removed Orcid intersection on DOIBoost

This commit is contained in:
Sandro La Bruzzo 2023-11-30 14:36:50 +01:00
parent 279100fa52
commit 7b5e04f37e
7 changed files with 128 additions and 177 deletions

View File

@ -133,32 +133,6 @@
<arg>--targetPath</arg><arg>${inputPathMAG}/dataset</arg>
<arg>--master</arg><arg>yarn-cluster</arg>
</spark>
<ok to="PreProcessORCID"/>
<error to="Kill"/>
</action>
<!-- ORCID SECTION -->
<action name="PreProcessORCID">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Convert ORCID to Dataset</name>
<class>eu.dnetlib.doiboost.orcid.SparkPreprocessORCID</class>
<jar>dhp-doiboost-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.sql.shuffle.partitions=3840
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--sourcePath</arg><arg>${inputPathOrcid}</arg>
<arg>--workingPath</arg><arg>${workingPathOrcid}</arg>
<arg>--master</arg><arg>yarn-cluster</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>
</action>

View File

@ -59,10 +59,10 @@
</property>
<!-- ORCID Parameters -->
<property>
<name>workingPathOrcid</name>
<description>the ORCID working path</description>
</property>
<!-- <property>-->
<!-- <name>workingPathOrcid</name>-->
<!-- <description>the ORCID working path</description>-->
<!-- </property>-->
</parameters>
@ -170,32 +170,6 @@
<arg>--targetPath</arg><arg>${workingPath}/uwPublication</arg>
<arg>--master</arg><arg>yarn-cluster</arg>
</spark>
<ok to="ProcessORCID"/>
<error to="Kill"/>
</action>
<!-- ORCID SECTION -->
<action name="ProcessORCID">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>
<mode>cluster</mode>
<name>Convert ORCID to Dataset</name>
<class>eu.dnetlib.doiboost.orcid.SparkConvertORCIDToOAF</class>
<jar>dhp-doiboost-${projectVersion}.jar</jar>
<spark-opts>
--executor-memory=${sparkExecutorMemory}
--executor-cores=${sparkExecutorCores}
--driver-memory=${sparkDriverMemory}
--conf spark.sql.shuffle.partitions=3840
--conf spark.extraListeners=${spark2ExtraListeners}
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts>
<arg>--workingPath</arg><arg>${workingPathOrcid}</arg>
<arg>--targetPath</arg><arg>${workingPath}/orcidPublication</arg>
<arg>--master</arg><arg>yarn-cluster</arg>
</spark>
<ok to="CreateDOIBoost"/>
<error to="Kill"/>
</action>

View File

@ -66,7 +66,7 @@ object SparkGenerateDoiBoost {
Encoders.tuple(Encoders.STRING, mapEncoderPub)
implicit val mapEncoderRel: Encoder[Relation] = Encoders.kryo[Relation]
logger.info("Phase 2) Join Crossref with UnpayWall")
logger.info("Phase 1) Join Crossref with UnpayWall")
val crossrefPublication: Dataset[(String, Publication)] =
spark.read.load(s"$workingDirPath/crossrefPublication").as[Publication].map(p => (p.getId, p))
@ -91,20 +91,11 @@ object SparkGenerateDoiBoost {
.write
.mode(SaveMode.Overwrite)
.save(s"$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"$workingDirPath/secondJoin")
logger.info("Phase 4) Join Result with MAG")
logger.info("Phase 2) Join Result with MAG")
val sj: Dataset[(String, Publication)] =
spark.read.load(s"$workingDirPath/secondJoin").as[Publication].map(p => (p.getId, p))
spark.read.load(s"$workingDirPath/firstJoin").as[Publication].map(p => (p.getId, p))
val magPublication: Dataset[(String, Publication)] =
spark.read.load(s"$workingDirPath/magPublication").as[Publication].map(p => (p.getId, p))

View File

@ -40,7 +40,7 @@
</spark-opts>
<arg>--orcidPath</arg><arg>${orcidPath}</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg>
<arg>--graphPath</arg><arg>${graphPath}/publication</arg>
<arg>--graphPath</arg><arg>${graphPath}</arg>
<arg>--master</arg><arg>yarn</arg>
</spark>
<ok to="End"/>

