This commit is contained in:
Claudio Atzori 2022-10-19 15:21:02 +02:00
commit ec39b84898
1 changed files with 64 additions and 68 deletions

View File

@ -24,11 +24,11 @@ import scala.io.Source
object DataciteToOAFTransformation {
case class HostedByMapType(
openaire_id: String,
datacite_name: String,
official_name: String,
similarity: Option[Float]
) {}
openaire_id: String,
datacite_name: String,
official_name: String,
similarity: Option[Float]
) {}
val mapper = new ObjectMapper()
@ -47,12 +47,12 @@ object DataciteToOAFTransformation {
}
/** This method should skip record if json contains invalid text
* defined in file datacite_filter
*
* @param record : not parsed Datacite record
* @param json : parsed record
* @return True if the record should be skipped
*/
* defined in file datacite_filter
*
* @param record : not parsed Datacite record
* @param json : parsed record
* @return True if the record should be skipped
*/
def skip_record(record: String, json: org.json4s.JValue): Boolean = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
datacite_filter.exists(f => record.contains(f)) || (json \\ "publisher")
@ -98,11 +98,10 @@ object DataciteToOAFTransformation {
}
/**
* This utility method indicates whether the embargo date has been reached
* @param embargo_end_date
* @return True if the embargo date has been reached, false otherwise
*/
/** This utility method indicates whether the embargo date has been reached
* @param embargo_end_date
* @return True if the embargo date has been reached, false otherwise
*/
def embargo_end(embargo_end_date: String): Boolean = {
val dt = LocalDate.parse(embargo_end_date, DateTimeFormatter.ofPattern("[yyyy-MM-dd]"))
val td = LocalDate.now()
@ -147,27 +146,27 @@ object DataciteToOAFTransformation {
}
}
/***
* Use the vocabulary dnet:publication_resource to find a synonym to one of these terms and get the instance.type.
* Using the dnet:result_typologies vocabulary, we look up the instance.type synonym
* to generate one of the following main entities:
* - publication
* - dataset
* - software
* - otherresearchproduct
* @param resourceType
* @param resourceTypeGeneral
* @param schemaOrg
* @param vocabularies
* @return
*/
/** *
* Use the vocabulary dnet:publication_resource to find a synonym to one of these terms and get the instance.type.
* Using the dnet:result_typologies vocabulary, we look up the instance.type synonym
* to generate one of the following main entities:
* - publication
* - dataset
* - software
* - otherresearchproduct
*
* @param resourceType
* @param resourceTypeGeneral
* @param schemaOrg
* @param vocabularies
* @return
*/
def getTypeQualifier(
resourceType: String,
resourceTypeGeneral: String,
schemaOrg: String,
vocabularies: VocabularyGroup
): (Qualifier, Qualifier) = {
resourceType: String,
resourceTypeGeneral: String,
schemaOrg: String,
vocabularies: VocabularyGroup
): (Qualifier, Qualifier) = {
if (resourceType != null && resourceType.nonEmpty) {
val typeQualifier =
vocabularies.getSynonymAsQualifier(ModelConstants.DNET_PUBLICATION_RESOURCE, resourceType)
@ -212,11 +211,11 @@ object DataciteToOAFTransformation {
}
def getResult(
resourceType: String,
resourceTypeGeneral: String,
schemaOrg: String,
vocabularies: VocabularyGroup
): Result = {
resourceType: String,
resourceTypeGeneral: String,
schemaOrg: String,
vocabularies: VocabularyGroup
): Result = {
val typeQualifiers: (Qualifier, Qualifier) =
getTypeQualifier(resourceType, resourceTypeGeneral, schemaOrg, vocabularies)
if (typeQualifiers == null)
@ -258,11 +257,11 @@ object DataciteToOAFTransformation {
}
/** As describe in ticket #6377
* when the result come from figshare we need to remove subject
* and set Access rights OPEN.
*
* @param r
*/
* when the result come from figshare we need to remove subject
* and set Access rights OPEN.
*
* @param r
*/
def fix_figshare(r: Result): Unit = {
if (r.getInstance() != null) {
@ -289,12 +288,12 @@ object DataciteToOAFTransformation {
}
def generateRelation(
sourceId: String,
targetId: String,
relClass: String,
cf: KeyValue,
di: DataInfo
): Relation = {
sourceId: String,
targetId: String,
relClass: String,
cf: KeyValue,
di: DataInfo
): Relation = {
val r = new Relation
r.setSource(sourceId)
@ -323,12 +322,12 @@ object DataciteToOAFTransformation {
}
def generateOAF(
input: String,
ts: Long,
dateOfCollection: Long,
vocabularies: VocabularyGroup,
exportLinks: Boolean
): List[Oaf] = {
input: String,
ts: Long,
dateOfCollection: Long,
vocabularies: VocabularyGroup,
exportLinks: Boolean
): List[Oaf] = {
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
lazy val json = parse(input)
@ -413,7 +412,6 @@ object DataciteToOAFTransformation {
return List()
result.setAuthor(authors.asJava)
val titles: List[TitleType] = (json \\ "titles").extractOrElse[List[TitleType]](List())
result.setTitle(
@ -437,8 +435,6 @@ object DataciteToOAFTransformation {
.asJava
)
val dates = (json \\ "dates").extract[List[DateType]]
val publication_year = (json \\ "publicationYear").extractOrElse[String](null)
@ -645,16 +641,16 @@ object DataciteToOAFTransformation {
}
private def generateRelations(
rels: List[RelatedIdentifierType],
id: String,
date: String
): List[Relation] = {
rels: List[RelatedIdentifierType],
id: String,
date: String
): List[Relation] = {
rels
.filter(r =>
subRelTypeMapping
.contains(r.relationType) && (r.relatedIdentifierType.equalsIgnoreCase("doi") ||
r.relatedIdentifierType.equalsIgnoreCase("pmid") ||
r.relatedIdentifierType.equalsIgnoreCase("arxiv"))
r.relatedIdentifierType.equalsIgnoreCase("pmid") ||
r.relatedIdentifierType.equalsIgnoreCase("arxiv"))
)
.map(r => {
val rel = new Relation
@ -686,4 +682,4 @@ object DataciteToOAFTransformation {
s"10|$b::${DHPUtils.md5(a)}"
}
}
}