D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

- Removed ModelConstants and update Relation enum

This commit is contained in:
Sandro La Bruzzo 2023-05-03 14:03:01 +02:00
parent f8f4b9a018
commit 88fffa6dbd
58 changed files with 306 additions and 312 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/opencitations/CreateActionSetSparkJob.java
View File

@ -161,9 +161,9 @@ public class CreateActionSetSparkJob implements Serializable {
r.setProvenance(PROVENANCE);
r.setSource(source);
r.setTarget(target);
r.setRelType(ModelConstants.RESULT_RESULT);
r.setSubRelType(ModelConstants.CITATION);
r.setRelClass(ModelConstants.CITES);
r.setRelType(Relation.RELTYPE.resultResult);
r.setSubRelType(Relation.SUBRELTYPE.citation);
r.setRelClass(Relation.RELCLASS.Cites);
return r;
}

6
dhp-workflows/dhp-aggregation/src/main/scala/eu/dnetlib/dhp/datacite/DataciteToOAFTransformation.scala
View File

@ -648,11 +648,11 @@ object DataciteToOAFTransformation {
(r.relationType.toLowerCase.contains("cite") || r.relationType.toLowerCase.contains("reference"))
)
.map(r => {
r.relationType match {
case Relation.RELCLASS.Cites.toString | Relation.RELCLASS.References.toString =>
Relation.RELCLASS.valueOf(r.relationType) match {
case Relation.RELCLASS.Cites | Relation.RELCLASS.References =>
val target = DHPUtils.generateUnresolvedIdentifier(r.relatedIdentifier, r.relatedIdentifierType)
relation(id, target, Relation.SUBRELTYPE.citation, Relation.RELCLASS.Cites, date)
case Relation.RELCLASS.IsCitedBy.toString | Relation.RELCLASS.IsReferencedBy.toString =>
case Relation.RELCLASS.IsCitedBy | Relation.RELCLASS.IsReferencedBy =>
val source = DHPUtils.generateUnresolvedIdentifier(r.relatedIdentifier, r.relatedIdentifierType)
relation(source, id, Relation.SUBRELTYPE.citation, Relation.RELCLASS.Cites, date)
}

31
dhp-workflows/dhp-aggregation/src/main/scala/eu/dnetlib/dhp/sx/bio/BioDBToOAF.scala
View File

@ -2,13 +2,13 @@ package eu.dnetlib.dhp.sx.bio
import com.google.common.collect.Lists
import eu.dnetlib.dhp.schema.common.ModelConstants
import eu.dnetlib.dhp.schema.oaf.utils.{GraphCleaningFunctions, IdentifierFactory, OafMapperUtils}
import eu.dnetlib.dhp.schema.oaf._
import eu.dnetlib.dhp.schema.oaf.utils.{GraphCleaningFunctions, IdentifierFactory, OafMapperUtils}
import org.json4s.DefaultFormats
import org.json4s.JsonAST.{JField, JObject, JString}
import org.json4s.jackson.JsonMethods.{compact, parse, render}
import collection.JavaConverters._
import scala.collection.JavaConverters._
object BioDBToOAF {
@ -127,8 +127,8 @@ object BioDBToOAF {
target_pid_type,
generate_unresolved_id(source_pid, source_pid_type),
collectedFromMap("elsevier"),
"relationship",
relation_semantic,
Relation.SUBRELTYPE.relationship,
Relation.RELCLASS.lookUp(relation_semantic),
date
)
@ -323,8 +323,8 @@ object BioDBToOAF {
"pmid",
d.getId,
collectedFromMap("uniprot"),
ModelConstants.RELATIONSHIP,
ModelConstants.IS_RELATED_TO,
Relation.SUBRELTYPE.relationship,
Relation.RELCLASS.IsRelatedTo,
if (i_date.isDefined) i_date.get.date else null
)
rel.getProvenance.asScala.map(p => p.getCollectedfrom)
@ -335,8 +335,8 @@ object BioDBToOAF {
"doi",
d.getId,
collectedFromMap("uniprot"),
ModelConstants.RELATIONSHIP,
ModelConstants.IS_RELATED_TO,
Relation.SUBRELTYPE.relationship,
Relation.RELCLASS.IsRelatedTo,
if (i_date.isDefined) i_date.get.date else null
)
List(d, rel)
@ -353,8 +353,8 @@ object BioDBToOAF {
pidType: String,
sourceId: String,
collectedFrom: KeyValue,
subRelType: String,
relClass: String,
subRelType: Relation.SUBRELTYPE,
relClass: Relation.RELCLASS,
date: String
): Relation = {
@ -370,7 +370,7 @@ object BioDBToOAF {
rel.setProvenance(provenance)
rel.setRelType(ModelConstants.RESULT_RESULT)
rel.setRelType(Relation.RELTYPE.resultResult)
rel.setSubRelType(subRelType)
rel.setRelClass(relClass)
@ -398,10 +398,11 @@ object BioDBToOAF {
pidType,
sourceId,
collectedFrom,
ModelConstants.SUPPLEMENT,
ModelConstants.IS_SUPPLEMENT_TO,
Relation.SUBRELTYPE.supplement,
Relation.RELCLASS.IsSupplementTo,
date
)
}
def pdbTOOaf(input: String): List[Oaf] = {
@ -573,8 +574,8 @@ object BioDBToOAF {
"pmid",
d.getId,
collectedFromMap("ebi"),
ModelConstants.RELATIONSHIP,
ModelConstants.IS_RELATED_TO,
Relation.SUBRELTYPE.relationship,
Relation.RELCLASS.IsRelatedTo,
GraphCleaningFunctions.cleanDate(input.date)
)
)

2
dhp-workflows/dhp-aggregation/src/test/java/eu/dnetlib/dhp/actionmanager/opencitations/CreateOpenCitationsASTest.java
View File

@ -326,7 +326,7 @@ public class CreateOpenCitationsASTest {
});
assertEquals(5, check.filter(r -> r.getSource().equals(doi1)).count());
check.filter(r -> r.getSource().equals(doi1)).foreach(r -> assertEquals(ModelConstants.CITES, r.getRelClass()));
check.filter(r -> r.getSource().equals(doi1)).foreach(r -> assertEquals(Relation.RELCLASS.Cites, r.getRelClass()));
}
}

2
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/PrepareGroupsJob.java
View File

@ -65,7 +65,7 @@ public class PrepareGroupsJob {
final Dataset<Relation> mergedRels = ClusterUtils
.loadRelations(graphPath, spark)
.filter((FilterFunction<Relation>) r -> r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS));
.filter((FilterFunction<Relation>) r -> r.getRelClass().equals(Relation.RELCLASS.isMergedIn));
final TypedColumn<Tuple2<OaBrokerMainEntity, Relation>, ResultGroup> aggr = new ResultAggregator()
.toColumn();

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/PrepareRelatedDatasetsJob.java
View File

@ -68,7 +68,7 @@ public class PrepareRelatedDatasetsJob {
final Dataset<Relation> rels = ClusterUtils
.loadRelations(graphPath, spark)
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(ModelConstants.RESULT_RESULT))
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(Relation.RELTYPE.resultResult))
.filter((FilterFunction<Relation>) r -> ClusterUtils.isValidResultResultClass(r.getRelClass()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getSource()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getTarget()));
@ -78,7 +78,7 @@ public class PrepareRelatedDatasetsJob {
.map((MapFunction<Tuple2<Relation, OaBrokerRelatedDataset>, RelatedDataset>) t -> {
final RelatedDataset rel = new RelatedDataset(t._1.getSource(),
t._2);
rel.getRelDataset().setRelType(t._1.getRelClass());
rel.getRelDataset().setRelType(t._1.getRelClass().toString());
return rel;
}, Encoders.bean(RelatedDataset.class));

5
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/PrepareRelatedProjectsJob.java
View File

@ -68,10 +68,11 @@ public class PrepareRelatedProjectsJob {
(MapFunction<Project, OaBrokerProject>) ConversionUtils::oafProjectToBrokerProject,
Encoders.bean(OaBrokerProject.class));
final Dataset<Relation> rels = ClusterUtils
.loadRelations(graphPath, spark)
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(ModelConstants.RESULT_PROJECT))
.filter((FilterFunction<Relation>) r -> !r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS))
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(Relation.RELTYPE.resultProject))
.filter((FilterFunction<Relation>) r -> !r.getRelClass().equals(Relation.RELCLASS.isMergedIn))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getSource()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getTarget()));

