Xml record serialization for author PIDs: 1) only one value per PID type is allowed; 2) orcid prevails over orcid_pending

dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/utils/AuthorPidTypeComparator.java

@@ -0,0 +1,52 @@
+
+package eu.dnetlib.dhp.oa.provision.utils;
+
+import java.util.Comparator;
+import java.util.Optional;
+
+import eu.dnetlib.dhp.schema.common.ModelConstants;
+import eu.dnetlib.dhp.schema.common.ModelSupport;
+import eu.dnetlib.dhp.schema.oaf.Qualifier;
+import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
+
+public class AuthorPidTypeComparator implements Comparator<StructuredProperty> {
+
+	@Override
+	public int compare(StructuredProperty left, StructuredProperty right) {
+
+		String lClass = Optional
+			.ofNullable(left)
+			.map(StructuredProperty::getQualifier)
+			.map(Qualifier::getClassid)
+			.orElse(null);
+
+		String rClass = Optional
+			.ofNullable(right)
+			.map(StructuredProperty::getQualifier)
+			.map(Qualifier::getClassid)
+			.orElse(null);
+
+		if (lClass == null && rClass == null)
+			return 0;
+		if (lClass == null)
+			return 1;
+		if (rClass == null)
+			return -1;
+
+		if (lClass.equals(rClass))
+			return 0;
+
+		if (lClass.equals(ModelConstants.ORCID))
+			return -1;
+		if (rClass.equals(ModelConstants.ORCID))
+			return 1;
+
+		if (lClass.equals(ModelConstants.ORCID_PENDING))
+			return -1;
+		if (rClass.equals(ModelConstants.ORCID_PENDING))
+			return 1;
+
+		return 0;
+	}
+
+}

dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/utils/XmlRecordFactory.java

@@ -254,6 +254,18 @@ public class XmlRecordFactory implements Serializable {
 														p -> p,
 														(p1, p2) -> p1))
 											.values()
+											.stream()
+											.collect(
+												Collectors
+													.groupingBy(
+														p -> p.getValue(),
+														Collectors
+															.mapping(
+																p -> p,
+																Collectors.minBy(new AuthorPidTypeComparator()))))
+											.values()
+											.stream()
+											.map(op -> op.get())
 											.forEach(
 												sp -> {
 													String pidType = getAuthorPidType(sp.getQualifier().getClassid());

dhp-workflows/dhp-graph-provision/src/test/java/eu/dnetlib/dhp/oa/provision/XmlRecordFactoryTest.java

@@ -14,6 +14,7 @@ import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 
+import com.fasterxml.jackson.databind.DeserializationFeature;
 import com.fasterxml.jackson.databind.ObjectMapper;
 
 import eu.dnetlib.dhp.oa.provision.model.JoinedEntity;
@@ -25,6 +26,9 @@ public class XmlRecordFactoryTest {
 
 	private static final String otherDsTypeId = "scholarcomminfra,infospace,pubsrepository::mock,entityregistry,entityregistry::projects,entityregistry::repositories,websource";
 
+	private static ObjectMapper OBJECT_MAPPER = new ObjectMapper()
+		.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
+
 	@Test
 	public void testXMLRecordFactory() throws IOException, DocumentException {
 
@@ -33,7 +37,7 @@ public class XmlRecordFactoryTest {
 		XmlRecordFactory xmlRecordFactory = new XmlRecordFactory(contextMapper, false, XmlConverterJob.schemaLocation,
 			otherDsTypeId);
 
-		Publication p = new ObjectMapper()
+		Publication p = OBJECT_MAPPER
 			.readValue(IOUtils.toString(getClass().getResourceAsStream("publication.json")), Publication.class);
 
 		String xml = xmlRecordFactory.build(new JoinedEntity<>(p));
@@ -44,10 +48,14 @@ public class XmlRecordFactoryTest {
 
 		assertNotNull(doc);
 
-		// System.out.println(doc.asXML());
+		System.out.println(doc.asXML());
 
-		Assertions.assertEquals("0000-0001-9613-6639", doc.valueOf("//creator[@rank = '1']/@orcid"));
+		Assertions.assertEquals("0000-0001-9613-6638", doc.valueOf("//creator[@rank = '1']/@orcid"));
 		Assertions.assertEquals("0000-0001-9613-6639", doc.valueOf("//creator[@rank = '1']/@orcid_pending"));
+
+		Assertions.assertEquals("0000-0001-9613-9956", doc.valueOf("//creator[@rank = '2']/@orcid"));
+		Assertions.assertEquals("", doc.valueOf("//creator[@rank = '2']/@orcid_pending"));
+
 		// TODO add assertions based of values extracted from the XML record
 	}
 }

dhp-workflows/dhp-graph-provision/src/test/resources/eu/dnetlib/dhp/oa/provision/publication.json

@@ -6,41 +6,15 @@
       "name": "Jaehyun",
       "pid": [
         {
-          "dataInfo": {
-            "deletedbyinference": false,
-            "inferenceprovenance": "",
-            "inferred": false,
-            "invisible": false,
-            "provenanceaction": {
-              "classid": "",
-              "classname": "",
-              "schemeid": "",
-              "schemename": ""
-            },
-            "trust": ""
-          },
           "qualifier": {
             "classid": "orcid",
             "classname": "Open Researcher and Contributor ID",
             "schemeid": "dnet:pid_types",
             "schemename": "dnet:pid_types"
           },
-          "value": "0000-0001-9613-6639"
+          "value": "0000-0001-9613-6638"
         },
         {
-          "dataInfo": {
-            "deletedbyinference": false,
-            "inferenceprovenance": "",
-            "inferred": false,
-            "invisible": false,
-            "provenanceaction": {
-              "classid": "",
-              "classname": "",
-              "schemeid": "",
-              "schemename": ""
-            },
-            "trust": ""
-          },
           "qualifier": {
             "classid": "orcid_pending",
             "classname": "Open Researcher and Contributor ID",
@@ -57,7 +31,26 @@
       "affiliation": [],
       "fullname": "Berrada, Salim",
       "name": "Salim",
-      "pid": [],
+      "pid": [
+        {
+          "qualifier": {
+            "classid": "orcid",
+            "classname": "Open Researcher and Contributor ID",
+            "schemeid": "dnet:pid_types",
+            "schemename": "dnet:pid_types"
+          },
+          "value": "0000-0001-9613-9956"
+        },
+        {
+          "qualifier": {
+            "classid": "orcid_pending",
+            "classname": "Open Researcher and Contributor ID",
+            "schemeid": "dnet:pid_types",
+            "schemename": "dnet:pid_types"
+          },
+          "value": "0000-0001-9613-9956"
+        }
+      ],
       "rank": 2,
       "surname": "Berrada"
     },