lsmyrnaios
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

fixed organization id created when mapping APC affiliations. Factored out ROR constants in dhp-common

This commit is contained in:
Claudio Atzori 2023-05-15 15:44:46 +02:00
parent d8882c4481
commit 8a463cc3e8
4 changed files with 18 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
dhp-common/src/main/java/eu/dnetlib/dhp/common/Constants.java
View File

@ -10,6 +10,12 @@ public class Constants {
public static final Map<String, String> accessRightsCoarMap = Maps.newHashMap(); public static final Map<String, String> accessRightsCoarMap = Maps.newHashMap();
public static final Map<String, String> coarCodeLabelMap = Maps.newHashMap(); public static final Map<String, String> coarCodeLabelMap = Maps.newHashMap();
public static final String ROR_NS_PREFIX = "ror_________";
public static final String ROR_OPENAIRE_ID = "10|openaire____::993a7ae7a863813cf95028b50708e222";
public static final String ROR_DATASOURCE_NAME = "Research Organization Registry (ROR)";
public static String COAR_ACCESS_RIGHT_SCHEMA = "http://vocabularies.coar-repositories.org/documentation/access_rights/"; public static String COAR_ACCESS_RIGHT_SCHEMA = "http://vocabularies.coar-repositories.org/documentation/access_rights/";
private Constants() { private Constants() {

9
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/ror/GenerateRorActionSetJob.java
View File

@ -40,6 +40,7 @@ import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.actionmanager.ror.model.ExternalIdType; import eu.dnetlib.dhp.actionmanager.ror.model.ExternalIdType;
import eu.dnetlib.dhp.actionmanager.ror.model.RorOrganization; import eu.dnetlib.dhp.actionmanager.ror.model.RorOrganization;
import eu.dnetlib.dhp.application.ArgumentApplicationParser; import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.Constants;
import eu.dnetlib.dhp.common.HdfsSupport; import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.schema.action.AtomicAction; import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.dhp.schema.common.ModelConstants; import eu.dnetlib.dhp.schema.common.ModelConstants;
@ -59,10 +60,8 @@ public class GenerateRorActionSetJob {
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper(); private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static final String ROR_NS_PREFIX = "ror_________";
private static final List<KeyValue> ROR_COLLECTED_FROM = listKeyValues( private static final List<KeyValue> ROR_COLLECTED_FROM = listKeyValues(
"10|openaire____::993a7ae7a863813cf95028b50708e222", "ROR"); Constants.ROR_OPENAIRE_ID, Constants.ROR_DATASOURCE_NAME);
private static final DataInfo ROR_DATA_INFO = dataInfo( private static final DataInfo ROR_DATA_INFO = dataInfo(
false, "", false, false, ENTITYREGISTRY_PROVENANCE_ACTION, "0.92"); false, "", false, false, ENTITYREGISTRY_PROVENANCE_ACTION, "0.92");
@ -126,7 +125,7 @@ public class GenerateRorActionSetJob {
final Organization o = new Organization(); final Organization o = new Organization();
o.setId(calculateOpenaireId(r.getId())); o.setId(calculateOpenaireId(r.getId()));
o.setOriginalId(Arrays.asList(String.format("%s::%s", ROR_NS_PREFIX, r.getId()))); o.setOriginalId(Arrays.asList(String.format("%s::%s", Constants.ROR_NS_PREFIX, r.getId())));
o.setCollectedfrom(ROR_COLLECTED_FROM); o.setCollectedfrom(ROR_COLLECTED_FROM);
o.setPid(pids(r)); o.setPid(pids(r));
o.setDateofcollection(now.toString()); o.setDateofcollection(now.toString());
@ -170,7 +169,7 @@ public class GenerateRorActionSetJob {
} }
private static String calculateOpenaireId(final String rorId) { private static String calculateOpenaireId(final String rorId) {
return String.format("20|%s::%s", ROR_NS_PREFIX, DHPUtils.md5(rorId)); return String.format("20|%s::%s", Constants.ROR_NS_PREFIX, DHPUtils.md5(rorId));
} }
private static List<StructuredProperty> pids(final RorOrganization r) { private static List<StructuredProperty> pids(final RorOrganization r) {

5
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java
View File

@ -20,6 +20,7 @@ import com.google.common.collect.Lists;
import com.google.common.collect.Maps; import com.google.common.collect.Maps;
import com.google.common.collect.Sets; import com.google.common.collect.Sets;
import eu.dnetlib.dhp.common.Constants;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup; import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants; import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.common.ModelSupport; import eu.dnetlib.dhp.schema.common.ModelSupport;
@ -326,8 +327,10 @@ public abstract class AbstractMdRecordToOafMapper {
String rorId = element.attributeValue("affiliationIdentifier"); String rorId = element.attributeValue("affiliationIdentifier");
if (StringUtils.isNotBlank(rorId)) { if (StringUtils.isNotBlank(rorId)) {
String fullRorId = Constants.ROR_NS_PREFIX + "::" + rorId;
String resultId = entity.getId(); String resultId = entity.getId();
String orgId = createOpenaireId("organization", rorId, true); String orgId = createOpenaireId("organization", fullRorId, true);
List<KeyValue> properties = Lists.newArrayList(); List<KeyValue> properties = Lists.newArrayList();

5
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java
View File

@ -23,6 +23,7 @@ import org.mockito.junit.jupiter.MockitoExtension;
import com.fasterxml.jackson.databind.ObjectMapper; import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.common.Constants;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup; import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants; import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*; import eu.dnetlib.dhp.schema.oaf.*;
@ -922,7 +923,7 @@ class MappersTest {
} }
@Test @Test
void testOpenAPC() throws IOException, DocumentException { void testOpenAPC() throws IOException {
final String xml = IOUtils.toString(Objects.requireNonNull(getClass().getResourceAsStream("oaf_openapc.xml"))); final String xml = IOUtils.toString(Objects.requireNonNull(getClass().getResourceAsStream("oaf_openapc.xml")));
final List<Oaf> list = new OafToOafMapper(vocs, true, true).processMdRecord(xml); final List<Oaf> list = new OafToOafMapper(vocs, true, true).processMdRecord(xml);
@ -955,6 +956,8 @@ class MappersTest {
if (StringUtils.startsWith(source, "50")) { if (StringUtils.startsWith(source, "50")) {
assertEquals(ModelConstants.HAS_AUTHOR_INSTITUTION, r.getRelClass()); assertEquals(ModelConstants.HAS_AUTHOR_INSTITUTION, r.getRelClass());
} else if (StringUtils.startsWith(source, "20")) { } else if (StringUtils.startsWith(source, "20")) {
assertTrue(StringUtils.contains(source, "::"));
assertEquals("20|" + Constants.ROR_NS_PREFIX, StringUtils.substringBefore(source, "::"));
assertEquals(ModelConstants.IS_AUTHOR_INSTITUTION_OF, r.getRelClass()); assertEquals(ModelConstants.IS_AUTHOR_INSTITUTION_OF, r.getRelClass());
} else { } else {
throw new IllegalArgumentException("invalid source / target prefixes for affiliation relations"); throw new IllegalArgumentException("invalid source / target prefixes for affiliation relations");