Dataset {
                 log.info("WFS with unkown output format: {}", url);
                 return null;
             }
+        } else if (lowerCaseURL.endsWith(".html")) {
+            return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/HTML");
         }
 
         return null;
     }
 
     /**
-     * Convert a {@Document} containin a gmd:MD_Metadata element into a DCAT-AP.de compliant dcat:Dataset {@link Resource}.
+     * Convert a {@link Document} containin a gmd:MD_Metadata element into a DCAT-AP.de compliant dcat:Dataset {@link Resource}.
      */
     public Resource convert(Document input) throws DocumentException, IOException {
         return convert(input.getRootElement());
@@ -612,13 +653,26 @@ public class MDMetadata2Dataset {
         dataset.addProperty(RDF.type, DCAT.Dataset);
 
         if (scopeCode != null) {
+            try {
                 dataset.addProperty(DCTerms.type, model.createResource(Mapping.mapScopeCode(scopeCode)));
+            } catch (IllegalArgumentException ignore) {
+                log.warn("Dataset {} has an invalid scopeCode.", id);
+            }
         }
 
         final String title = getTextOrNull(metadata.selectSingleNode("gmd:identificationInfo/*/gmd:citation/gmd:CI_Citation/gmd:title/gco:CharacterString"));
         if (title != null) {
+
+            if (title.contains("#locale-eng:")) {
+                // In some source systems, German and English titles are strangely combined in one text.
+                String germanTitle = StringUtils.substringBefore(title, "#locale-eng:");
+                String englishTitle = StringUtils.substringAfter(title, "#locale-eng:");
+                dataset.addLiteral(DCTerms.title, model.createLiteral(germanTitle.trim(), LANGUAGE_DE));
+                dataset.addLiteral(DCTerms.title, model.createLiteral(englishTitle.trim(), LANGUAGE_EN));
+            } else {
                 dataset.addLiteral(DCTerms.title, model.createLiteral(title.trim(), LANGUAGE_DE));
             }
+        }
 
         final String description = getTextOrNull(metadata.selectSingleNode("gmd:identificationInfo/*/gmd:abstract/gco:CharacterString"));
         if (description != null) {
@@ -643,13 +697,21 @@ public class MDMetadata2Dataset {
 
         final Element characterSet = (Element) metadata.selectSingleNode("gmd:characterSet/gmd:MD_CharacterSetCode");
         if (characterSet != null) {
+            try {
                 dataset.addLiteral(Content.characterEncoding,
                         model.createTypedLiteral(Mapping.mapCharacterSetCode(characterSet), XSDDatatype.XSDstring));
+            } catch (IllegalArgumentException ignore) {
+                log.warn("Dataset {} has an invalid characterEncoding.", id);
+            }
         }
 
         final Element maintenanceFrequencyCode = (Element) metadata.selectSingleNode("gmd:identificationInfo/*/gmd:resourceMaintenance/gmd:MD_MaintenanceInformation/gmd:maintenanceAndUpdateFrequency/gmd:MD_MaintenanceFrequencyCode");
         if (maintenanceFrequencyCode != null) {
+            try {
                 dataset.addProperty(DCTerms.accrualPeriodicity, model.createResource(Mapping.mapMaintenanceFrequencyCode(maintenanceFrequencyCode)));
+            } catch (IllegalArgumentException ignore) {
+                log.warn("Dataset {} has an invalid maintenanceFrequencyCode.", id);
+            }
         }
 
         final List<Node> topicCategories = metadata.selectNodes("gmd:identificationInfo/*/gmd:topicCategory/*");
@@ -752,14 +814,35 @@ public class MDMetadata2Dataset {
                 dataset.addProperty(DCTerms.conformsTo, model.createResource("http://data.europa.eu/eli/reg/2010/1089/oj"));
             } else if ("VERORDNUNG (EG) Nr. 976/2009 DER KOMMISSION vom 19. Oktober 2009 zur Durchführung der Richtlinie 2007/2/EG des Europäischen Parlaments und des Rates hinsichtlich der Netzdienste".equals(identifier)) {
                 dataset.addProperty(DCTerms.conformsTo, model.createResource("http://data.europa.eu/eli/reg/2009/976/oj"));
+            } else if (StringUtils.contains(identifier, "Verordnung (EG) Nr. 1205/2008")) {
+                dataset.addProperty(DCTerms.conformsTo, model.createResource("http://data.europa.eu/eli/reg/2008/1205/oj"));
             }
         }
 
