master #59

Closed
claudio.atzori wants to merge 3221 commits from master into stable_ids
9 changed files with 148 additions and 61 deletions
Showing only changes of commit 27e96767e0 - Show all commits

View File

@ -62,7 +62,7 @@ object SparkGenerateDoiBoost {
val orcidPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/orcidPublication").as[Publication].map(p => (p.getId, p)) val orcidPublication: Dataset[(String, Publication)] = spark.read.load(s"$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"$workingDirPath/secondJoin") fj.joinWith(orcidPublication, fj("_1").equalTo(orcidPublication("_1")), "left").map(applyMerge).write.mode(SaveMode.Overwrite).save(s"$workingDirPath/secondJoin")
logger.info("Phase 3) Join Result with MAG") logger.info("Phase 4) Join Result with MAG")
val sj: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/secondJoin").as[Publication].map(p => (p.getId, p)) val sj: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/secondJoin").as[Publication].map(p => (p.getId, p))
val magPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/magPublication").as[Publication].map(p => (p.getId, p)) val magPublication: Dataset[(String, Publication)] = spark.read.load(s"$workingDirPath/magPublication").as[Publication].map(p => (p.getId, p))

View File

@ -21,15 +21,17 @@ object SparkImportMagIntoDataset {
val stream = Map( val stream = Map(
"Affiliations" -> Tuple2("mag/Affiliations.txt", Seq("AffiliationId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "GridId:string", "OfficialPage:string", "WikiPage:string", "PaperCount:long", "CitationCount:long", "Latitude:float?", "Longitude:float?", "CreatedDate:DateTime")), "Affiliations" -> Tuple2("mag/Affiliations.txt", Seq("AffiliationId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "GridId:string", "OfficialPage:string", "WikiPage:string", "PaperCount:long", "PaperFamilyCount:long", "CitationCount:long", "Iso3166Code:string", "Latitude:float?", "Longitude:float?", "CreatedDate:DateTime")),
"Authors" -> Tuple2("mag/Authors.txt", Seq("AuthorId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "LastKnownAffiliationId:long?", "PaperCount:long", "CitationCount:long", "CreatedDate:DateTime")), "AuthorExtendedAttributes" -> Tuple2("mag/AuthorExtendedAttributes.txt", Seq("AuthorId:long", "AttributeType:int", "AttributeValue:string")),
"ConferenceInstances" -> Tuple2("mag/ConferenceInstances.txt", Seq("ConferenceInstanceId:long", "NormalizedName:string", "DisplayName:string", "ConferenceSeriesId:long", "Location:string", "OfficialUrl:string", "StartDate:DateTime?", "EndDate:DateTime?", "AbstractRegistrationDate:DateTime?", "SubmissionDeadlineDate:DateTime?", "NotificationDueDate:DateTime?", "FinalVersionDueDate:DateTime?", "PaperCount:long", "CitationCount:long", "Latitude:float?", "Longitude:float?", "CreatedDate:DateTime")), "Authors" -> Tuple2("mag/Authors.txt", Seq("AuthorId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "LastKnownAffiliationId:long?", "PaperCount:long", "PaperFamilyCount:long", "CitationCount:long", "CreatedDate:DateTime")),
