D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

defaults to 0000 - Unknown in case the instance type lookup in the dnet:result_typologies doesn't find a corresponding result type binding

This commit is contained in:
Claudio Atzori 2024-11-13 16:27:00 +01:00
parent 07f267bb10
commit 4a3b173ca2
4 changed files with 36 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/MergeUtils.java
View File

@ -74,15 +74,9 @@ public class MergeUtils {
if (!vocs.vocabularyExists(ModelConstants.DNET_RESULT_TYPOLOGIES)) {
return (T) mergedResult;
} else {
final Qualifier expectedResultType = vocs.lookupTermBySynonym(
ModelConstants.DNET_RESULT_TYPOLOGIES,
i.getInstancetype().getClassid());
if (Objects.isNull(expectedResultType)) {
throw new IllegalArgumentException(
"instance type not bound to any result type in dnet:result_typologies: " +
i.getInstancetype().getClassid());
}
final Qualifier expectedResultType = Optional
.ofNullable(vocs.lookupTermBySynonym(ModelConstants.DNET_RESULT_TYPOLOGIES, i.getInstancetype().getClassid()))
.orElse(OafMapperUtils.unknown(ModelConstants.DNET_RESULT_TYPOLOGIES, ModelConstants.DNET_RESULT_TYPOLOGIES));
// there is a clash among the result types
if (!expectedResultType.getClassid().equals(mergedResult.getResulttype().getClassid())) {

49
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/AbstractMdRecordToOafMapper.java
View File