         // convert INSPIRE keywords into dcatap:applicableLegislation properties
-        if (descriptiveKeywords.contains("inspireidentifiziert")
-                || descriptiveKeywords.contains("INSPIRE")
-                || descriptiveKeywords.contains("inspire")) {
-            dataset.addProperty(PROPERTY_applicable_Legislation, model.createResource("http://data.europa.eu/eli/dir/2007/2/2019-06-26"));
+        if (descriptiveKeywords.contains("inspireidentifiziert")) {
+            dataset.addProperty(DCATAP.applicableLegislation, model.createResource("http://data.europa.eu/eli/dir/2007/2/2019-06-26"));
+        }
+
+        // Create the HVD properties in accordance with the GDI-DE mapping rules of 2024-03-20
+        final List<Node> hvdRelevantKeywordNodes =
+                metadata.selectNodes("gmd:identificationInfo/*/gmd:descriptiveKeywords/gmd:MD_Keywords[" +
+                        "gmd:thesaurusName/gmd:CI_Citation/gmd:title/gco:CharacterString = 'High-value dataset categories'" +
+                        " or gmd:thesaurusName/gmd:CI_Citation/gmd:title/gmx:Anchor/@xlink:href = 'http://data.europa.eu/bna/asd487ae75']");
+        if (!hvdRelevantKeywordNodes.isEmpty()) {
+            dataset.addProperty(DCATAP.applicableLegislation, model.createResource("http://data.europa.eu/eli/reg_impl/2023/138/oj"));
+        }
+        for (Node keyword : hvdRelevantKeywordNodes) {
+            Node characterString = keyword.selectSingleNode("./gmd:keyword/gco:CharacterString/text()");
+            Node anchorRef = keyword.selectSingleNode("./gmd:keyword/gmx:Anchor/@xlink:href");
+
+            if (characterString != null) {
+                dataset.addProperty(DCATAP.hvdCategory, Mapping.mapHvdCategory(characterString.getStringValue()));
+            }
+
+            if (anchorRef != null) {
+                dataset.addProperty(DCATAP.hvdCategory, model.createResource(anchorRef.getStringValue()));
+            }
         }
 
         // convert preview image
@@ -781,7 +864,13 @@ public class MDMetadata2Dataset {
 
         // add coupled services
         if (cswInterface != null && searchId != null) {
-            final List<Element> services = cswInterface.findOperatesOn(searchId);
+            List<Element> services;
+            try {
+                services = cswInterface.findOperatesOn(searchId);
+            } catch (Exception e) {
+                log.error("Error searching for coupled services of dataset {}: {}", id, e.getMessage());
+                services = Collections.emptyList();
+            }
             for (Element service : services) {
                 try {
                     for (Resource dist : convertServiceToDistributions(service)) {
@@ -843,6 +932,7 @@ public class MDMetadata2Dataset {
         return dataset;
     }
 
+
     private boolean isMultiFormatDataset(Element metadata) {
         final List<Node> formatElements = metadata.selectNodes("gmd:distributionInfo/*/gmd:distributionFormat");
         return formatElements.size() > 1;
@@ -915,6 +1005,10 @@ public class MDMetadata2Dataset {
             }
         }
 