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/PrepareRelatedPublicationsJob.java
View File

@ -69,7 +69,7 @@ public class PrepareRelatedPublicationsJob {
final Dataset<Relation> rels = ClusterUtils
.loadRelations(graphPath, spark)
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(ModelConstants.RESULT_RESULT))
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(Relation.RELTYPE.resultResult))
.filter((FilterFunction<Relation>) r -> ClusterUtils.isValidResultResultClass(r.getRelClass()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getSource()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getTarget()));
@ -79,7 +79,7 @@ public class PrepareRelatedPublicationsJob {
.map((MapFunction<Tuple2<Relation, OaBrokerRelatedPublication>, RelatedPublication>) t -> {
final RelatedPublication rel = new RelatedPublication(
t._1.getSource(), t._2);
rel.getRelPublication().setRelType(t._1.getRelClass());
rel.getRelPublication().setRelType(t._1.getRelClass().toString());
return rel;
}, Encoders.bean(RelatedPublication.class));

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/PrepareRelatedSoftwaresJob.java
View File

@ -73,8 +73,8 @@ public class PrepareRelatedSoftwaresJob {
final Dataset<Relation> rels;
rels = ClusterUtils
.loadRelations(graphPath, spark)
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(ModelConstants.RESULT_RESULT))
.filter((FilterFunction<Relation>) r -> !r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS))
.filter((FilterFunction<Relation>) r -> r.getRelType().equals(Relation.RELTYPE.resultResult))
.filter((FilterFunction<Relation>) r -> !r.getRelClass().equals(Relation.RELCLASS.isMergedIn))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getSource()))
.filter((FilterFunction<Relation>) r -> !ClusterUtils.isDedupRoot(r.getTarget()));

5
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedDatasets/EnrichMissingDatasetIsReferencedBy.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedDatasets;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingDatasetIsReferencedBy extends AbstractEnrichMissingDataset {
@ -12,7 +12,8 @@ public class EnrichMissingDatasetIsReferencedBy extends AbstractEnrichMissingDat
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_REFERENCED_BY);
return relType.equals(Relation.RELCLASS.IsReferencedBy);
}
}

5
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedDatasets/EnrichMissingDatasetIsRelatedTo.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedDatasets;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingDatasetIsRelatedTo extends AbstractEnrichMissingDataset {
@ -12,7 +12,8 @@ public class EnrichMissingDatasetIsRelatedTo extends AbstractEnrichMissingDatase
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_RELATED_TO);
return relType.equals(Relation.RELCLASS.IsRelatedTo);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedDatasets/EnrichMissingDatasetIsSupplementedBy.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedDatasets;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingDatasetIsSupplementedBy extends AbstractEnrichMissingDataset {
@ -12,7 +12,7 @@ public class EnrichMissingDatasetIsSupplementedBy extends AbstractEnrichMissingD
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_SUPPLEMENTED_BY);
return relType.equals(Relation.RELCLASS.IsSupplementedBy);
}
}

5
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedDatasets/EnrichMissingDatasetIsSupplementedTo.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedDatasets;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingDatasetIsSupplementedTo extends AbstractEnrichMissingDataset {
@ -12,7 +12,8 @@ public class EnrichMissingDatasetIsSupplementedTo extends AbstractEnrichMissingD
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_SUPPLEMENT_TO);
return relType.equals(Relation.RELCLASS.IsSupplementTo);
}
}

6
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedDatasets/EnrichMissingDatasetReferences.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedDatasets;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingDatasetReferences extends AbstractEnrichMissingDataset {
@ -12,7 +12,9 @@ public class EnrichMissingDatasetReferences extends AbstractEnrichMissingDataset
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.REFERENCES);
return relType.equals(Relation.RELCLASS.References);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedPublications/EnrichMissingPublicationIsReferencedBy.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedPublications;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingPublicationIsReferencedBy extends AbstractEnrichMissingPublication {
@ -12,6 +12,6 @@ public class EnrichMissingPublicationIsReferencedBy extends AbstractEnrichMissin
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_REFERENCED_BY);
return relType.equals(Relation.RELCLASS.IsReferencedBy);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedPublications/EnrichMissingPublicationIsRelatedTo.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedPublications;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingPublicationIsRelatedTo extends AbstractEnrichMissingPublication {
@ -12,7 +12,7 @@ public class EnrichMissingPublicationIsRelatedTo extends AbstractEnrichMissingPu
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_RELATED_TO);
return relType.equals(Relation.RELCLASS.IsRelatedTo);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedPublications/EnrichMissingPublicationIsSupplementedBy.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedPublications;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingPublicationIsSupplementedBy extends AbstractEnrichMissingPublication {
@ -12,6 +12,6 @@ public class EnrichMissingPublicationIsSupplementedBy extends AbstractEnrichMiss
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_SUPPLEMENTED_BY);
return relType.equals(Relation.RELCLASS.IsSupplementedBy);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedPublications/EnrichMissingPublicationIsSupplementedTo.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedPublications;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingPublicationIsSupplementedTo extends AbstractEnrichMissingPublication {
@ -12,7 +12,7 @@ public class EnrichMissingPublicationIsSupplementedTo extends AbstractEnrichMiss
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.IS_SUPPLEMENT_TO);
return relType.equals(Relation.RELCLASS.IsSupplementTo);
}
}

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/matchers/relatedPublications/EnrichMissingPublicationReferences.java
View File

@ -2,7 +2,7 @@
package eu.dnetlib.dhp.broker.oa.matchers.relatedPublications;
import eu.dnetlib.dhp.broker.model.Topic;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
public class EnrichMissingPublicationReferences extends AbstractEnrichMissingPublication {
@ -12,7 +12,7 @@ public class EnrichMissingPublicationReferences extends AbstractEnrichMissingPub
@Override
protected boolean filterByType(final String relType) {
return relType.equals(ModelConstants.REFERENCES);
return relType.equals(Relation.RELCLASS.References);
}
}

1
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/util/BrokerConstants.java
View File

@ -16,7 +16,6 @@ public class BrokerConstants {
}
public static final String OPEN_ACCESS = "OPEN";
public static final String IS_MERGED_IN_CLASS = ModelConstants.IS_MERGED_IN;
public static final String COLLECTED_FROM_REL = "collectedFrom";

29
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/util/ClusterUtils.java
View File

@ -1,10 +1,10 @@
package eu.dnetlib.dhp.broker.oa.util;
import java.util.Arrays;
import java.util.HashSet;
import java.util.Set;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.schema.oaf.Relation;
import org.apache.commons.lang3.StringUtils;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.sql.Dataset;
@ -13,12 +13,9 @@ import org.apache.spark.sql.SaveMode;
import org.apache.spark.sql.SparkSession;
import org.apache.spark.util.LongAccumulator;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.HdfsSupport;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
import java.util.Arrays;
import java.util.HashSet;
import java.util.Set;
public class ClusterUtils {
@ -59,12 +56,12 @@ public class ClusterUtils {
return id.contains("dedup");
}
public static final boolean isValidResultResultClass(final String s) {
return s.equals(ModelConstants.IS_REFERENCED_BY)
|| s.equals(ModelConstants.IS_RELATED_TO)
|| s.equals(ModelConstants.REFERENCES)
|| s.equals(ModelConstants.IS_SUPPLEMENTED_BY)
|| s.equals(ModelConstants.IS_SUPPLEMENT_TO);
public static final boolean isValidResultResultClass(final Relation.RELCLASS r) {
return r.equals(Relation.RELCLASS.IsReferencedBy)
|| r.equals(Relation.RELCLASS.References)
|| r.equals(Relation.RELCLASS.IsRelatedTo)
|| r.equals(Relation.RELCLASS.IsSupplementTo)
|| r.equals(Relation.RELCLASS.IsSupplementedBy);
}
public static <T> T incrementAccumulator(final T o, final LongAccumulator acc) {

4
dhp-workflows/dhp-broker-events/src/main/java/eu/dnetlib/dhp/broker/oa/util/ConversionUtils.java
View File

