implementation of the whitelist for similarity relations #144

Merged
claudio.atzori merged 3 commits from dedup_whitelist into beta 2021-09-27 16:47:41 +02:00
5 changed files with 856 additions and 548 deletions

View File

@ -0,0 +1,151 @@
package eu.dnetlib.dhp.oa.dedup;
import java.io.IOException;
import java.util.Optional;
import org.apache.commons.io.IOUtils;
import org.apache.spark.SparkConf;
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.FilterFunction;
import org.apache.spark.api.java.function.MapFunction;
import org.apache.spark.api.java.function.PairFunction;
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.dom4j.DocumentException;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.xml.sax.SAXException;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.oa.dedup.model.Block;
import eu.dnetlib.dhp.schema.oaf.DataInfo;
import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.utils.ISLookupClientFactory;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
import eu.dnetlib.pace.config.DedupConfig;
import eu.dnetlib.pace.model.MapDocument;
import eu.dnetlib.pace.util.MapDocumentUtil;
import scala.Tuple2;
import scala.Tuple3;
public class SparkWhitelistSimRels extends AbstractSparkAction {
private static final Logger log = LoggerFactory.getLogger(SparkCreateSimRels.class);
private static final String WHITELIST_SEPARATOR = "####";
public SparkWhitelistSimRels(ArgumentApplicationParser parser, SparkSession spark) {
super(parser, spark);
}
public static void main(String[] args) throws Exception {
ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils
.toString(
SparkCreateSimRels.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/dedup/whitelistSimRels_parameters.json")));
parser.parseArgument(args);
SparkConf conf = new SparkConf();
new SparkWhitelistSimRels(parser, getSparkSession(conf))
.run(ISLookupClientFactory.getLookUpService(parser.get("isLookUpUrl")));
}
@Override
public void run(ISLookUpService isLookUpService)
throws DocumentException, IOException, ISLookUpException, SAXException {
// read oozie parameters
final String graphBasePath = parser.get("graphBasePath");
final String isLookUpUrl = parser.get("isLookUpUrl");
final String actionSetId = parser.get("actionSetId");
final String workingPath = parser.get("workingPath");
final int numPartitions = Optional
.ofNullable(parser.get("numPartitions"))
.map(Integer::valueOf)
.orElse(NUM_PARTITIONS);
final String whiteListPath = parser.get("whiteListPath");
log.info("numPartitions: '{}'", numPartitions);
log.info("graphBasePath: '{}'", graphBasePath);
log.info("isLookUpUrl: '{}'", isLookUpUrl);
log.info("actionSetId: '{}'", actionSetId);
log.info("workingPath: '{}'", workingPath);
log.info("whiteListPath: '{}'", whiteListPath);
JavaSparkContext sc = JavaSparkContext.fromSparkContext(spark.sparkContext());
//file format: source####target
Dataset<Tuple2<String, String>> whiteListRels = spark.createDataset(sc
.textFile(whiteListPath)
//check if the line is in the correct format: id1####id2
.filter(s -> s.contains(WHITELIST_SEPARATOR) && s.split(WHITELIST_SEPARATOR).length == 2)
.map(s -> new Tuple2<>(s.split(WHITELIST_SEPARATOR)[0], s.split(WHITELIST_SEPARATOR)[1]))
.rdd(),
Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
// for each dedup configuration
for (DedupConfig dedupConf : getConfigurations(isLookUpService, actionSetId)) {
final String entity = dedupConf.getWf().getEntityType();
final String subEntity = dedupConf.getWf().getSubEntityValue();
log.info("Adding whitelist simrels for: '{}'", subEntity);
final String outputPath = DedupUtility.createSimRelPath(workingPath, actionSetId, subEntity);
Dataset<Tuple2<String, String>> entities = spark.createDataset(sc
.textFile(DedupUtility.createEntityPath(graphBasePath, subEntity))
.repartition(numPartitions)
.mapToPair(
(PairFunction<String, String, String>) s -> {
MapDocument d = MapDocumentUtil.asMapDocumentWithJPath(dedupConf, s);
return new Tuple2<>(d.getIdentifier(), "present");
})
.rdd(),
Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
Dataset<Tuple2<String, String>> whiteListRels1 = whiteListRels
.joinWith(entities, whiteListRels.col("_1").equalTo(entities.col("_1")), "inner")
.map((MapFunction<Tuple2<Tuple2<String, String>, Tuple2<String, String>>, Tuple2<String, String>>) Tuple2::_1, Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
Dataset<Tuple2<String, String>> whiteListRels2 = whiteListRels1
.joinWith(entities, whiteListRels1.col("_2").equalTo(entities.col("_1")), "inner")
.map((MapFunction<Tuple2<Tuple2<String, String>, Tuple2<String, String>>, Tuple2<String, String>>) Tuple2::_1, Encoders.tuple(Encoders.STRING(), Encoders.STRING()));
Dataset<Relation> whiteListSimRels = whiteListRels2
.map((MapFunction<Tuple2<String, String>, Relation>)
r -> createSimRel(r._1(), r._2(), entity),
Encoders.bean(Relation.class)
);
saveParquet(whiteListSimRels, outputPath, SaveMode.Append);
}
}
private Relation createSimRel(String source, String target, String entity) {
final Relation r = new Relation();
r.setSource(source);
r.setTarget(target);
r.setSubRelType("dedupSimilarity");
r.setRelClass("isSimilarTo");
r.setDataInfo(new DataInfo());
switch (entity) {
case "result":
r.setRelType("resultResult");
break;
case "organization":
r.setRelType("organizationOrganization");
break;
default:
throw new IllegalArgumentException("unmanaged entity type: " + entity);
}
return r;
}
}

