dimitris.pierrakos
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

extended mapping applied to datacite records to produce affiliations using the ROR ids. Inc ase of APCs it includes the amount and the currently in the relation

This commit is contained in:
Claudio Atzori 2023-05-02 11:56:51 +02:00
parent d02916ef82
commit d8882c4481
2 changed files with 87 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java
View File

@ -1,14 +1,9 @@
package eu.dnetlib.dhp.oa.graph.raw;
import static eu.dnetlib.dhp.schema.common.ModelConstants.DNET_PID_TYPES;
import static eu.dnetlib.dhp.schema.common.ModelConstants.IS_PRODUCED_BY;
import static eu.dnetlib.dhp.schema.common.ModelConstants.OUTCOME;
import static eu.dnetlib.dhp.schema.common.ModelConstants.PRODUCES;
import static eu.dnetlib.dhp.schema.common.ModelConstants.REPOSITORY_PROVENANCE_ACTIONS;
import static eu.dnetlib.dhp.schema.common.ModelConstants.RESULT_PROJECT;
import static eu.dnetlib.dhp.schema.common.ModelConstants.UNKNOWN;
import static eu.dnetlib.dhp.schema.common.ModelConstants.*;
import static eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils.*;
import static eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils.createOpenaireId;
import java.net.MalformedURLException;
import java.net.URL;
@ -22,6 +17,7 @@ import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.google.common.collect.Lists;
import com.google.common.collect.Maps;
import com.google.common.collect.Sets;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
@ -195,6 +191,7 @@ public abstract class AbstractMdRecordToOafMapper {
rels.addAll(addProjectRels(doc, entity));
rels.addAll(addOtherResultRels(doc, entity));
rels.addAll(addRelations(doc, entity));
rels.addAll(addAffiliations(doc, entity));
oafs.addAll(rels);
}
@ -295,7 +292,7 @@ public abstract class AbstractMdRecordToOafMapper {
final String relClassInverse = ModelSupport
.findInverse(ModelSupport.rel(relType, subRelType, relClass))
.getInverseRelClass();
final String validationdDate = ((Node) o).valueOf("@validationDate");
final String validationDate = ((Node) o).valueOf("@validationDate");
if (StringUtils.isNotBlank(target)) {
final String targetType = element.attributeValue("targetType");
@ -306,13 +303,13 @@ public abstract class AbstractMdRecordToOafMapper {
OafMapperUtils
.getRelation(
entity.getId(), targetId, relType, subRelType, relClass, entity,
validationdDate));
validationDate));
rels
.add(
OafMapperUtils
.getRelation(
targetId, entity.getId(), relType, subRelType, relClassInverse, entity,
validationdDate));
validationDate));
}
}
}
@ -320,6 +317,47 @@ public abstract class AbstractMdRecordToOafMapper {
return rels;
}
private List<Oaf> addAffiliations(Document doc, OafEntity entity) {
final List<Oaf> rels = Lists.newArrayList();
for (Object o : doc.selectNodes("//datacite:affiliation[@affiliationIdentifierScheme='ROR']")) {
Element element = (Element) o;
String rorId = element.attributeValue("affiliationIdentifier");
if (StringUtils.isNotBlank(rorId)) {
String resultId = entity.getId();
String orgId = createOpenaireId("organization", rorId, true);
List<KeyValue> properties = Lists.newArrayList();
String apcAmount = doc.valueOf("//oaf:processingchargeamount");
String apcCurrency = doc.valueOf("//oaf:processingchargeamount/@currency");
if (StringUtils.isNotBlank(apcAmount) && StringUtils.isNotBlank(apcCurrency)) {
properties.add(OafMapperUtils.keyValue("apc_amount", apcAmount));
properties.add(OafMapperUtils.keyValue("apc_currency", apcCurrency));
}
rels
.add(
OafMapperUtils
.getRelation(
resultId, orgId, RESULT_ORGANIZATION, AFFILIATION, HAS_AUTHOR_INSTITUTION,
entity.getCollectedfrom(), entity.getDataInfo(), entity.getLastupdatetimestamp(), null,
properties));
rels
.add(
OafMapperUtils
.getRelation(
orgId, resultId, RESULT_ORGANIZATION, AFFILIATION, IS_AUTHOR_INSTITUTION_OF,
entity.getCollectedfrom(), entity.getDataInfo(), entity.getLastupdatetimestamp(), null,
properties));
}
}
return rels;
}
protected abstract List<Oaf> addOtherResultRels(
final Document doc,
final OafEntity entity);

42
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java
View File

@ -930,16 +930,52 @@ class MappersTest {
System.out.println(new ObjectMapper().writeValueAsString(list));
System.out.println("***************");
final Publication p = (Publication) list.get(0);
final Optional<Oaf> o = list.stream().filter(r -> r instanceof Publication).findFirst();
assertTrue(o.isPresent());
Publication p = (Publication) o.get();
assertTrue(p.getInstance().size() > 0);
assertEquals("https://doi.org/10.1155/2015/439379", p.getInstance().get(0).getUrl().get(0));
assertTrue(p.getProcessingchargeamount() != null);
assertTrue(p.getProcessingchargecurrency() != null);
assertNotNull(p.getProcessingchargeamount());
assertNotNull(p.getProcessingchargecurrency());
assertEquals("1721.47", p.getProcessingchargeamount().getValue());
assertEquals("EUR", p.getProcessingchargecurrency().getValue());
List<Oaf> affiliations = list.stream().filter(r -> r instanceof Relation).collect(Collectors.toList());
assertEquals(2, affiliations.size());
for (Oaf aff : affiliations) {
Relation r = (Relation) aff;
assertEquals(ModelConstants.AFFILIATION, r.getSubRelType());
assertEquals(ModelConstants.RESULT_ORGANIZATION, r.getRelType());
String source = r.getSource();
if (StringUtils.startsWith(source, "50")) {
assertEquals(ModelConstants.HAS_AUTHOR_INSTITUTION, r.getRelClass());
} else if (StringUtils.startsWith(source, "20")) {
assertEquals(ModelConstants.IS_AUTHOR_INSTITUTION_OF, r.getRelClass());
} else {
throw new IllegalArgumentException("invalid source / target prefixes for affiliation relations");
}
List<KeyValue> apcInfo = r.getProperties();
assertEquals(
"EUR", apcInfo
.stream()
.filter(kv -> "apc_currency".equals(kv.getKey()))
.map(KeyValue::getValue)
.findFirst()
.orElse(""));
assertEquals(
"1721.47", apcInfo
.stream()
.filter(kv -> "apc_amount".equals(kv.getKey()))
.map(KeyValue::getValue)
.findFirst()
.orElse(""));
}
}
@Test