@ -100,7 +100,7 @@ public class ConversionUtils {
res.setOpenaireId(cleanOpenaireId(result.getId()));
res.setOriginalId(first(result.getOriginalId()));
res.setTypology(result.getResulttype());
res.setTypology(result.getResulttype().toString());
res.setTitles(structPropList(result.getTitle()));
res.setAbstracts(result.getDescription());
res.setLanguage(classId(result.getLanguage()));
@ -112,7 +112,7 @@ public class ConversionUtils {
res.setContributor(result.getContributor());
res
.setJournal(
result instanceof Publication ? oafJournalToBrokerJournal(((Publication) result).getJournal()) : null);
result instanceof Publication ? oafJournalToBrokerJournal(result.getJournal()) : null);
res.setPids(allResultPids(result));
res.setInstances(flatMappedList(result.getInstance(), ConversionUtils::oafInstanceToBrokerInstances));
res

10
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/AbstractSparkAction.java
View File

@ -153,10 +153,12 @@ abstract class AbstractSparkAction implements Serializable {
}
private boolean isOpenOrgsDedupMergeRelation(Relation rel) {
return ModelConstants.ORG_ORG_RELTYPE.equals(rel.getRelType()) &&
ModelConstants.DEDUP.equals(rel.getSubRelType())
&& (ModelConstants.IS_MERGED_IN.equals(rel.getRelClass()) ||
ModelConstants.MERGES.equals(rel.getRelClass()));
return Relation.RELTYPE.organizationOrganization.equals(rel.getRelType()) &&
Relation.SUBRELTYPE.dedup.equals(rel.getSubRelType())
&& (Relation.RELCLASS.isMergedIn.equals(rel.getRelClass()) ||
Relation.RELCLASS.merges.equals(rel.getRelClass()));
}
protected static Boolean parseECField(String field) {

9
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/DedupUtility.java
View File

@ -157,17 +157,18 @@ public class DedupUtility {
public static Relation createSimRel(String source, String target, String entity) {
final Relation r = new Relation();
r.setSource(source);
r.setTarget(target);
r.setSubRelType("dedupSimilarity");
r.setRelClass(ModelConstants.IS_SIMILAR_TO);
r.setSubRelType(Relation.SUBRELTYPE.dedup);
r.setRelClass(Relation.RELCLASS.isSimilarTo);
switch (entity) {
case "result":
r.setRelType(ModelConstants.RESULT_RESULT);
r.setRelType(Relation.RELTYPE.resultResult);
break;
case "organization":
r.setRelType(ModelConstants.ORG_ORG_RELTYPE);
r.setRelType(Relation.RELTYPE.organizationOrganization);
break;
default:
throw new IllegalArgumentException("unmanaged entity type: " + entity);

8
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkCopyOpenorgsMergeRels.java
View File

@ -90,9 +90,9 @@ public class SparkCopyOpenorgsMergeRels extends AbstractSparkAction {
}
private boolean isMergeRel(Relation rel) {
return (rel.getRelClass().equals(ModelConstants.MERGES)
|| rel.getRelClass().equals(ModelConstants.IS_MERGED_IN))
&& rel.getRelType().equals(ModelConstants.ORG_ORG_RELTYPE)
&& rel.getSubRelType().equals(ModelConstants.DEDUP);
return (rel.getRelClass().equals(Relation.RELCLASS.merges)
|| rel.getRelClass().equals(Relation.RELCLASS.isMergedIn))
&& rel.getRelType().equals(Relation.RELTYPE.organizationOrganization)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup);
}
}

6
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkCopyOpenorgsSimRels.java
View File

@ -81,9 +81,9 @@ public class SparkCopyOpenorgsSimRels extends AbstractSparkAction {
}
private boolean filterOpenorgsRels(Relation rel) {
return rel.getRelClass().equals(ModelConstants.IS_SIMILAR_TO)
&& rel.getRelType().equals(ModelConstants.ORG_ORG_RELTYPE)
&& rel.getSubRelType().equals(ModelConstants.DEDUP) && isOpenorgs(rel);
return rel.getRelClass().equals(Relation.RELCLASS.isSimilarTo)
&& rel.getRelType().equals(Relation.RELTYPE.organizationOrganization)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup) && isOpenorgs(rel);
}
}

11
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkCreateMergeRels.java
View File

@ -110,7 +110,7 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
.load(DedupUtility.createSimRelPath(workingPath, actionSetId, subEntity))
.as(Encoders.bean(Relation.class))
.javaRDD()
.map(it -> new Edge<>(hash(it.getSource()), hash(it.getTarget()), it.getRelClass()))
.map(it -> new Edge<>(hash(it.getSource()), hash(it.getTarget()), it.getRelClass().toString()))
.rdd();
Dataset<Tuple2<String, String>> rawMergeRels = spark
@ -199,14 +199,15 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
id -> {
List<Relation> rels = new ArrayList<>();
rels.add(rel(cc.getCcId(), id, ModelConstants.MERGES, dedupConf));
rels.add(rel(cc.getCcId(), id, Relation.RELCLASS.merges, dedupConf));
return rels.stream();
})
.iterator();
}
private Relation rel(String source, String target, String relClass, DedupConfig dedupConf) {
// TODO NEED to REVIEW THIS FUNCTION, THE UTILITY FUNCTION SHOULD BE MOVED ON SOME SUPPORT CLASS OR REUSE OTHER FUNCTION
private Relation rel(String source, String target, Relation.RELCLASS relClass, DedupConfig dedupConf) {
String entityType = dedupConf.getWf().getEntityType();
@ -214,8 +215,8 @@ public class SparkCreateMergeRels extends AbstractSparkAction {
r.setSource(source);
r.setTarget(target);
r.setRelClass(relClass);
r.setRelType(entityType + entityType.substring(0, 1).toUpperCase() + entityType.substring(1));
r.setSubRelType(ModelConstants.DEDUP);
r.setRelType(Relation.RELTYPE.valueOf(entityType + entityType.substring(0, 1).toUpperCase() + entityType.substring(1)));
r.setSubRelType(Relation.SUBRELTYPE.dedup);
DataInfo info = new DataInfo();

12
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkPrepareNewOrgs.java
View File

@ -191,15 +191,15 @@ public class SparkPrepareNewOrgs extends AbstractSparkAction {
switch (entityType) {
case "result":
if (rel.getRelClass().equals(ModelConstants.IS_DIFFERENT_FROM)
&& rel.getRelType().equals(ModelConstants.RESULT_RESULT)
&& rel.getSubRelType().equals(ModelConstants.DEDUP))
if (rel.getRelClass().equals(Relation.RELCLASS.isDifferentFrom)
&& rel.getRelType().equals(Relation.RELTYPE.resultResult)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup))
return true;
break;
case "organization":
if (rel.getRelClass().equals(ModelConstants.IS_DIFFERENT_FROM)
&& rel.getRelType().equals(ModelConstants.ORG_ORG_RELTYPE)
&& rel.getSubRelType().equals(ModelConstants.DEDUP))
if (rel.getRelClass().equals(Relation.RELCLASS.isDifferentFrom)
&& rel.getRelType().equals(Relation.RELTYPE.organizationOrganization)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup))
return true;
break;
default:

12
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkPrepareOrgRels.java
View File

@ -106,15 +106,15 @@ public class SparkPrepareOrgRels extends AbstractSparkAction {
switch (entityType) {
case "result":
if (rel.getRelClass().equals(ModelConstants.IS_DIFFERENT_FROM)
&& rel.getRelType().equals(ModelConstants.RESULT_RESULT)
&& rel.getSubRelType().equals(ModelConstants.DEDUP))
if (rel.getRelClass().equals(Relation.RELCLASS.isDifferentFrom)
&& rel.getRelType().equals(Relation.RELTYPE.resultResult)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup))
return true;
break;
case "organization":
if (rel.getRelClass().equals(ModelConstants.IS_DIFFERENT_FROM)
&& rel.getRelType().equals(ModelConstants.ORG_ORG_RELTYPE)
&& rel.getSubRelType().equals(ModelConstants.DEDUP))
if (rel.getRelClass().equals(Relation.RELCLASS.isDifferentFrom)
&& rel.getRelType().equals(Relation.RELTYPE.organizationOrganization)
&& rel.getSubRelType().equals(Relation.SUBRELTYPE.dedup))
return true;
break;
default:

10
dhp-workflows/dhp-dedup-openaire/src/main/java/eu/dnetlib/dhp/oa/dedup/SparkPropagateRelation.java
View File

