GH-5227: fix binding assigner optimizer in FedX (#5228)

eclipse-rdf4j · Jan 24, 2025 · 98d8944 · 98d8944
2 parents 230a437 + 99d7ba3
commit 98d8944
Show file tree

Hide file tree

Showing 2 changed files with 74 additions and 2 deletions.
diff --git a/...deration/src/main/java/org/eclipse/rdf4j/federated/evaluation/FederationEvalStrategy.java b/...deration/src/main/java/org/eclipse/rdf4j/federated/evaluation/FederationEvalStrategy.java
@@ -107,6 +107,7 @@
 import org.eclipse.rdf4j.query.algebra.ValueExpr;
 import org.eclipse.rdf4j.query.algebra.Var;
 import org.eclipse.rdf4j.query.algebra.evaluation.QueryEvaluationStep;
+import org.eclipse.rdf4j.query.algebra.evaluation.QueryOptimizer;
 import org.eclipse.rdf4j.query.algebra.evaluation.QueryValueEvaluationStep;
 import org.eclipse.rdf4j.query.algebra.evaluation.ValueExprEvaluationException;
 import org.eclipse.rdf4j.query.algebra.evaluation.federation.FederatedService;
@@ -118,6 +119,7 @@
 import org.eclipse.rdf4j.query.algebra.evaluation.iterator.HashJoinIteration;
 import org.eclipse.rdf4j.query.algebra.evaluation.optimizer.ConstantOptimizer;
 import org.eclipse.rdf4j.query.algebra.evaluation.optimizer.DisjunctiveConstraintOptimizer;
+import org.eclipse.rdf4j.query.algebra.evaluation.optimizer.StandardQueryOptimizerPipeline;
 import org.eclipse.rdf4j.query.algebra.evaluation.util.QueryEvaluationUtil;
 import org.eclipse.rdf4j.query.algebra.helpers.TupleExprs;
 import org.eclipse.rdf4j.query.algebra.helpers.collectors.VarNameCollector;
@@ -147,6 +149,14 @@ public abstract class FederationEvalStrategy extends StrictEvaluationStrategy {
 
 	protected FederationContext federationContext;
 
+	/**
+	 * List of standard {@link QueryOptimizer}s applicable to federation
+	 */
+	private static final List<QueryOptimizer> standardOptimizers = List.of(
+			StandardQueryOptimizerPipeline.BINDING_ASSIGNER,
+			StandardQueryOptimizerPipeline.BINDING_SET_ASSIGNMENT_INLINER,
+			StandardQueryOptimizerPipeline.DISJUNCTIVE_CONSTRAINT_OPTIMIZER);
+
 	public FederationEvalStrategy(FederationContext federationContext) {
 		super(new org.eclipse.rdf4j.query.algebra.evaluation.TripleSource() {
 
@@ -209,9 +219,11 @@ public TupleExpr optimize(TupleExpr expr, EvaluationStatistics evaluationStatist
 		}
 
 		/* original RDF4J optimizers */
-		new ConstantOptimizer(this).optimize(query, dataset, bindings); // maybe remove this optimizer later
+		for (QueryOptimizer optimizer : standardOptimizers) {
+			optimizer.optimize(query, dataset, bindings);
+		}
 
-		new DisjunctiveConstraintOptimizer().optimize(query, dataset, bindings);
+		new ConstantOptimizer(this).optimize(query, dataset, bindings); // maybe remove this optimizer later
 
 		/*
 		 * TODO add some generic optimizers: - FILTER ?s=1 && ?s=2 => EmptyResult - Remove variables that are not

diff --git a/...tion/src/test/java/org/eclipse/rdf4j/federated/evaluation/FederationEvalStrategyTest.java b/...tion/src/test/java/org/eclipse/rdf4j/federated/evaluation/FederationEvalStrategyTest.java
@@ -10,9 +10,23 @@
  *******************************************************************************/
 package org.eclipse.rdf4j.federated.evaluation;
 
+import java.util.Arrays;
 import java.util.List;
+import java.util.Set;
+import java.util.stream.Collectors;
 
 import org.eclipse.rdf4j.federated.SPARQLBaseTest;
+import org.eclipse.rdf4j.federated.cache.SourceSelectionCache;
+import org.eclipse.rdf4j.federated.cache.SourceSelectionCache.StatementSourceAssurance;
+import org.eclipse.rdf4j.federated.endpoint.Endpoint;
+import org.eclipse.rdf4j.federated.structures.SubQuery;
+import org.eclipse.rdf4j.model.util.Values;
+import org.eclipse.rdf4j.model.vocabulary.FOAF;
+import org.eclipse.rdf4j.model.vocabulary.OWL;
+import org.eclipse.rdf4j.model.vocabulary.RDF;
+import org.eclipse.rdf4j.query.TupleQuery;
+import org.eclipse.rdf4j.repository.Repository;
+import org.eclipse.rdf4j.repository.RepositoryConnection;
 import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.Test;
 
@@ -46,4 +60,50 @@ public void testOptimize_SingleMember_Service() throws Exception {
 
 		Assertions.assertTrue(queryPlan.startsWith("QueryRoot"));
 	}
+
+	@Test
+	public void testSourceSelectionCache_setBindings() throws Exception {
+
+		var bob = Values.iri("http://example.com/bob");
+
+		List<Endpoint> endpoints = prepareTest(
+				Arrays.asList("/tests/basic/data_emptyStore.ttl", "/tests/basic/data_emptyStore.ttl"));
+
+		Repository repo1 = getRepository(1);
+		Repository repo2 = getRepository(2);
+
+		String repo1Id = endpoints.get(0).getId();
+
+		try (RepositoryConnection con = repo1.getConnection()) {
+			con.add(bob, RDF.TYPE, FOAF.PERSON);
+		}
+
+		try (RepositoryConnection con = repo2.getConnection()) {
+			con.add(FOAF.PERSON, RDF.TYPE, OWL.CLASS);
+		}
+
+		Repository fedxRepo = fedxRule.getRepository();
+
+		fedxRule.enableDebug();
+
+		try (var conn = fedxRepo.getConnection()) {
+
+			TupleQuery tq = conn.prepareTupleQuery("SELECT * WHERE { ?s a ?type }");
+			tq.setBinding("s", bob);
+
+			try (var tqr = tq.evaluate()) {
+				// just consume the result
+				Assertions.assertEquals(Set.of(FOAF.PERSON),
+						tqr.stream().map(bs -> bs.getValue("type")).collect(Collectors.toSet()));
+			}
+		}
+
+		SourceSelectionCache cache = federationContext().getSourceSelectionCache();
+
+		var assurance = cache.getAssurance(new SubQuery(bob, RDF.TYPE, null),
+				federationContext().getEndpointManager().getEndpoint(repo1Id));
+
+		// we expect that the source selection cache can assure statements
+		Assertions.assertEquals(StatementSourceAssurance.HAS_REMOTE_STATEMENTS, assurance);
+	}
 }