"ConferenceInstances" -> Tuple2("mag/ConferenceInstances.txt", Seq("ConferenceInstanceId:long", "NormalizedName:string", "DisplayName:string", "ConferenceSeriesId:long", "Location:string", "OfficialUrl:string", "StartDate:DateTime?", "EndDate:DateTime?", "AbstractRegistrationDate:DateTime?", "SubmissionDeadlineDate:DateTime?", "NotificationDueDate:DateTime?", "FinalVersionDueDate:DateTime?", "PaperCount:long", "PaperFamilyCount:long" ,"CitationCount:long", "Latitude:float?", "Longitude:float?", "CreatedDate:DateTime")),
"ConferenceSeries" -> Tuple2("mag/ConferenceSeries.txt", Seq("ConferenceSeriesId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "PaperCount:long", "CitationCount:long", "CreatedDate:DateTime")), "ConferenceSeries" -> Tuple2("mag/ConferenceSeries.txt", Seq("ConferenceSeriesId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "PaperCount:long", "CitationCount:long", "CreatedDate:DateTime")),
"EntityRelatedEntities" -> Tuple2("advanced/EntityRelatedEntities.txt", Seq("EntityId:long", "EntityType:string", "RelatedEntityId:long", "RelatedEntityType:string", "RelatedType:int", "Score:float")), "EntityRelatedEntities" -> Tuple2("advanced/EntityRelatedEntities.txt", Seq("EntityId:long", "EntityType:string", "RelatedEntityId:long", "RelatedEntityType:string", "RelatedType:int", "Score:float")),
"FieldOfStudyChildren" -> Tuple2("advanced/FieldOfStudyChildren.txt", Seq("FieldOfStudyId:long", "ChildFieldOfStudyId:long")), "FieldOfStudyChildren" -> Tuple2("advanced/FieldOfStudyChildren.txt", Seq("FieldOfStudyId:long", "ChildFieldOfStudyId:long")),
"FieldOfStudyExtendedAttributes" -> Tuple2("advanced/FieldOfStudyExtendedAttributes.txt", Seq("FieldOfStudyId:long", "AttributeType:int", "AttributeValue:string")), "FieldOfStudyExtendedAttributes" -> Tuple2("advanced/FieldOfStudyExtendedAttributes.txt", Seq("FieldOfStudyId:long", "AttributeType:int", "AttributeValue:string")),
"FieldsOfStudy" -> Tuple2("advanced/FieldsOfStudy.txt", Seq("FieldOfStudyId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "MainType:string", "Level:int", "PaperCount:long", "CitationCount:long", "CreatedDate:DateTime")), // ['FieldOfStudyId:long', 'Rank:uint', 'NormalizedName:string', 'DisplayName:string', 'MainType:string', 'Level:int', 'PaperCount:long', 'PaperFamilyCount:long', 'CitationCount:long', 'CreatedDate:DateTime']
"Journals" -> Tuple2("mag/Journals.txt", Seq("JournalId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "Issn:string", "Publisher:string", "Webpage:string", "PaperCount:long", "CitationCount:long", "CreatedDate:DateTime")), "FieldsOfStudy" -> Tuple2("advanced/FieldsOfStudy.txt", Seq("FieldOfStudyId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "MainType:string", "Level:int", "PaperCount:long", "PaperFamilyCount:long", "CitationCount:long", "CreatedDate:DateTime")),
"Journals" -> Tuple2("mag/Journals.txt", Seq("JournalId:long", "Rank:uint", "NormalizedName:string", "DisplayName:string", "Issn:string", "Publisher:string", "Webpage:string", "PaperCount:long", "PaperFamilyCount:long" ,"CitationCount:long", "CreatedDate:DateTime")),
