continue mapping from crossref to OAF

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

@@ -1,15 +1,30 @@
 package eu.dnetlib.doiboost.crossref
 
 import eu.dnetlib.dhp.schema.oaf._
+import eu.dnetlib.dhp.utils.DHPUtils
 import org.json4s
 import org.json4s.DefaultFormats
 import org.json4s.jackson.JsonMethods._
 import org.slf4j.Logger
 
 import scala.collection.JavaConverters._
+
 class Crossref2Oaf {
 
+//STATIC STRING
+  val MAG = "MAG"
+  val ORCID = "ORCID"
+  val CROSSREF = "Crossref"
+  val UNPAYWALL = "UnpayWall"
+  val GRID_AC = "grid.ac"
+  val WIKPEDIA = "wikpedia"
+  val doiBoostNSPREFIX = "doiboost____"
+  val OPENAIRE_PREFIX = "openaire____"
+  val SEPARATOR = "::"
+  val DNET_LANGUAGES = "dnet:languages"
   val PID_TYPES = "dnet:pid_types"
+
+
   val mappingCrossrefType = Map(
     "book-section" -> "publication",
     "book" -> "publication",
@@ -32,7 +47,7 @@ class Crossref2Oaf {
     "report-series" -> "publication",
     "standard" -> "publication",
     "standard-series" -> "publication",
-      "posted-content"-> "publication",
+    "posted-content" -> "publication",
     "dataset" -> "dataset"
   )
 
@@ -59,12 +74,12 @@ class Crossref2Oaf {
     "report-series" -> "0017 Report",
     "standard" -> "0038 Other literature type",
     "standard-series" -> "0038 Other literature type",
-    "dataset"->                       "0021 Dataset",
+    "dataset" -> "0021 Dataset",
-    "preprint"->                      "0016 Preprint",
+    "preprint" -> "0016 Preprint",
-    "report"->                        "0017 Report"
+    "report" -> "0017 Report"
   )
 
-  def convert(input: String, logger:Logger): Result = {
+  def convert(input: String, logger: Logger): Result = {
     implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
     lazy val json: json4s.JValue = parse(input)
     val objectType = (json \ "type").extractOrElse[String](null)
@@ -74,38 +89,68 @@ class Crossref2Oaf {
     val result = generateItemFromType(objectType, objectSubType)
     if (result == null)
       return result
-    val cOBJCategory = mappingCrossrefSubType.getOrElse(objectType,mappingCrossrefSubType.getOrElse(objectSubType,"0038 Other literature type"));
+    val cOBJCategory = mappingCrossrefSubType.getOrElse(objectType, mappingCrossrefSubType.getOrElse(objectSubType, "0038 Other literature type"));
+    logger.debug(mappingCrossrefType(objectType))
+    logger.debug(cOBJCategory)
 
-    logger.info(mappingCrossrefType(objectType))
+    //MAPPING Crossref DOI into PID
-    logger.info(cOBJCategory)
+    val doi: String = (json \ "DOI").extract[String]
-    val doi:String = (json \ "DOI").extract[String]
-    val pid = new StructuredProperty()
-    pid.setValue(doi)
-    pid.setQualifier(new Qualifier)
-    result.setPid(List(createSP(doi,"doi", PID_TYPES)).asJava)
 
-    logger.info(doi)
+    result.setPid(List(createSP(doi, "doi", PID_TYPES)).asJava)
+    //MAPPING Crossref DOI into OriginalId
+    result.setOriginalId(List(doi).asJava)
+    //Set identifier as {50|60} | doiboost____::md5(DOI)
+    result.setId(generateIdentifier(result, doi))
+
+    // Add DataInfo
+    result.setDataInfo(generateDataInfo())
+
+    result.setLastupdatetimestamp((json \"indexed" \"timestamp").extract[Long])
+
+    result.setDateofcollection((json \"indexed" \"date-time").extract[String])
 
     result
   }
 
 
-  def createSP(value:String, classId:String, schemeId:String ):StructuredProperty = {
+  def generateIdentifier(oaf: Result, doi:String): String = {
+    val id = DHPUtils.md5(doi.toLowerCase)
+    if (oaf.isInstanceOf[Dataset])
+      return s"60|${doiBoostNSPREFIX}${SEPARATOR}${id}"
+    s"50|${doiBoostNSPREFIX}${SEPARATOR}${id}"
+  }
+
+  def generateDataInfo(): DataInfo = {
+    val di =new DataInfo
+    di.setDeletedbyinference(false)
+    di.setInferred(false)
+    di.setInvisible(false)
+    di.setTrust("0.9")
+    di.setProvenanceaction(createQualifier("sysimport:actionset", "dnet:provenanceActions"))
+    di
+  }
+
+
+  def createSP(value: String, classId: String, schemeId: String): StructuredProperty = {
     val sp = new StructuredProperty
-    val q = new Qualifier
+    sp.setQualifier(createQualifier(classId, schemeId))
-    q.setClassid(classId)
-    q.setClassname(classId)
-    q.setSchemeid(schemeId)
-    q.setSchemename(schemeId  )
     sp.setValue(value)
-    sp.setQualifier(q)
     sp
 
   }
 
+  def createQualifier(cls:String, sch:String):Qualifier = {
+    val q = new Qualifier
+    q.setClassid(cls)
+    q.setClassname(cls)
+    q.setSchemeid(sch)
+    q.setSchemename(sch)
+    q
+  }
 
-  def generateItemFromType (objectType:String, objectSubType:String):Result = {
+
-    if (mappingCrossrefType.contains(objectType)){
+  def generateItemFromType(objectType: String, objectSubType: String): Result = {
+    if (mappingCrossrefType.contains(objectType)) {
       if (mappingCrossrefType(objectType).equalsIgnoreCase("publication"))
         return new Publication()
       if (mappingCrossrefType(objectType).equalsIgnoreCase("dataset"))

dhp-workflows/dhp-doiboost/src/test/resources/log4j.properties

@@ -6,5 +6,6 @@ log4j.appender.A1=org.apache.log4j.ConsoleAppender
 
 # A1 uses PatternLayout.
 log4j.logger.org = ERROR
+log4j.logger.eu.dnetlib = DEBUG
 log4j.appender.A1.layout=org.apache.log4j.PatternLayout
 log4j.appender.A1.layout.ConversionPattern=%-4r [%t] %-5p %c %x - %m%n