2020-09-29 12:01:25 +02:00
|
|
|
package eu.dnetlib.jobs;
|
|
|
|
|
|
|
|
import eu.dnetlib.Deduper;
|
2022-04-19 15:29:29 +02:00
|
|
|
import eu.dnetlib.graph.JavaGraphProcessor;
|
2020-12-04 15:41:31 +01:00
|
|
|
import eu.dnetlib.pace.config.DedupConfig;
|
|
|
|
import eu.dnetlib.pace.util.MapDocumentUtil;
|
2020-09-29 12:01:25 +02:00
|
|
|
import eu.dnetlib.pace.utils.Utility;
|
|
|
|
import eu.dnetlib.support.ArgumentApplicationParser;
|
2020-12-04 15:41:31 +01:00
|
|
|
import eu.dnetlib.support.ConnectedComponent;
|
|
|
|
import eu.dnetlib.support.Relation;
|
2020-09-29 12:01:25 +02:00
|
|
|
import org.apache.spark.SparkConf;
|
2020-12-04 15:41:31 +01:00
|
|
|
import org.apache.spark.api.java.JavaPairRDD;
|
|
|
|
import org.apache.spark.api.java.JavaRDD;
|
|
|
|
import org.apache.spark.api.java.JavaSparkContext;
|
|
|
|
import org.apache.spark.api.java.function.PairFunction;
|
|
|
|
import org.apache.spark.graphx.Edge;
|
|
|
|
import org.apache.spark.rdd.RDD;
|
|
|
|
import org.apache.spark.sql.Dataset;
|
|
|
|
import org.apache.spark.sql.Encoders;
|
|
|
|
import org.apache.spark.sql.SaveMode;
|
2020-09-29 12:01:25 +02:00
|
|
|
import org.apache.spark.sql.SparkSession;
|
|
|
|
import org.slf4j.Logger;
|
|
|
|
import org.slf4j.LoggerFactory;
|
2020-12-04 15:41:31 +01:00
|
|
|
import scala.Tuple2;
|
2020-09-29 12:01:25 +02:00
|
|
|
|
2020-12-04 15:41:31 +01:00
|
|
|
import java.io.IOException;
|
2023-04-17 11:06:27 +02:00
|
|
|
import java.util.List;
|
2020-09-29 12:01:25 +02:00
|
|
|
import java.util.Optional;
|
|
|
|
|
2020-12-04 15:41:31 +01:00
|
|
|
import static eu.dnetlib.Deduper.hash;
|
|
|
|
|
2020-09-29 12:01:25 +02:00
|
|
|
public class SparkCreateMergeRels extends AbstractSparkJob {
|
|
|
|
|
|
|
|
private static final Logger log = LoggerFactory.getLogger(SparkCreateMergeRels.class);
|
|
|
|
|
|
|
|
public SparkCreateMergeRels(ArgumentApplicationParser parser, SparkSession spark) {
|
|
|
|
super(parser, spark);
|
|
|
|
}
|
|
|
|
|
|
|
|
public static void main(String[] args) throws Exception {
|
|
|
|
|
|
|
|
ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
2020-12-04 15:41:31 +01:00
|
|
|
Utility.readResource("/jobs/parameters/createMergeRels_parameters.json", SparkCreateMergeRels.class)
|
2020-09-29 12:01:25 +02:00
|
|
|
);
|
|
|
|
|
|
|
|
parser.parseArgument(args);
|
|
|
|
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
|
2020-12-04 15:41:31 +01:00
|
|
|
new SparkCreateMergeRels(
|
2020-09-29 12:01:25 +02:00
|
|
|
parser,
|
|
|
|
getSparkSession(conf)
|
|
|
|
).run();
|
|
|
|
}
|
|
|
|
|
|
|
|
@Override
|
2020-12-04 15:41:31 +01:00
|
|
|
public void run() throws IOException {
|
2020-09-29 12:01:25 +02:00
|
|
|
|
|
|
|
// read oozie parameters
|
|
|
|
final String entitiesPath = parser.get("entitiesPath");
|
|
|
|
final String workingPath = parser.get("workingPath");
|
|
|
|
final String dedupConfPath = parser.get("dedupConfPath");
|
|
|
|
final int numPartitions = Optional
|
|
|
|
.ofNullable(parser.get("numPartitions"))
|
|
|
|
.map(Integer::valueOf)
|
|
|
|
.orElse(NUM_PARTITIONS);
|
|
|
|
|
|
|
|
log.info("entitiesPath: '{}'", entitiesPath);
|
|
|
|
log.info("workingPath: '{}'", workingPath);
|
|
|
|
log.info("dedupConfPath: '{}'", dedupConfPath);
|
|
|
|
log.info("numPartitions: '{}'", numPartitions);
|
|
|
|
|
2020-12-04 15:41:31 +01:00
|
|
|
DedupConfig dedupConf = DedupConfig.load(readFileFromHDFS(dedupConfPath));
|
|
|
|
|
|
|
|
JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
|
|
|
|
|
|
|
|
final JavaPairRDD<Object, String> vertexes = sc
|
|
|
|
.textFile(entitiesPath)
|
|
|
|
.map(s -> MapDocumentUtil.getJPathString(dedupConf.getWf().getIdPath(), s))
|
|
|
|
.mapToPair((PairFunction<String, Object, String>) s -> new Tuple2<>(hash(s), s));
|
|
|
|
|
2022-04-19 15:29:29 +02:00
|
|
|
final JavaRDD<Edge<String>> edgeRdd = spark
|
2020-12-04 15:41:31 +01:00
|
|
|
.read()
|
|
|
|
.load(workingPath + "/simrels")
|
|
|
|
.as(Encoders.bean(Relation.class))
|
|
|
|
.javaRDD()
|
2022-04-19 15:29:29 +02:00
|
|
|
.map(Relation::toEdgeRdd);
|
2020-12-04 15:41:31 +01:00
|
|
|
|
2023-04-17 11:06:27 +02:00
|
|
|
JavaPairRDD<String, List<String>> ccs = JavaGraphProcessor
|
|
|
|
.findCCs(vertexes, edgeRdd, dedupConf.getWf().getMaxIterations());
|
2020-12-04 15:41:31 +01:00
|
|
|
|
|
|
|
JavaRDD<Relation> mergeRel = ccs
|
2023-04-17 11:06:27 +02:00
|
|
|
.filter(cc -> cc._2().size() > 1)
|
2020-12-04 15:41:31 +01:00
|
|
|
.flatMap(cc -> Deduper.ccToMergeRel(cc, dedupConf))
|
|
|
|
.map(it -> new Relation(it._1(), it._2(), "mergeRel"));
|
|
|
|
|
|
|
|
final Dataset<Relation> mergeRels = spark
|
|
|
|
.createDataset(
|
|
|
|
mergeRel.rdd(),
|
|
|
|
Encoders.bean(Relation.class));
|
|
|
|
|
|
|
|
mergeRels.write().mode(SaveMode.Overwrite).parquet(workingPath + "/mergerels");
|
|
|
|
|
2020-09-29 12:01:25 +02:00
|
|
|
}
|
|
|
|
}
|