"PaperAbstractsInvertedIndex" -> Tuple2("nlp/PaperAbstractsInvertedIndex.txt.*", Seq("PaperId:long", "IndexedAbstract:string")), "PaperAbstractsInvertedIndex" -> Tuple2("nlp/PaperAbstractsInvertedIndex.txt.*", Seq("PaperId:long", "IndexedAbstract:string")),
"PaperAuthorAffiliations" -> Tuple2("mag/PaperAuthorAffiliations.txt", Seq("PaperId:long", "AuthorId:long", "AffiliationId:long?", "AuthorSequenceNumber:uint", "OriginalAuthor:string", "OriginalAffiliation:string")), "PaperAuthorAffiliations" -> Tuple2("mag/PaperAuthorAffiliations.txt", Seq("PaperId:long", "AuthorId:long", "AffiliationId:long?", "AuthorSequenceNumber:uint", "OriginalAuthor:string", "OriginalAffiliation:string")),
"PaperCitationContexts" -> Tuple2("nlp/PaperCitationContexts.txt", Seq("PaperId:long", "PaperReferenceId:long", "CitationContext:string")), "PaperCitationContexts" -> Tuple2("nlp/PaperCitationContexts.txt", Seq("PaperId:long", "PaperReferenceId:long", "CitationContext:string")),
@ -39,7 +41,7 @@ object SparkImportMagIntoDataset {
"PaperReferences" -> Tuple2("mag/PaperReferences.txt", Seq("PaperId:long", "PaperReferenceId:long")), "PaperReferences" -> Tuple2("mag/PaperReferences.txt", Seq("PaperId:long", "PaperReferenceId:long")),
"PaperResources" -> Tuple2("mag/PaperResources.txt", Seq("PaperId:long", "ResourceType:int", "ResourceUrl:string", "SourceUrl:string", "RelationshipType:int")), "PaperResources" -> Tuple2("mag/PaperResources.txt", Seq("PaperId:long", "ResourceType:int", "ResourceUrl:string", "SourceUrl:string", "RelationshipType:int")),
"PaperUrls" -> Tuple2("mag/PaperUrls.txt", Seq("PaperId:long", "SourceType:int?", "SourceUrl:string", "LanguageCode:string")), "PaperUrls" -> Tuple2("mag/PaperUrls.txt", Seq("PaperId:long", "SourceType:int?", "SourceUrl:string", "LanguageCode:string")),
"Papers" -> Tuple2("mag/Papers.txt", Seq("PaperId:long", "Rank:uint", "Doi:string", "DocType:string", "PaperTitle:string", "OriginalTitle:string", "BookTitle:string", "Year:int?", "Date:DateTime?", "Publisher:string", "JournalId:long?", "ConferenceSeriesId:long?", "ConferenceInstanceId:long?", "Volume:string", "Issue:string", "FirstPage:string", "LastPage:string", "ReferenceCount:long", "CitationCount:long", "EstimatedCitation:long", "OriginalVenue:string", "FamilyId:long?", "CreatedDate:DateTime")), "Papers" -> Tuple2("mag/Papers.txt", Seq("PaperId:long", "Rank:uint", "Doi:string", "DocType:string", "PaperTitle:string", "OriginalTitle:string", "BookTitle:string", "Year:int?", "Date:DateTime?", "OnlineDate:DateTime?", "Publisher:string", "JournalId:long?", "ConferenceSeriesId:long?", "ConferenceInstanceId:long?", "Volume:string", "Issue:string", "FirstPage:string", "LastPage:string", "ReferenceCount:long", "CitationCount:long", "EstimatedCitation:long", "OriginalVenue:string", "FamilyId:long?", "FamilyRank:uint?", "CreatedDate:DateTime")),
"RelatedFieldOfStudy" -> Tuple2("advanced/RelatedFieldOfStudy.txt", Seq("FieldOfStudyId1:long", "Type1:string", "FieldOfStudyId2:long", "Type2:string", "Rank:float")) "RelatedFieldOfStudy" -> Tuple2("advanced/RelatedFieldOfStudy.txt", Seq("FieldOfStudyId1:long", "Type1:string", "FieldOfStudyId2:long", "Type2:string", "Rank:float"))
) )

