Clear fileName of index objects when hasName triple is removed

src/main/scala/se/lu/nateko/cp/meta/instanceserver/InstanceServer.scala

@@ -64,13 +64,19 @@ trait InstanceServer extends AutoCloseable:
 
 end InstanceServer
 
-trait TriplestoreConnection extends AutoCloseable:
+trait StatementSource {
+	def getStatements(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): CloseableIterator[Statement]
+	def hasStatement(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): Boolean
+
+	final def hasStatement(st: Statement): Boolean = st.getSubject() match
+		case subj: IRI => hasStatement(subj, st.getPredicate(), st.getObject())
+		case _ => false
+}
+
+trait TriplestoreConnection extends AutoCloseable, StatementSource:
 	def primaryContext: IRI
 	def readContexts: Seq[IRI]
 	def factory: ValueFactory
-
-	def getStatements(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): CloseableIterator[Statement]
-	def hasStatement(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): Boolean
 	def withContexts(primary: IRI, read: Seq[IRI]): TriplestoreConnection
 
 	final def withReadContexts(read: Seq[IRI]): TriplestoreConnection =
@@ -81,10 +87,6 @@ trait TriplestoreConnection extends AutoCloseable:
 		if readContexts.length == 1 && readContexts.head == primaryContext then this
 		else withContexts(primaryContext, Seq(primaryContext))
 
-	final def hasStatement(st: Statement): Boolean = st.getSubject() match
-		case subj: IRI => hasStatement(subj, st.getPredicate(), st.getObject())
-		case _ => false
-
 
 object TriplestoreConnection:
 	import Validated.CardinalityExpectation.{AtMostOne, ExactlyOne}

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

@@ -1,6 +1,6 @@
 package se.lu.nateko.cp.meta.services.sparql.magic
 
-import org.eclipse.rdf4j.model.{IRI, Value, ValueFactory}
+import org.eclipse.rdf4j.model.{IRI, ValueFactory}
 import org.eclipse.rdf4j.sail.Sail
 import org.roaringbitmap.buffer.{BufferFastAggregation, ImmutableRoaringBitmap, MutableRoaringBitmap}
 import org.slf4j.LoggerFactory
@@ -232,13 +232,10 @@ class CpIndex(sail: Sail, geo: Future[GeoIndex], data: IndexData) extends ReadWr
 			sail.accessEagerly:
 				list.forEach:
 					case RdfUpdate(Rdf4jStatement(subj, pred, obj), isAssertion) =>
-						processUpdate(subj, pred, obj, isAssertion)
+						data.processTriple(subj, pred, obj, isAssertion, vocab)
 					case _ => ()
 			list.clear()
 
-	private def processUpdate(subj: IRI, pred: IRI, obj: Value, isAssertion: Boolean)(using GlobConn): Unit =
-		data.processTriple(subj, pred, obj, isAssertion, vocab)
-
 end CpIndex
 
 

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

@@ -7,7 +7,7 @@ import org.slf4j.LoggerFactory
 import se.lu.nateko.cp.meta.core.algo.DatetimeHierarchicalBitmap.DateTimeGeo
 import se.lu.nateko.cp.meta.core.algo.{DatetimeHierarchicalBitmap, HierarchicalBitmap}
 import se.lu.nateko.cp.meta.core.crypto.Sha256Sum
-import se.lu.nateko.cp.meta.instanceserver.TriplestoreConnection as TSC
+import se.lu.nateko.cp.meta.instanceserver.StatementSource
 import se.lu.nateko.cp.meta.services.linkeddata.UriSerializer.Hash
 import se.lu.nateko.cp.meta.services.sparql.index.*
 import se.lu.nateko.cp.meta.services.sparql.index.StringHierarchicalBitmap.StringGeo
@@ -69,7 +69,7 @@ class IndexData(nObjects: Int)(
 		.getOrElseUpdate(prop, new AnyRefMap[prop.ValueType, MutableRoaringBitmap])
 		.asInstanceOf[AnyRefMap[prop.ValueType, MutableRoaringBitmap]]
 
-	def processTriple(subj: IRI, pred: IRI, obj: Value, isAssertion: Boolean, vocab: CpmetaVocab)(using TSC): Unit = {
+	def processTriple(subj: IRI, pred: IRI, obj: Value, isAssertion: Boolean, vocab: CpmetaVocab)(using statements: StatementSource): Unit = {
 		import vocab.*
 		import vocab.prov.{wasAssociatedWith, startedAtTime, endedAtTime}
 		import vocab.dcterms.hasPart
@@ -93,10 +93,10 @@ class IndexData(nObjects: Int)(
 				}
 
 			case `hasName` =>
-				modForDobj(subj) { oe =>
+				val _ = modForDobj(subj) { oe =>
 					val fName = obj.stringValue
 					if (isAssertion) oe.fName = fName
-					else if (oe.fName == fName) oe.fileName == null
+					else if (oe.fName == fName) { oe.fName = null }
 					handleContinuousPropUpdate(FileName, fName, oe.idx, isAssertion)
 				}
 
