2020-06-22 08:51:31 +02:00
|
|
|
|
|
|
|
package eu.dnetlib.dhp.broker.oa;
|
|
|
|
|
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
|
|
|
|
|
|
|
import java.util.Optional;
|
|
|
|
|
|
|
|
import org.apache.commons.io.IOUtils;
|
|
|
|
import org.apache.spark.SparkConf;
|
|
|
|
import org.apache.spark.sql.Dataset;
|
|
|
|
import org.apache.spark.sql.Encoders;
|
|
|
|
import org.apache.spark.sql.SaveMode;
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
|
2020-06-23 08:37:35 +02:00
|
|
|
import eu.dnetlib.broker.objects.OaBrokerRelatedSoftware;
|
2020-06-22 08:51:31 +02:00
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
2020-06-23 09:32:32 +02:00
|
|
|
import eu.dnetlib.dhp.broker.oa.util.BrokerConstants;
|
2020-06-22 08:51:31 +02:00
|
|
|
import eu.dnetlib.dhp.broker.oa.util.ClusterUtils;
|
|
|
|
import eu.dnetlib.dhp.broker.oa.util.ConversionUtils;
|
|
|
|
import eu.dnetlib.dhp.broker.oa.util.aggregators.withRels.RelatedSoftware;
|
2020-06-23 09:32:32 +02:00
|
|
|
import eu.dnetlib.dhp.schema.common.ModelConstants;
|
2020-06-22 08:51:31 +02:00
|
|
|
import eu.dnetlib.dhp.schema.oaf.Relation;
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.Software;
|
|
|
|
|
2020-06-22 11:45:14 +02:00
|
|
|
public class PrepareRelatedSoftwaresJob {
|
2020-06-22 08:51:31 +02:00
|
|
|
|
2020-06-22 11:45:14 +02:00
|
|
|
private static final Logger log = LoggerFactory.getLogger(PrepareRelatedSoftwaresJob.class);
|
2020-06-22 08:51:31 +02:00
|
|
|
|
|
|
|
public static void main(final String[] args) throws Exception {
|
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
|
|
|
IOUtils
|
2020-06-23 09:32:32 +02:00
|
|
|
.toString(
|
|
|
|
PrepareRelatedSoftwaresJob.class
|
|
|
|
.getResourceAsStream("/eu/dnetlib/dhp/broker/oa/common_params.json")));
|
2020-06-22 08:51:31 +02:00
|
|
|
parser.parseArgument(args);
|
|
|
|
|
|
|
|
final Boolean isSparkSessionManaged = Optional
|
|
|
|
.ofNullable(parser.get("isSparkSessionManaged"))
|
|
|
|
.map(Boolean::valueOf)
|
|
|
|
.orElse(Boolean.TRUE);
|
|
|
|
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
|
|
|
|
|
|
|
final String graphPath = parser.get("graphPath");
|
|
|
|
log.info("graphPath: {}", graphPath);
|
|
|
|
|
2020-06-22 11:45:14 +02:00
|
|
|
final String workingPath = parser.get("workingPath");
|
|
|
|
log.info("workingPath: {}", workingPath);
|
|
|
|
|
|
|
|
final String relsPath = workingPath + "/relatedSoftwares";
|
2020-06-22 08:51:31 +02:00
|
|
|
log.info("relsPath: {}", relsPath);
|
|
|
|
|
|
|
|
final SparkConf conf = new SparkConf();
|
|
|
|
|
|
|
|
runWithSparkSession(conf, isSparkSessionManaged, spark -> {
|
|
|
|
|
|
|
|
ClusterUtils.removeDir(spark, relsPath);
|
2020-06-22 11:45:14 +02:00
|
|
|
|
2020-06-23 08:37:35 +02:00
|
|
|
final Dataset<OaBrokerRelatedSoftware> softwares = ClusterUtils
|
|
|
|
.readPath(spark, graphPath + "/software", Software.class)
|
|
|
|
.filter(sw -> !ClusterUtils.isDedupRoot(sw.getId()))
|
|
|
|
.map(ConversionUtils::oafSoftwareToBrokerSoftware, Encoders.bean(OaBrokerRelatedSoftware.class));
|
2020-06-22 08:51:31 +02:00
|
|
|
|
2020-06-23 08:37:35 +02:00
|
|
|
final Dataset<Relation> rels = ClusterUtils
|
|
|
|
.readPath(spark, graphPath + "/relation", Relation.class)
|
2020-06-24 15:47:06 +02:00
|
|
|
.filter(r -> r.getDataInfo().getDeletedbyinference())
|
2020-06-23 09:32:32 +02:00
|
|
|
.filter(r -> r.getRelType().equals(ModelConstants.RESULT_RESULT))
|
|
|
|
.filter(r -> !r.getRelClass().equals(BrokerConstants.IS_MERGED_IN_CLASS))
|
2020-06-23 08:37:35 +02:00
|
|
|
.filter(r -> !ClusterUtils.isDedupRoot(r.getSource()))
|
|
|
|
.filter(r -> !ClusterUtils.isDedupRoot(r.getTarget()));
|
2020-06-22 08:51:31 +02:00
|
|
|
|
|
|
|
rels
|
2020-06-23 08:37:35 +02:00
|
|
|
.joinWith(softwares, softwares.col("openaireId").equalTo(rels.col("target")), "inner")
|
2020-06-23 09:32:32 +02:00
|
|
|
.map(t -> new RelatedSoftware(t._1.getSource(), t._2), Encoders.bean(RelatedSoftware.class))
|
2020-06-22 08:51:31 +02:00
|
|
|
.write()
|
|
|
|
.mode(SaveMode.Overwrite)
|
|
|
|
.json(relsPath);
|
|
|
|
|
|
|
|
});
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|