removing jena non-compliant attribute during harvesting

geonetwork · Nov 25, 2024 · 7ba4d47 · 7ba4d47
1 parent 97bdd7b
commit 7ba4d47
Showing 1 changed file with 4 additions and 0 deletions.
diff --git a/harvesters/src/main/java/org/fao/geonet/kernel/harvest/harvester/simpleurl/RDFUtils.java b/harvesters/src/main/java/org/fao/geonet/kernel/harvest/harvester/simpleurl/RDFUtils.java
@@ -7,6 +7,7 @@
 import org.apache.jena.riot.Lang;
 import org.apache.jena.riot.RDFDataMgr;
 import org.fao.geonet.Constants;
+import org.fao.geonet.constants.Geonet;
 import org.fao.geonet.domain.ISODate;
 import org.fao.geonet.domain.Pair;
 import org.fao.geonet.utils.Log;
@@ -81,6 +82,9 @@ public static HashMap<String, Element> getAllUuids(String feedUrl) throws Except
     public static HashMap<String, Element> getAllUuids(Element feed) throws Exception {
         Element rdfDocument = checkForMissingRdfAbout(feed);
 
+        // if xsi:schemaLocation is present on the rdf:RDF element Jena fails to parse the Element.
+        rdfDocument.removeAttribute("schemaLocation", Geonet.Namespaces.XSI);
+
         Model model = ModelFactory.createMemModelMaker().createDefaultModel();
         RDFDataMgr.read(model,
             IOUtils.toInputStream(Xml.getString(rdfDocument), StandardCharsets.UTF_8),