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

minor fix

This commit is contained in:
Sandro La Bruzzo 2020-05-22 20:51:42 +02:00
parent 72278b9375
commit 147dd389bf
2 changed files with 41 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/SparkGenerateDoiBoost.scala
View File

@ -23,21 +23,20 @@ object SparkGenerateDoiBoost {
.master(parser.get("master")).getOrCreate()
val crossrefPublicationPath =parser.get("crossrefPublicationPath")
val crossrefDatasetPath =parser.get("crossrefDatasetPath")
val uwPublicationPath =parser.get("uwPublicationPath")
val magPublicationPath =parser.get("magPublicationPath")
val orcidPublicationPath =parser.get("orcidPublicationPath")
val workingDirPath =parser.get("workingDirPath")
val crossrefPublicationPath = parser.get("crossrefPublicationPath")
val crossrefDatasetPath = parser.get("crossrefDatasetPath")
val uwPublicationPath = parser.get("uwPublicationPath")
val magPublicationPath = parser.get("magPublicationPath")
val orcidPublicationPath = parser.get("orcidPublicationPath")
val workingDirPath = parser.get("workingDirPath")
logger.info("Phase 1) repartition and move all the dataset in a same working folder")
spark.read.load(crossrefPublicationPath).as(Encoders.bean(classOf[Publication])).map(s=>s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/crossrefPublication")
spark.read.load(crossrefDatasetPath).as(Encoders.bean(classOf[OafDataset])).map(s=>s)(Encoders.kryo[OafDataset]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/crossrefDataset")
spark.read.load(uwPublicationPath).as(Encoders.bean(classOf[Publication])).map(s=>s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/uwPublication")
spark.read.load(orcidPublicationPath).as(Encoders.bean(classOf[Publication])).map(s=>s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/orcidPublication")
spark.read.load(magPublicationPath).as(Encoders.bean(classOf[Publication])).map(s=>s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/magPublication")
spark.read.load(crossrefPublicationPath).as(Encoders.bean(classOf[Publication])).map(s => s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/crossrefPublication")
spark.read.load(crossrefDatasetPath).as(Encoders.bean(classOf[OafDataset])).map(s => s)(Encoders.kryo[OafDataset]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/crossrefDataset")
spark.read.load(uwPublicationPath).as(Encoders.bean(classOf[Publication])).map(s => s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/uwPublication")
spark.read.load(orcidPublicationPath).as(Encoders.bean(classOf[Publication])).map(s => s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/orcidPublication")
spark.read.load(magPublicationPath).as(Encoders.bean(classOf[Publication])).map(s => s)(Encoders.kryo[Publication]).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/magPublication")
implicit val mapEncoderPub: Encoder[Publication] = Encoders.kryo[Publication]
implicit val mapEncoderDataset: Encoder[OafDataset] = Encoders.kryo[OafDataset]
@ -45,17 +44,39 @@ object SparkGenerateDoiBoost {
logger.info("Phase 2) Join Crossref with UnpayWall")
val crossrefPublication:Dataset[(String,Publication)] = spark.read.load(s"$workingDirPath/crossrefPublication").as[Publication].map(p =>(p.getId, p) )
val uwPublication:Dataset[(String,Publication)] = spark.read.load(s"$workingDirPath/crossrefPublication").as[Publication].map(p =>(p.getId, p) )
val crossrefPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/crossrefPublication").as[Publication].map(p => (p.getId, p))
val uwPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/uwPublication").as[Publication].map(p => (p.getId, p))
crossrefPublication.joinWith(uwPublication, crossrefPublication("_1").equalTo(uwPublication("_1")),"left").map(item => {
def applyMerge(item:((String, Publication), (String, Publication))) : Publication =
{
val crossrefPub = item._1._2
val unpayWallPub = item._1._2
if(unpayWallPub!= null) {
crossrefPub.mergeFrom(unpayWallPub)
if (item._2!= null) {
val otherPub = item._2._2
if (otherPub != null) {
crossrefPub.mergeFrom(otherPub)
}
}
crossrefPub
}).write.save(s"$workingDirPath/firstJoin")
}
crossrefPublication.joinWith(uwPublication, crossrefPublication("_1").equalTo(uwPublication("_1")), "left").map(applyMerge).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/firstJoin")
logger.info("Phase 3) Join Result with ORCID")
val fj: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/firstJoin").as[Publication].map(p => (p.getId, p))
val orcidPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/orcidPublication").as[Publication].map(p => (p.getId, p))
fj.joinWith(orcidPublication, fj("_1").equalTo(orcidPublication("_1")), "left").map(applyMerge).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/secondJoin")
logger.info("Phase 3) Join Result with MAG")
val sj: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/secondJoin").as[Publication].map(p => (p.getId, p))
sj.where(sj("_1").like())
val magPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/magPublication").as[Publication].map(p => (p.getId, p))
sj.joinWith(magPublication, sj("_1").equalTo(magPublication("_1")), "left").map(applyMerge).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/doiBoostPublication")
}

2
dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/oozie_app/workflow.xml
View File

@ -40,7 +40,7 @@
<start to="ResetWorkingPath"/>
<start to="CreateDOIBoost"/>
<kill name="Kill">