alessia.bardi
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Make the node of doiBoost to point to the correct OpenAire Organization in relations

This commit is contained in:
Sandro La Bruzzo 2021-09-28 11:31:19 +02:00
parent c48c43fa9e
commit 8f99d2af86
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/SparkGenerateDoiBoost.scala
View File

@ -208,7 +208,7 @@ object SparkGenerateDoiBoost {
(r.getTarget,r)
else
("resolved", r)
})
})(Encoders.tuple(Encoders.STRING, mapEncoderRel))
val openaireOrganization:Dataset[(String,String)] = spark.read.text(openaireOrganizationPath).as[String].flatMap(s => extractIdGRID(s)).groupByKey(_._2).reduceGroups((x,y) => if (x != null) x else y ).map(_._2)
@ -222,7 +222,7 @@ object SparkGenerateDoiBoost {
else
currentRels.setTarget(currentOrgs._1)
currentRels
}.write.save(s"$workingDirPath/doiBoostPublicationAffiliation")
}.filter(r=> !r.getSource.startsWith("unresolved") && !r.getTarget.startsWith("unresolved")).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/doiBoostPublicationAffiliation")
magPubs.joinWith(a,magPubs("_1").equalTo(a("PaperId"))).map( item => {
val affiliation = item._2