View File

@ -1,5 +1,6 @@
package eu.dnetlib.dhp.enrich.orcid
import eu.dnetlib.dhp.schema.common.ModelConstants
import eu.dnetlib.dhp.schema.oaf.{Author, Publication}
import eu.dnetlib.dhp.schema.sx.OafUtils
import org.apache.spark.sql.Row
@ -13,9 +14,11 @@ object AuthorEnricher extends Serializable {
a.setName(givenName)
a.setSurname(familyName)
a.setFullname(s"$givenName $familyName")
a.setPid(List(OafUtils.createSP(orcid, "ORCID", "ORCID")).asJava)
val pid = OafUtils.createSP(orcid, ModelConstants.ORCID, ModelConstants.ORCID)
pid.setDataInfo(OafUtils.generateDataInfo())
pid.getDataInfo.setProvenanceaction(OafUtils.createQualifier("ORCID_ENRICHMENT", "ORCID_ENRICHMENT"))
a.setPid(List(pid).asJava)
a
}
def toOAFAuthor(r: Row): java.util.List[Author] = {

View File

@ -1,13 +1,11 @@
package eu.dnetlib.dhp.enrich.orcid
import com.fasterxml.jackson.databind.ObjectMapper
import eu.dnetlib.dhp.application.AbstractScalaApplication
import eu.dnetlib.dhp.oa.merge.AuthorMerger
import eu.dnetlib.dhp.schema.oaf.{Author, DataInfo, Instance, Publication, StructuredProperty}
import org.apache.spark.sql.{Dataset, Encoder, Encoders, Row, SaveMode, SparkSession}
import org.apache.spark.sql.functions.{col, collect_set, concat, explode, expr, first, flatten, lower, size, struct}
import eu.dnetlib.dhp.schema.oaf.{OtherResearchProduct, Publication, Result, Software}
import org.apache.spark.sql.functions._
import org.apache.spark.sql._
import org.slf4j.{Logger, LoggerFactory}
import org.apache.spark.sql.types._
class SparkEnrichGraphWithOrcidAuthors(propertyPath: String, args: Array[String], log: Logger)
extends AbstractScalaApplication(propertyPath, args, log: Logger) {
@ -22,33 +20,49 @@ class SparkEnrichGraphWithOrcidAuthors(propertyPath: String, args: Array[String]
log.info(s"orcidPath is '$orcidPath'")
val targetPath = parser.get("targetPath")
log.info(s"targetPath is '$targetPath'")
enrichResult(spark, graphPath, orcidPath, targetPath)
val orcidPublication: Dataset[Row] = generateOrcidTable(spark, orcidPath)
enrichResult(
spark,
s"$graphPath/publication",
orcidPublication,
s"$targetPath/publication",
Encoders.bean(classOf[Publication])
)
enrichResult(
spark,
s"$graphPath/dataset",
orcidPublication,
s"$targetPath/dataset",
Encoders.bean(classOf[eu.dnetlib.dhp.schema.oaf.Dataset])
)
enrichResult(
spark,
s"$graphPath/software",
orcidPublication,
s"$targetPath/software",
Encoders.bean(classOf[Software])
)
enrichResult(
spark,
s"$graphPath/otherresearchproduct",
orcidPublication,
s"$targetPath/otherresearchproduct",
Encoders.bean(classOf[OtherResearchProduct])
)
}
def enrichResult(spark: SparkSession, graphPath: String, orcidPath: String, outputPath: String): Unit = {
val orcidPublication = generateOrcidTable(spark, orcidPath)
private def enrichResult[T <: Result](
spark: SparkSession,
graphPath: String,
orcidPublication: Dataset[Row],
outputPath: String,
enc: Encoder[T]
): Unit = {
implicit val publicationEncoder = Encoders.bean(classOf[Publication])
val aschema = new StructType()
.add("id", StringType)
.add("dataInfo", Encoders.bean(classOf[DataInfo]).schema)
.add(
"author",Encoders.bean(classOf[Author]).schema
)
val schema = new StructType()
.add("id", StringType)