View File

@ -26,12 +26,15 @@ object SparkPreProcessMAG {
.master(parser.get("master")).getOrCreate() .master(parser.get("master")).getOrCreate()
val sourcePath = parser.get("sourcePath") val sourcePath = parser.get("sourcePath")
val workingPath = parser.get("workingPath")
val targetPath = parser.get("targetPath")
import spark.implicits._ import spark.implicits._
implicit val mapEncoderPubs: Encoder[Publication] = org.apache.spark.sql.Encoders.kryo[Publication] implicit val mapEncoderPubs: Encoder[Publication] = org.apache.spark.sql.Encoders.kryo[Publication]
implicit val tupleForJoinEncoder: Encoder[(String, Publication)] = Encoders.tuple(Encoders.STRING, mapEncoderPubs) implicit val tupleForJoinEncoder: Encoder[(String, Publication)] = Encoders.tuple(Encoders.STRING, mapEncoderPubs)
logger.info("Phase 1) make uninque DOI in Papers:") logger.info("Phase 1) make uninque DOI in Papers:")
val d: Dataset[MagPapers] = spark.read.load(s"${parser.get("sourcePath")}/Papers").as[MagPapers] val d: Dataset[MagPapers] = spark.read.load(s"$sourcePath/Papers").as[MagPapers]
// Filtering Papers with DOI, and since for the same DOI we have multiple version of item with different PapersId we get the last one // Filtering Papers with DOI, and since for the same DOI we have multiple version of item with different PapersId we get the last one
val result: RDD[MagPapers] = d.where(col("Doi").isNotNull) val result: RDD[MagPapers] = d.where(col("Doi").isNotNull)
@ -41,11 +44,12 @@ object SparkPreProcessMAG {
.map(_._2) .map(_._2)
val distinctPaper: Dataset[MagPapers] = spark.createDataset(result) val distinctPaper: Dataset[MagPapers] = spark.createDataset(result)
distinctPaper.write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/Papers_distinct")
distinctPaper.write.mode(SaveMode.Overwrite).save(s"$workingPath/Papers_distinct")
logger.info("Phase 0) Enrich Publication with description") logger.info("Phase 0) Enrich Publication with description")
val pa = spark.read.load(s"${parser.get("sourcePath")}/PaperAbstractsInvertedIndex").as[MagPaperAbstract] val pa = spark.read.load(s"$sourcePath/PaperAbstractsInvertedIndex").as[MagPaperAbstract]
pa.map(ConversionUtil.transformPaperAbstract).write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/PaperAbstract") pa.map(ConversionUtil.transformPaperAbstract).write.mode(SaveMode.Overwrite).save(s"$workingPath/PaperAbstract")
logger.info("Phase 3) Group Author by PaperId") logger.info("Phase 3) Group Author by PaperId")
val authors = spark.read.load(s"$sourcePath/Authors").as[MagAuthor] val authors = spark.read.load(s"$sourcePath/Authors").as[MagAuthor]
@ -64,24 +68,24 @@ object SparkPreProcessMAG {
} else } else
mpa mpa
}).groupBy("PaperId").agg(collect_list(struct($"author", $"affiliation")).as("authors")) }).groupBy("PaperId").agg(collect_list(struct($"author", $"affiliation")).as("authors"))
.write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/merge_step_1_paper_authors") .write.mode(SaveMode.Overwrite).save(s"$workingPath/merge_step_1_paper_authors")
logger.info("Phase 4) create First Version of publication Entity with Paper Journal and Authors") logger.info("Phase 4) create First Version of publication Entity with Paper Journal and Authors")
val journals = spark.read.load(s"$sourcePath/Journals").as[MagJournal] val journals = spark.read.load(s"$sourcePath/Journals").as[MagJournal]
val papers = spark.read.load((s"${parser.get("targetPath")}/Papers_distinct")).as[MagPapers] val papers = spark.read.load((s"$workingPath/Papers_distinct")).as[MagPapers]
val paperWithAuthors = spark.read.load(s"${parser.get("targetPath")}/merge_step_1_paper_authors").as[MagPaperWithAuthorList] val paperWithAuthors = spark.read.load(s"$workingPath/merge_step_1_paper_authors").as[MagPaperWithAuthorList]
val firstJoin = papers.joinWith(journals, papers("JournalId").equalTo(journals("JournalId")), "left") val firstJoin = papers.joinWith(journals, papers("JournalId").equalTo(journals("JournalId")), "left")
firstJoin.joinWith(paperWithAuthors, firstJoin("_1.PaperId").equalTo(paperWithAuthors("PaperId")), "left") firstJoin.joinWith(paperWithAuthors, firstJoin("_1.PaperId").equalTo(paperWithAuthors("PaperId")), "left")
.map { a => ConversionUtil.createOAFFromJournalAuthorPaper(a) } .map { a => ConversionUtil.createOAFFromJournalAuthorPaper(a) }
.write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/merge_step_2") .write.mode(SaveMode.Overwrite).save(s"$workingPath/merge_step_2")
var magPubs: Dataset[(String, Publication)] = var magPubs: Dataset[(String, Publication)] =
spark.read.load(s"${parser.get("targetPath")}/merge_step_2").as[Publication] spark.read.load(s"$workingPath/merge_step_2").as[Publication]
.map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)] .map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)]
