203 lines
8.3 KiB
Java
203 lines
8.3 KiB
Java
|
|
package eu.dnetlib.dhp.oa.graph.dump.skgif;
|
|
|
|
import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;
|
|
|
|
import java.io.Serializable;
|
|
import java.util.*;
|
|
import java.util.stream.Collectors;
|
|
|
|
import eu.dnetlib.dhp.oa.graph.dump.skgif.beans.EncloseMinElement;
|
|
import eu.dnetlib.dhp.schema.oaf.Organization;
|
|
import eu.dnetlib.dhp.schema.oaf.Relation;
|
|
import eu.dnetlib.dhp.skgif.model.MinOrganization;
|
|
import eu.dnetlib.dhp.skgif.model.RelationType;
|
|
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.MapFunction;
|
|
import org.apache.spark.api.java.function.MapGroupsFunction;
|
|
import org.apache.spark.sql.Dataset;
|
|
import org.apache.spark.sql.Encoders;
|
|
import org.apache.spark.sql.SaveMode;
|
|
import org.apache.spark.sql.SparkSession;
|
|
import org.slf4j.Logger;
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
|
import eu.dnetlib.dhp.schema.oaf.Datasource;
|
|
import eu.dnetlib.dhp.skgif.model.Identifier;
|
|
import eu.dnetlib.dhp.skgif.model.Prefixes;
|
|
import scala.Tuple2;
|
|
|
|
/**
|
|
* @author miriam.baglioni
|
|
* @Date 21/02/24
|
|
*/
|
|
public class DumpDatasource implements Serializable {
|
|
private static final Logger log = LoggerFactory.getLogger(DumpDatasource.class);
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
String jsonConfiguration = IOUtils
|
|
.toString(
|
|
DumpDatasource.class
|
|
.getResourceAsStream(
|
|
"/eu/dnetlib/dhp/oa/graph/dump/dump_datasource_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 workingDir = parser.get("workingDir");
|
|
log.info("workingDir: {}", workingDir);
|
|
|
|
final String outputPath = parser.get("outputPath");
|
|
log.info("outputPath: {}", outputPath);
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
runWithSparkSession(
|
|
conf,
|
|
isSparkSessionManaged,
|
|
spark -> {
|
|
Utils.removeOutputDir(spark, outputPath + "Datasources");
|
|
|
|
mapDatasource(spark, inputPath, outputPath, workingDir);
|
|
});
|
|
}
|
|
|
|
private static void mapDatasource(SparkSession spark, String inputPath, String outputPath, String workingDir) {
|
|
Dataset<Relation> relation = Utils.readPath(spark, inputPath + "relation", Relation.class)
|
|
.filter((FilterFunction<Relation>) r -> !r.getDataInfo().getDeletedbyinference())
|
|
.filter((FilterFunction<Relation>) r -> r.getRelClass().equalsIgnoreCase(RelationType.DATASOURCE_PROVIDED_BY_ORGANIZATION.label));
|
|
|
|
Dataset<EncloseMinElement> eme = Utils.readPath(spark, workingDir + "minEntity", EncloseMinElement.class)
|
|
.filter((FilterFunction<EncloseMinElement>) e -> Optional.ofNullable(e.getMinOrganization()).isPresent());
|
|
|
|
Dataset<Datasource> datasourceDataset = Utils
|
|
.readPath(spark, inputPath + "datasource", Datasource.class)
|
|
.filter(
|
|
(FilterFunction<Datasource>) d -> !d.getDataInfo().getInvisible()
|
|
&& !d.getDataInfo().getDeletedbyinference());
|
|
Dataset<Tuple2<String, EncloseMinElement>> datasourceOrganization = relation.joinWith(eme, relation.col("target").equalTo(eme.col("enclosedEntityId")))
|
|
.map((MapFunction<Tuple2<Relation, EncloseMinElement>, Tuple2<String, EncloseMinElement>>) t2 -> new Tuple2<>(t2._1().getSource(), t2._2()), Encoders.tuple(Encoders.STRING(), Encoders.bean(EncloseMinElement.class)));
|
|
|
|
datasourceDataset.joinWith(datasourceOrganization, datasourceDataset.col("id").equalTo(datasourceOrganization.col("_1")), "left")
|
|
.groupByKey((MapFunction<Tuple2<Datasource, Tuple2<String, EncloseMinElement>>, String>) t2 -> t2._1().getId(), Encoders.STRING() )
|
|
.mapGroups((MapGroupsFunction<String, Tuple2<Datasource, Tuple2<String, EncloseMinElement>>, eu.dnetlib.dhp.skgif.model.Datasource>) (k,vs) -> {
|
|
eu.dnetlib.dhp.skgif.model.Datasource datasource = new eu.dnetlib.dhp.skgif.model.Datasource();
|
|
Tuple2<Datasource, Tuple2<String, EncloseMinElement>> first = vs.next();
|
|
Datasource d = first._1();
|
|