@ -74,7 +74,7 @@ public class SparkPropagateRelation extends AbstractSparkAction {
// <mergedObjectID, dedupID>
Dataset<Tuple2<String, String>> mergedIds = mergeRels
.where(col("relClass").equalTo(ModelConstants.MERGES))
.where(col("relClass").equalTo(Relation.RELCLASS.merges))
.select(col("source"), col("target"))
.distinct()
.map(
@ -111,7 +111,7 @@ public class SparkPropagateRelation extends AbstractSparkAction {
.filter(getRelationFilterFunction())
.groupByKey(
(MapFunction<Relation, String>) r -> String
.join(r.getSource(), r.getTarget(), r.getRelType(), r.getSubRelType(), r.getRelClass()),
.join(r.getSource(), r.getTarget(), r.getRelType().toString(), r.getSubRelType().toString(), r.getRelClass().toString()),
Encoders.STRING())
.agg(new RelationAggregator().toColumn())
.map((MapFunction<Tuple2<String, Relation>, Relation>) Tuple2::_2, Encoders.bean(Relation.class));
@ -150,9 +150,9 @@ public class SparkPropagateRelation extends AbstractSparkAction {
private FilterFunction<Relation> getRelationFilterFunction() {
return r -> StringUtils.isNotBlank(r.getSource()) ||
StringUtils.isNotBlank(r.getTarget()) ||
StringUtils.isNotBlank(r.getRelType()) ||
StringUtils.isNotBlank(r.getSubRelType()) ||
StringUtils.isNotBlank(r.getRelClass());
r.getRelType() != null ||
r.getSubRelType()!=null ||
r.getRelClass()!=null;
}
private static String getId(Relation r, FieldType type) {

12
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkDedupTest.java
View File

@ -415,9 +415,9 @@ public class SparkDedupTest implements Serializable {
"50|doi_________::d5021b53204e4fdeab6ff5d5bc468032",
"50|arXiv_______::c93aeb433eb90ed7a86e29be00791b7c");
merges.forEach(r -> {
assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
assertEquals(ModelConstants.DEDUP, r.getSubRelType());
assertEquals(ModelConstants.MERGES, r.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r.getRelType());
assertEquals(Relation.SUBRELTYPE.dedup, r.getSubRelType());
assertEquals(Relation.RELCLASS.merges, r.getRelClass());
assertTrue(dups.contains(r.getTarget()));
});
@ -426,9 +426,9 @@ public class SparkDedupTest implements Serializable {
.collectAsList();
assertEquals(3, mergedIn.size());
mergedIn.forEach(r -> {
assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
assertEquals(ModelConstants.DEDUP, r.getSubRelType());
assertEquals(ModelConstants.IS_MERGED_IN, r.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r.getRelType());
assertEquals(Relation.SUBRELTYPE.dedup, r.getSubRelType());
assertEquals(Relation.RELCLASS.isMergedIn, r.getRelClass());
assertTrue(dups.contains(r.getSource()));
});

12
dhp-workflows/dhp-dedup-openaire/src/test/java/eu/dnetlib/dhp/oa/dedup/SparkPublicationRootsTest.java
View File

@ -190,9 +190,9 @@ public class SparkPublicationRootsTest implements Serializable {
"50|doi_________::d5021b53204e4fdeab6ff5d5bc468032",
"50|arXiv_______::c93aeb433eb90ed7a86e29be00791b7c");
mergeList.forEach(r -> {
assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
assertEquals(ModelConstants.DEDUP, r.getSubRelType());
assertEquals(ModelConstants.MERGES, r.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r.getRelType());
assertEquals(Relation.SUBRELTYPE.dedup, r.getSubRelType());
assertEquals(Relation.RELCLASS.merges, r.getRelClass());
assertTrue(dups.contains(r.getTarget()));
});
@ -201,9 +201,9 @@ public class SparkPublicationRootsTest implements Serializable {
.collectAsList();
assertEquals(3, mergedIn.size());
mergedIn.forEach(r -> {
assertEquals(ModelConstants.RESULT_RESULT, r.getRelType());
assertEquals(ModelConstants.DEDUP, r.getSubRelType());
assertEquals(ModelConstants.IS_MERGED_IN, r.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r.getRelType());
assertEquals(Relation.SUBRELTYPE.dedup, r.getSubRelType());
assertEquals(Relation.RELCLASS.isMergedIn, r.getRelClass());
assertTrue(dups.contains(r.getSource()));
});

49
dhp-workflows/dhp-doiboost/src/main/scala/eu/dnetlib/doiboost/crossref/Crossref2Oaf.scala
View File

@ -384,9 +384,9 @@ case object Crossref2Oaf {
val rel = new Relation
rel.setSource(sourceId)
rel.setTarget(targetId)
rel.setRelType(ModelConstants.RESULT_RESULT)
rel.setRelClass(ModelConstants.CITES)
rel.setSubRelType(ModelConstants.CITATION)
rel.setRelType(Relation.RELTYPE.resultResult)
rel.setRelClass(Relation.RELCLASS.Cites)
rel.setSubRelType(Relation.SUBRELTYPE.citation)
rel.setProvenance(Lists.newArrayList(OafMapperUtils.getProvenance(collectedFrom, dataInfo)))
List(rel)
@ -417,14 +417,14 @@ case object Crossref2Oaf {
null
}
def generateRelation(sourceId: String, targetId: String, relClass: String): Relation = {
def generateRelation(sourceId: String, targetId: String, relClass: Relation.RELCLASS): Relation = {
val r = new Relation
r.setSource(sourceId)
r.setTarget(targetId)
r.setRelType(ModelConstants.RESULT_PROJECT)
r.setRelType(Relation.RELTYPE.resultProject)
r.setRelClass(relClass)
r.setSubRelType(ModelConstants.OUTCOME)
r.setSubRelType(Relation.SUBRELTYPE.outcome)
r.setProvenance(Lists.newArrayList(OafMapperUtils.getProvenance(collectedFrom, dataInfo)))
r
@ -435,14 +435,15 @@ case object Crossref2Oaf {
nsPrefix: String,
extractField: String => String
): Unit = {
if (funder.award.isDefined && funder.award.get.nonEmpty)
funder.award.get
.map(extractField)
.filter(a => a != null && a.nonEmpty)
.foreach(award => {
val targetId = getProjectId(nsPrefix, DHPUtils.md5(award))
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
})
}
@ -471,21 +472,21 @@ case object Crossref2Oaf {
case "10.13039/501100000923" => generateSimpleRelationFromAward(funder, "arc_________", a => a)
case "10.13039/501100000038" =>
val targetId = getProjectId("nserc_______", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case "10.13039/501100000155" =>
val targetId = getProjectId("sshrc_______", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case "10.13039/501100000024" =>
val targetId = getProjectId("cihr________", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case "10.13039/100020031" =>
val targetId = getProjectId("tara________", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case "10.13039/501100005416" => generateSimpleRelationFromAward(funder, "rcn_________", a => a)
case "10.13039/501100002848" => generateSimpleRelationFromAward(funder, "conicytf____", a => a)
@ -495,8 +496,8 @@ case object Crossref2Oaf {
case "10.13039/501100003407" =>
generateSimpleRelationFromAward(funder, "miur________", a => a)
val targetId = getProjectId("miur________", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case "10.13039/501100006588" | "10.13039/501100004488" =>
generateSimpleRelationFromAward(
funder,
@ -509,15 +510,15 @@ case object Crossref2Oaf {
case "10.13039/100004440" =>
generateSimpleRelationFromAward(funder, "wt__________", a => a)
val targetId = getProjectId("wt__________", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
//ASAP
case "10.13039/100018231" => generateSimpleRelationFromAward(funder, "asap________", a => a)
//CHIST-ERA
case "10.13039/501100001942" =>
val targetId = getProjectId("chistera____", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
//HE
case "10.13039/100018693" | "10.13039/100018694" | "10.13039/100019188" | "10.13039/100019180" |
"10.13039/100018695" | "10.13039/100019185" | "10.13039/100019186" | "10.13039/100019187" =>
@ -559,8 +560,8 @@ case object Crossref2Oaf {
case "Wellcome Trust Masters Fellowship" =>
generateSimpleRelationFromAward(funder, "wt__________", a => a)
val targetId = getProjectId("wt__________", "1e5e62235d094afd01cd56e65112fc63")
queue += generateRelation(sourceId, targetId, ModelConstants.IS_PRODUCED_BY)
queue += generateRelation(targetId, sourceId, ModelConstants.PRODUCES)
queue += generateRelation(sourceId, targetId, Relation.RELCLASS.isProducedBy)
queue += generateRelation(targetId, sourceId, Relation.RELCLASS.produces)
case _ => logger.debug("no match for " + funder.name)
}

