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

when migrating actionsets from DM cluster, populate the AtomicAction.targetValue when empty (dedup similarities)

This commit is contained in:
Claudio Atzori 2020-03-15 18:07:59 +01:00
parent 9c84e21b87
commit af835f2f98
1 changed files with 41 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/migration/actions/TransformActions.java
View File

@ -8,7 +8,10 @@ import com.google.protobuf.InvalidProtocolBufferException;
import eu.dnetlib.actionmanager.actions.AtomicAction; import eu.dnetlib.actionmanager.actions.AtomicAction;
import eu.dnetlib.data.proto.OafProtos; import eu.dnetlib.data.proto.OafProtos;
import eu.dnetlib.dhp.application.ArgumentApplicationParser; import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.schema.oaf.DataInfo;
import eu.dnetlib.dhp.schema.oaf.Oaf; import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.schema.oaf.Qualifier;
import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.utils.ISLookupClientFactory; import eu.dnetlib.dhp.utils.ISLookupClientFactory;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException; import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService; import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
@ -82,19 +85,55 @@ public class TransformActions implements Serializable {
sc.sequenceFile(sourcePath, Text.class, Text.class) sc.sequenceFile(sourcePath, Text.class, Text.class)
.mapToPair(a -> new Tuple2<>(a._1(), AtomicAction.fromJSON(a._2().toString()))) .mapToPair(a -> new Tuple2<>(a._1(), AtomicAction.fromJSON(a._2().toString())))
.mapToPair(a -> new Tuple2<>(a._1(), transformAction(a._2()))) .mapToPair(a -> new Tuple2<>(a._1(), transformAction(a._1().toString(), a._2())))
.saveAsHadoopFile(targetDirectory.toString(), Text.class, Text.class, SequenceFileOutputFormat.class, GzipCodec.class); .saveAsHadoopFile(targetDirectory.toString(), Text.class, Text.class, SequenceFileOutputFormat.class, GzipCodec.class);
} }
} }
} }
private Text transformAction(AtomicAction aa) throws InvalidProtocolBufferException, JsonProcessingException { private Text transformAction(String atomicaActionId, AtomicAction aa) throws InvalidProtocolBufferException, JsonProcessingException {
final ObjectMapper mapper = new ObjectMapper(); final ObjectMapper mapper = new ObjectMapper();
if (aa.getTargetValue() != null && aa.getTargetValue().length > 0) { if (aa.getTargetValue() != null && aa.getTargetValue().length > 0) {
Oaf oaf = ProtoConverter.convert(OafProtos.Oaf.parseFrom(aa.getTargetValue())); Oaf oaf = ProtoConverter.convert(OafProtos.Oaf.parseFrom(aa.getTargetValue()));
aa.setTargetValue(mapper.writeValueAsString(oaf).getBytes()); aa.setTargetValue(mapper.writeValueAsString(oaf).getBytes());
} else {
if (atomicaActionId.contains("dedupSimilarity")) {
final String[] splitId = atomicaActionId.split("@");
String source = splitId[0];
String target = splitId[2];
String[] relSemantic = splitId[1].split("_");
Relation rel = new Relation();
rel.setSource(source);
rel.setTarget(target);
rel.setRelType(relSemantic[0]);
rel.setSubRelType(relSemantic[1]);
rel.setRelClass(relSemantic[2]);
DataInfo d = new DataInfo();
d.setDeletedbyinference(false);
d.setInferenceprovenance("deduplication");
d.setInferred(true);
d.setInvisible(false);
Qualifier provenanceaction = new Qualifier();
provenanceaction.setClassid("deduplication");
provenanceaction.setClassname("deduplication");
provenanceaction.setSchemeid("dnet:provenanceActions");
provenanceaction.setSchemename("dnet:provenanceActions");
d.setProvenanceaction(provenanceaction);
rel.setDataInfo(d);
aa.setTargetValue(mapper.writeValueAsString(rel).getBytes());
}
} }
return new Text(mapper.writeValueAsString(aa)); return new Text(mapper.writeValueAsString(aa));