Benefit from explicit-nulls in CpIndex and IndexData

src/main/scala/se/lu/nateko/cp/meta/services/sparql/magic/CpIndex.scala

@@ -1,7 +1,5 @@
 package se.lu.nateko.cp.meta.services.sparql.magic
 
-import scala.language.unsafeNulls
-
 import org.eclipse.rdf4j.model.IRI
 import org.eclipse.rdf4j.model.ValueFactory
 import org.eclipse.rdf4j.sail.Sail
@@ -36,10 +34,10 @@ trait ObjSpecific{
 }
 
 trait ObjInfo extends ObjSpecific{
-	def spec: IRI
-	def submitter: IRI
-	def station: IRI
-	def site: IRI
+	def spec: IRI | Null
+	def submitter: IRI | Null
+	def station: IRI | Null
+	def site: IRI | Null
 	def fileName: Option[String]
 	def sizeInBytes: Option[Long]
 	def samplingHeightMeters: Option[Float]
@@ -51,7 +49,7 @@ trait ObjInfo extends ObjSpecific{
 }
 
 class CpIndex(sail: Sail, geo: Future[GeoIndex], data: IndexData)(using EnvriConfigs) extends ReadWriteLocking:
-	private val log = LoggerFactory.getLogger(getClass())
+	private val log = LoggerFactory.getLogger(getClass()).nn
 	private val filtering = Filtering(data, geo)
 
 	import data.{contMap, stats, objs, initOk, idLookup}
@@ -89,7 +87,7 @@ class CpIndex(sail: Sail, geo: Future[GeoIndex], data: IndexData)(using EnvriCon
 		log.info("CpIndex got initialized with non-empty index data to use")
 		reportDebugInfo()
 
-	given factory: ValueFactory = sail.getValueFactory
+	given factory: ValueFactory = sail.getValueFactory.nn
 	val vocab = new CpmetaVocab(factory)
 
 	private val queue = new ArrayBlockingQueue[RdfUpdate](UpdateQueueSize)
@@ -100,11 +98,11 @@ class CpIndex(sail: Sail, geo: Future[GeoIndex], data: IndexData)(using EnvriCon
 	def fetch(req: DataObjectFetch): Iterator[ObjInfo] = readLocked{
 		//val start = System.currentTimeMillis
 
-		val filter = filtering(req.filter).fold(initOk)(BufferFastAggregation.and(_, initOk))
+		val filter = filtering(req.filter).fold(initOk)(BufferFastAggregation.and(_, initOk).nn)
 
 		val idxIter: Iterator[Int] = req.sort match{
 			case None =>
-				filter.iterator.asScala.drop(req.offset).map(_.intValue)
+				filter.iterator.nn.asScala.drop(req.offset).map(_.intValue)
 			case Some(SortBy(prop, descending)) =>
 				data.bitmap(prop).iterateSorted(Some(filter), req.offset, descending)
 		}
@@ -131,7 +129,7 @@ class CpIndex(sail: Sail, geo: Future[GeoIndex], data: IndexData)(using EnvriCon
 	}
 
 	def getUniqueKeywords(req: DataObjectFetch): Iterable[String] = readLocked {
-		val objectIds = filtering(req.filter).fold(initOk)(BufferFastAggregation.and(_, initOk))
+		val objectIds = filtering(req.filter).fold(initOk)(BufferFastAggregation.and(_, initOk).nn)
 		data.getObjectKeywords(objectIds)
 	}
 

src/main/scala/se/lu/nateko/cp/meta/services/sparql/magic/Filtering.scala

@@ -1,7 +1,5 @@
 package se.lu.nateko.cp.meta.services.sparql.magic
 
-import scala.language.unsafeNulls
-
 import org.eclipse.rdf4j.model.IRI
 import org.roaringbitmap.buffer.{BufferFastAggregation, ImmutableRoaringBitmap, MutableRoaringBitmap}
 import se.lu.nateko.cp.meta.services.sparql.index.*
@@ -34,7 +32,7 @@ class Filtering(data: IndexData, geo: Future[GeoIndex]) {
 
 		case Exists(prop) => prop match {
 				case cp: ContProp => Some(data.bitmap(cp).all)
-				case optUriProp: OptUriProperty => 
+				case optUriProp: OptUriProperty =>
 					// TODO: Not covered by tests, so unsure if it's correct right now.
 					data.categoryBitmap(optUriProp, Seq(None)) match {
 							case bm if bm.isEmpty => None
@@ -53,7 +51,7 @@ class Filtering(data: IndexData, geo: Future[GeoIndex]) {
 				.collect { case iri: IRI => iri }
 				.collect { case CpVocab.DataObject(hash, _) => idLookup.get(hash) }
 				.flatten
-			Some(ImmutableRoaringBitmap.bitmapOf(objIndices*))
+			Some(ImmutableRoaringBitmap.bitmapOf(objIndices*).nn)
 
 		case CategFilter(category, values) =>
 			Some(data.categoryBitmap(category, values))
@@ -94,8 +92,8 @@ class Filtering(data: IndexData, geo: Future[GeoIndex]) {
 			case Some(Failure(exc)) =>
 				throw Exception("Geo indexing failed", exc)
 