2
dhp-workflows/dhp-doiboost/src/test/scala/eu/dnetlib/dhp/doiboost/crossref/CrossrefMappingTest.scala
View File

@ -144,7 +144,7 @@ class CrossrefMappingTest {
val relationList: List[Relation] = result
.filter(s => s.isInstanceOf[Relation])
.map(r => r.asInstanceOf[Relation])
.filter(r => r.getSubRelType.equalsIgnoreCase(ModelConstants.CITATION))
.filter(r => r.getSubRelType.equalsIgnoreCase(Relation.SUBRELTYPE.citation))
assertNotNull(relationList)
assertFalse(relationList.isEmpty)

19
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/PropagationConstant.java
View File

@ -112,15 +112,16 @@ public class PropagationConstant {
String className
) {
ArrayList<Relation> newRelations = new ArrayList();
newRelations
.add(
getRelation(
orgId,
resultId,
ModelConstants.IS_AUTHOR_INSTITUTION_OF,
ModelConstants.RESULT_ORGANIZATION,
ModelConstants.AFFILIATION,
Relation.RELCLASS.isAuthorInstitutionOf,
Relation.RELTYPE.resultOrganization,
Relation.SUBRELTYPE.affiliation,
PROPAGATION_DATA_INFO_TYPE,
classID,
className));
@ -129,9 +130,9 @@ public class PropagationConstant {
getRelation(
resultId,
orgId,
ModelConstants.HAS_AUTHOR_INSTITUTION,
ModelConstants.RESULT_ORGANIZATION,
ModelConstants.AFFILIATION,
Relation.RELCLASS.hasAuthorInstitution,
Relation.RELTYPE.resultOrganization,
Relation.SUBRELTYPE.affiliation,
PROPAGATION_DATA_INFO_TYPE,
classID,
className));
@ -142,9 +143,9 @@ public class PropagationConstant {
public static Relation getRelation(
String source,
String target,
String rel_class,
String rel_type,
String subrel_type,
Relation.RELCLASS rel_class,
Relation.RELTYPE rel_type,
Relation.SUBRELTYPE subrel_type,
String inference_provenance,
String inference_class_id,
String inference_class_name) {

2
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/PrepareDatasourceCountryAssociation.java
View File

@ -96,7 +96,7 @@ public class PrepareDatasourceCountryAssociation {
// filtering of the relations taking the non deleted by inference and those with IsProvidedBy as relclass
Dataset<Relation> relation = readPath(spark, inputPath + "/relation", Relation.class)
.filter(
(FilterFunction<Relation>) rel -> rel.getRelClass().equalsIgnoreCase(ModelConstants.IS_PROVIDED_BY));
(FilterFunction<Relation>) rel -> rel.getRelClass() == Relation.RELCLASS.isProvidedBy);
// filtering of the organization taking only the non deleted by inference and those with information about the
// country

2
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/projecttoresult/PrepareProjectResultsAssociation.java
View File

@ -81,7 +81,7 @@ public class PrepareProjectResultsAssociation {
+ " FROM relation "
+ " WHERE datainfo.deletedbyinference = false "
+ " AND lower(relClass) = '"
+ ModelConstants.IS_PRODUCED_BY.toLowerCase()
+ Relation.RELCLASS.isProducedBy.toString().toLowerCase()
+ "'";
Dataset<Row> resproj_relation = spark.sql(resproj_relation_query);

6
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/projecttoresult/SparkResultToProjectThroughSemRelJob.java
View File

@ -112,9 +112,9 @@ public class SparkResultToProjectThroughSemRelJob {
getRelation(
projectId,
resId,
ModelConstants.PRODUCES,
ModelConstants.RESULT_PROJECT,
ModelConstants.OUTCOME,
Relation.RELCLASS.produces,
Relation.RELTYPE.resultProject,
Relation.SUBRELTYPE.outcome,
PROPAGATION_DATA_INFO_TYPE,
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_ID,
PROPAGATION_RELATION_RESULT_PROJECT_SEM_REL_CLASS_NAME)));

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttocommunityfromorganization/PrepareResultCommunitySet.java
View File

@ -78,13 +78,13 @@ public class PrepareResultCommunitySet {
+ " FROM relation "
+ " WHERE datainfo.deletedbyinference = false "
+ " AND lower(relClass) = '"
+ ModelConstants.HAS_AUTHOR_INSTITUTION.toLowerCase()
+ Relation.RELCLASS.hasAuthorInstitution.toString().toLowerCase()
+ "') result_organization "
+ "LEFT JOIN (SELECT source, collect_set(target) org_set "
+ " FROM relation "
+ " WHERE datainfo.deletedbyinference = false "
+ " AND lower(relClass) = '"
+ ModelConstants.MERGES.toLowerCase()
+ Relation.RELCLASS.merges.toString().toLowerCase()
+ "' "
+ " GROUP BY source) organization_organization "
+ "ON result_organization.target = organization_organization.source ";

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttoorganizationfrominstrepo/PrepareResultInstRepoAssociation.java
View File

@ -109,7 +109,7 @@ public class PrepareResultInstRepoAssociation {
+ "JOIN ( SELECT source, target "
+ "FROM relation "
+ "WHERE lower(relclass) = '"
+ ModelConstants.IS_PROVIDED_BY.toLowerCase()
+ Relation.RELCLASS.isProvidedBy.toString().toLowerCase()
+ "' "
+ "AND datainfo.deletedbyinference = false ) rel "
+ "ON d.id = rel.source ";
@ -129,7 +129,7 @@ public class PrepareResultInstRepoAssociation {
+ "from relation "
+ "where datainfo.deletedbyinference = false "
+ "and lower(relClass) = '"
+ ModelConstants.HAS_AUTHOR_INSTITUTION.toLowerCase()
+ Relation.RELCLASS.hasAuthorInstitution.toString().toLowerCase()
+ "' "
+ "group by source";

6
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttoorganizationfrominstrepo/SparkResultToOrganizationFromIstRepoJob.java
View File

@ -142,9 +142,9 @@ public class SparkResultToOrganizationFromIstRepoJob {
.add(
getRelation(
resultId, orgId,
ModelConstants.HAS_AUTHOR_INSTITUTION,
ModelConstants.RESULT_ORGANIZATION,
ModelConstants.AFFILIATION, PROPAGATION_DATA_INFO_TYPE,
Relation.RELCLASS.hasAuthorInstitution,
Relation.RELTYPE.resultOrganization,
Relation.SUBRELTYPE.affiliation, PROPAGATION_DATA_INFO_TYPE,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_ID,
PROPAGATION_RELATION_RESULT_ORGANIZATION_INST_REPO_CLASS_NAME))

10
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttoorganizationfromsemrel/PrepareInfo.java
View File

@ -36,14 +36,14 @@ public class PrepareInfo implements Serializable {
private static final String ORGANIZATION_ORGANIZATION_QUERY = "SELECT target key, collect_set(source) as valueSet "
+
"FROM relation " +
"WHERE lower(relclass) = '" + ModelConstants.IS_PARENT_OF.toLowerCase() +
"WHERE lower(relclass) = '" + Relation.RELCLASS.IsParentOf.toString().toLowerCase() +
"' and datainfo.deletedbyinference = false " +
"GROUP BY target";
// associates results with all the orgs they are affiliated to
private static final String RESULT_ORGANIZATION_QUERY = "SELECT source key, collect_set(target) as valueSet " +
"FROM relation " +
"WHERE lower(relclass) = '" + ModelConstants.HAS_AUTHOR_INSTITUTION.toLowerCase() +
"WHERE lower(relclass) = '" + Relation.RELCLASS.hasAuthorInstitution.toString().toLowerCase() +
"' and datainfo.deletedbyinference = false " +
"GROUP BY source";
@ -115,7 +115,7 @@ public class PrepareInfo implements Serializable {
relation
.filter(
(FilterFunction<Relation>) r -> r.getRelClass().equals(ModelConstants.HAS_AUTHOR_INSTITUTION))
(FilterFunction<Relation>) r -> r.getRelClass().equals(Relation.RELCLASS.hasAuthorInstitution))
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
@ -124,14 +124,14 @@ public class PrepareInfo implements Serializable {
Dataset<String> children = spark
.sql(
"Select distinct target as child from relation where " +
"lower(relclass)='" + ModelConstants.IS_PARENT_OF.toLowerCase() +
"lower(relclass)='" + Relation.RELCLASS.IsParentOf.toString().toLowerCase() +
"' and datainfo.deletedbyinference = false")
.as(Encoders.STRING());
Dataset<String> parent = spark
.sql(
"Select distinct source as parent from relation " +
"where lower(relclass)='" + ModelConstants.IS_PARENT_OF.toLowerCase() +
"where lower(relclass)='" + Relation.RELCLASS.IsParentOf.toString().toLowerCase() +
"' and datainfo.deletedbyinference = false")
.as(Encoders.STRING());