.add("dataInfo", Encoders.bean(classOf[DataInfo]).schema)
.add(
"instance",
ArrayType(new StructType().add("pid", ArrayType(Encoders.bean(classOf[StructuredProperty]).schema)))
)
val entities = spark.read
.schema(schema)
.schema(enc.schema)
.json(graphPath)
.select(col("id"), col("datainfo"), col("instance"))
.where("datainfo.deletedbyinference = false")
.drop("datainfo")
.withColumn("instances", explode(col("instance")))
@ -58,7 +72,8 @@ class SparkEnrichGraphWithOrcidAuthors(propertyPath: String, args: Array[String]
col("pids.value").alias("pid_value"),
col("id").alias("dnet_id")
)
val orcidDnet = orcidPublication
val orcidDnet = orcidPublication
.join(
entities,
lower(col("schema")).equalTo(lower(col("pid_schema"))) &&
@ -69,36 +84,25 @@ class SparkEnrichGraphWithOrcidAuthors(propertyPath: String, args: Array[String]
.agg(collect_set(orcidPublication("author")).alias("orcid_authors"))
.select("dnet_id", "orcid_authors")
.cache()
orcidDnet.count()
val result = spark.read.schema(enc.schema).json(graphPath).as[T](enc)
orcidPublication
.join(
entities,
lower(col("schema")).equalTo(lower(col("pid_schema"))) &&
lower(col("value")).equalTo(lower(col("pid_value"))),
"inner"
)
.groupBy(col("dnet_id")).agg(collect_set(struct(col("pid_schema"), col("pid_value")))).write.mode("Overwrite").save("/user/sandro.labruzzo/enrich_pub")
val publication = spark.read.schema(publicationEncoder.schema).json(graphPath).as[Publication]
publication
.joinWith(orcidDnet, publication("id").equalTo(orcidDnet("dnet_id")), "left")
result
.joinWith(orcidDnet, result("id").equalTo(orcidDnet("dnet_id")), "left")
.map {
case (p: Publication, null) => {
p
}
case (p: Publication, r: Row) =>
case (r: T, null) =>
r
case (p: T, r: Row) =>
p.setAuthor(AuthorMerger.enrichOrcid(p.getAuthor, AuthorEnricher.toOAFAuthor(r)))
p
}
}(enc)
.write
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(outputPath)
}
def generateOrcidTable(spark: SparkSession, inputPath: String): Dataset[Row] = {
private def generateOrcidTable(spark: SparkSession, inputPath: String): Dataset[Row] = {
val orcidAuthors =
spark.read.load(s"$inputPath/Authors").select("orcid", "familyName", "givenName", "creditName", "otherNames")
val orcidWorks = spark.read
@ -107,14 +111,14 @@ class SparkEnrichGraphWithOrcidAuthors(propertyPath: String, args: Array[String]
.where(
"identifier.schema = 'doi' or identifier.schema ='pmid' or identifier.schema ='pmc' or identifier.schema ='arxiv' or identifier.schema ='handle'"
)
val orcidPublication =orcidAuthors
val orcidPublication = orcidAuthors
.join(orcidWorks, orcidAuthors("orcid").equalTo(orcidWorks("orcid")))
.select(
col("identifier.schema").alias("schema"),
col("identifier.value").alias("value"),
struct(orcidAuthors("orcid").alias("orcid"), col("givenName"), col("familyName")).alias("author")
)
orcidPublication
orcidPublication.cache()
}
}
@ -123,10 +127,8 @@ object SparkEnrichGraphWithOrcidAuthors {
val log: Logger = LoggerFactory.getLogger(SparkEnrichGraphWithOrcidAuthors.getClass)
def main(args: Array[String]): Unit = {
new SparkEnrichGraphWithOrcidAuthors("/eu/dnetlib/dhp/enrich/orcid/enrich_graph_orcid_parameters.json", args, log)
.initialize()
.run()
}
}