View File

@ -20,6 +20,10 @@
<name>workingPath</name> <name>workingPath</name>
<description>path for the working directory</description> <description>path for the working directory</description>
</property> </property>
<property>
<name>whiteListPath</name>
<description>path for the whitelist of similarity relations</description>
</property>
<property> <property>
<name>dedupGraphPath</name> <name>dedupGraphPath</name>
<description>path for the output graph</description> <description>path for the output graph</description>
@ -130,6 +134,34 @@
<arg>--workingPath</arg><arg>${workingPath}</arg> <arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--numPartitions</arg><arg>8000</arg> <arg>--numPartitions</arg><arg>8000</arg>
</spark> </spark>
<ok to="WhitelistSimRels"/>
<error to="Kill"/>
</action>
<action name="WhitelistSimRels">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Add Whitelist Similarity Relations</name>
<class>eu.dnetlib.dhp.oa.dedup.SparkWhitelistSimRels</class>
<jar>dhp-dedup-openaire-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=3840
</spark-opts>
<arg>--graphBasePath</arg><arg>${graphBasePath}</arg>
<arg>--isLookUpUrl</arg><arg>${isLookUpUrl}</arg>
<arg>--actionSetId</arg><arg>${actionSetId}</arg>
<arg>--workingPath</arg><arg>${workingPath}</arg>
<arg>--whiteListPath</arg><arg>${whiteListPath}</arg>
<arg>--numPartitions</arg><arg>8000</arg>
</spark>
<ok to="CreateMergeRel"/> <ok to="CreateMergeRel"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>

View File

@ -0,0 +1,38 @@
[
{
"paramName": "la",
"paramLongName": "isLookUpUrl",
"paramDescription": "address for the LookUp",
"paramRequired": true
},
{
"paramName": "asi",
"paramLongName": "actionSetId",
"paramDescription": "action set identifier (name of the orchestrator)",
"paramRequired": true
},
{
"paramName": "i",
"paramLongName": "graphBasePath",
"paramDescription": "the base path of the raw graph",
"paramRequired": true
},
{
"paramName": "w",
"paramLongName": "workingPath",
"paramDescription": "path of the working directory",
"paramRequired": true
},
{
"paramName": "np",
"paramLongName": "numPartitions",
"paramDescription": "number of partitions for the similarity relations intermediate phases",
"paramRequired": false
},
{
"paramName": "wl",
"paramLongName": "whiteListPath",
"paramDescription": "whitelist file path for the addition of custom simrels",
"paramRequired": true
}
]

View File

@ -0,0 +1,2 @@
50|r37b0ad08687::f645b9729d1e1025a72c57883f0f2cac####50|r37b0ad08687::4c55b436743b5c49fa32cd582fd9e1aa
50|datacite____::a90f49f9fde5393c00633bea6e4e374a####50|datacite____::5f55cdee77303ba8a2bf9996c32a330c