6
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/resulttoorganizationfromsemrel/StepActions.java
View File

@ -193,9 +193,9 @@ public class StepActions implements Serializable {
orgId -> getRelation(
v.getKey(),
orgId,
ModelConstants.HAS_AUTHOR_INSTITUTION,
ModelConstants.RESULT_ORGANIZATION,
ModelConstants.AFFILIATION,
Relation.RELCLASS.hasAuthorInstitution,
Relation.RELTYPE.resultOrganization,
Relation.SUBRELTYPE.affiliation,
PROPAGATION_DATA_INFO_TYPE,
PROPAGATION_RELATION_RESULT_ORGANIZATION_SEM_REL_CLASS_ID,
PROPAGATION_RELATION_RESULT_ORGANIZATION_SEM_REL_CLASS_NAME))

8
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/resulttoorganizationfromsemrel/SparkJobTest.java
View File

@ -119,8 +119,8 @@ public class SparkJobTest {
tmp.foreach(r -> System.out.println(OBJECT_MAPPER.writeValueAsString(r)));
Assertions.assertEquals(18, tmp.count());
tmp.foreach(r -> Assertions.assertEquals(ModelConstants.AFFILIATION, r.getSubRelType()));
tmp.foreach(r -> Assertions.assertEquals(ModelConstants.RESULT_ORGANIZATION, r.getRelType()));
tmp.foreach(r -> Assertions.assertEquals(Relation.SUBRELTYPE.affiliation, r.getSubRelType()));
tmp.foreach(r -> Assertions.assertEquals(Relation.RELTYPE.resultOrganization, r.getRelType()));
tmp
.foreach(
r -> Assertions
@ -148,7 +148,7 @@ public class SparkJobTest {
Assertions.assertEquals(9, tmp.filter(r -> r.getSource().substring(0, 3).equals("50|")).count());
tmp
.filter(r -> r.getSource().substring(0, 3).equals("50|"))
.foreach(r -> Assertions.assertEquals(ModelConstants.HAS_AUTHOR_INSTITUTION, r.getRelClass()));
.foreach(r -> Assertions.assertEquals(Relation.RELCLASS.hasAuthorInstitution, r.getRelClass()));
Assertions
.assertEquals(
2, tmp.filter(r -> r.getSource().equals("50|doajarticles::1cae0b82b56ccd97c2db1f698def7074")).count());
@ -168,7 +168,7 @@ public class SparkJobTest {
Assertions.assertEquals(9, tmp.filter(r -> r.getSource().substring(0, 3).equals("20|")).count());
tmp
.filter(r -> r.getSource().substring(0, 3).equals("20|"))
.foreach(r -> Assertions.assertEquals(ModelConstants.IS_AUTHOR_INSTITUTION_OF, r.getRelClass()));
.foreach(r -> Assertions.assertEquals(Relation.RELCLASS.isAuthorInstitutionOf, r.getRelClass()));
Assertions
.assertEquals(
1, tmp.filter(r -> r.getSource().equals("20|doajarticles::1cae0b82b56ccd97c2db1f698def7074")).count());

6
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/resulttoorganizationfromsemrel/StepActionsTest.java
View File

@ -117,17 +117,17 @@ public class StepActionsTest {
verificationDs
.foreach(
(ForeachFunction<Relation>) r -> Assertions
.assertEquals(ModelConstants.HAS_AUTHOR_INSTITUTION, r.getRelClass()));
.assertEquals(Relation.RELCLASS.hasAuthorInstitution, r.getRelClass()));
verificationDs
.foreach(
(ForeachFunction<Relation>) r -> Assertions
.assertEquals(ModelConstants.RESULT_ORGANIZATION, r.getRelType()));
.assertEquals(Relation.RELTYPE.resultOrganization, r.getRelType()));
verificationDs
.foreach(
(ForeachFunction<Relation>) r -> Assertions
.assertEquals(ModelConstants.AFFILIATION, r.getSubRelType()));
.assertEquals(Relation.SUBRELTYPE.affiliation, r.getSubRelType()));
verificationDs
.foreach(

2
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/GetDatasourceFromCountry.java
View File

@ -89,7 +89,7 @@ public class GetDatasourceFromCountry implements Serializable {
(MapFunction<String, Relation>) value -> OBJECT_MAPPER.readValue(value, Relation.class),
Encoders.bean(Relation.class))
.filter(
(FilterFunction<Relation>) rel -> rel.getRelClass().equalsIgnoreCase(ModelConstants.IS_PROVIDED_BY));
(FilterFunction<Relation>) rel -> rel.getRelClass().equalsIgnoreCase(Relation.RELCLASS.isProvidedBy));
organization
.joinWith(relation, organization.col("id").equalTo(relation.col("target")))

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

@ -1,33 +1,26 @@
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.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.oaf.utils.OafMapperUtils.*;
import static eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory.*;
import java.util.*;
import java.util.stream.Collectors;
import com.google.common.collect.Lists;
import com.google.common.collect.Sets;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Entity;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory;
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.validator.routines.UrlValidator;
import org.dom4j.*;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.google.common.collect.Lists;
import com.google.common.collect.Sets;
import java.util.*;
import java.util.stream.Collectors;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory;
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils;
import static eu.dnetlib.dhp.schema.common.ModelConstants.*;
import static eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory.createOpenaireId;
import static eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils.*;
public abstract class AbstractMdRecordToOafMapper {
@ -261,7 +254,7 @@ public abstract class AbstractMdRecordToOafMapper {
res
.add(
OafMapperUtils
.getRelation(projectId, docId, RESULT_PROJECT, OUTCOME, PRODUCES, entity, validationdDate));
.getRelation(projectId, docId, Relation.RELTYPE.resultProject, Relation.SUBRELTYPE.outcome, Relation.RELCLASS.produces, entity, validationdDate));
}
}
@ -276,12 +269,11 @@ public abstract class AbstractMdRecordToOafMapper {
Element element = (Element) o;
final String target = StringUtils.trim(element.getText());
final String relType = element.attributeValue("relType");
final String subRelType = element.attributeValue("subRelType");
final String relClass = element.attributeValue("relClass");
final Relation.RELTYPE relType = Relation.RELTYPE.valueOf(element.attributeValue("relType"));
final Relation.SUBRELTYPE subRelType = Relation.SUBRELTYPE.valueOf(element.attributeValue("subRelType"));
final Relation.RELCLASS relClass = Relation.RELCLASS.lookUp(element.attributeValue("relClass"));
if (StringUtils.isNotBlank(target) && StringUtils.isNotBlank(relType) && StringUtils.isNotBlank(subRelType)
&& StringUtils.isNotBlank(relClass)) {
if (StringUtils.isNotBlank(target)) {
final String validationdDate = ((Node) o).valueOf("@validationDate");

2
dhp-workflows/dhp-graph-mapper/src/main/scala/eu/dnetlib/dhp/oa/graph/resolution/SparkResolveEntities.scala
View File

@ -90,6 +90,8 @@ object SparkResolveEntities {
case EntityType.dataset => mapper.readValue(input, classOf[OafDataset])
case EntityType.software => mapper.readValue(input, classOf[Software])
case EntityType.otherresearchproduct => mapper.readValue(input, classOf[OtherResearchProduct])
case _ => throw new IllegalArgumentException(s"Unexpected entity type $entity")
}
}

22
dhp-workflows/dhp-graph-mapper/src/main/scala/eu/dnetlib/dhp/sx/graph/SparkConvertRDDtoDataset.scala
View File