@ -95,10 +99,10 @@ object SparkPreProcessMAG {
.map(item => ConversionUtil.updatePubsWithConferenceInfo(item)) .map(item => ConversionUtil.updatePubsWithConferenceInfo(item))
.write .write
.mode(SaveMode.Overwrite) .mode(SaveMode.Overwrite)
.save(s"${parser.get("targetPath")}/merge_step_2_conference") .save(s"$workingPath/merge_step_2_conference")
magPubs= spark.read.load(s"${parser.get("targetPath")}/merge_step_2_conference").as[Publication] magPubs= spark.read.load(s"$workingPath/merge_step_2_conference").as[Publication]
.map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)] .map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)]
val paperUrlDataset = spark.read.load(s"$sourcePath/PaperUrls").as[MagPaperUrl].groupBy("PaperId").agg(collect_list(struct("sourceUrl")).as("instances")).as[MagUrl] val paperUrlDataset = spark.read.load(s"$sourcePath/PaperUrls").as[MagPaperUrl].groupBy("PaperId").agg(collect_list(struct("sourceUrl")).as("instances")).as[MagUrl]
@ -108,27 +112,27 @@ object SparkPreProcessMAG {
magPubs.joinWith(paperUrlDataset, col("_1").equalTo(paperUrlDataset("PaperId")), "left") magPubs.joinWith(paperUrlDataset, col("_1").equalTo(paperUrlDataset("PaperId")), "left")
.map { a: ((String, Publication), MagUrl) => ConversionUtil.addInstances((a._1._2, a._2)) } .map { a: ((String, Publication), MagUrl) => ConversionUtil.addInstances((a._1._2, a._2)) }
.write.mode(SaveMode.Overwrite) .write.mode(SaveMode.Overwrite)
.save(s"${parser.get("targetPath")}/merge_step_3") .save(s"$workingPath/merge_step_3")
// logger.info("Phase 6) Enrich Publication with description") // logger.info("Phase 6) Enrich Publication with description")
// val pa = spark.read.load(s"${parser.get("sourcePath")}/PaperAbstractsInvertedIndex").as[MagPaperAbstract] // val pa = spark.read.load(s"${parser.get("sourcePath")}/PaperAbstractsInvertedIndex").as[MagPaperAbstract]
// pa.map(ConversionUtil.transformPaperAbstract).write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/PaperAbstract") // pa.map(ConversionUtil.transformPaperAbstract).write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/PaperAbstract")
val paperAbstract = spark.read.load((s"${parser.get("targetPath")}/PaperAbstract")).as[MagPaperAbstract] val paperAbstract = spark.read.load((s"$workingPath/PaperAbstract")).as[MagPaperAbstract]
magPubs = spark.read.load(s"${parser.get("targetPath")}/merge_step_3").as[Publication] magPubs = spark.read.load(s"$workingPath/merge_step_3").as[Publication]
.map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)] .map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)]
magPubs.joinWith(paperAbstract, col("_1").equalTo(paperAbstract("PaperId")), "left") magPubs.joinWith(paperAbstract, col("_1").equalTo(paperAbstract("PaperId")), "left")
.map(item => ConversionUtil.updatePubsWithDescription(item) .map(item => ConversionUtil.updatePubsWithDescription(item)
).write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/merge_step_4") ).write.mode(SaveMode.Overwrite).save(s"$workingPath/merge_step_4")
logger.info("Phase 7) Enrich Publication with FieldOfStudy") logger.info("Phase 7) Enrich Publication with FieldOfStudy")
magPubs = spark.read.load(s"${parser.get("targetPath")}/merge_step_4").as[Publication] magPubs = spark.read.load(s"$workingPath/merge_step_4").as[Publication]
.map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)] .map(p => (ConversionUtil.extractMagIdentifier(p.getOriginalId.asScala), p)).as[(String, Publication)]
val fos = spark.read.load(s"$sourcePath/FieldsOfStudy").select($"FieldOfStudyId".alias("fos"), $"DisplayName", $"MainType") val fos = spark.read.load(s"$sourcePath/FieldsOfStudy").select($"FieldOfStudyId".alias("fos"), $"DisplayName", $"MainType")
@ -144,14 +148,14 @@ object SparkPreProcessMAG {
.equalTo(paperField("PaperId")), "left") .equalTo(paperField("PaperId")), "left")
.map(item => ConversionUtil.updatePubsWithSubject(item)) .map(item => ConversionUtil.updatePubsWithSubject(item))
.write.mode(SaveMode.Overwrite) .write.mode(SaveMode.Overwrite)
.save(s"${parser.get("targetPath")}/mag_publication") .save(s"$workingPath/mag_publication")
val s:RDD[Publication] = spark.read.load(s"${parser.get("targetPath")}/mag_publication").as[Publication] val s:RDD[Publication] = spark.read.load(s"$workingPath/mag_publication").as[Publication]
.map(p=>Tuple2(p.getId, p)).rdd.reduceByKey((a:Publication, b:Publication) => ConversionUtil.mergePublication(a,b)) .map(p=>Tuple2(p.getId, p)).rdd.reduceByKey((a:Publication, b:Publication) => ConversionUtil.mergePublication(a,b))
.map(_._2) .map(_._2)
spark.createDataset(s).as[Publication].write.mode(SaveMode.Overwrite).save(s"${parser.get("targetPath")}/mag_publication_u") spark.createDataset(s).as[Publication].write.mode(SaveMode.Overwrite).save(s"$targetPath/magPublication")
} }
} }