View File

@ -1,78 +1,85 @@
package eu.dnetlib.dhp.enrich.orcid
import eu.dnetlib.dhp.schema.oaf.Publication
import eu.dnetlib.dhp.schema.oaf.{Author, Publication}
import org.apache.spark.sql.{Column, Encoder, Encoders, Row, SparkSession}
import org.junit.jupiter.api.Test
import org.slf4j.{Logger, LoggerFactory}
import org.apache.spark.sql.functions._
case class Pid(pidScheme: String, pidValue: String) {}
case class AuthorPid(fullName: String, pids: List[Pid]) {}
case class PubSummary(id: String, authorWithPids: List[AuthorPid])
class EnrichOrcidTest {
val log: Logger = LoggerFactory.getLogger(getClass)
def orcid_intersection_wrong(p: PubSummary): PubSummary = {
if (p.authorWithPids.isEmpty)
null
else {
val incorrectAuthor = p.authorWithPids.filter(a => a.pids.filter(p => p.pidScheme != null && p.pidScheme.toLowerCase.contains("orcid")).map(p => p.pidValue.toLowerCase).distinct.size > 1)
if (incorrectAuthor.nonEmpty) {
PubSummary(p.id, incorrectAuthor)
}
else {
null
}
}
}
def test() = {
val spark = SparkSession.builder().master("local[*]").getOrCreate()
spark.sparkContext.setLogLevel("ERROR")
// spark.sparkContext.setLogLevel("ERROR")
// new SparkEnrichGraphWithOrcidAuthors(null, null, null)
// .enrichResult(
// spark,
// "/Users/sandro/orcid_test/publication",
// "",
// "/tmp/graph/",
// Encoders.bean(classOf[Publication])
// )
val schema = Encoders.bean(classOf[Publication]).schema
//
// val simplifyAuthor = udf((r: Seq[Row]) => {
// r
// .map(k =>
// AuthorPid(
// k.getAs[String]("fullname"),
// k.getAs[Seq[Row]]("pid")
// .map(p => Pid(p.getAs[Row]("qualifier").getAs[String]("classid"), p.getAs[String]("value")))
// .toList
// )
// )
// .filter(l => l.pids.nonEmpty)
// .toList
// })
//
// val wrong_orcid_intersection = udf((a: Seq[Row]) => {
// a.map(author => {
// val pids_with_orcid: Seq[Row] = author
// .getAs[Seq[Row]]("pids")
// .filter(p =>
// p.getAs[String]("pidScheme") != null && p.getAs[String]("pidScheme").toLowerCase.contains("orcid")
// )
// if (pids_with_orcid.exists(p => p.getAs[String]("pidScheme").equals("ORCID"))) {
// if (pids_with_orcid.map(p => p.getAs[String]("pidValue").toLowerCase).distinct.size > 1) {
// AuthorPid(
// author.getAs[String]("fullName"),
// pids_with_orcid.map(p => Pid(p.getAs[String]("pidScheme"), p.getAs[String]("pidValue"))).toList
// )
//
// } else
// null
// } else
// null
// }).filter(author => author != null)
// })
Encoders
import spark.implicits._
// val enriched = spark.read
// .schema(schema)
// .json("/Users/sandro/orcid_test/publication_enriched")
// .select(col("id"), explode(col("author")).as("authors"))
// .withColumn("ap", col("authors.pid.qualifier.classid"))
// .withColumn("dp", col("authors.pid.datainfo.provenanceAction.classid"))
//
// .show()
val simplifyAuthor = udf((r: Seq[Row]) => {
r
.map(k =>
AuthorPid(k.getAs[String]("fullname"),
k.getAs[Seq[Row]]("pid")
.map(
p => Pid(p.getAs[Row]("qualifier").getAs[String]("classid"), p.getAs[String]("value"))
).toList)
).filter(l => l.pids.nonEmpty)
.toList
}
)
val wrong_orcid_intersection = udf((a: Seq[Row]) => {
a.map(author => {
val pids_with_orcid: Seq[Row] = author.getAs[Seq[Row]]("pids").filter(p => p.getAs[String]("pidScheme")!= null && p.getAs[String]("pidScheme").toLowerCase.contains("orcid"))
if (pids_with_orcid.exists(p => p.getAs[String]("pidScheme").equals("ORCID"))) {
if (pids_with_orcid.map(p => p.getAs[String]("pidValue").toLowerCase).distinct.size > 1) {
AuthorPid(author.getAs[String]("fullName"),pids_with_orcid.map(p => Pid(p.getAs[String]("pidScheme"),p.getAs[String]("pidValue"))).toList )
}
else
null
} else
null
}).filter(author => author != null)
})
val enriched = spark.read.schema(schema).json("/Users/sandro/orcid_test/publication_enriched").select(col("id"), simplifyAuthor(col("author")).alias("authors"))
.select(col("id"), wrong_orcid_intersection(col("authors")).alias("wi")).where("wi is not null")
enriched.show(20, 1000, true)
}
}