@ -153,29 +153,32 @@ public abstract class AbstractMdRecordToOafMapper {
final DataInfo entityInfo = prepareDataInfo(doc, this.invisible);
final long lastUpdateTimestamp = new Date().getTime();
final List<Instance> instances = prepareInstances(doc, entityInfo, collectedFrom, hostedBy);
final Instance instance = prepareInstances(doc, entityInfo, collectedFrom, hostedBy);
final String type = getResultType(instances);
if (!Optional.ofNullable(instance.getInstancetype())
.map(Qualifier::getClassid)
.filter(StringUtils::isNotBlank)
.isPresent()) {
return Lists.newArrayList();
}
return createOafs(doc, type, instances, collectedFrom, entityInfo, lastUpdateTimestamp);
final String type = getResultType(instance);
return createOafs(doc, type, instance, collectedFrom, entityInfo, lastUpdateTimestamp);
} catch (final DocumentException e) {
log.error("Error with record:\n" + xml);
return Lists.newArrayList();
}
}
protected String getResultType(final List<Instance> instances) {
protected String getResultType(final Instance instance) {
if (this.vocs.vocabularyExists(ModelConstants.DNET_RESULT_TYPOLOGIES)) {
final String instanceType = instances
.stream()
.map(i -> i.getInstancetype().getClassid())
.findFirst()
.filter(s -> !UNKNOWN.equalsIgnoreCase(s))
.orElse("0000"); // Unknown
return Optional
return Optional.ofNullable(instance.getInstancetype())
.map(Qualifier::getClassid)
.map(instanceType -> Optional
.ofNullable(this.vocs.getSynonymAsQualifier(ModelConstants.DNET_RESULT_TYPOLOGIES, instanceType))
.map(Qualifier::getClassid)
.orElse("0000"))
.orElse("0000");
} else {
throw new IllegalStateException("Missing vocabulary: " + ModelConstants.DNET_RESULT_TYPOLOGIES);
@ -196,12 +199,12 @@ public abstract class AbstractMdRecordToOafMapper {
protected List<Oaf> createOafs(
final Document doc,
final String type,
final List<Instance> instances,
final Instance instance,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp) {
final OafEntity entity = createEntity(doc, type, instances, collectedFrom, info, lastUpdateTimestamp);
final OafEntity entity = createEntity(doc, type, instance, collectedFrom, info, lastUpdateTimestamp);
final Set<String> originalId = Sets.newHashSet(entity.getOriginalId());
originalId.add(entity.getId());
@ -234,19 +237,19 @@ public abstract class AbstractMdRecordToOafMapper {
private OafEntity createEntity(final Document doc,
final String type,
final List<Instance> instances,
final Instance instance,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp) {
switch (type.toLowerCase()) {
case "publication":
final Publication p = new Publication();
populateResultFields(p, doc, instances, collectedFrom, info, lastUpdateTimestamp);
populateResultFields(p, doc, instance, collectedFrom, info, lastUpdateTimestamp);
p.setJournal(prepareJournal(doc, info));
return p;
case "dataset":
final Dataset d = new Dataset();
populateResultFields(d, doc, instances, collectedFrom, info, lastUpdateTimestamp);
populateResultFields(d, doc, instance, collectedFrom, info, lastUpdateTimestamp);
d.setStoragedate(prepareDatasetStorageDate(doc, info));
d.setDevice(prepareDatasetDevice(doc, info));
d.setSize(prepareDatasetSize(doc, info));
@ -257,7 +260,7 @@ public abstract class AbstractMdRecordToOafMapper {
return d;
case "software":
final Software s = new Software();
populateResultFields(s, doc, instances, collectedFrom, info, lastUpdateTimestamp);
populateResultFields(s, doc, instance, collectedFrom, info, lastUpdateTimestamp);
s.setDocumentationUrl(prepareSoftwareDocumentationUrls(doc, info));
s.setLicense(prepareSoftwareLicenses(doc, info));
s.setCodeRepositoryUrl(prepareSoftwareCodeRepositoryUrl(doc, info));
@ -267,7 +270,7 @@ public abstract class AbstractMdRecordToOafMapper {
case "otherresearchproducts":
default:
final OtherResearchProduct o = new OtherResearchProduct();
populateResultFields(o, doc, instances, collectedFrom, info, lastUpdateTimestamp);
populateResultFields(o, doc, instance, collectedFrom, info, lastUpdateTimestamp);
o.setContactperson(prepareOtherResearchProductContactPersons(doc, info));
o.setContactgroup(prepareOtherResearchProductContactGroups(doc, info));
o.setTool(prepareOtherResearchProductTools(doc, info));
@ -414,7 +417,7 @@ public abstract class AbstractMdRecordToOafMapper {
private void populateResultFields(
final Result r,
final Document doc,
final List<Instance> instances,
final Instance instance,
final KeyValue collectedFrom,
final DataInfo info,
final long lastUpdateTimestamp) {
@ -448,8 +451,8 @@ public abstract class AbstractMdRecordToOafMapper {
r.setExternalReference(new ArrayList<>()); // NOT PRESENT IN MDSTORES
r.setProcessingchargeamount(field(doc.valueOf("//oaf:processingchargeamount"), info));
r.setProcessingchargecurrency(field(doc.valueOf("//oaf:processingchargeamount/@currency"), info));
r.setInstance(instances);
r.setBestaccessright(OafMapperUtils.createBestAccessRights(instances));
r.setInstance(Arrays.asList(instance));
r.setBestaccessright(OafMapperUtils.createBestAccessRights(Arrays.asList(instance)));
r.setEoscifguidelines(prepareEOSCIfGuidelines(doc, info));
}
@ -508,7 +511,7 @@ public abstract class AbstractMdRecordToOafMapper {
protected abstract Qualifier prepareResourceType(Document doc, DataInfo info);
protected abstract List<Instance> prepareInstances(
protected abstract Instance prepareInstances(
Document doc,
DataInfo info,
KeyValue collectedfrom,

4
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/OafToOafMapper.java
View File

@ -135,7 +135,7 @@ public class OafToOafMapper extends AbstractMdRecordToOafMapper {
}
@Override
protected List<Instance> prepareInstances(
protected Instance prepareInstances(
final Document doc,
final DataInfo info,
final KeyValue collectedfrom,
@ -197,7 +197,7 @@ public class OafToOafMapper extends AbstractMdRecordToOafMapper {
instance.getUrl().addAll(validUrl);
}
return Lists.newArrayList(instance);
return instance;
}
/**

4
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/OdfToOafMapper.java
View File

@ -126,7 +126,7 @@ public class OdfToOafMapper extends AbstractMdRecordToOafMapper {
}
@Override
protected List<Instance> prepareInstances(
protected Instance prepareInstances(
final Document doc,
final DataInfo info,
final KeyValue collectedfrom,
@ -210,7 +210,7 @@ public class OdfToOafMapper extends AbstractMdRecordToOafMapper {
instance.setUrl(new ArrayList<>());
instance.getUrl().addAll(validUrl);
}
return Arrays.asList(instance);
return instance;
}
protected String trimAndDecodeUrl(String url) {