View File

@ -1,6 +1,7 @@
package eu.dnetlib.doiboost.orcid package eu.dnetlib.doiboost.orcid
import eu.dnetlib.dhp.schema.oaf.{Author, Publication} import eu.dnetlib.dhp.schema.oaf.{Author, DataInfo, Publication}
import eu.dnetlib.dhp.schema.orcid.OrcidDOI
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}
import org.apache.commons.lang.StringUtils import org.apache.commons.lang.StringUtils
@ -43,16 +44,19 @@ object ORCIDToOAF {
} }
def convertTOOAF(input:ORCIDElement) :Publication = { def convertTOOAF(input:OrcidDOI) :Publication = {
val doi = input.doi val doi = input.getDoi
val pub:Publication = new Publication val pub:Publication = new Publication
pub.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava) pub.setPid(List(createSP(doi.toLowerCase, "doi", PID_TYPES)).asJava)
pub.setDataInfo(generateDataInfo()) pub.setDataInfo(generateDataInfo())
pub.setId(generateIdentifier(pub, doi.toLowerCase)) pub.setId(generateIdentifier(pub, doi.toLowerCase))
try{ try{
pub.setAuthor(input.authors.map(a=> {
generateAuthor(a.name, a.surname, a.creditName, a.oid) val l:List[Author]= input.getAuthors.asScala.map(a=> {
}).asJava) generateAuthor(a.getName, a.getSurname, a.getCreditName, a.getOid)
})(collection.breakOut)
pub.setAuthor(l.asJava)
pub.setCollectedfrom(List(DoiBoostMappingUtil.createORIDCollectedFrom()).asJava) pub.setCollectedfrom(List(DoiBoostMappingUtil.createORIDCollectedFrom()).asJava)
pub.setDataInfo(DoiBoostMappingUtil.generateDataInfo()) pub.setDataInfo(DoiBoostMappingUtil.generateDataInfo())
pub pub
@ -63,6 +67,13 @@ object ORCIDToOAF {
} }
} }
def generateOricPIDDatainfo():DataInfo = {
val di =DoiBoostMappingUtil.generateDataInfo("0.91")
di.getProvenanceaction.setClassid("sysimport:crosswalk:entityregistry")
di.getProvenanceaction.setClassname("Harvested")
di
}
def generateAuthor(given: String, family: String, fullName:String, orcid: String): Author = { def generateAuthor(given: String, family: String, fullName:String, orcid: String): Author = {
val a = new Author val a = new Author
a.setName(given) a.setName(given)
@ -72,7 +83,7 @@ object ORCIDToOAF {
else else
a.setFullname(s"$given $family") a.setFullname(s"$given $family")
if (StringUtils.isNotBlank(orcid)) if (StringUtils.isNotBlank(orcid))
a.setPid(List(createSP(orcid, ORCID, PID_TYPES)).asJava) a.setPid(List(createSP(orcid, ORCID, PID_TYPES, generateOricPIDDatainfo())).asJava)
a a
} }