+        if (formatResource == null) {
+            formatResource = guessFormat(url);
+        }
+
         if (formatResource != null) {
             distribution.addProperty(DCTerms.format, formatResource);
         } else if (!ignoreInvalidMapping) {

src/main/java/de/landsh/opendata/csw2dcat/Mapping.java

@@ -204,9 +204,30 @@ public class Mapping {
             return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/CSV");
         } else if ("HTML".equals(applicationProfile)) {
             return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/HTML");
+        } else if ("GML".equals(applicationProfile)) {
+            return ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/GML");
         } else {
             log.info("{} is not an allowed applicationProfile.", applicationProfile);
             throw new IllegalArgumentException(applicationProfile + " is not an allowed applicationProfile.");
         }
     }
+
+    public static Resource mapHvdCategory(String germanName) {
+        if ("Georaum".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_ac64a52d");
+        } else if ("Erdbeobachtung und Umwelt".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_dd313021");
+        } else if ("Meteorologie".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_164e0bf5");
+        } else if ("Statistik".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_e1da4e07");
+        } else if ("Unternehmen und Eigentümerschaft von Unternehmen".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_a9135398");
+        } else if ("Mobilität".equals(germanName)) {
+            return ResourceFactory.createResource("http://data.europa.eu/bna/c_b79e35eb");
+        } else {
+            log.info("{} is not an allowed HVD category.", germanName);
+            throw new IllegalArgumentException(germanName + " is not an allowed HVD category.");
+        }
+    }
 }

src/main/java/de/landsh/opendata/csw2dcat/vocabulary/DCATAP.java

+package de.landsh.opendata.csw2dcat.vocabulary;
+
+import org.apache.jena.rdf.model.Model;
+import org.apache.jena.rdf.model.ModelFactory;
+import org.apache.jena.rdf.model.Property;
+import org.apache.jena.rdf.model.Resource;
+
+public class DCATAP {
+    public static final String NS = "http://data.europa.eu/r5r/";
+    public static final Resource NAMESPACE;
+    public static final Property applicableLegislation;
+    public static final Property hvdCategory;
+
+    private static final Model m_model = ModelFactory.createDefaultModel();
+
+    static {
+        NAMESPACE = m_model.createResource(NS);
+        applicableLegislation = m_model.createProperty(NS, "applicableLegislation");
+        hvdCategory = m_model.createProperty(NS, "hvdCategory");
+    }
+
+    private DCATAP() {
+    }
+
+    public static String getURI() {
+        return NS;
+    }
+}

src/test/java/de/landsh/opendata/csw2dcat/MDMetadata2DatasetTests.java

@@ -30,6 +30,7 @@ public class MDMetadata2DatasetTests {
     private static final Resource FILE_TYPE_WFS_SRC = ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/WFS_SRVC");
     private static final Resource FILE_TYPE_CSV = ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/CSV");
     private static final Resource FILE_TYPE_SHP = ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/SHP");
+    private static final Resource FILE_TYPE_GML = ResourceFactory.createResource("http://publications.europa.eu/resource/authority/file-type/GML");
 
     private static final Property PROPERTY_applicable_Legislation = ResourceFactory.createProperty("http://data.europa.eu/r5r/", "applicableLegislation");
     static Namespace NS_GMD = Namespace.get("gmd", "http://www.isotc211.org/2005/gmd");
@@ -1288,6 +1289,88 @@ public class MDMetadata2DatasetTests {
 