-	private def negate(bm: ImmutableRoaringBitmap) =
-		if objs.length == 0 then emptyBitmap else ImmutableRoaringBitmap.flip(bm, 0, objs.length.toLong)
+	private def negate(bm: ImmutableRoaringBitmap): MutableRoaringBitmap =
+		if objs.length == 0 then emptyBitmap else ImmutableRoaringBitmap.flip(bm, 0, objs.length.toLong).nn
 
 	private def collectUnless[T](iter: Iterator[T])(cond: T => Boolean): Option[Seq[T]] = {
 		var condHappened = false
@@ -107,8 +105,8 @@ class Filtering(data: IndexData, geo: Future[GeoIndex]) {
 	}
 
 	private def or(bms: Seq[ImmutableRoaringBitmap]): Option[MutableRoaringBitmap] =
-		if (bms.isEmpty) Some(emptyBitmap) else Some(BufferFastAggregation.or(bms*))
+		if (bms.isEmpty) Some(emptyBitmap) else Some(BufferFastAggregation.or(bms*).nn)
 
 	private def and(bms: Seq[ImmutableRoaringBitmap]): Option[MutableRoaringBitmap] =
-		if (bms.isEmpty) None else Some(BufferFastAggregation.and(bms*))
+		if (bms.isEmpty) None else Some(BufferFastAggregation.and(bms*).nn)
 }

src/main/scala/se/lu/nateko/cp/meta/services/sparql/magic/index/IndexData.scala

@@ -1,7 +1,5 @@
 package se.lu.nateko.cp.meta.services.sparql.magic.index
 
-import scala.language.unsafeNulls
-
 import org.eclipse.rdf4j.model.IRI
 import org.eclipse.rdf4j.model.Literal
 import org.eclipse.rdf4j.model.Statement
@@ -27,7 +25,6 @@ import se.lu.nateko.cp.meta.services.sparql.index.StringHierarchicalBitmap.Strin
 import se.lu.nateko.cp.meta.services.sparql.magic.ObjInfo
 import se.lu.nateko.cp.meta.utils.parseCommaSepList
 import se.lu.nateko.cp.meta.utils.parseJsonStringArray
-import se.lu.nateko.cp.meta.utils.rdf4j.===
 import se.lu.nateko.cp.meta.utils.rdf4j.Rdf4jStatement
 import se.lu.nateko.cp.meta.utils.rdf4j.asString
 import se.lu.nateko.cp.meta.utils.rdf4j.toJava
@@ -46,12 +43,12 @@ final class DataStartGeo(objs: IndSeq[ObjEntry]) extends DateTimeGeo(objs(_).dat
 final class DataEndGeo(objs: IndSeq[ObjEntry]) extends DateTimeGeo(objs(_).dataEnd)
 final class SubmStartGeo(objs: IndSeq[ObjEntry]) extends DateTimeGeo(objs(_).submissionStart)
 final class SubmEndGeo(objs: IndSeq[ObjEntry]) extends DateTimeGeo(objs(_).submissionEnd)
-final class FileNameGeo(objs: IndSeq[ObjEntry]) extends StringGeo(objs.apply(_).fName)
+final class FileNameGeo(objs: IndSeq[ObjEntry]) extends StringGeo(objs.apply(_).nn.fName.nn)
 
 final case class StatKey(spec: IRI, submitter: IRI, station: Option[IRI], site: Option[IRI])
 final case class StatEntry(key: StatKey, count: Int)
 
