D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

Implemented ID generation using IdentifierRecordFactory on DOIBoost

This commit is contained in:
Sandro La Bruzzo 2020-11-09 11:53:55 +01:00
parent 2d76497488
commit 8e1d43aab2
4 changed files with 26 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/crossref/Crossref2Oaf.scala
View File

@ -1,6 +1,7 @@
package eu.dnetlib.doiboost.crossref package eu.dnetlib.doiboost.crossref
import eu.dnetlib.dhp.schema.oaf._ import eu.dnetlib.dhp.schema.oaf._
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory
import eu.dnetlib.dhp.utils.DHPUtils import eu.dnetlib.dhp.utils.DHPUtils
import eu.dnetlib.doiboost.DoiBoostMappingUtil._ import eu.dnetlib.doiboost.DoiBoostMappingUtil._
import org.apache.commons.lang.StringUtils import org.apache.commons.lang.StringUtils
@ -96,7 +97,12 @@ case object Crossref2Oaf {
result.setOriginalId(tmp.filter(id => id != null).asJava) result.setOriginalId(tmp.filter(id => id != null).asJava)
//Set identifier as 50 | doiboost____::md5(DOI) //Set identifier as 50 | doiboost____::md5(DOI)
//IMPORTANT
//The old method result.setId(generateIdentifier(result, doi))
//will be replaced using IdentifierFactory
result.setId(generateIdentifier(result, doi)) result.setId(generateIdentifier(result, doi))
result.setId(IdentifierFactory.createIdentifier(result))
// Add DataInfo // Add DataInfo
result.setDataInfo(generateDataInfo()) result.setDataInfo(generateDataInfo())

6
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/mag/MagDataModel.scala
View File

@ -1,6 +1,7 @@
package eu.dnetlib.doiboost.mag package eu.dnetlib.doiboost.mag
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory
import eu.dnetlib.dhp.schema.oaf.{Instance, Journal, Publication, StructuredProperty} import eu.dnetlib.dhp.schema.oaf.{Instance, Journal, Publication, StructuredProperty}
import eu.dnetlib.doiboost.DoiBoostMappingUtil import eu.dnetlib.doiboost.DoiBoostMappingUtil
import org.json4s import org.json4s
@ -190,8 +191,11 @@ case object ConversionUtil {
pub.setPid(List(createSP(paper.Doi.toLowerCase, "doi", PID_TYPES)).asJava) pub.setPid(List(createSP(paper.Doi.toLowerCase, "doi", PID_TYPES)).asJava)
pub.setOriginalId(List(paper.PaperId.toString, paper.Doi.toLowerCase).asJava) pub.setOriginalId(List(paper.PaperId.toString, paper.Doi.toLowerCase).asJava)
//Set identifier as 50|doiboost____::md5(DOI) //IMPORTANT
//The old method result.setId(generateIdentifier(result, doi))
//will be replaced using IdentifierFactory
pub.setId(generateIdentifier(pub, paper.Doi.toLowerCase)) pub.setId(generateIdentifier(pub, paper.Doi.toLowerCase))
pub.setId(IdentifierFactory.createIdentifier(pub))
val mainTitles = createSP(paper.PaperTitle, "main title", "dnet:dataCite_title") val mainTitles = createSP(paper.PaperTitle, "main title", "dnet:dataCite_title")
val originalTitles = createSP(paper.OriginalTitle, "alternative title", "dnet:dataCite_title") val originalTitles = createSP(paper.OriginalTitle, "alternative title", "dnet:dataCite_title")

8
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/ORCIDToOAF.scala
View File

@ -1,5 +1,6 @@
package eu.dnetlib.doiboost.orcid package eu.dnetlib.doiboost.orcid
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory
import eu.dnetlib.dhp.schema.oaf.{Author, Publication} import eu.dnetlib.dhp.schema.oaf.{Author, Publication}
import eu.dnetlib.doiboost.DoiBoostMappingUtil import eu.dnetlib.doiboost.DoiBoostMappingUtil
import eu.dnetlib.doiboost.DoiBoostMappingUtil.{ORCID, PID_TYPES, createSP, generateDataInfo, generateIdentifier} import eu.dnetlib.doiboost.DoiBoostMappingUtil.{ORCID, PID_TYPES, createSP, generateDataInfo, generateIdentifier}
@ -48,7 +49,14 @@ object ORCIDToOAF {
val pub:Publication = new Publication val pub:Publication = new Publication
pub.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava) pub.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava)
pub.setDataInfo(generateDataInfo()) pub.setDataInfo(generateDataInfo())
//IMPORTANT
//The old method pub.setId(IdentifierFactory.createIdentifier(pub))
//will be replaced using IdentifierFactory
pub.setId(generateIdentifier(pub, doi.toLowerCase)) pub.setId(generateIdentifier(pub, doi.toLowerCase))
pub.setId(IdentifierFactory.createIdentifier(pub))
try{ try{
pub.setAuthor(input.authors.map(a=> { pub.setAuthor(input.authors.map(a=> {
generateAuthor(a.name, a.surname, a.creditName, a.oid) generateAuthor(a.name, a.surname, a.creditName, a.oid)

7
dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/uw/UnpayWallToOAF.scala
View File

@ -1,5 +1,6 @@
package eu.dnetlib.doiboost.uw package eu.dnetlib.doiboost.uw
import eu.dnetlib.dhp.schema.oaf.utils.IdentifierFactory
import eu.dnetlib.dhp.schema.oaf.{Instance, Publication} import eu.dnetlib.dhp.schema.oaf.{Instance, Publication}
import org.json4s import org.json4s
import org.json4s.DefaultFormats import org.json4s.DefaultFormats
@ -33,7 +34,13 @@ object UnpayWallToOAF {
val oaLocation:OALocation = (json \ "best_oa_location").extractOrElse[OALocation](null) val oaLocation:OALocation = (json \ "best_oa_location").extractOrElse[OALocation](null)
pub.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava) pub.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava)
//IMPORTANT
//The old method pub.setId(IdentifierFactory.createIdentifier(pub))
//will be replaced using IdentifierFactory
pub.setId(generateIdentifier(pub, doi.toLowerCase)) pub.setId(generateIdentifier(pub, doi.toLowerCase))
pub.setId(IdentifierFactory.createIdentifier(pub))
pub.setCollectedfrom(List(createUnpayWallCollectedFrom()).asJava) pub.setCollectedfrom(List(createUnpayWallCollectedFrom()).asJava)
pub.setDataInfo(generateDataInfo()) pub.setDataInfo(generateDataInfo())