@ -104,24 +104,12 @@ object SparkConvertRDDtoDataset {
log.info("Converting Relation")
val relClassFilter = List(
ModelConstants.MERGES,
ModelConstants.IS_MERGED_IN,
ModelConstants.HAS_AMONG_TOP_N_SIMILAR_DOCS,
ModelConstants.IS_AMONG_TOP_N_SIMILAR_DOCS
)
val rddRelation = spark.sparkContext
.textFile(s"$sourcePath/relation")
.map(s => mapper.readValue(s, classOf[Relation]))
.filter(r => r.getSource.startsWith("50") && r.getTarget.startsWith("50"))
.filter(r => filterRelations(r))
//filter OpenCitations relations
// .filter(r =>
// r.getDataInfo.getProvenanceaction != null &&
// !"sysimport:crosswalk:opencitations".equals(r.getDataInfo.getProvenanceaction.getClassid)
// )
spark.createDataset(rddRelation).as[Relation].write.mode(SaveMode.Overwrite).save(s"$relPath")
}
@ -133,12 +121,12 @@ object SparkConvertRDDtoDataset {
*/
val relClassFilter = List(
ModelConstants.MERGES,
ModelConstants.IS_MERGED_IN,
ModelConstants.HAS_AMONG_TOP_N_SIMILAR_DOCS,
ModelConstants.IS_AMONG_TOP_N_SIMILAR_DOCS
Relation.RELCLASS.merges,
Relation.RELCLASS.isMergedIn,
Relation.RELCLASS.HasAmongTopNSimilarDocuments,
Relation.RELCLASS.IsAmongTopNSimilarDocuments
)
if (relClassFilter.exists(k => k.equalsIgnoreCase(r.getRelClass)))
if (relClassFilter.contains(r.getRelClass))
false
else {
if (r.getProvenance == null || r.getProvenance.isEmpty)

1
dhp-workflows/dhp-graph-mapper/src/main/scala/eu/dnetlib/dhp/sx/graph/SparkCreateInputGraph.scala
View File

@ -112,7 +112,6 @@ object SparkCreateInputGraph {
log.info(s"Extract ${clazz.getSimpleName}")
oafDs
.filter(o => o.isInstanceOf[T])
.map(p => p.asInstanceOf[T])
.write
.mode(SaveMode.Overwrite)
.save(targetPath)

2
dhp-workflows/dhp-graph-mapper/src/main/scala/eu/dnetlib/dhp/sx/graph/SparkCreateScholix.scala
View File

@ -51,7 +51,7 @@ object SparkCreateScholix {
val relationDS: Dataset[(String, Relation)] = spark.read
.load(relationPath)
.as[Relation]
.filter(r => !r.getRelClass.toLowerCase.contains("merge"))
.filter(r => !r.getRelClass.toString.toLowerCase.contains("merge"))
.map(r => (r.getSource, r))(Encoders.tuple(Encoders.STRING, relEncoder))
val summaryDS: Dataset[(String, ScholixSummary)] = spark.read

62
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java
View File

@ -1,17 +1,13 @@
package eu.dnetlib.dhp.oa.graph.raw;
import static eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions.cleanup;
import static eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions.fixVocabularyNames;
import static org.junit.jupiter.api.Assertions.*;
import static org.mockito.Mockito.lenient;
import java.io.IOException;
import java.util.List;
import java.util.Objects;
import java.util.Optional;
import java.util.stream.Collectors;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory;
import eu.dnetlib.dhp.schema.oaf.utils.PidType;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.dom4j.DocumentException;
@ -21,14 +17,16 @@ import org.junit.jupiter.api.extension.ExtendWith;
import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension;
import com.fasterxml.jackson.databind.ObjectMapper;
import java.io.IOException;
import java.util.List;
import java.util.Objects;
import java.util.Optional;
import java.util.stream.Collectors;
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory;
import eu.dnetlib.dhp.schema.oaf.utils.PidType;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import static eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions.cleanup;
import static eu.dnetlib.dhp.schema.oaf.utils.GraphCleaningFunctions.fixVocabularyNames;
import static org.junit.jupiter.api.Assertions.*;
import static org.mockito.Mockito.lenient;
@ExtendWith(MockitoExtension.class)
class MappersTest {
@ -129,7 +127,7 @@ class MappersTest {
.stream()
.filter(o -> o instanceof Relation)
.map(o -> (Relation) o)
.filter(r -> ModelConstants.RESULT_PROJECT.equals(r.getRelType()))
.filter(r -> Relation.RELTYPE.resultProject.equals(r.getRelType()))
.collect(Collectors.toList());
assertEquals(2, resultProject.size());
@ -152,7 +150,7 @@ class MappersTest {
.stream()
.filter(o -> o instanceof Relation)
.map(o -> (Relation) o)
.filter(r -> ModelConstants.RESULT_ORGANIZATION.equals(r.getRelType()))
.filter(r -> Relation.RELTYPE.resultOrganization.equals(r.getRelType()))
.collect(Collectors.toList());
assertEquals(2, affiliation.size());
@ -297,17 +295,17 @@ class MappersTest {
assertEquals(d.getId(), r1.getSource());
assertEquals("40|corda_______::e06332dee33bec6c2ba4c98601053229", r1.getTarget());
assertEquals(ModelConstants.RESULT_PROJECT, r1.getRelType());
assertEquals(ModelConstants.OUTCOME, r1.getSubRelType());
assertEquals(ModelConstants.IS_PRODUCED_BY, r1.getRelClass());
assertEquals(Relation.RELTYPE.resultProject, r1.getRelType());
assertEquals(Relation.SUBRELTYPE.outcome, r1.getSubRelType());
assertEquals(Relation.RELCLASS.isProducedBy, r1.getRelClass());
assertTrue(r1.getValidated());
assertEquals("2020-01-01", r1.getValidationDate());
assertEquals(d.getId(), r2.getTarget());
assertEquals("40|corda_______::e06332dee33bec6c2ba4c98601053229", r2.getSource());
assertEquals(ModelConstants.RESULT_PROJECT, r2.getRelType());
assertEquals(ModelConstants.OUTCOME, r2.getSubRelType());
assertEquals(ModelConstants.PRODUCES, r2.getRelClass());
assertEquals(Relation.RELTYPE.resultProject, r2.getRelType());
assertEquals(Relation.SUBRELTYPE.outcome, r2.getSubRelType());
assertEquals(Relation.RELCLASS.produces, r2.getRelClass());
assertTrue(r2.getValidated());
assertEquals("2020-01-01", r2.getValidationDate());
@ -597,15 +595,15 @@ class MappersTest {
assertEquals(s.getId(), r1.getSource());
assertEquals("50|doi_________::b453e7b4b2130ace57ff0c3db470a982", r1.getTarget());
assertEquals(ModelConstants.RESULT_RESULT, r1.getRelType());
assertEquals(ModelConstants.RELATIONSHIP, r1.getSubRelType());
assertEquals(ModelConstants.IS_REFERENCED_BY, r1.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r1.getRelType());
assertEquals(Relation.SUBRELTYPE.relationship, r1.getSubRelType());
assertEquals(Relation.RELCLASS.IsReferencedBy, r1.getRelClass());
assertEquals(s.getId(), r2.getTarget());
assertEquals("50|doi_________::b453e7b4b2130ace57ff0c3db470a982", r2.getSource());
assertEquals(ModelConstants.RESULT_RESULT, r2.getRelType());
assertEquals(ModelConstants.RELATIONSHIP, r2.getSubRelType());
assertEquals(ModelConstants.REFERENCES, r2.getRelClass());
assertEquals(Relation.RELTYPE.resultResult, r2.getRelType());
assertEquals(Relation.SUBRELTYPE.relationship, r2.getSubRelType());
assertEquals(Relation.RELCLASS.References, r2.getRelClass());
}

8
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MigrateDbEntitiesApplicationTest.java
View File

@ -246,10 +246,10 @@ class MigrateDbEntitiesApplicationTest {
assertEquals(ModelConstants.DATASOURCE_ORGANIZATION, r1.getRelType());
assertEquals(ModelConstants.DATASOURCE_ORGANIZATION, r2.getRelType());
assertEquals(ModelConstants.PROVISION, r1.getSubRelType());
assertEquals(ModelConstants.PROVISION, r2.getSubRelType());
assertEquals(Relation.SUBRELTYPE.provision, r1.getSubRelType());
assertEquals(Relation.SUBRELTYPE.provision, r2.getSubRelType());
assertEquals(ModelConstants.IS_PROVIDED_BY, r1.getRelClass());
assertEquals(Relation.RELCLASS.isProvidedBy, r1.getRelClass());
assertEquals(ModelConstants.PROVIDES, r2.getRelClass());
}
@ -272,7 +272,7 @@ class MigrateDbEntitiesApplicationTest {
assertValidId(rel.getProvenance().get(0).getCollectedfrom().getKey());
assertEquals(ModelConstants.PROJECT_ORGANIZATION, rel.getRelType());
assertEquals(ModelConstants.PARTICIPATION, rel.getSubRelType());
assertEquals(Relation.SUBRELTYPE.participation, rel.getSubRelType());
assertEquals(ModelConstants.IS_PARTICIPANT, rel.getRelClass());
assertNotNull(rel.getProperties());

