2020-03-27 10:42:17 +01:00
|
|
|
package eu.dnetlib.dhp.oa.dedup;
|
2019-12-06 13:38:00 +01:00
|
|
|
|
2019-12-18 16:28:22 +01:00
|
|
|
import com.google.common.hash.Hashing;
|
2020-03-27 10:42:17 +01:00
|
|
|
import eu.dnetlib.dhp.oa.dedup.graph.ConnectedComponent;
|
2019-12-06 13:38:00 +01:00
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
|
2020-03-27 10:42:17 +01:00
|
|
|
import eu.dnetlib.dhp.oa.dedup.graph.GraphProcessor;
|
2019-12-06 13:38:00 +01:00
|
|
|
import eu.dnetlib.dhp.schema.oaf.Relation;
|
2020-03-20 13:01:56 +01:00
|
|
|
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
|
2019-12-06 13:38:00 +01:00
|
|
|
import eu.dnetlib.pace.config.DedupConfig;
|
|
|
|
import eu.dnetlib.pace.util.MapDocumentUtil;
|
|
|
|
import org.apache.commons.io.IOUtils;
|
2020-03-20 13:01:56 +01:00
|
|
|
import org.apache.spark.SparkConf;
|
2019-12-06 13:38:00 +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.FlatMapFunction;
|
|
|
|
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.SparkSession;
|
2020-03-20 13:01:56 +01:00
|
|
|
import org.dom4j.DocumentException;
|
2019-12-06 13:38:00 +01:00
|
|
|
import scala.Tuple2;
|
|
|
|
|
|
|
|
import java.util.ArrayList;
|
|
|
|
import java.util.List;
|
|
|
|
|
|
|
|
public class SparkCreateConnectedComponent {
|
|
|
|
|
|
|
|
public static void main(String[] args) throws Exception {
|
2020-03-27 10:42:17 +01:00
|
|
|
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
|
|
|
|
IOUtils.toString(
|
|
|
|
SparkCreateConnectedComponent.class.getResourceAsStream("/eu/dnetlib/dhp/oa/dedup/createCC_parameters.json")));
|
2019-12-06 13:38:00 +01:00
|
|
|
parser.parseArgument(args);
|
|
|
|
|
2020-03-20 13:01:56 +01:00
|
|
|
new SparkCreateConnectedComponent().run(parser);
|
|
|
|
}
|
|
|
|
|
|
|
|
private void run(ArgumentApplicationParser parser) throws ISLookUpException, DocumentException {
|
|
|
|
|
|
|
|
final String graphBasePath = parser.get("graphBasePath");
|
|
|
|
final String workingPath = parser.get("workingPath");
|
|
|
|
final String isLookUpUrl = parser.get("isLookUpUrl");
|
|
|
|
final String actionSetId = parser.get("actionSetId");
|
|
|
|
|
|
|
|
try (SparkSession spark = getSparkSession(parser)) {
|
|
|
|
|
|
|
|
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
|
|
|
|
|
|
|
|
for (DedupConfig dedupConf: DedupUtility.getConfigurations(isLookUpUrl, actionSetId)) {
|
|
|
|
|
|
|
|
final String entity = dedupConf.getWf().getEntityType();
|
|
|
|
final String subEntity = dedupConf.getWf().getSubEntityValue();
|
|
|
|
|
|
|
|
final JavaPairRDD<Object, String> vertexes = sc.textFile(graphBasePath + "/" + subEntity)
|
|
|
|
.map(s -> MapDocumentUtil.getJPathString(dedupConf.getWf().getIdPath(), s))
|
|
|
|
.mapToPair((PairFunction<String, Object, String>)
|
|
|
|
s -> new Tuple2<Object, String>(getHashcode(s), s)
|
|
|
|
);
|
|
|
|
|
|
|
|
final Dataset<Relation> similarityRelations = spark.read().load(DedupUtility.createSimRelPath(workingPath, actionSetId, subEntity)).as(Encoders.bean(Relation.class));
|
|
|
|
final RDD<Edge<String>> edgeRdd = similarityRelations.javaRDD().map(it -> new Edge<>(getHashcode(it.getSource()), getHashcode(it.getTarget()), it.getRelClass())).rdd();
|
|
|
|
final JavaRDD<ConnectedComponent> cc = GraphProcessor.findCCs(vertexes.rdd(), edgeRdd, dedupConf.getWf().getMaxIterations()).toJavaRDD();
|
|
|
|
final Dataset<Relation> mergeRelation = spark.createDataset(cc.filter(k -> k.getDocIds().size() > 1).flatMap((FlatMapFunction<ConnectedComponent, Relation>) c ->
|
|
|
|
c.getDocIds()
|
|
|
|
.stream()
|
|
|
|
.flatMap(id -> {
|
|
|
|
List<Relation> tmp = new ArrayList<>();
|
|
|
|
Relation r = new Relation();
|
|
|
|
r.setSource(c.getCcId());
|
|
|
|
r.setTarget(id);
|
|
|
|
r.setRelClass("merges");
|
|
|
|
tmp.add(r);
|
|
|
|
r = new Relation();
|
|
|
|
r.setTarget(c.getCcId());
|
|
|
|
r.setSource(id);
|
|
|
|
r.setRelClass("isMergedIn");
|
|
|
|
tmp.add(r);
|
|
|
|
return tmp.stream();
|
|
|
|
}).iterator()).rdd(), Encoders.bean(Relation.class));
|
|
|
|
mergeRelation.write().mode("overwrite").save(DedupUtility.createMergeRelPath(workingPath, actionSetId, entity));
|
|
|
|
}
|
|
|
|
}
|
2019-12-18 16:28:22 +01:00
|
|
|
}
|
2019-12-06 13:38:00 +01:00
|
|
|
|
2019-12-18 16:28:22 +01:00
|
|
|
public static long getHashcode(final String id) {
|
2020-03-20 16:48:14 +01:00
|
|
|
return Hashing.murmur3_128().hashString(id).asLong();
|
2019-12-06 13:38:00 +01:00
|
|
|
}
|
2020-03-20 13:01:56 +01:00
|
|
|
|
|
|
|
private static SparkSession getSparkSession(ArgumentApplicationParser parser) {
|
|
|
|
SparkConf conf = new SparkConf();
|
|
|
|
|
|
|
|
return SparkSession
|
|
|
|
.builder()
|
|
|
|
.appName(SparkCreateSimRels.class.getSimpleName())
|
|
|
|
.master(parser.get("master"))
|
|
|
|
.config(conf)
|
|
|
|
.getOrCreate();
|
|
|
|
}
|
2019-12-06 13:38:00 +01:00
|
|
|
}
|