added scholix generation

This commit is contained in:
Sandro La Bruzzo 2021-07-06 17:18:06 +02:00
parent 4c54bd8742
commit 8535506c22
5 changed files with 125 additions and 40 deletions

View File

@ -4,9 +4,10 @@ import eu.dnetlib.dhp.application.ArgumentApplicationParser
import eu.dnetlib.dhp.schema.oaf.Relation import eu.dnetlib.dhp.schema.oaf.Relation
import eu.dnetlib.dhp.schema.sx.scholix.Scholix import eu.dnetlib.dhp.schema.sx.scholix.Scholix
import eu.dnetlib.dhp.schema.sx.summary.ScholixSummary import eu.dnetlib.dhp.schema.sx.summary.ScholixSummary
import eu.dnetlib.dhp.sx.graph.scholix.ScholixUtils
import org.apache.commons.io.IOUtils import org.apache.commons.io.IOUtils
import org.apache.spark.SparkConf import org.apache.spark.SparkConf
import org.apache.spark.sql.{Dataset, Encoder, Encoders, SparkSession} import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession}
import org.slf4j.{Logger, LoggerFactory} import org.slf4j.{Logger, LoggerFactory}
object SparkCreateScholix { object SparkCreateScholix {
@ -31,40 +32,45 @@ object SparkCreateScholix {
log.info(s"targetPath -> $targetPath") log.info(s"targetPath -> $targetPath")
implicit val relEncoder:Encoder[Relation] = Encoders.kryo[Relation] implicit val relEncoder: Encoder[Relation] = Encoders.kryo[Relation]
implicit val summaryEncoder :Encoder[ScholixSummary] = Encoders.kryo[ScholixSummary] implicit val summaryEncoder: Encoder[ScholixSummary] = Encoders.kryo[ScholixSummary]
implicit val scholixEncoder :Encoder[Scholix] = Encoders.kryo[Scholix] implicit val scholixEncoder: Encoder[Scholix] = Encoders.kryo[Scholix]
import spark.implicits._
val relationDS:Dataset[(String, Relation)] = spark.read.load(relationPath).as[Relation] val relationDS: Dataset[(String, Relation)] = spark.read.load(relationPath).as[Relation]
.map(r => (r.getSource, r))(Encoders.tuple(Encoders.STRING, relEncoder)) .map(r => (r.getSource, r))(Encoders.tuple(Encoders.STRING, relEncoder))
val summaryDS:Dataset[(String, ScholixSummary)] = spark.read.load(summaryPath).as[ScholixSummary]
.map(r => (r.getId, r))(Encoders.tuple(Encoders.STRING, summaryEncoder))
val res: Array[((String, Relation), (String, ScholixSummary))] =relationDS.joinWith(summaryDS, relationDS("_1").equalTo(summaryDS("_1")), "left").take(10)
res.foreach(r =>println(r._1._2))
// relationDS.joinWith(summaryDS, relationDS("_1").equalTo(summaryDS("_1")), "left")
// .map {input:((String,Relation), (String, ScholixSummary)) =>
// val rel:Relation = input._1._2
// val source:ScholixSummary = input._2._2
//
//
// val s = new Scholix
//
//
// }
val summaryDS: Dataset[(String, ScholixSummary)] = spark.read.load(summaryPath).as[ScholixSummary]
.map(r => (r.getId, r))(Encoders.tuple(Encoders.STRING, summaryEncoder))
relationDS.joinWith(summaryDS, relationDS("_1").equalTo(summaryDS("_1")), "left")
.map { input: ((String, Relation), (String, ScholixSummary)) =>
val rel: Relation = input._1._2
val source: ScholixSummary = input._2._2
(rel.getTarget, ScholixUtils.scholixFromSource(rel, source))
}(Encoders.tuple(Encoders.STRING, scholixEncoder))
.write.mode(SaveMode.Overwrite).save(s"$targetPath/scholix_from_source")
val scholixSource: Dataset[(String, Scholix)] = spark.read.load(s"$targetPath/scholix_from_source").as[(String, Scholix)](Encoders.tuple(Encoders.STRING, scholixEncoder))
scholixSource.joinWith(summaryDS, scholixSource("_1").equalTo(summaryDS("_1")), "left")
.map { input: ((String, Scholix), (String, ScholixSummary)) =>
val s: Scholix = input._1._2
val target: ScholixSummary = input._2._2
ScholixUtils.generateCompleteScholix(s, target)
}.write.mode(SaveMode.Overwrite).save(s"$targetPath/scholix_one_verse")
val scholix_o_v: Dataset[Scholix] = spark.read.load(s"$targetPath/scholix_one_verse").as[Scholix]
scholix_o_v.flatMap(s => List(s, ScholixUtils.createInverseScholixRelation(s))).groupByKey(_.getIdentifier).reduceGroups { (x, y) =>
if (x != null)
x
else
y
}.write.mode(SaveMode.Overwrite).save(s"$targetPath/scholix")
} }
} }

