enrico.ottonello
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

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

This commit is contained in:
Claudio Atzori 2021-03-23 09:59:12 +01:00
parent 8db248aa13 625e4c29c4
commit b75ad76f79
1 changed files with 18 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/DoiBoostMappingUtil.scala
View File

@ -5,6 +5,7 @@ import eu.dnetlib.dhp.schema.oaf.{AccessRight, DataInfo, Dataset, Field, Instanc
import eu.dnetlib.dhp.utils.DHPUtils import eu.dnetlib.dhp.utils.DHPUtils
import org.apache.commons.lang3.StringUtils import org.apache.commons.lang3.StringUtils
import com.fasterxml.jackson.databind.ObjectMapper import com.fasterxml.jackson.databind.ObjectMapper
import eu.dnetlib.dhp.schema.common.ModelConstants
import eu.dnetlib.dhp.schema.scholexplorer.OafUtils import eu.dnetlib.dhp.schema.scholexplorer.OafUtils
import org.json4s import org.json4s
import org.json4s.DefaultFormats import org.json4s.DefaultFormats
@ -107,23 +108,17 @@ object DoiBoostMappingUtil {
def fixResult(result: Dataset) :Dataset = { def fixResult(result: Dataset) :Dataset = {
val instanceType = result.getInstance().asScala.find(i => i.getInstancetype != null && i.getInstancetype.getClassid.nonEmpty) val instanceType = extractInstance(result)
if (instanceType.isDefined) { if (instanceType.isDefined) {
result.getInstance().asScala.foreach(i => i.setInstancetype(instanceType.get.getInstancetype)) result.getInstance().asScala.foreach(i => i.setInstancetype(instanceType.get.getInstancetype))
} }
result.getInstance().asScala.foreach(i => { result.getInstance().asScala.foreach(i => {
i.setHostedby(getUnknownHostedBy()) i.setHostedby(ModelConstants.UNKNOWN_REPOSITORY)
}) })
result result
} }
def getUnknownHostedBy():KeyValue = {
val hb = new KeyValue
hb.setValue("Unknown Repository")
hb.setKey(s"10|$OPENAIRE_PREFIX::55045bd2a65019fd8e6741a755395c8c")
hb
}
def getOpenAccessQualifier():AccessRight = { def getOpenAccessQualifier():AccessRight = {
@ -135,6 +130,11 @@ object DoiBoostMappingUtil {
} }
def extractInstance(r:Result):Option[Instance] = {
r.getInstance().asScala.find(i => i.getInstancetype != null && i.getInstancetype.getClassid.nonEmpty)
}
def fixPublication(input:((String,Publication), (String,HostedByItemType))): Publication = { def fixPublication(input:((String,Publication), (String,HostedByItemType))): Publication = {
val publication = input._1._2 val publication = input._1._2
@ -142,7 +142,7 @@ object DoiBoostMappingUtil {
val item = if (input._2 != null) input._2._2 else null val item = if (input._2 != null) input._2._2 else null
val instanceType = publication.getInstance().asScala.find(i => i.getInstancetype != null && i.getInstancetype.getClassid.nonEmpty) val instanceType:Option[Instance] = extractInstance(publication)
if (instanceType.isDefined) { if (instanceType.isDefined) {
publication.getInstance().asScala.foreach(i => i.setInstancetype(instanceType.get.getInstancetype)) publication.getInstance().asScala.foreach(i => i.setInstancetype(instanceType.get.getInstancetype))
@ -150,28 +150,30 @@ object DoiBoostMappingUtil {
publication.getInstance().asScala.foreach(i => { publication.getInstance().asScala.foreach(i => {
val hb = new KeyValue var hb = new KeyValue
if (item != null) { if (item != null) {
hb.setValue(item.officialname) hb.setValue(item.officialname)
hb.setKey(generateDSId(item.id)) hb.setKey(generateDSId(item.id))
if (item.openAccess) if (item.openAccess)
i.setAccessright(getOpenAccessQualifier()) i.setAccessright(getOpenAccessQualifier())
publication.setBestaccessright(getOpenAccessQualifier()) val ar = getOpenAccessQualifier()
publication.setBestaccessright(OafUtils.createQualifier(ar.getClassid, ar.getClassname, ar.getSchemeid, ar.getSchemename))
} }
else { else {
hb.setValue("Unknown Repository") hb = ModelConstants.UNKNOWN_REPOSITORY
hb.setKey(s"10|$OPENAIRE_PREFIX::55045bd2a65019fd8e6741a755395c8c")
} }
i.setHostedby(hb) i.setHostedby(hb)
}) })
val ar = publication.getInstance().asScala.filter(i => i.getInstancetype != null && i.getAccessright!= null && i.getAccessright.getClassid!= null).map(f=> f.getAccessright.getClassid) val ar = publication.getInstance().asScala.filter(i => i.getInstancetype != null && i.getAccessright!= null && i.getAccessright.getClassid!= null).map(f=> f.getAccessright.getClassid)
if (ar.nonEmpty) { if (ar.nonEmpty) {
if(ar.contains("OPEN")){ if(ar.contains(ModelConstants.ACCESS_RIGHT_OPEN)){
publication.setBestaccessright(getOpenAccessQualifier()) val ar = getOpenAccessQualifier()
publication.setBestaccessright(OafUtils.createQualifier(ar.getClassid, ar.getClassname, ar.getSchemeid, ar.getSchemename))
} }
else { else {
publication.setBestaccessright(getRestrictedQualifier()) val ar = getRestrictedQualifier()
publication.setBestaccessright(OafUtils.createQualifier(ar.getClassid, ar.getClassname, ar.getSchemeid, ar.getSchemename))
} }
} }
publication publication