-def emptyBitmap = MutableRoaringBitmap.bitmapOf()
+def emptyBitmap: MutableRoaringBitmap = MutableRoaringBitmap.bitmapOf().nn
 
 final class IndexData(nObjects: Int)(
 	// These members are public only because of serialization, and should not be accessed directly.
@@ -65,7 +62,7 @@ final class IndexData(nObjects: Int)(
 	val stats: AnyRefMap[StatKey, MutableRoaringBitmap] = AnyRefMap.empty,
 	val initOk: MutableRoaringBitmap = emptyBitmap
 ) extends Serializable:
-	private val log = LoggerFactory.getLogger(getClass())
+	private val log = LoggerFactory.getLogger(getClass()).nn
 
 	private def dataStartBm = DatetimeHierarchicalBitmap(DataStartGeo(objs))
 	private def dataEndBm = DatetimeHierarchicalBitmap(DataEndGeo(objs))
@@ -83,7 +80,7 @@ final class IndexData(nObjects: Int)(
 
 	def getObjectKeywords(objectIds: ImmutableRoaringBitmap): Iterable[String] = {
 		categoryKeys(Keyword).collect {
-			case keyword if !BufferFastAggregation.and(objectIds, keywordBitmap(Seq(keyword))).isEmpty() =>
+			case keyword if !BufferFastAggregation.and(objectIds, keywordBitmap(Seq(keyword))).nn.isEmpty() =>
 				keyword
 		}
 	}
@@ -109,7 +106,7 @@ final class IndexData(nObjects: Int)(
 				keywordBitmap(values.asInstanceOf[Iterable[Keyword.ValueType]])
 			case _ => {
 				val category = categMap(prop)
-				BufferFastAggregation.or(values.map(v => category.getOrElse(v, emptyBitmap)).toSeq*)
+				BufferFastAggregation.or(values.map(v => category.getOrElse(v, emptyBitmap)).toSeq*).nn
 			}
 		}
 	}
@@ -141,7 +138,7 @@ final class IndexData(nObjects: Int)(
 		val objectMap = categMap(Keyword)
 		val objects = keywords.flatMap(objectMap.get)
 
-		BufferFastAggregation.or(LazyList(specObjects, objects).flatten*)
+		BufferFastAggregation.or(LazyList(specObjects, objects).flatten*).nn
 	}
 
 	def processUpdate(
@@ -164,10 +161,13 @@ final class IndexData(nObjects: Int)(
 							if (filterByEnvri) EnvriResolver.infer(subj.toJava).foreach: envri =>
 								updateCategSet(EnvriProp, envri, oe.idx, isAssertion)
 							if (isAssertion) {
-								if (oe.spec != null) removeStat(oe, initOk)
+								if (oe.spec != null) {
+									removeStat(oe, initOk)
+								}
+
 								oe.spec = spec
 								addStat(oe, initOk)
-							} else if (spec === oe.spec) {
+							} else if (spec == oe.spec) {
 								removeStat(oe, initOk)
 								oe.spec = null
 							}
@@ -183,7 +183,7 @@ final class IndexData(nObjects: Int)(
 
 			case `hasName` =>
 				getDataObject(subj).foreach { oe =>
-					val fName = obj.stringValue
+					val fName: String = obj.stringValue.nn
 					if (isAssertion) oe.fName = fName
 					else if (oe.fName == fName) { oe.fName = null }
 					handleContinuousPropUpdate(FileName, fName, oe.idx, isAssertion)
@@ -310,7 +310,7 @@ final class IndexData(nObjects: Int)(
 
 							val directPrevVers: IndexedSeq[Int] =
 								StatementSource.getStatements(subj, isNextVersionOf, null)
-									.flatMap(st => getDataObject(st.getObject).map(_.idx))
+									.flatMap(st => getDataObject(st.getObject.nn).map(_.idx))
 									.toIndexedSeq
 
 							directPrevVers.foreach { oldIdx =>
@@ -437,7 +437,6 @@ final class IndexData(nObjects: Int)(
 				val _ = mappings.remove(categ)
 			}
 	}
-
 	private def getSpecProjectKeywords(spec: IRI)(using CpmetaVocab, StatementSource): Set[String] = {
 		StatementSource
 			.getUriValues(spec, summon[CpmetaVocab].hasAssociatedProject)
@@ -541,7 +540,7 @@ final class IndexData(nObjects: Int)(
 		case CpVocab.DataObject(hash, prefix) =>
 			val entry = getObjEntry(hash)
 			if (entry.prefix == "") {
-				entry.prefix = prefix.intern()
+				entry.prefix = prefix.nn.intern().nn
 			}
 			Some(entry)
 
@@ -575,27 +574,41 @@ private def targetUri(obj: Value, isAssertion: Boolean) =
 	then obj.asInstanceOf[IRI]
 	else null
 