View File

@ -1,21 +1,72 @@
package eu.dnetlib.doiboost.orcid package eu.dnetlib.doiboost.orcid
import com.fasterxml.jackson.databind.{DeserializationFeature, ObjectMapper}
import eu.dnetlib.dhp.application.ArgumentApplicationParser import eu.dnetlib.dhp.application.ArgumentApplicationParser
import eu.dnetlib.dhp.oa.merge.AuthorMerger
import eu.dnetlib.dhp.schema.oaf.Publication import eu.dnetlib.dhp.schema.oaf.Publication
import eu.dnetlib.dhp.schema.orcid.OrcidDOI
import eu.dnetlib.doiboost.mag.ConversionUtil import eu.dnetlib.doiboost.mag.ConversionUtil
import org.apache.commons.io.IOUtils import org.apache.commons.io.IOUtils
import org.apache.spark.SparkConf import org.apache.spark.SparkConf
import org.apache.spark.rdd.RDD import org.apache.spark.rdd.RDD
import org.apache.spark.sql.expressions.Aggregator
import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession} import org.apache.spark.sql.{Dataset, Encoder, Encoders, SaveMode, SparkSession}
import org.slf4j.{Logger, LoggerFactory} import org.slf4j.{Logger, LoggerFactory}
object SparkConvertORCIDToOAF { object SparkConvertORCIDToOAF {
val logger: Logger = LoggerFactory.getLogger(SparkConvertORCIDToOAF.getClass)
def getPublicationAggregator(): Aggregator[(String, Publication), Publication, Publication] = new Aggregator[(String, Publication), Publication, Publication]{
override def zero: Publication = new Publication()
override def reduce(b: Publication, a: (String, Publication)): Publication = {
b.mergeFrom(a._2)
b.setAuthor(AuthorMerger.mergeAuthor(a._2.getAuthor, b.getAuthor))
if (b.getId == null)
b.setId(a._2.getId)
b
}
override def merge(wx: Publication, wy: Publication): Publication = {
wx.mergeFrom(wy)
wx.setAuthor(AuthorMerger.mergeAuthor(wy.getAuthor, wx.getAuthor))
if(wx.getId == null && wy.getId.nonEmpty)
wx.setId(wy.getId)
wx
}
override def finish(reduction: Publication): Publication = reduction
override def bufferEncoder: Encoder[Publication] =
Encoders.kryo(classOf[Publication])
override def outputEncoder: Encoder[Publication] =
Encoders.kryo(classOf[Publication])
}
def run(spark:SparkSession,sourcePath:String, targetPath:String):Unit = {
implicit val mapEncoderPubs: Encoder[Publication] = Encoders.kryo[Publication]
implicit val mapOrcid: Encoder[OrcidDOI] = Encoders.kryo[OrcidDOI]
implicit val tupleForJoinEncoder: Encoder[(String, Publication)] = Encoders.tuple(Encoders.STRING, mapEncoderPubs)
val mapper = new ObjectMapper()
mapper.getDeserializationConfig.withFeatures(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES)
val dataset:Dataset[OrcidDOI] = spark.createDataset(spark.sparkContext.textFile(sourcePath).map(s => mapper.readValue(s,classOf[OrcidDOI])))
logger.info("Converting ORCID to OAF")
dataset.map(o => ORCIDToOAF.convertTOOAF(o)).filter(p=>p!=null)
.map(d => (d.getId, d))
.groupByKey(_._1)(Encoders.STRING)
.agg(getPublicationAggregator().toColumn)
.map(p => p._2)
.write.mode(SaveMode.Overwrite).save(targetPath)
}
def main(args: Array[String]): Unit = { def main(args: Array[String]): Unit = {
val logger: Logger = LoggerFactory.getLogger(SparkConvertORCIDToOAF.getClass)
val conf: SparkConf = new SparkConf() val conf: SparkConf = new SparkConf()
val parser = new ArgumentApplicationParser(IOUtils.toString(SparkConvertORCIDToOAF.getClass.getResourceAsStream("/eu/dnetlib/dhp/doiboost/convert_map_to_oaf_params.json"))) val parser = new ArgumentApplicationParser(IOUtils.toString(SparkConvertORCIDToOAF.getClass.getResourceAsStream("/eu/dnetlib/dhp/doiboost/convert_map_to_oaf_params.json")))
parser.parseArgument(args) parser.parseArgument(args)
@ -26,19 +77,12 @@ object SparkConvertORCIDToOAF {
.appName(getClass.getSimpleName) .appName(getClass.getSimpleName)
.master(parser.get("master")).getOrCreate() .master(parser.get("master")).getOrCreate()
implicit val mapEncoderPubs: Encoder[Publication] = Encoders.kryo[Publication]
implicit val tupleForJoinEncoder: Encoder[(String, Publication)] = Encoders.tuple(Encoders.STRING, mapEncoderPubs)
import spark.implicits._
val sourcePath = parser.get("sourcePath") val sourcePath = parser.get("sourcePath")
val targetPath = parser.get("targetPath") val targetPath = parser.get("targetPath")
val dataset:Dataset[ORCIDElement] = spark.read.json(sourcePath).as[ORCIDElement] run(spark, sourcePath, targetPath)
logger.info("Converting ORCID to OAF")
val d:RDD[Publication] = dataset.map(o => ORCIDToOAF.convertTOOAF(o)).filter(p=>p!=null).map(p=>(p.getId,p)).rdd.reduceByKey(ConversionUtil.mergePublication)
.map(_._2)
spark.createDataset(d).as[Publication].write.mode(SaveMode.Overwrite).save(targetPath)
} }
} }

View File

@ -39,14 +39,7 @@
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message> <message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill> </kill>
<action name="ResetWorkingPath">
<fs>
<delete path='${workingDirPath}'/>
<mkdir path='${workingDirPath}'/>
</fs>
<ok to="CreateDOIBoost"/>
<error to="Kill"/>
</action>
<action name="CreateDOIBoost"> <action name="CreateDOIBoost">
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">

