146 lines
5.0 KiB
Java
146 lines
5.0 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.Optional;
|
|
import java.util.stream.Collectors;
|
|
|
|
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.Organization;
|
|
import eu.dnetlib.dhp.schema.oaf.Project;
|
|
import eu.dnetlib.dhp.skgif.model.Identifier;
|
|
import eu.dnetlib.dhp.skgif.model.OrganizationTypes;
|
|
import eu.dnetlib.dhp.skgif.model.Prefixes;
|
|
|
|
/**
|
|
* @author miriam.baglioni
|
|
* @Date 21/02/24
|
|
*/
|
|
public class DumpOrganization implements Serializable {
|
|
private static final Logger log = LoggerFactory.getLogger(DumpOrganization.class);
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
String jsonConfiguration = IOUtils
|
|
.toString(
|
|
DumpOrganization.class
|
|
.getResourceAsStream(
|
|
"/eu/dnetlib/dhp/oa/graph/dump/skgif/dump_organization_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 outputPath = parser.get("outputPath");
|
|
log.info("outputPath: {}", outputPath);
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
runWithSparkSession(
|
|
conf,
|
|
isSparkSessionManaged,
|
|
spark -> {
|
|
Utils.removeOutputDir(spark, outputPath + "organizations");
|
|
|
|
mapOrganization(spark, inputPath, outputPath);
|
|
});
|
|
}
|
|
|
|
private static void mapOrganization(SparkSession spark, String inputPath, String outputPath) {
|
|
Dataset<Organization> organizations = Utils.readPath(spark, inputPath + "organization", Organization.class);
|
|
organizations = organizations
|
|
.filter(
|
|
(FilterFunction<Organization>) o -> !o.getDataInfo().getDeletedbyinference()
|
|
&& !o.getDataInfo().getInvisible())
|
|
.groupByKey((MapFunction<Organization, String>) p -> p.getId(), Encoders.STRING())
|
|
.mapGroups(
|
|
(MapGroupsFunction<String, Organization, Organization>) (k, v) -> v.next(),
|
|
Encoders.bean(Organization.class));
|
|
|
|
organizations.map((MapFunction<Organization, eu.dnetlib.dhp.skgif.model.Organization>) o -> {
|
|
if (!Optional.ofNullable(o.getPid()).isPresent() || o.getPid().size() == 0)
|
|
return null;
|
|
eu.dnetlib.dhp.skgif.model.Organization organization = new eu.dnetlib.dhp.skgif.model.Organization();
|
|
// organization.setLocal_identifier(Utils.getIdentifier(Prefixes.ORGANIZATION, o.getId()));
|
|
organization.setLocal_identifier(o.getId());
|
|
organization
|
|
.setCountry(
|
|
Optional
|
|
.ofNullable(o.getCountry().getClassid())
|
|
.orElse(new String()));
|
|
organization
|
|
.setName(
|
|
Optional
|
|
.ofNullable(o.getLegalname().getValue())
|
|
.orElse(new String()));
|
|
organization
|
|
.setShort_name(
|
|
Optional
|
|
.ofNullable(o.getLegalshortname())
|
|
.map(v -> v.getValue())
|
|
.orElse(new String()));
|
|
organization
|
|
.setIdentifiers(
|
|
o
|
|
.getPid()
|
|
.stream()
|
|
.map(p -> Identifier.newInstance(p.getQualifier().getClassid(), p.getValue()))
|
|
.collect(Collectors.toList()));
|
|
organization
|
|
.setOther_names(
|
|
o
|
|
.getAlternativeNames()
|
|
.stream()
|
|
.map(a -> a.getValue())
|
|
.collect(Collectors.toList()));
|
|
organization.setType(getOrganizationType(o));
|
|
return organization;
|
|
}, Encoders.bean(eu.dnetlib.dhp.skgif.model.Organization.class))
|
|
.filter((FilterFunction<eu.dnetlib.dhp.skgif.model.Organization>) o -> o != null)
|
|
.write()
|
|
.mode(SaveMode.Overwrite)
|
|
.option("compression", "gzip")
|
|
.json(outputPath + "organizations");
|
|
}
|
|
|
|
private static String getOrganizationType(Organization o) {
|
|
if (Optional.ofNullable(o.getEcenterprise()).isPresent()
|
|
&& o.getEcenterprise().getValue().equalsIgnoreCase("true"))
|
|
return OrganizationTypes.COMPANY.label;
|
|
if (Optional.ofNullable(o.getEchighereducation()).isPresent()
|
|
&& o.getEchighereducation().getValue().equalsIgnoreCase("true"))
|
|
return OrganizationTypes.EDUCATION.label;
|
|
if (Optional.ofNullable(o.getEcresearchorganization()).isPresent()
|
|
&& o.getEcresearchorganization().getValue().equalsIgnoreCase("true"))
|
|
return OrganizationTypes.EDUCATION.label;
|
|
if (Optional.ofNullable(o.getEcnonprofit()).isPresent()
|
|
&& o.getEcnonprofit().getValue().equalsIgnoreCase("true"))
|
|
return OrganizationTypes.NONPROFIT.label;
|
|
|
|
return OrganizationTypes.OTHER.label;
|
|
|
|
}
|
|
}
|