enrico.ottonello
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge branch 'master' into dhp_oaf_model

This commit is contained in:
Claudio Atzori 2020-06-16 16:11:19 +02:00
parent 603b1bd0bb 9e2c23e391
commit 631fef12a7
4 changed files with 48 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/GenerateEventsApplication.java
View File

@ -140,7 +140,8 @@ public class GenerateEventsApplication {
final Dataset<Publication> publications = readPath(spark, graphPath + "/publication", Publication.class);
final Dataset<Relation> rels = readPath(spark, graphPath + "/relation", Relation.class)
.filter(r -> !r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS));
.filter(r -> !r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS))
.cache();
final Dataset<OpenaireBrokerResult> r0 = readPath(
spark, graphPath + "/" + sourceClass.getSimpleName().toLowerCase(), Result.class)

16
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/util/ConversionUtils.java
View File

@ -68,6 +68,22 @@ public class ConversionUtils {
: null;
}
public static eu.dnetlib.broker.objects.Publication oafPublicationToBrokerPublication(final Publication p) {
return p != null ? new eu.dnetlib.broker.objects.Publication()
.setOriginalId(p.getOriginalId().get(0))
.setTitle(structPropValue(p.getTitle()))
.setPids(p.getPid().stream().map(ConversionUtils::oafPidToBrokerPid).collect(Collectors.toList()))
.setInstances(
p
.getInstance()
.stream()
.map(ConversionUtils::oafInstanceToBrokerInstances)
.flatMap(List::stream)
.collect(Collectors.toList()))
.setCollectedFrom(p.getCollectedfrom().stream().map(KeyValue::getValue).findFirst().orElse(null))
: null;
}
public static final OpenaireBrokerResult oafResultToBrokerResult(final Result result) {
return result != null ? new OpenaireBrokerResult()

34
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/util/aggregators/withRels/RelatedEntityFactory.java
View File

@ -1,10 +1,11 @@
package eu.dnetlib.dhp.broker.oa.util.aggregators.withRels;
import eu.dnetlib.broker.objects.Dataset;
import eu.dnetlib.broker.objects.Project;
import eu.dnetlib.broker.objects.Publication;
import eu.dnetlib.broker.objects.Software;
import eu.dnetlib.dhp.broker.oa.util.ConversionUtils;
import eu.dnetlib.dhp.schema.oaf.Dataset;
import eu.dnetlib.dhp.schema.oaf.Project;
import eu.dnetlib.dhp.schema.oaf.Publication;
import eu.dnetlib.dhp.schema.oaf.Software;
public class RelatedEntityFactory {
@ -13,18 +14,21 @@ public class RelatedEntityFactory {
final String relType,
final T target,
final Class<RT> clazz) {
if (clazz == RelatedProject.class) {
return (RT) new RelatedProject(sourceId, relType, (Project) target);
return (RT) new RelatedProject(sourceId, relType,
ConversionUtils.oafProjectToBrokerProject((Project) target));
} else if (clazz == RelatedSoftware.class) {
return (RT) new RelatedSoftware(sourceId, relType,
ConversionUtils.oafSoftwareToBrokerSoftware((Software) target));
} else if (clazz == RelatedDataset.class) {
return (RT) new RelatedDataset(sourceId, relType,
ConversionUtils.oafDatasetToBrokerDataset((Dataset) target));
} else if (clazz == RelatedPublication.class) {
return (RT) new RelatedPublication(sourceId, relType,
ConversionUtils.oafPublicationToBrokerPublication((Publication) target));
} else {
return null;
}
if (clazz == RelatedSoftware.class) {
return (RT) new RelatedSoftware(sourceId, relType, (Software) target);
}
if (clazz == RelatedDataset.class) {
return (RT) new RelatedDataset(sourceId, relType, (Dataset) target);
}
if (clazz == RelatedPublication.class) {
return (RT) new RelatedPublication(sourceId, relType, (Publication) target);
}
return null;
}
}

13
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/OdfToOafMapper.java
View File

@ -126,9 +126,16 @@ public class OdfToOafMapper extends AbstractMdRecordToOafMapper {
for (final Object o : doc.selectNodes("//datacite:alternateIdentifier[@alternateIdentifierType='URL']")) {
url.add(((Node) o).getText().trim());
}
for (final Object o : doc
.selectNodes("//datacite:alternateIdentifier[@alternateIdentifierType='landingPage']")) {
url.add(((Node) o).getText().trim());
}
for (final Object o : doc.selectNodes("//datacite:identifier[@identifierType='URL']")) {
url.add(((Node) o).getText().trim());
}
for (final Object o : doc.selectNodes("//datacite:identifier[@identifierType='landingPage']")) {
url.add(((Node) o).getText().trim());
}
for (final Object o : doc.selectNodes("//datacite:alternateIdentifier[@alternateIdentifierType='DOI']")) {
url.add(HTTP_DX_DOI_PREIFX + ((Node) o).getText().trim());
}
@ -367,11 +374,13 @@ public class OdfToOafMapper extends AbstractMdRecordToOafMapper {
res
.addAll(
prepareListStructPropsWithValidQualifier(
doc, "//datacite:identifier[@identifierType != 'URL']", "@identifierType", DNET_PID_TYPES, info));
doc, "//datacite:identifier[@identifierType != 'URL' and @identifierType != 'landingPage']",
"@identifierType", DNET_PID_TYPES, info));
res
.addAll(
prepareListStructPropsWithValidQualifier(
doc, "//datacite:alternateIdentifier[@alternateIdentifierType != 'URL']",
doc,
"//datacite:alternateIdentifier[@alternateIdentifierType != 'URL' and @alternateIdentifierType != 'landingPage']",
"@alternateIdentifierType", DNET_PID_TYPES, info));
return res;
}