19
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/PrepareRelationsJob.java
View File

@ -3,11 +3,9 @@ package eu.dnetlib.dhp.oa.provision;
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
import java.util.HashSet;
import java.util.Optional;
import java.util.PriorityQueue;
import java.util.Set;
import java.util.*;
import java.util.stream.Collectors;
import java.util.stream.StreamSupport;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
@ -80,10 +78,15 @@ public class PrepareRelationsJob {
.orElse(DEFAULT_NUM_PARTITIONS);
log.info("relPartitions: {}", relPartitions);
Set<String> relationFilter = Optional
Set<Relation.RELCLASS> relationFilter = Optional
.ofNullable(parser.get("relationFilter"))
.map(String::toLowerCase)
.map(s -> Sets.newHashSet(Splitter.on(",").split(s)))
.map(s -> Sets.newHashSet(
StreamSupport.stream(
Splitter.on(",").split(s).spliterator(), false)
.map(Relation.RELCLASS::valueOf)
.collect(Collectors.toList())
) )
.orElse(new HashSet<>());
log.info("relationFilter: {}", relationFilter);
@ -128,11 +131,11 @@ public class PrepareRelationsJob {
* @param relPartitions number of partitions for the output RDD
*/
private static void prepareRelationsRDD(SparkSession spark, String inputRelationsPath, String outputPath,
Set<String> relationFilter, int sourceMaxRelations, int targetMaxRelations, int relPartitions) {
Set<Relation.RELCLASS> relationFilter, int sourceMaxRelations, int targetMaxRelations, int relPartitions) {
JavaRDD<Relation> rels = readPathRelationRDD(spark, inputRelationsPath)
.filter(rel -> !(rel.getSource().startsWith("unresolved") || rel.getTarget().startsWith("unresolved")))
.filter(rel -> !relationFilter.contains(StringUtils.lowerCase(rel.getRelClass())));
.filter(rel -> !relationFilter.contains(rel.getRelClass()));
JavaRDD<Relation> pruned = pruneRels(
pruneRels(

26
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/RelationComparator.java
View File

@ -13,21 +13,23 @@ import eu.dnetlib.dhp.schema.oaf.Relation;
public class RelationComparator implements Comparator<Relation> {
private static final Map<String, Integer> weights = Maps.newHashMap();
private static final Map<Relation.SUBRELTYPE, Integer> weights = Maps.newHashMap();
static {
weights.put(ModelConstants.OUTCOME, 0);
weights.put(ModelConstants.SUPPLEMENT, 1);
weights.put(ModelConstants.REVIEW, 2);
weights.put(ModelConstants.CITATION, 3);
weights.put(ModelConstants.AFFILIATION, 4);
weights.put(ModelConstants.RELATIONSHIP, 5);
weights.put(ModelConstants.PUBLICATION_DATASET, 6);
weights.put(ModelConstants.SIMILARITY, 7);
weights.put(Relation.SUBRELTYPE.outcome, 0);
weights.put(Relation.SUBRELTYPE.supplement, 1);
weights.put(Relation.SUBRELTYPE.review, 2);
weights.put(Relation.SUBRELTYPE.citation, 3);
weights.put(Relation.SUBRELTYPE.affiliation, 4);
//TODO CLAUDIO PLEASE CHECK IF the SUBSTITUTION OF publicationDataset WITH RELATIONSHIPS IS OK
// weights.put(Relation.SUBRELTYPE.relationship, 5);
weights.put(Relation.SUBRELTYPE.relationship, 6);
weights.put(Relation.SUBRELTYPE.similarity, 7);
weights.put(ModelConstants.PROVISION, 8);
weights.put(ModelConstants.PARTICIPATION, 9);
weights.put(ModelConstants.DEDUP, 10);
weights.put(Relation.SUBRELTYPE.provision, 8);
weights.put(Relation.SUBRELTYPE.participation, 9);
weights.put(Relation.SUBRELTYPE.dedup, 10);
}
private Integer getWeight(Relation o) {

27
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/SortableRelation.java
View File

@ -14,23 +14,24 @@ import eu.dnetlib.dhp.schema.oaf.Relation;
public class SortableRelation extends Relation implements Comparable<SortableRelation>, Serializable {
private static final Map<String, Integer> weights = Maps.newHashMap();
private static final Map<Relation.SUBRELTYPE, Integer> weights = Maps.newHashMap();
//TODO version and part missing why publication is there?
static {
weights.put(ModelConstants.OUTCOME, 0);
weights.put(ModelConstants.SUPPLEMENT, 1);
weights.put(ModelConstants.REVIEW, 2);
weights.put(ModelConstants.CITATION, 3);
weights.put(ModelConstants.AFFILIATION, 4);
weights.put(ModelConstants.RELATIONSHIP, 5);
weights.put(ModelConstants.PUBLICATION_RESULTTYPE_CLASSID, 6);
weights.put(ModelConstants.SIMILARITY, 7);
weights.put(Relation.SUBRELTYPE.outcome, 0);
weights.put(Relation.SUBRELTYPE.supplement, 1);
weights.put(Relation.SUBRELTYPE.review, 2);
weights.put(Relation.SUBRELTYPE.citation, 3);
weights.put(Relation.SUBRELTYPE.affiliation, 4);
weights.put(Relation.SUBRELTYPE.relationship, 5);
//weights.put(ModelConstants.PUBLICATION_RESULTTYPE_CLASSID, 6);
weights.put(Relation.SUBRELTYPE.similarity, 7);
weights.put(ModelConstants.PROVISION, 8);
weights.put(ModelConstants.PARTICIPATION, 9);
weights.put(ModelConstants.DEDUP, 10);
weights.put(Relation.SUBRELTYPE.provision, 8);
weights.put(Relation.SUBRELTYPE.participation, 9);
weights.put(Relation.SUBRELTYPE.dedup, 10);
}
private static final long serialVersionUID = 34753984579L;
private String groupingKey;

20
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/model/SortableRelationKey.java
View File

@ -17,19 +17,19 @@ public class SortableRelationKey implements Comparable<SortableRelationKey>, Ser
private static final Map<String, Integer> weights = Maps.newHashMap();
static {
weights.put(ModelConstants.PARTICIPATION, 0);
weights.put(ModelConstants.OUTCOME, 1);
weights.put(ModelConstants.AFFILIATION, 2);
weights.put(ModelConstants.DEDUP, 3);
weights.put(Relation.SUBRELTYPE.participation, 0);
weights.put(Relation.SUBRELTYPE.outcome, 1);
weights.put(Relation.SUBRELTYPE.affiliation, 2);
weights.put(Relation.SUBRELTYPE.dedup, 3);
weights.put(ModelConstants.PUBLICATION_DATASET, 4);
weights.put(ModelConstants.SUPPLEMENT, 5);
weights.put(ModelConstants.REVIEW, 6);
weights.put(ModelConstants.RELATIONSHIP, 7);
weights.put(Relation.SUBRELTYPE.supplement, 5);
weights.put(Relation.SUBRELTYPE.review, 6);
weights.put(Relation.SUBRELTYPE.relationship, 7);
weights.put(ModelConstants.PART, 8);
weights.put(ModelConstants.PROVISION, 9);
weights.put(Relation.SUBRELTYPE.provision, 9);
weights.put(ModelConstants.VERSION, 10);
weights.put(ModelConstants.SIMILARITY, 11);
weights.put(ModelConstants.CITATION, 12);
weights.put(Relation.SUBRELTYPE.similarity, 11);
weights.put(Relation.SUBRELTYPE.citation, 12);
}
private static final long serialVersionUID = 3232323;

2
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/utils/XmlRecordFactory.java
View File

@ -1459,7 +1459,7 @@ public class XmlRecordFactory implements Serializable {
}
private boolean isDuplicate(final RelatedEntityWrapper link) {
return ModelConstants.DEDUP.equalsIgnoreCase(link.getRelation().getSubRelType());
return Relation.SUBRELTYPE.dedup.equalsIgnoreCase(link.getRelation().getSubRelType());
}
private List<String> listExtraInfo(final Entity entity) {