datasource.setLocal_identifier(Utils.getIdentifier(Prefixes.DATASOURCE, d.getId()));
|
|
datasource
|
|
.setIdentifiers(
|
|
d
|
|
.getPid()
|
|
.stream()
|
|
.map(p -> Identifier.newInstance(p.getQualifier().getClassid(), p.getValue()))
|
|
.collect(Collectors.toList()));
|
|
|
|
datasource.setName(d.getOfficialname().getValue());
|
|
datasource.setSubmission_policy_url(d.getSubmissionpolicyurl());
|
|
datasource
|
|
.setJurisdiction(
|
|
Optional
|
|
.ofNullable(d.getJurisdiction())
|
|
.map(v -> v.getClassid())
|
|
.orElse(new String()));
|
|
datasource.setPreservation_policy_url(d.getPreservationpolicyurl());
|
|
datasource.setVersion_control(d.getVersioncontrol());
|
|
|
|
datasource
|
|
.setData_source_classification(
|
|
Optional
|
|
.ofNullable(d.getEoscdatasourcetype())
|
|
.map(v -> v.getClassname())
|
|
.orElse(new String()));
|
|
datasource.setResearch_product_type(getEoscProductType(d.getResearchentitytypes()));
|
|
datasource.setThematic(d.getThematic());
|
|
datasource
|
|
.setResearch_product_access_policy(
|
|
Optional
|
|
.ofNullable(d.getDatabaseaccesstype())
|
|
.map(v -> getResearchProductAccessPolicy(d.getDatabaseaccesstype().getValue()))
|
|
.orElse(new ArrayList<>()));
|
|
datasource
|
|
.setResearch_product_metadata_access_policy(
|
|
Optional
|
|
.ofNullable(d.getResearchproductmetadataaccesspolicies())
|
|
.map(v -> getResearchProductAccessPolicy(d.getResearchproductmetadataaccesspolicies()))
|
|
.orElse(new ArrayList<>()));
|
|
if(Optional.ofNullable(first._2()).isPresent()){
|
|
List<MinOrganization> organizations = new ArrayList<>();
|
|
organizations.add(first._2()._2().getMinOrganization());
|
|
vs.forEachRemaining(org -> organizations.add(org._2()._2().getMinOrganization()));
|
|
datasource.setOrganization(organizations);
|
|
}
|
|
return datasource;
|
|
|
|
}, Encoders.bean( eu.dnetlib.dhp.skgif.model.Datasource.class))
|
|
|
|
.write()
|
|
.mode(SaveMode.Overwrite)
|
|
.option("compression", "gzip")
|
|
.json(outputPath + "Datasource");
|
|
}
|
|
|
|
private static List<String> getResearchProductAccessPolicy(List<String> value) {
|
|
|
|
return value
|
|
.stream()
|
|
.map(v -> getResearchProductAccessPolicy(v))
|
|
.filter(Objects::nonNull)
|
|
.map(v -> v.get(0))
|
|
.distinct()
|
|
.collect(Collectors.toList());
|
|
}
|
|
|
|
private static List<String> getResearchProductAccessPolicy(String value) {
|
|
// "databaseaccesstype if open => open access (https://vocabularies.coar-repositories.org/access_rights/c_abf2/)
|
|
// if restricted => restricted access (https://vocabularies.coar-repositories.org/access_rights/c_16ec/)
|
|
// if closed => metadata only access (https://vocabularies.coar-repositories.org/access_rights/c_14cb/) "
|
|
switch (value) {
|
|
case "open":// (https://vocabularies.coar-repositories.org/access_rights/c_abf2/)
|
|
return Arrays.asList("open access");
|
|
case "restricted":// (https://vocabularies.coar-repositories.org/access_rights/c_16ec/)
|
|
return Arrays.asList("restricted access");
|
|
case "closed":// (https://vocabularies.coar-repositories.org/access_rights/c_14cb/)
|
|
return Arrays.asList("metadata only access");
|
|
default:
|
|
return null;
|
|
}
|
|
}
|
|
|
|
private static List<String> getEoscProductType(List<String> researchentitytypes) {
|
|
|
|
List<String> eoscProductType = new ArrayList<>();
|
|
if (researchentitytypes != null) {
|
|
|
|
if (researchentitytypes.contains("Software"))
|
|
eoscProductType.add("Research Software");
|
|
if (researchentitytypes.contains("Research Publications") || researchentitytypes.contains("Literature"))
|
|
eoscProductType.add("Research Literature");
|
|
if (researchentitytypes.contains("Research Data"))
|
|
eoscProductType.add("Research Data");
|
|
if (researchentitytypes.contains("Organization") ||
|
|
researchentitytypes.contains("Organizations") ||
|
|
researchentitytypes.contains("Services") ||
|
|
researchentitytypes.contains("Projects"))
|
|
eoscProductType.add("Other research product");
|
|
}
|
|
return eoscProductType;
|
|
}
|
|
}
|