fix!: fix valueHasUri bad values and missing types (DEV-1036)

.gitignore

@@ -21,6 +21,7 @@ sipi/test
 /docs/architecture/.structurizr
 /docs/architecture/workspace.json
 
+
 **/target/
 *.aux
 *.bbl
@@ -39,11 +40,11 @@ sipi/test
 
 dependencies.txt
 /client-test-data.zip
+/db_ls_test_server_dump.trig
 /sipi/images/082E/*
 /sipi/images/originals/082E/*
 sipi/images/1111/*
 sipi/images/originals/1111/*
-/dependencies.bzl
 
 .idea/
 .metals/

test_data/upgrade/pr2079.trig

@@ -0,0 +1,29 @@
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+@prefix owl: <http://www.w3.org/2002/07/owl#> .
+@prefix knora-base: <http://www.knora.org/ontology/knora-base#> .
+@prefix knora-admin: <http://www.knora.org/ontology/knora-admin#> .
+@prefix anything: <http://www.knora.org/ontology/0001/anything#> .
+
+# value with missing valueHasUri datatype
+<http://rdfh.ch/0103/fN89IUgvSSyMxJ7XWssP9w/values/Rl2rfjDlRBWeuRr-EgIgCw>
+    rdf:type                      knora-base:UriValue ;
+    knora-base:attachedToUser     <http://rdfh.ch/users/user-unil-valentina-ponzetto> ;
+    knora-base:hasPermissions     "CR knora-admin:ProjectAdmin|D knora-admin:Creator|M knora-admin:ProjectMember|V knora-admin:KnownUser,knora-admin:UnknownUser" ;
+    knora-base:isDeleted          false ;
+    knora-base:valueCreationDate  "2018-01-12T19:21:37.991+00:00"^^xsd:dateTime ;
+    knora-base:valueHasOrder      1 ;
+    knora-base:valueHasString     "https://viaf.org/viaf/2571146" ;
+    knora-base:valueHasUUID       "PYQ5mqH8TYOKCytMozecWw" ;
+    knora-base:valueHasUri        "https://viaf.org/viaf/2571146" .
+
+# bad valueHasUri value and missing datatype
+<http://rdfh.ch/0103/5LE8P57nROClWUxEPJhiug/values/fEbt5NzaSe6GnCqKoF4Nhg/standoff/2>
+    rdf:type                                knora-base:StandoffUriTag ;
+    knora-base:standoffTagHasEnd            50 ;
+    knora-base:standoffTagHasStart          20 ;
+    knora-base:standoffTagHasStartIndex     2 ;
+    knora-base:standoffTagHasStartParent     <http://rdfh.ch/0103/5LE8P57nROClWUxEPJhiug/values/fEbt5NzaSe6GnCqKoF4Nhg/standoff/1> ;
+    knora-base:standoffTagHasUUID           "dd2e785b-041e-4bcb-a77d-570efe8a068c" ;
+    knora-base:valueHasUri                  <http://www.maison-george-sand.fr/> .

webapi/src/main/scala/org/knora/webapi/store/triplestore/upgrade/RepositoryUpdatePlan.scala

@@ -54,6 +54,11 @@ object RepositoryUpdatePlan {
         versionNumber = 20,
         plugin = new UpgradePluginPR2018(featureFactoryConfig, log),
         prBasedVersionString = Some("PR 2018")
+      ),
+      PluginForKnoraBaseVersion(
+        versionNumber = 20,
+        plugin = new UpgradePluginPR2079(featureFactoryConfig, log),
+        prBasedVersionString = Some("PR 2079")
       )
     )
 

webapi/src/main/scala/org/knora/webapi/store/triplestore/upgrade/plugins/UpgradePluginPR2079.scala