         assertEquals(1, distributionHTML, "one HTML distribution (the Atom feed viewer)");
         assertEquals(1, distributionATOM, "one ATOM distribution (the feed itself)");
+    }
+
+    /**
+     * An incorrect code list specification (in this case for MD_MaintenanceFrequencyCode) must not stop the process.
+     */
+    @Test
+    public void convert_incorrectMaintenanceFrequencyCode() throws DocumentException, IOException {
+        settings.findWMSinInfo = true;
+
+        final Document inputDocument = saxReader.read(getClass().getResourceAsStream("/bb585473-0987-436a-bc7a-132880a4df0d.xml"));
+        final Resource dataset = service.convert(inputDocument);
+
+        assertNotNull(dataset);
+        assertEquals("bb585473-0987-436a-bc7a-132880a4df0d", dataset.getProperty(DCTerms.identifier).getString());
+    }
+
+    /**
+     * Detect multiple formats (SHP and GML) in one dataset.
+     * The WMS link is so broken that it won't be detected as a WMS.
+     */
+    @Test
+    public void convert_multiFormat() throws DocumentException, IOException {
+        settings.findWMSinInfo = true;
+
+        final Document inputDocument = saxReader.read(getClass().getResourceAsStream("/6782583C-9747-4C16-B94D-3EAF2BFDB171.xml"));
+
+        final Resource result = service.convert(inputDocument);
+
+        final Map<String, Resource> distributionMap = collectDistributions(result);
+        assertEquals(2, distributionMap.size());
+        
+        Resource distribution = distributionMap.get("https://opendata.schleswig-holstein.de/dataset/d501a7ab-a85e-4988-a98e-8f3ae3ab5df2/resource/1bcad7cd-9c94-4a06-bbbc-c7929bb37751/download/lsg.zip");
+        assertEquals(FILE_TYPE_SHP, distribution.getProperty(DCTerms.format).getResource());
+        assertEquals("Landschaftsschutzgebiete (1:25.000)", distribution.getProperty(DCTerms.title).getString());
+
+        Resource distribution2 = distributionMap.get("https://umweltgeodienste.schleswig-holstein.de/WFS_UWAT?service=wfs&version=2.0.0&request=GetFeature&typeNames=app:lsg");
+        assertEquals(FILE_TYPE_GML, distribution2.getProperty(DCTerms.format).getResource());
+        assertEquals("WFS Landschaftsschutzgebiete (1:25.000)", distribution2.getProperty(DCTerms.title).getString());
+    }
+
+    /**
+     * Specification of the HVD category as free text in characterString elements
+     * (in accordance with the GDI-DE mapping rules of 2024-03-20 variant a)
+     */
+    @Test
+    public void convert_highValueDataset_a() throws DocumentException, IOException {
+        final Document inputDocument = saxReader.read(getClass().getResourceAsStream("/hvd1.xml"));
+        final Resource result = service.convert(inputDocument);
 
+        assertEquals("http://data.europa.eu/eli/reg_impl/2023/138/oj",
+                result.getPropertyResourceValue(DCATAP.applicableLegislation).getURI());
+        assertEquals("http://data.europa.eu/bna/c_ac64a52d",
+                result.getPropertyResourceValue(DCATAP.hvdCategory).getURI());
     }
+
+    /**
+     *  Referencing the EU vocabulary for HVD categories as gmx:Anchor elements
+     *  (in accordance with the GDI-DE mapping rules of 2024-03-20 variant b)
+     */
+    @Test
+    public void convert_highValueDataset_b() throws DocumentException, IOException {
+        final Document inputDocument = saxReader.read(getClass().getResourceAsStream("/hvd2.xml"));
+        final Resource result = service.convert(inputDocument);
+
+        assertEquals("http://data.europa.eu/eli/reg_impl/2023/138/oj",
+                result.getPropertyResourceValue(DCATAP.applicableLegislation).getURI());
+        assertEquals("http://data.europa.eu/bna/c_ac64a52d",
+                result.getPropertyResourceValue(DCATAP.hvdCategory).getURI());
+    }
+
+    /**
+     *  In some source systems, German and English titles are strangely combined in one text.
+     */
+    @Test
+    public void convert_mixedGermanEnglishTitles() throws DocumentException, IOException {
+        final Document inputDocument = saxReader.read(getClass().getResourceAsStream("/00729c12-e689-446d-83bd-dd8cfd0eeeb8.xml"));
+        final Resource result = service.convert(inputDocument);
+
+        assertTrue(result.hasProperty( DCTerms.title, ResourceFactory.createLangLiteral(
+                "Robben im Schleswig-Holsteinischen Wattenmeer 1995", "de") ));
+        assertTrue(result.hasProperty( DCTerms.title, ResourceFactory.createLangLiteral(
+                "Seals in the Schleswig-Holstein Wadden Sea in 1995", "en") ));
+    }
+
 }