sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

updated CreateInputGraph because ggenerics don't work on Spark Dataset

This commit is contained in:
Sandro La Bruzzo 2021-07-09 10:29:24 +02:00
parent cd17e19044
commit 0ea576745f
2 changed files with 24 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/sx/graph/SparkCreateInputGraph.scala
View File

@ -34,6 +34,15 @@ object SparkCreateInputGraph {
)
implicit val oafEncoder: Encoder[Oaf] = Encoders.kryo(classOf[Oaf])
implicit val publicationEncoder: Encoder[Publication] = Encoders.kryo(classOf[Publication])
implicit val datasetEncoder: Encoder[OafDataset] = Encoders.kryo(classOf[OafDataset])
implicit val softwareEncoder: Encoder[Software] = Encoders.kryo(classOf[Software])
implicit val orpEncoder: Encoder[OtherResearchProduct] = Encoders.kryo(classOf[OtherResearchProduct])
implicit val relEncoder: Encoder[Relation] = Encoders.kryo(classOf[Relation])
val sourcePath = parser.get("sourcePath")
log.info(s"sourcePath -> $sourcePath")
@ -43,9 +52,21 @@ object SparkCreateInputGraph {
val oafDs:Dataset[Oaf] = spark.read.load(s"$sourcePath/*").as[Oaf]
resultObject.foreach(r => extractEntities(oafDs,s"$targetPath/extracted/${r._1}", r._2, log))
extractEntities(oafDs,s"$targetPath/extracted/relation", classOf[Relation], log)
log.info("Extract Publication")
oafDs.filter(o => o.isInstanceOf[Publication]).map(p => p.asInstanceOf[Publication]).write.mode(SaveMode.Overwrite).save(s"$targetPath/extracted/publication")
log.info("Extract dataset")
oafDs.filter(o => o.isInstanceOf[OafDataset]).map(p => p.asInstanceOf[OafDataset]).write.mode(SaveMode.Overwrite).save(s"$targetPath/extracted/dataset")
log.info("Extract software")
oafDs.filter(o => o.isInstanceOf[Software]).map(p => p.asInstanceOf[Software]).write.mode(SaveMode.Overwrite).save(s"$targetPath/extracted/software")
log.info("Extract otherResearchProduct")
oafDs.filter(o => o.isInstanceOf[OtherResearchProduct]).map(p => p.asInstanceOf[OtherResearchProduct]).write.mode(SaveMode.Overwrite).save(s"$targetPath/extracted/otherResearchProduct")
log.info("Extract Relation")
oafDs.filter(o => o.isInstanceOf[Relation]).map(p => p.asInstanceOf[Relation]).write.mode(SaveMode.Overwrite).save(s"$targetPath/extracted/relation")
resultObject.foreach { r =>
log.info(s"Make ${r._1} unique")

2
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/sx/graph/scholix/ScholixUtils.scala
View File

@ -331,7 +331,7 @@ object ScholixUtils {
s.setDate(dt.distinct.asJava)
}
if (r.getDescription!= null && !r.getDescription.isEmpty) {
val d = r.getDescription.asScala.find(f => f.getValue!=null)
val d = r.getDescription.asScala.find(f => f!= null && f.getValue!=null)
if (d.isDefined)
s.setDescription(d.get.getValue)
}