@@ -0,0 +1,67 @@
+/*
+ * Copyright © 2021 - 2022 Swiss National Data and Service Center for the Humanities and/or DaSCH Service Platform contributors.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.knora.webapi.store.triplestore.upgrade.plugins
+
+import com.typesafe.scalalogging.Logger
+import org.knora.webapi.feature.FeatureFactoryConfig
+import org.knora.webapi.messages.OntologyConstants
+import org.knora.webapi.messages.util.rdf._
+import org.knora.webapi.store.triplestore.upgrade.UpgradePlugin
+
+/**
+ * Transforms a repository for Knora PR 2079.
+ * Adds missing datatype ^^<http://www.w3.org/2001/XMLSchema#anyURI> and/or value to valueHasUri
+ */
+class UpgradePluginPR2079(featureFactoryConfig: FeatureFactoryConfig, log: Logger) extends UpgradePlugin {
+  private val nodeFactory: RdfNodeFactory = RdfFeatureFactory.getRdfNodeFactory(featureFactoryConfig)
+
+  override def transform(model: RdfModel): Unit = {
+    val statementsToRemove: collection.mutable.Set[Statement] = collection.mutable.Set.empty
+    val statementsToAdd: collection.mutable.Set[Statement]    = collection.mutable.Set.empty
+
+    for (statement: Statement <- model) {
+      if (statement.pred.iri == OntologyConstants.KnoraBase.ValueHasUri) {
+        statement.obj match {
+          case literal: DatatypeLiteral =>
+            if (literal.datatype != OntologyConstants.Xsd.Uri) {
+              statementsToRemove += statement
+
+              statementsToAdd += nodeFactory.makeStatement(
+                subj = statement.subj,
+                pred = statement.pred,
+                obj = nodeFactory.makeDatatypeLiteral(literal.value, OntologyConstants.Xsd.Uri),
+                context = statement.context
+              )
+
+              log.info(
+                s"Transformed valueHasIri: $literal to ${nodeFactory
+                  .makeDatatypeLiteral(literal.value, OntologyConstants.Xsd.Uri)}."
+              )
+            }
+
+          case node: IriNode =>
+            statementsToRemove += statement
+
+            statementsToAdd += nodeFactory.makeStatement(
+              subj = statement.subj,
+              pred = statement.pred,
+              obj = nodeFactory.makeDatatypeLiteral(node.iri, OntologyConstants.Xsd.Uri),
+              context = statement.context
+            )
+
+            log.info(
+              s"Transformed valueHasIri $node to ${nodeFactory.makeDatatypeLiteral(node.iri, OntologyConstants.Xsd.Uri)}."
+            )
+
+          case _ => ()
+        }
+      }
+    }
+
+    model.removeStatements(statementsToRemove.toSet)
+    model.addStatements(statementsToAdd.toSet)
+  }
+}

webapi/src/test/scala/org/knora/webapi/store/triplestore/upgrade/plugins/UpgradePluginPR2079Spec.scala

@@ -0,0 +1,84 @@
+/*
+ * Copyright © 2021 - 2022 Swiss National Data and Service Center for the Humanities and/or DaSCH Service Platform contributors.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.knora.webapi.store.triplestore.upgrade.plugins
+
+import com.typesafe.scalalogging.LazyLogging
+import dsp.errors.AssertionException
+import org.knora.webapi.messages.OntologyConstants
+import org.knora.webapi.messages.util.rdf._
+
+class UpgradePluginPR2079Spec extends UpgradePluginSpec with LazyLogging {
+  private val nodeFactory: RdfNodeFactory = RdfFeatureFactory.getRdfNodeFactory(defaultFeatureFactoryConfig)
+
+  "Upgrade plugin PR2079" should {
+    "fix the missing datatype of valueHasUri" in {
+      // Parse the input file.
+      val model: RdfModel = trigFileToModel("../test_data/upgrade/pr2079.trig")
+
+      // Use the plugin to transform the input.
+      val plugin = new UpgradePluginPR2079(defaultFeatureFactoryConfig, log)
+      plugin.transform(model)
+
+      // Check that the datatype was fixed.
+      val subj = nodeFactory.makeIriNode("http://rdfh.ch/0103/fN89IUgvSSyMxJ7XWssP9w/values/Rl2rfjDlRBWeuRr-EgIgCw")
+      val pred = nodeFactory.makeIriNode(OntologyConstants.KnoraBase.ValueHasUri)
+
+      model
+        .find(
+          subj = Some(subj),
+          pred = Some(pred),
+          obj = None
+        )
+        .toSet
+        .headOption match {
+        case Some(statement: Statement) =>
+          statement.obj match {
+            case datatypeLiteral: DatatypeLiteral =>
+              assert(datatypeLiteral.datatype == OntologyConstants.Xsd.Uri)
+
+            case other =>
+              throw AssertionException(s"Unexpected object for $pred: $other")
+          }
+
+        case None => throw AssertionException(s"No statement found with subject $subj and predicate $pred")
+      }
+    }
+
+    "fix value valueHasUri etered ad node w/o datatype" in {
+      // Parse the input file.
+      val model: RdfModel = trigFileToModel("../test_data/upgrade/pr2079.trig")
+
+      // Use the plugin to transform the input.
+      val plugin = new UpgradePluginPR2079(defaultFeatureFactoryConfig, log)
+      plugin.transform(model)
+
+      // Check that the value amd datatype was fixed.
+      val subj =
+        nodeFactory.makeIriNode("http://rdfh.ch/0103/5LE8P57nROClWUxEPJhiug/values/fEbt5NzaSe6GnCqKoF4Nhg/standoff/2")
+      val pred = nodeFactory.makeIriNode(OntologyConstants.KnoraBase.ValueHasUri)
+
+      model
+        .find(
+          subj = Some(subj),
+          pred = Some(pred),
+          obj = None
+        )
+        .toSet
+        .headOption match {
+        case Some(statement: Statement) =>
+          statement.obj match {
+            case datatypeLiteral: DatatypeLiteral =>
+              assert(datatypeLiteral.datatype == OntologyConstants.Xsd.Uri)
+
+            case other =>
+              throw AssertionException(s"Unexpected object for $pred: $other")
+          }
+
+        case None => throw AssertionException(s"No statement found with subject $subj and predicate $pred")
+      }
+    }
+  }
+}