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

Merge branch 'master' of https://code-repo.d4science.org/D-Net/dnet-hadoop

This commit is contained in:
Claudio Atzori 2020-07-30 16:25:06 +02:00
parent 4ff8007518 c97c8f0c44
commit 4bbfcf1ac6
1 changed files with 15 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/sx/graph/SparkSplitOafTODLIEntities.scala
View File

@ -41,28 +41,37 @@ object SparkSplitOafTODLIEntities {
.getOrCreate()
val OAFDataset:Dataset[Oaf] = spark.read.load(s"$workingPath/OAFDataset").as[Oaf]
val OAFDataset:Dataset[Oaf] = spark.read.load(s"$workingPath/input/OAFDataset").as[Oaf]
val ebi_dataset:Dataset[DLIDataset] = spark.read.load(s"$workingPath/ebi/baseline_dataset_ebi").as[DLIDataset]
val ebi_publication:Dataset[DLIPublication] = spark.read.load(s"$workingPath/ebi/baseline_publication_ebi").as[DLIPublication]
val ebi_relation:Dataset[DLIRelation] = spark.read.load(s"$workingPath/ebi/baseline_relation_ebi").as[DLIRelation]
OAFDataset
.filter(s => s != null && s.isInstanceOf[DLIPublication])
.map(s =>s.asInstanceOf[DLIPublication])
.union(ebi_publication)
.map(d => (d.getId, d))(Encoders.tuple(Encoders.STRING, pubEncoder))
.groupByKey(_._1)(Encoders.STRING)
.agg(EBIAggregator.getDLIPublicationAggregator().toColumn)
.map(p => p._2)
.repartition(1000)
.write.mode(SaveMode.Overwrite).save(s"$workingPath/publication")
.write.mode(SaveMode.Overwrite).save(s"$workingPath/graph/publication")
OAFDataset
.filter(s => s != null && s.isInstanceOf[DLIDataset])
.map(s =>s.asInstanceOf[DLIDataset])
.union(ebi_dataset)
.map(d => (d.getId, d))(Encoders.tuple(Encoders.STRING, datEncoder))
.groupByKey(_._1)(Encoders.STRING)
.agg(EBIAggregator.getDLIDatasetAggregator().toColumn)
.map(p => p._2)
.repartition(1000)
.write.mode(SaveMode.Overwrite).save(s"$workingPath/dataset")
.write.mode(SaveMode.Overwrite).save(s"$workingPath/graph/dataset")
OAFDataset
@ -73,18 +82,19 @@ object SparkSplitOafTODLIEntities {
.agg(EBIAggregator.getDLIUnknownAggregator().toColumn)
.map(p => p._2)
.repartition(1000)
.write.mode(SaveMode.Overwrite).save(s"$workingPath/unknown")
.write.mode(SaveMode.Overwrite).save(s"$workingPath/graph/unknown")
OAFDataset
.filter(s => s != null && s.isInstanceOf[DLIRelation])
.map(s =>s.asInstanceOf[DLIRelation])
.union(ebi_relation)
.map(d => (getKeyRelation(d), d))(Encoders.tuple(Encoders.STRING, relEncoder))
.groupByKey(_._1)(Encoders.STRING)
.agg(EBIAggregator.getDLIRelationAggregator().toColumn)
.map(p => p._2)
.repartition(1000)
.write.mode(SaveMode.Overwrite).save(s"$workingPath/relation")
.write.mode(SaveMode.Overwrite).save(s"$workingPath/graph/relation")