218 lines
8.7 KiB
Java
218 lines
8.7 KiB
Java
|
|
package eu.dnetlib.dhp.oa.graph.dump.filterentities;
|
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
|
|
|
import java.io.Serializable;
|
|
import java.util.Arrays;
|
|
import java.util.Optional;
|
|
import java.util.function.Predicate;
|
|
import java.util.stream.Collectors;
|
|
import java.util.stream.Stream;
|
|
|
|
import org.apache.commons.io.IOUtils;
|
|
import org.apache.spark.SparkConf;
|
|
import org.apache.spark.api.java.function.FilterFunction;
|
|
import org.apache.spark.api.java.function.FlatMapFunction;
|
|
import org.apache.spark.api.java.function.MapFunction;
|
|
import org.apache.spark.sql.*;
|
|
import org.apache.spark.sql.Dataset;
|
|
import org.slf4j.Logger;
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
import com.fasterxml.jackson.core.JsonProcessingException;
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
import eu.dnetlib.dhp.oa.graph.dump.skgif.EmitFromEntities;
|
|
import eu.dnetlib.dhp.oa.graph.dump.skgif.Utils;
|
|
import eu.dnetlib.dhp.schema.common.EntityType;
|
|
import eu.dnetlib.dhp.schema.common.ModelSupport;
|
|
import eu.dnetlib.dhp.schema.oaf.*;
|
|
import scala.Tuple2;
|
|
|
|
/**
|
|
* @author miriam.baglioni
|
|
* @Date 12/03/24
|
|
*/
|
|
public class SelectConnectedEntities implements Serializable {
|
|
private static final Logger log = LoggerFactory.getLogger(EmitFromEntities.class);
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
String jsonConfiguration = IOUtils
|
|
.toString(
|
|
SelectConnectedEntities.class
|
|
.getResourceAsStream(
|
|
"/eu/dnetlib/dhp/oa/graph/dump/select_connected_entities_parameters.json"));
|
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
|
|
parser.parseArgument(args);
|
|
|
|
Boolean isSparkSessionManaged = Optional
|
|
.ofNullable(parser.get("isSparkSessionManaged"))
|
|
.map(Boolean::valueOf)
|
|
.orElse(Boolean.TRUE);
|
|
|
|
log.info("isSparkSessionManaged: {}", isSparkSessionManaged);
|
|
|
|
final String inputPath = parser.get("sourcePath");
|
|
log.info("inputPath: {}", inputPath);
|
|
final String filterPath = parser.get("filterPath");
|
|
log.info("filterPath: {}", filterPath);
|
|
final String workingDir = parser.get("workingDir");
|
|
log.info("workingDir: {}", workingDir);
|
|
SparkConf conf = new SparkConf();
|
|
|
|
runWithSparkSession(
|
|
conf,
|
|
isSparkSessionManaged,
|
|
spark -> {
|
|
|
|
selectConnectedEntities(spark, inputPath, filterPath, workingDir);
|
|
|
|
});
|
|
}
|
|
|
|
private static <R extends Result> void selectConnectedEntities(SparkSession spark, String inputPath,
|
|
String filterPath,
|
|
String workingDir) throws JsonProcessingException {
|
|
|
|
Dataset<String> resultIds = spark.emptyDataset(Encoders.STRING());
|
|
for (EntityType entity : ModelSupport.entityTypes.keySet())
|
|
if (ModelSupport.isResult(entity))
|
|
resultIds = resultIds
|
|
.union(
|
|
spark
|
|
.read()
|
|
.parquet(filterPath + entity.name() + "_ids")
|
|
.select("id")
|
|
.as(Encoders.STRING()));
|
|
|
|
Dataset<Relation> relation = Utils
|
|
.readPath(spark, inputPath + "relation", Relation.class)
|
|
.filter((FilterFunction<Relation>) r -> !r.getDataInfo().getDeletedbyinference());
|
|
Dataset<Organization> organizations = Utils
|
|
.readPath(spark, inputPath + "organization", Organization.class)
|
|
.filter((FilterFunction<Organization>) o -> !o.getDataInfo().getDeletedbyinference());
|
|
Dataset<Project> projects = Utils
|
|
.readPath(spark, inputPath + "project", Project.class)
|
|
.filter((FilterFunction<Project>) p -> !p.getDataInfo().getDeletedbyinference())
|
|
.filter(
|
|
(FilterFunction<Project>) p -> Optional.ofNullable(p.getFundingtree()).isPresent() &&
|
|
p.getFundingtree().size() > 0 &&
|
|
Utils
|
|
.getFunderName(p.getFundingtree().get(0).getValue())
|
|
.equalsIgnoreCase("European Commission"));
|
|
Dataset<Datasource> datasources = Utils
|
|
.readPath(spark, inputPath + "datasource", Datasource.class)
|
|
.filter((FilterFunction<Datasource>) d -> !d.getDataInfo().getDeletedbyinference());
|
|
|
|
// select relations having source in the set of identifiers selected for eosc
|
|
Dataset<Relation> resultSource = resultIds
|
|
.joinWith(relation, resultIds.col("value").equalTo(relation.col("source")))
|
|
.map((MapFunction<Tuple2<String, Relation>, Relation>) t2 -> t2._2(), Encoders.bean(Relation.class));
|
|
|
|
// write relations having source and target in the set
|
|
resultIds
|
|
.joinWith(resultSource, resultIds.col("value").equalTo(resultSource.col("target")))
|
|