+// TODO: Option.scala isn't currently written for explicit-nulls. Maybe changed in later scalac versions?
+private def makeOption[A](arg: A | Null) = {
+	if (arg == null) { None }
+	else { Some(arg.nn) }
+}
+
 private def keyForDobj(obj: ObjEntry): Option[StatKey] =
-	if obj.spec == null || obj.submitter == null then None
-	else
-		Some(
-			StatKey(obj.spec, obj.submitter, Option(obj.station), Option(obj.site))
-		)
+	(obj.spec, obj.submitter) match {
+		case (spec: IRI, submitter: IRI) => {
+			Some(
+				StatKey(
+					spec,
+					submitter,
+					makeOption(obj.station),
+					makeOption(obj.site)
+				)
+			)
+		}
+		case _ => None
+	}
 
 private def ifDateTime(dt: Value)(mod: Long => Unit): Unit = dt match
-	case lit: Literal if lit.getDatatype === XSD.DATETIME =>
-		try mod(Instant.parse(lit.stringValue).toEpochMilli)
+	case lit: Literal if lit.getDatatype == XSD.DATETIME =>
+		try mod(Instant.parse(lit.stringValue).nn.toEpochMilli)
 		catch case _: Throwable => () // ignoring wrong dateTimes
 	case _ =>
 
 private def ifLong(dt: Value)(mod: Long => Unit): Unit = dt match
-	case lit: Literal if lit.getDatatype === XSD.LONG =>
+	case lit: Literal if lit.getDatatype == XSD.LONG =>
 		try mod(lit.longValue)
 		catch case _: Throwable => () // ignoring wrong longs
 	case _ =>
 
 private def ifFloat(dt: Value)(mod: Float => Unit): Unit = dt match
-	case lit: Literal if lit.getDatatype === XSD.FLOAT =>
+	case lit: Literal if lit.getDatatype == XSD.FLOAT =>
 		try mod(lit.floatValue)
 		catch case _: Throwable => () // ignoring wrong floats
 	case _ =>

src/main/scala/se/lu/nateko/cp/meta/services/sparql/magic/index/ObjEntry.scala

@@ -1,7 +1,5 @@
 package se.lu.nateko.cp.meta.services.sparql.magic.index
 
-import scala.language.unsafeNulls
-
 import org.eclipse.rdf4j.model.{IRI, ValueFactory}
 import se.lu.nateko.cp.meta.core.crypto.Sha256Sum
 import se.lu.nateko.cp.meta.services.sparql.magic.ObjInfo
@@ -10,12 +8,12 @@ import java.time.Instant
 import scala.compiletime.uninitialized
 
 final class ObjEntry(val hash: Sha256Sum, val idx: Int, var prefix: String) extends ObjInfo with Serializable {
-	var spec: IRI = uninitialized
-	var submitter: IRI = uninitialized
-	var station: IRI = uninitialized
-	var site: IRI = uninitialized
+	var spec: IRI | Null = uninitialized
+	var submitter: IRI | Null = uninitialized
+	var station: IRI | Null = uninitialized
+	var site: IRI | Null = uninitialized
 	var size: Long = -1
-	var fName: String = ""
+	var fName: String | Null = ""
 	var samplingHeight: Float = Float.NaN
 	var dataStart: Long = Long.MinValue
 	var dataEnd: Long = Long.MinValue
@@ -25,15 +23,20 @@ final class ObjEntry(val hash: Sha256Sum, val idx: Int, var prefix: String) exte
 
 	private final def dateTimeFromLong(dt: Long): Option[Instant] =
 		if (dt == Long.MinValue) None
-		else Some(Instant.ofEpochMilli(dt))
+		else Some(Instant.ofEpochMilli(dt).nn)
 
 	def sizeInBytes: Option[Long] = if (size >= 0) Some(size) else None
-	def fileName: Option[String] = Option(fName)
+	def fileName: Option[String] = makeOption(fName)
 	def samplingHeightMeters: Option[Float] = if (samplingHeight == Float.NaN) None else Some(samplingHeight)
 	def dataStartTime: Option[Instant] = dateTimeFromLong(dataStart)
 	def dataEndTime: Option[Instant] = dateTimeFromLong(dataEnd)
 	def submissionStartTime: Option[Instant] = dateTimeFromLong(submissionStart)
 	def submissionEndTime: Option[Instant] = dateTimeFromLong(submissionEnd)
 
-	def uri(factory: ValueFactory): IRI = factory.createIRI(prefix + hash.base64Url)
+	def uri(factory: ValueFactory): IRI = factory.createIRI(prefix + hash.base64Url).nn
+
+	private def makeOption[T](arg: T | Null) : Option[T] = {
+		if (arg == null) { None }
+		else { Some(arg.nn) }
+	}
 }