2021-08-02 19:35:23 +02:00
|
|
|
package eu.dnetlib.dhp.oa.graph.hostedbymap
|
|
|
|
|
|
|
|
import com.fasterxml.jackson.databind.ObjectMapper
|
|
|
|
import eu.dnetlib.dhp.application.ArgumentApplicationParser
|
|
|
|
import eu.dnetlib.dhp.oa.graph.hostedbymap.model.EntityInfo
|
|
|
|
import eu.dnetlib.dhp.schema.common.ModelConstants
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.utils.OafMapperUtils
|
|
|
|
import eu.dnetlib.dhp.schema.oaf.{Datasource, Instance, OpenAccessRoute, Publication}
|
|
|
|
import org.apache.commons.io.IOUtils
|
|
|
|
import org.apache.spark.SparkConf
|
|
|
|
import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession}
|
|
|
|
import org.json4s.DefaultFormats
|
|
|
|
import org.slf4j.{Logger, LoggerFactory}
|
|
|
|
|
|
|
|
import scala.collection.JavaConverters._
|
|
|
|
|
|
|
|
|
|
|
|
object SparkApplyHostedByMapToResult {
|
|
|
|
|
|
|
|
def applyHBtoPubs(join: Dataset[EntityInfo], pubs: Dataset[Publication]) = {
|
|
|
|
pubs.joinWith(join, pubs.col("id").equalTo(join.col("id")), "left")
|
|
|
|
.map(t2 => {
|
|
|
|
val p: Publication = t2._1
|
|
|
|
if (t2._2 != null) {
|
|
|
|
val ei: EntityInfo = t2._2
|
|
|
|
val i = p.getInstance().asScala
|
|
|
|
if (i.size == 1) {
|
|
|
|
val inst: Instance = i(0)
|
2021-08-13 12:00:42 +02:00
|
|
|
inst.getHostedby.setKey(ei.getHostedById)
|
2021-08-02 19:35:23 +02:00
|
|
|
inst.getHostedby.setValue(ei.getName)
|
2021-08-13 12:00:42 +02:00
|
|
|
if (ei.getOpenAccess) {
|
2021-08-02 19:35:23 +02:00
|
|
|
inst.setAccessright(OafMapperUtils.accessRight(ModelConstants.ACCESS_RIGHT_OPEN, "Open Access", ModelConstants.DNET_ACCESS_MODES, ModelConstants.DNET_ACCESS_MODES))
|
|
|
|
inst.getAccessright.setOpenAccessRoute(OpenAccessRoute.hybrid)
|
2021-08-10 17:49:45 +02:00
|
|
|
p.setBestaccessright(OafMapperUtils.createBestAccessRights(p.getInstance()));
|
2021-08-02 19:35:23 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
}
|
|
|
|
p
|
|
|
|
})(Encoders.bean(classOf[Publication]))
|
|
|
|
}
|
|
|
|
def main(args: Array[String]): Unit = {
|
|
|
|
|
|
|
|
|
|
|
|
val logger: Logger = LoggerFactory.getLogger(getClass)
|
|
|
|
val conf: SparkConf = new SparkConf()
|
|
|
|
val parser = new ArgumentApplicationParser(IOUtils.toString(getClass.getResourceAsStream("/eu/dnetlib/dhp/oa/graph/hostedbymap/hostedby_apply_params.json")))
|
|
|
|
parser.parseArgument(args)
|
|
|
|
val spark: SparkSession =
|
|
|
|
SparkSession
|
|
|
|
.builder()
|
|
|
|
.config(conf)
|
|
|
|
.appName(getClass.getSimpleName)
|
|
|
|
.master(parser.get("master")).getOrCreate()
|
|
|
|
|
|
|
|
|
|
|
|
val graphPath = parser.get("graphPath")
|
|
|
|
|
|
|
|
val outputPath = parser.get("outputPath")
|
2021-08-04 10:14:57 +02:00
|
|
|
val preparedInfoPath = parser.get("preparedInfoPath")
|
2021-08-02 19:35:23 +02:00
|
|
|
|
|
|
|
|
|
|
|
implicit val formats = DefaultFormats
|
|
|
|
|
|
|
|
|
|
|
|
implicit val mapEncoderPubs: Encoder[Publication] = Encoders.bean(classOf[Publication])
|
|
|
|
implicit val mapEncoderEinfo: Encoder[EntityInfo] = Encoders.bean(classOf[EntityInfo])
|
|
|
|
val mapper = new ObjectMapper()
|
|
|
|
|
2021-08-04 10:14:57 +02:00
|
|
|
val pubs : Dataset[Publication] = spark.read.textFile(graphPath + "/publication")
|
2021-08-02 19:35:23 +02:00
|
|
|
.map(r => mapper.readValue(r, classOf[Publication]))
|
|
|
|
|
2021-08-04 10:14:57 +02:00
|
|
|
val pinfo : Dataset[EntityInfo] = spark.read.textFile(preparedInfoPath)
|
2021-08-02 19:35:23 +02:00
|
|
|
.map(ei => mapper.readValue(ei, classOf[EntityInfo]))
|
|
|
|
|
2021-08-04 10:14:57 +02:00
|
|
|
applyHBtoPubs(pinfo, pubs).write.mode(SaveMode.Overwrite).option("compression","gzip").json(outputPath)
|
2021-08-02 19:35:23 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
}
|