.map((MapFunction<Tuple2<String, Relation>, Relation>) t2 -> t2._2(), Encoders.bean(Relation.class))
|
|
.write()
|
|
.option("compression", "gzip")
|
|
.mode(SaveMode.Overwrite)
|
|
.json(workingDir + "relation");
|
|
|
|
// write relations between results and organizations
|
|
resultSource
|
|
.joinWith(organizations, resultSource.col("target").equalTo(organizations.col("id")))
|
|
.map((MapFunction<Tuple2<Relation, Organization>, Relation>) t2 -> t2._1(), Encoders.bean(Relation.class))
|
|
.write()
|
|
.mode(SaveMode.Append)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "relation");
|
|
|
|
resultSource
|
|
.joinWith(projects, resultSource.col("target").equalTo(projects.col("id")))
|
|
.map((MapFunction<Tuple2<Relation, Project>, Relation>) t2 -> t2._1(), Encoders.bean(Relation.class))
|
|
.write()
|
|
.mode(SaveMode.Append)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "relation");
|
|
|
|
// write organizations linked to results in the set
|
|
resultSource
|
|
.joinWith(organizations, resultSource.col("target").equalTo(organizations.col("id")))
|
|
.map(
|
|
(MapFunction<Tuple2<Relation, Organization>, Organization>) t2 -> t2._2(),
|
|
Encoders.bean(Organization.class))
|
|
.write()
|
|
.mode(SaveMode.Overwrite)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "organization");
|
|
|
|
// write projects linked to results in the set
|
|
resultSource
|
|
.joinWith(projects, resultSource.col("target").equalTo(projects.col("id")))
|
|
.map((MapFunction<Tuple2<Relation, Project>, Project>) t2 -> t2._2(), Encoders.bean(Project.class))
|
|
.write()
|
|
.mode(SaveMode.Append)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "project");
|
|
|
|
// read the results and select all the distinct instance.hostedbykey
|
|
Dataset<String> datasourceReferencedIds = spark.emptyDataset(Encoders.STRING());
|
|
for (EntityType entity : ModelSupport.entityTypes.keySet())
|
|
if (ModelSupport.isResult(entity)) {
|
|
Class<R> resultClazz = ModelSupport.entityTypes.get(entity);
|
|
datasourceReferencedIds = datasourceReferencedIds
|
|
.union(
|
|
Utils
|
|
.readPath(spark, workingDir + entity.name(), resultClazz)
|
|
.flatMap(
|
|
(FlatMapFunction<R, String>) r -> r
|
|
.getInstance()
|
|
.stream()
|
|
.flatMap(i -> Stream.of(i.getHostedby().getKey(), i.getCollectedfrom().getKey()))
|
|
.collect(Collectors.toList())
|
|
.iterator(),
|
|
Encoders.STRING()));
|
|
}
|
|
datasourceReferencedIds = datasourceReferencedIds.distinct();
|
|
// join with the datasources and write the datasource in the join
|
|
datasourceReferencedIds
|
|
.joinWith(datasources, datasourceReferencedIds.col("value").equalTo(datasources.col("id")))
|
|
.map((MapFunction<Tuple2<String, Datasource>, Datasource>) t2 -> t2._2(), Encoders.bean(Datasource.class))
|
|
.write()
|
|
.mode(SaveMode.Overwrite)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "datasource");
|
|
|
|
// selecting relations between organizations and projects in the selected set
|
|
Dataset<Organization> organizationSbs = Utils.readPath(spark, workingDir + "organization", Organization.class);
|
|
Dataset<Project> projectSbs = Utils.readPath(spark, workingDir + "project", Project.class);
|
|
Dataset<Relation> orgSourceRels = organizationSbs
|
|
.joinWith(relation, organizationSbs.col("id").equalTo(relation.col("source")))
|
|
.map((MapFunction<Tuple2<Organization, Relation>, Relation>) t2 -> t2._2(), Encoders.bean(Relation.class));
|
|
orgSourceRels
|
|
.joinWith(projectSbs, orgSourceRels.col("target").equalTo(projectSbs.col("id")))
|
|
.map((MapFunction<Tuple2<Relation, Project>, Relation>) t2 -> t2._1(), Encoders.bean(Relation.class))
|
|
.write()
|
|
.mode(SaveMode.Append)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "relation");
|
|
|
|
// selecting relations between datasources and organizations in the selected set
|
|
Dataset<Datasource> datasourceSbs = Utils.readPath(spark, workingDir + "datasource", Datasource.class);
|
|
Dataset<Relation> dsSourceRels = datasourceSbs
|
|
.joinWith(relation, datasourceSbs.col("id").as("dsId").equalTo(relation.col("source")))
|
|
.map((MapFunction<Tuple2<Datasource, Relation>, Relation>) t2 -> t2._2(), Encoders.bean(Relation.class));
|
|
dsSourceRels
|
|
.joinWith(organizationSbs, dsSourceRels.col("target").equalTo(organizations.col("id").as("orgId")))
|
|
.map((MapFunction<Tuple2<Relation, Organization>, Relation>) t2 -> t2._1(), Encoders.bean(Relation.class))
|
|
.write()
|
|
.mode(SaveMode.Append)
|
|
.option("compression", "gzip")
|
|
.json(workingDir + "relation");
|
|
|
|
}
|
|
}
|