View File

@ -4,6 +4,7 @@ package eu.dnetlib.dhp.sx.graph.scholix
import eu.dnetlib.dhp.schema.oaf.{Dataset, Relation, Result, StructuredProperty} import eu.dnetlib.dhp.schema.oaf.{Dataset, Relation, Result, StructuredProperty}
import eu.dnetlib.dhp.schema.sx.scholix.{Scholix, ScholixCollectedFrom, ScholixEntityId, ScholixIdentifier, ScholixRelationship, ScholixResource} import eu.dnetlib.dhp.schema.sx.scholix.{Scholix, ScholixCollectedFrom, ScholixEntityId, ScholixIdentifier, ScholixRelationship, ScholixResource}
import eu.dnetlib.dhp.schema.sx.summary.{CollectedFromType, SchemeValue, ScholixSummary, Typology} import eu.dnetlib.dhp.schema.sx.summary.{CollectedFromType, SchemeValue, ScholixSummary, Typology}
import eu.dnetlib.dhp.utils.DHPUtils
import org.json4s import org.json4s
import org.json4s.DefaultFormats import org.json4s.DefaultFormats
import org.json4s.jackson.JsonMethods.parse import org.json4s.jackson.JsonMethods.parse
@ -45,13 +46,35 @@ object ScholixUtils {
def extractRelationDate(summary: ScholixSummary):String = { def extractRelationDate(summary: ScholixSummary):String = {
if(summary.getDate== null && !summary.getDate.isEmpty) if(summary.getDate== null || summary.getDate.isEmpty)
null null
else { else {
summary.getDate.get(0) summary.getDate.get(0)
} }
}
def inverseRelationShip(rel:ScholixRelationship):ScholixRelationship = {
new ScholixRelationship(rel.getInverse, rel.getSchema, rel.getName)
}
def createInverseScholixRelation(scholix: Scholix):Scholix = {
val s = new Scholix
s.setPublicationDate(scholix.getPublicationDate)
s.setPublisher(scholix.getPublisher)
s.setLinkprovider(scholix.getLinkprovider)
s.setRelationship(inverseRelationShip(scholix.getRelationship))
s.setSource(scholix.getTarget)
s.setTarget(scholix.getSource)
s.setIdentifier(DHPUtils.md5(s"${s.getSource.getIdentifier}::${s.getRelationship.getName}::${s.getTarget.getIdentifier}"))
s
} }
@ -78,6 +101,19 @@ object ScholixUtils {
} }
def generateCompleteScholix(scholix: Scholix, target:ScholixSummary): Scholix = {
val s = new Scholix
s.setPublicationDate(scholix.getPublicationDate)
s.setPublisher(scholix.getPublisher)
s.setLinkprovider(scholix.getLinkprovider)
s.setRelationship(scholix.getRelationship)
s.setSource(scholix.getSource)
s.setTarget(generateScholixResourceFromSummary(target))
s.setIdentifier(DHPUtils.md5(s"${s.getSource.getIdentifier}::${s.getRelationship.getName}::${s.getTarget.getIdentifier}"))
s
}
def generateScholixResourceFromSummary(summaryObject: ScholixSummary): ScholixResource = { def generateScholixResourceFromSummary(summaryObject: ScholixSummary): ScholixResource = {
val r = new ScholixResource val r = new ScholixResource
r.setIdentifier(summaryObject.getLocalIdentifier) r.setIdentifier(summaryObject.getLocalIdentifier)

View File

@ -0,0 +1,6 @@
[
{"paramName":"mt", "paramLongName":"master", "paramDescription": "should be local or yarn", "paramRequired": true},
{"paramName":"r", "paramLongName":"relationPath", "paramDescription": "the relation resolved Path", "paramRequired": true},
{"paramName":"s", "paramLongName":"summaryPath", "paramDescription": "the summary Path", "paramRequired": true},
{"paramName":"t", "paramLongName":"targetPath", "paramDescription": "the target base path of the scholix", "paramRequired": true}
]

View File

@ -10,7 +10,7 @@
</property> </property>
</parameters> </parameters>
<start to="CreateSummaries"/> <start to="CreateScholix"/>
<kill name="Kill"> <kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message> <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
@ -90,6 +90,33 @@
<arg>--sourcePath</arg><arg>${targetPath}/dedup</arg> <arg>--sourcePath</arg><arg>${targetPath}/dedup</arg>
<arg>--targetPath</arg><arg>${targetPath}/provision/summaries</arg> <arg>--targetPath</arg><arg>${targetPath}/provision/summaries</arg>
</spark> </spark>
<ok to="CreateScholix"/>
<error to="Kill"/>
</action>
<action name="CreateScholix">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn</master>
<mode>cluster</mode>
<name>Generate Scholix Dataset</name>
<class>eu.dnetlib.dhp.sx.graph.SparkCreateScholix</class>
<jar>dhp-graph-mapper-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.shuffle.partitions=6000
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--master</arg><arg>yarn</arg>
<arg>--summaryPath</arg><arg>${targetPath}/provision/summaries</arg>
<arg>--targetPath</arg><arg>${targetPath}/provision/scholix</arg>
<arg>--relationPath</arg><arg>${targetPath}/resolved/resolvedRelation</arg>
</spark>
<ok to="End"/> <ok to="End"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>

File diff suppressed because one or more lines are too long