View File

@ -8,6 +8,10 @@
<name>targetPath</name> <name>targetPath</name>
<description>the working dir base path</description> <description>the working dir base path</description>
</property> </property>
<property>
<name>workingPath</name>
<description>the working dir base path</description>
</property>
<property> <property>
<name>sparkDriverMemory</name> <name>sparkDriverMemory</name>
<description>memory for driver process</description> <description>memory for driver process</description>
@ -31,10 +35,10 @@
<action name="ResetWorkingPath"> <action name="ResetWorkingPath">
<fs> <fs>
<delete path='${targetPath}'/> <delete path='${workingPath}'/>
<mkdir path='${targetPath}'/> <mkdir path='${workingPath}'/>
</fs> </fs>
<ok to="PreprocessMag"/> <ok to="ConvertMagToDataset"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
@ -52,10 +56,10 @@
${sparkExtraOPT} ${sparkExtraOPT}
</spark-opts> </spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}</arg> <arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg> <arg>--targetPath</arg><arg>${workingPath}</arg>
<arg>--master</arg><arg>yarn-cluster</arg> <arg>--master</arg><arg>yarn-cluster</arg>
</spark> </spark>
<ok to="End"/> <ok to="PreprocessMag"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
@ -65,7 +69,7 @@
<spark xmlns="uri:oozie:spark-action:0.2"> <spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master> <master>yarn-cluster</master>
<mode>cluster</mode> <mode>cluster</mode>
<name>Convert Mag to Dataset</name> <name>Convert Mag to OAF Dataset</name>
<class>eu.dnetlib.doiboost.mag.SparkPreProcessMAG</class> <class>eu.dnetlib.doiboost.mag.SparkPreProcessMAG</class>
<jar>dhp-doiboost-${projectVersion}.jar</jar> <jar>dhp-doiboost-${projectVersion}.jar</jar>
<spark-opts> <spark-opts>
@ -75,7 +79,8 @@
--conf spark.sql.shuffle.partitions=3840 --conf spark.sql.shuffle.partitions=3840
${sparkExtraOPT} ${sparkExtraOPT}
</spark-opts> </spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}</arg> <arg>--sourcePath</arg><arg>${workingPath}</arg>
<arg>--workingPath</arg><arg>${workingPath}/process</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg> <arg>--targetPath</arg><arg>${targetPath}</arg>
<arg>--master</arg><arg>yarn-cluster</arg> <arg>--master</arg><arg>yarn-cluster</arg>
</spark> </spark>

View File

@ -1,6 +1,7 @@
[ [
{"paramName":"s", "paramLongName":"sourcePath", "paramDescription": "the base path of MAG input", "paramRequired": true}, {"paramName":"s", "paramLongName":"sourcePath", "paramDescription": "the base path of MAG input", "paramRequired": true},
{"paramName":"t", "paramLongName":"targetPath", "paramDescription": "the working dir path", "paramRequired": true}, {"paramName":"t", "paramLongName":"targetPath", "paramDescription": "the target dir path", "paramRequired": true},
{"paramName":"w", "paramLongName":"workingPath", "paramDescription": "the working dir path", "paramRequired": true},
{"paramName":"m", "paramLongName":"master", "paramDescription": "the master name", "paramRequired": true} {"paramName":"m", "paramLongName":"master", "paramDescription": "the master name", "paramRequired": true}
] ]

View File

@ -1,5 +1,8 @@
package eu.dnetlib.doiboost.orcid package eu.dnetlib.doiboost.orcid
import eu.dnetlib.dhp.schema.oaf.Publication
import eu.dnetlib.doiboost.orcid.SparkConvertORCIDToOAF.getClass
import org.apache.spark.sql.{Encoder, Encoders, SparkSession}
import org.codehaus.jackson.map.ObjectMapper import org.codehaus.jackson.map.ObjectMapper
import org.junit.jupiter.api.Assertions._ import org.junit.jupiter.api.Assertions._
import org.junit.jupiter.api.Test import org.junit.jupiter.api.Test
@ -21,6 +24,30 @@ class MappingORCIDToOAFTest {
}) })
} }
// @Test
// def testOAFConvert():Unit ={
//
// val spark: SparkSession =
// SparkSession
// .builder()
// .appName(getClass.getSimpleName)
// .master("local[*]").getOrCreate()
//
//
// SparkConvertORCIDToOAF.run( spark,"/Users/sandro/Downloads/orcid", "/Users/sandro/Downloads/orcid_oaf")
// implicit val mapEncoderPubs: Encoder[Publication] = Encoders.kryo[Publication]
//
// val df = spark.read.load("/Users/sandro/Downloads/orcid_oaf").as[Publication]
// println(df.first.getId)
// println(mapper.writeValueAsString(df.first()))
//
//
//
//
// }