dhp-graph-dump/dump/src/main/java/eu/dnetlib/dhp/oa/graph/dump/eosc/ExtendEoscResultWithOrganization.java

package eu.dnetlib.dhp.oa.graph.dump.eosc;

import static eu.dnetlib.dhp.common.SparkSessionSupport.runWithSparkSession;

import java.io.Serializable;
import java.util.*;

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.eosc.model.Affiliation;
import eu.dnetlib.dhp.eosc.model.OrganizationPid;
import eu.dnetlib.dhp.eosc.model.Result;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Organization;
import eu.dnetlib.dhp.schema.oaf.Relation;
import scala.Tuple2;

/**
 * @author miriam.baglioni
 * @Date 27/07/22
 */
public class ExtendEoscResultWithOrganization implements Serializable {
	private static final Logger log = LoggerFactory.getLogger(ExtendEoscResultWithOrganization.class);

	public static void main(String[] args) throws Exception {
		String jsonConfiguration = IOUtils
			.toString(
				ExtendEoscResultWithOrganization.class
					.getResourceAsStream(
						"/eu/dnetlib/dhp/oa/graph/dump/eosc_extend_result_with_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 workingPath = parser.get("workingPath");
		log.info("workingPath: {}", workingPath);

		SparkConf conf = new SparkConf();

		runWithSparkSession(
			conf,
			isSparkSessionManaged,
			spark -> {
				Utils.removeOutputDir(spark, workingPath + "/affiliation");
				addOrganizations(spark, inputPath, workingPath);
			});
	}

	private static void addOrganizations(SparkSession spark, String inputPath, String workingPath) {

		List<String> entities = Arrays.asList("publication", "dataset", "software", "otherresearchproduct");

		entities
			.parallelStream()
			.forEach(
				entity -> {
					Dataset<Result> results = Utils
						.readPath(spark, workingPath + "/" + entity, Result.class);
					Dataset<Relation> relations = Utils
						.readPath(spark, inputPath + "/relation", Relation.class)
						.filter(
							(FilterFunction<Relation>) r -> !r.getDataInfo().getDeletedbyinference() &&
								!r.getDataInfo().getInvisible()
								&& r.getSubRelType().equalsIgnoreCase(ModelConstants.AFFILIATION));

					Dataset<Organization> organizations = Utils
						.readPath(spark, inputPath + "/organization", Organization.class);
					Dataset<ResultOrganizations> resultOrganization = relations
						.joinWith(organizations, relations.col("source").equalTo(organizations.col("id")), "left")
						.map((MapFunction<Tuple2<Relation, Organization>, ResultOrganizations>) t2 -> {
							if (t2._2() != null) {
								ResultOrganizations rOrg = new ResultOrganizations();
								rOrg.setResultId(t2._1().getTarget());
								Affiliation org = new Affiliation();
								org.setId(t2._2().getId());
								if (Optional.ofNullable(t2._2().getLegalname()).isPresent()) {
									org.setName(t2._2().getLegalname().getValue());
								} else {
									org.setName("");
								}
								HashMap<String, Set<String>> organizationPids = new HashMap<>();
								if (Optional.ofNullable(t2._2().getPid()).isPresent())
									t2._2().getPid().forEach(p -> {
										if (!organizationPids.containsKey(p.getQualifier().getClassid()))
											organizationPids.put(p.getQualifier().getClassid(), new HashSet<>());
										organizationPids.get(p.getQualifier().getClassid()).add(p.getValue());
									});
								List<OrganizationPid> pids = new ArrayList<>();
								for (String key : organizationPids.keySet()) {
									for (String value : organizationPids.get(key)) {
										OrganizationPid pid = new OrganizationPid();
										pid.setValue(value);
										pid.setType(key);
										pids.add(pid);
									}
								}
								org.setPid(pids);
								rOrg.setAffiliation(org);
								return rOrg;
							}
							return null;

						}, Encoders.bean(ResultOrganizations.class))
						.filter(Objects::nonNull);
					results
						.joinWith(
							resultOrganization, results.col("id").equalTo(resultOrganization.col("resultId")), "left")
						.groupByKey(
							(MapFunction<Tuple2<Result, ResultOrganizations>, String>) t2 -> t2._1().getId(),
							Encoders.STRING())
						.mapGroups(
							(MapGroupsFunction<String, Tuple2<Result, ResultOrganizations>, Result>) (s, it) -> {
								Tuple2<Result, ResultOrganizations> first = it.next();
								if (first._2() == null) {
									return first._1();
								}
								Result ret = first._1();
								List<Affiliation> affiliation = new ArrayList<>();
								Set<String> alreadyInsertedAffiliations = new HashSet<>();
								affiliation.add(first._2().getAffiliation());
								alreadyInsertedAffiliations.add(first._2().getAffiliation().getId());
								it.forEachRemaining(res -> {
									if (!alreadyInsertedAffiliations.contains(res._2().getAffiliation().getId())) {
										affiliation.add(res._2().getAffiliation());
										alreadyInsertedAffiliations.add(res._2().getAffiliation().getId());
									}

								});
								ret.setAffiliation(affiliation);
								return ret;
							}, Encoders.bean(Result.class))
						.write()
						.mode(SaveMode.Overwrite)
						.option("compression", "gzip")
						.json(workingPath + "/affiliation/" + entity);
				});

	}

}