action migration workflow produces eu.dnetlib.dhp.schema.action.AtomicAction(s)

This commit is contained in:
Claudio Atzori 2020-03-19 14:00:38 +01:00
parent a0ab15a64c
commit 36236dd1c1
2 changed files with 96 additions and 50 deletions

View File

@ -5,17 +5,14 @@ import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.common.base.Splitter;
import com.google.common.collect.Lists;
import com.google.protobuf.InvalidProtocolBufferException;
import eu.dnetlib.actionmanager.actions.AtomicAction;
import eu.dnetlib.dhp.schema.action.AtomicAction;
import eu.dnetlib.data.proto.OafProtos;
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.Qualifier;
import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.schema.oaf.*;
import eu.dnetlib.dhp.utils.ISLookupClientFactory;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import org.apache.commons.codec.binary.Base64;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.logging.Log;
@ -84,24 +81,29 @@ public class TransformActions implements Serializable {
log.info(String.format("transforming actions from '%s' to '%s'", sourcePath, targetDirectory));
sc.sequenceFile(sourcePath, Text.class, Text.class)
.mapToPair(a -> new Tuple2<>(a._1(), AtomicAction.fromJSON(a._2().toString())))
.mapToPair(a -> new Tuple2<>(a._1(), eu.dnetlib.actionmanager.actions.AtomicAction.fromJSON(a._2().toString())))
.mapToPair(a -> new Tuple2<>(a._1(), transformAction(a._1().toString(), a._2())))
.filter(t -> StringUtils.isNotBlank(t._2().toString()))
.saveAsHadoopFile(targetDirectory.toString(), Text.class, Text.class, SequenceFileOutputFormat.class, GzipCodec.class);
}
}
}
private Text transformAction(String atomicaActionId, AtomicAction aa) throws InvalidProtocolBufferException, JsonProcessingException {
private Text transformAction(String atomicaActionId, eu.dnetlib.actionmanager.actions.AtomicAction aa) throws InvalidProtocolBufferException, JsonProcessingException {
final Text out = new Text();
final ObjectMapper mapper = new ObjectMapper();
if (aa.getTargetValue() != null && aa.getTargetValue().length > 0) {
Oaf oaf = ProtoConverter.convert(OafProtos.Oaf.parseFrom(aa.getTargetValue()));
aa.setTargetValue(mapper.writeValueAsString(oaf).getBytes());
out.set(mapper.writeValueAsString(doTransform(aa)));
} else {
if (atomicaActionId.contains("dedupSimilarity")) {
out.set(mapper.writeValueAsString(getRelationAtomicAction(atomicaActionId)));
}
}
return out;
}
private AtomicAction<Relation> getRelationAtomicAction(String atomicaActionId) {
final String[] splitId = atomicaActionId.split("@");
String source = splitId[0];
@ -132,11 +134,44 @@ public class TransformActions implements Serializable {
rel.setDataInfo(d);
aa.setTargetValue(mapper.writeValueAsString(rel).getBytes());
}
return new AtomicAction<>(Relation.class, rel);
}
return new Text(mapper.writeValueAsString(aa));
private AtomicAction doTransform(eu.dnetlib.actionmanager.actions.AtomicAction aa) throws InvalidProtocolBufferException {
final OafProtos.Oaf proto_oaf = OafProtos.Oaf.parseFrom(aa.getTargetValue());
final Oaf oaf = ProtoConverter.convert(proto_oaf);
switch (proto_oaf.getKind()) {
case entity:
switch (proto_oaf.getEntity().getType()) {
case datasource:
return new AtomicAction<>(Datasource.class, (Datasource) oaf);
case organization:
return new AtomicAction<>(Organization.class, (Organization) oaf);
case project:
return new AtomicAction<>(Project.class, (Project) oaf);
case result:
final String resulttypeid = proto_oaf.getEntity().getResult().getMetadata().getResulttype().getClassid();
switch (resulttypeid) {
case "publication":
return new AtomicAction<>(Publication.class, (Publication) oaf);
case "software":
return new AtomicAction<>(Software.class, (Software) oaf);
case "other":
return new AtomicAction<>(OtherResearchProduct.class, (OtherResearchProduct) oaf);
case "dataset":
return new AtomicAction<>(Dataset.class, (Dataset) oaf);
default:
// can be an update, where the resulttype is not specified
return new AtomicAction<>(Result.class, (Result) oaf);
}
default:
throw new IllegalArgumentException("invalid entity type: " + proto_oaf.getEntity().getType());
}
case relation:
return new AtomicAction<>(Relation.class, (Relation) oaf);
default:
throw new IllegalArgumentException("invalid kind: " + proto_oaf.getKind());
}
}
private String getTargetBaseDir(String isLookupUrl) throws ISLookUpException {

View File

@ -54,12 +54,25 @@
</property>
</parameters>
<global>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>mapreduce.job.queuename</name>
<value>${queueName}</value>
</property>
<property>
<name>oozie.launcher.mapred.job.queue.name</name>
<value>${oozieLauncherQueueName}</value>
</property>
</configuration>
</global>
<start to='migrate_actionsets' />
<action name='migrate_actionsets'>
<java>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<main-class>eu.dnetlib.dhp.migration.actions.MigrateActionSet</main-class>
<java-opt>-Dmapred.task.timeout=${distcp_task_timeout}</java-opt>
<arg>-is</arg><arg>${isLookupUrl}</arg>
@ -78,8 +91,6 @@
<action name="transform_actions">
<spark xmlns="uri:oozie:spark-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<master>yarn</master>
<mode>cluster</mode>
<name>transform_actions</name>