@@ -204,7 +204,7 @@ class IndexData(nObjects: Int)(
 											case _ =>
 					else if
 						deprecated.contains(oe.idx) && // this was to prevent needless repo access
-						!TSC.hasStatement(null, isNextVersionOf, obj)
+						!statements.hasStatement(null, isNextVersionOf, obj)
 					then deprecated.remove(oe.idx)
 				}
 
@@ -220,7 +220,7 @@ class IndexData(nObjects: Int)(
 							val deprecated = boolBitmap(DeprecationFlag)
 
 							val directPrevVers: IndexedSeq[Int] =
-								TSC.getStatements(subj, isNextVersionOf, null)
+								statements.getStatements(subj, isNextVersionOf, null)
 									.flatMap(st => modForDobj(st.getObject)(_.idx))
 									.toIndexedSeq
 
@@ -324,8 +324,8 @@ class IndexData(nObjects: Int)(
 		if (isAssertion) hasVarsBm.add(idx) else hasVarsBm.remove(idx)
 	}
 
-	private def nextVersCollIsComplete(obj: IRI, vocab: CpmetaVocab)(using TSC): Boolean =
-		TSC.getStatements(obj, vocab.dcterms.hasPart, null)
+	private def nextVersCollIsComplete(obj: IRI, vocab: CpmetaVocab)(using statements: StatementSource): Boolean =
+		statements.getStatements(obj, vocab.dcterms.hasPart, null)
 			.collect:
 				case Rdf4jStatement(_, _, member: IRI) => modForDobj(member): oe =>
 						oe.isNextVersion = true
@@ -334,8 +334,8 @@ class IndexData(nObjects: Int)(
 			.toIndexedSeq
 			.exists(identity)
 
-	private def getIdxsOfPrevVersThroughColl(deprecator: IRI, vocab: CpmetaVocab)(using TSC): Option[Int] =
-		TSC.getStatements(null, vocab.dcterms.hasPart, deprecator)
+	private def getIdxsOfPrevVersThroughColl(deprecator: IRI, vocab: CpmetaVocab)(using statements: StatementSource): Option[Int] =
+		statements.getStatements(null, vocab.dcterms.hasPart, deprecator)
 			.collect { case Rdf4jStatement(CpVocab.NextVersColl(oldHash), _, _) => getObjEntry(oldHash).idx }
 			.toIndexedSeq
 			.headOption

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

@@ -7,7 +7,7 @@ import se.lu.nateko.cp.meta.services.sparql.magic.ObjInfo
 import java.time.Instant
 import scala.compiletime.uninitialized
 
-class ObjEntry(val hash: Sha256Sum, val idx: Int, var prefix: String) extends ObjInfo with Serializable {
+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
@@ -21,7 +21,7 @@ class ObjEntry(val hash: Sha256Sum, val idx: Int, var prefix: String) extends Ob
 	var submissionEnd: Long = Long.MinValue
 	var isNextVersion: Boolean = false
 
-	private def dateTimeFromLong(dt: Long): Option[Instant] =
+	private final def dateTimeFromLong(dt: Long): Option[Instant] =
 		if (dt == Long.MinValue) None
 		else Some(Instant.ofEpochMilli(dt))
 

src/test/scala/se/lu/nateko/cp/meta/test/services/sparql/index/IndexDataTest.scala

@@ -0,0 +1,38 @@
+package se.lu.nateko.cp.meta.test.services.sparql.index
+
+import org.eclipse.rdf4j.model.{IRI, Statement, Value}
+import org.eclipse.rdf4j.model.impl.SimpleValueFactory
+import org.scalatest.funspec.AnyFunSpec
+import se.lu.nateko.cp.meta.api.CloseableIterator
+import se.lu.nateko.cp.meta.instanceserver.StatementSource
+import se.lu.nateko.cp.meta.services.sparql.magic.index.IndexData
+import se.lu.nateko.cp.meta.services.{CpVocab, CpmetaVocab}
+
+
+class IndexDataTest extends AnyFunSpec {
+	describe("processTriple") {
+		it("clears fileName of ObjEntry when hasName tuple is deleted") {
+			val factory = SimpleValueFactory.getInstance()
+			val vocab = CpmetaVocab(factory)
+			val subject: IRI = factory.createIRI("https://meta.icos-cp.eu/objects/oAzNtfjXddcnG_irI8fJT7W6")
+			val CpVocab.DataObject(hash, _) = subject : @unchecked
+			val data = IndexData(100)()
+
+			// IndexData requires a StatementSource but in this case we never pull any statements,
+			// hence we can leave things unimplemented.
+			given StatementSource with
+				def getStatements(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): CloseableIterator[Statement] = ???
+				def hasStatement(subject: IRI | Null, predicate: IRI | Null, obj: Value | Null): Boolean = ???
+
+			// Insert hasName triple
+			data.processTriple(subject, vocab.hasName, factory.createLiteral("test name"), true, vocab)
+			assert(data.objs.length == 1)
+			assert(data.getObjEntry(hash).fileName === Some("test name"))
+
+			// Remove it
+			data.processTriple(subject, vocab.hasName, factory.createLiteral("test name"), false, vocab)
+			assert(data.getObjEntry(hash).fileName === None)
+			assert(data.objs.length == 1)
+		}
+	}
+}