GH-4978 Index geosparql points using geo field (#4979)

eclipse-rdf4j · May 31, 2024 · 192b4ea · 192b4ea
2 parents 3d218d9 + 2082613
commit 192b4ea
Show file tree

Hide file tree

Showing 2 changed files with 100 additions and 1 deletion.
diff --git a/core/sail/lucene/src/main/java/org/eclipse/rdf4j/sail/lucene/impl/LuceneDocument.java b/core/sail/lucene/src/main/java/org/eclipse/rdf4j/sail/lucene/impl/LuceneDocument.java
@@ -174,7 +174,12 @@ private void indexShape(Object shape, String field) {
 					doc.add(f);
 				}
 			} else if (shape instanceof double[]) { // WKT:POINT
-				double point[] = (double[]) shape;
+				double[] point = (double[]) shape;
+
+				for (Field f : LatLonShape.createIndexableFields(GEO_FIELD_PREFIX + field, point[1],
+						point[0])) {
+					doc.add(f);
+				}
 				doc.add(new LatLonPoint(POINT_FIELD_PREFIX + field, point[1], point[0]));
 			} else if (shape instanceof Rectangle) { // WKT:ENVELOPE / RECTANGLE
 				Rectangle box = (Rectangle) shape;

diff --git a/core/sail/lucene/src/test/java/org/eclipse/rdf4j/sail/lucene/impl/LuceneIndexTest.java b/core/sail/lucene/src/test/java/org/eclipse/rdf4j/sail/lucene/impl/LuceneIndexTest.java
@@ -19,7 +19,9 @@
 
 import java.io.IOException;
 import java.util.HashSet;
+import java.util.Iterator;
 import java.util.Set;
+import java.util.function.Function;
 
 import org.apache.lucene.analysis.en.EnglishAnalyzer;
 import org.apache.lucene.analysis.standard.StandardAnalyzer;
@@ -39,12 +41,23 @@
 import org.eclipse.rdf4j.common.concurrent.locks.Properties;
 import org.eclipse.rdf4j.model.IRI;
 import org.eclipse.rdf4j.model.Literal;
+import org.eclipse.rdf4j.model.Model;
 import org.eclipse.rdf4j.model.Resource;
 import org.eclipse.rdf4j.model.Statement;
+import org.eclipse.rdf4j.model.Value;
 import org.eclipse.rdf4j.model.ValueFactory;
+import org.eclipse.rdf4j.model.base.CoreDatatype;
 import org.eclipse.rdf4j.model.impl.SimpleValueFactory;
+import org.eclipse.rdf4j.model.impl.TreeModel;
+import org.eclipse.rdf4j.model.vocabulary.GEO;
+import org.eclipse.rdf4j.model.vocabulary.GEOF;
+import org.eclipse.rdf4j.query.BindingSet;
+import org.eclipse.rdf4j.query.TupleQuery;
+import org.eclipse.rdf4j.query.TupleQueryResult;
 import org.eclipse.rdf4j.repository.sail.SailRepository;
 import org.eclipse.rdf4j.repository.sail.SailRepositoryConnection;
+import org.eclipse.rdf4j.repository.util.Repositories;
+import org.eclipse.rdf4j.sail.evaluation.TupleFunctionEvaluationMode;
 import org.eclipse.rdf4j.sail.lucene.LuceneSail;
 import org.eclipse.rdf4j.sail.lucene.SearchFields;
 import org.eclipse.rdf4j.sail.memory.MemoryStore;
@@ -478,4 +491,85 @@ private void assertNoStatement(Statement statement, Document document) {
 	 * for(String notFound : toFind) { assertEquals("Was the expected text value '" + notFound + "' found?", true,
 	 * false); } }
 	 */
+
+	@Test
+	public void geoSparqlQueryTest() {
+		final String prefix = "http://www.example.org/#";
+		final String prefixes = "PREFIX ex: <" + prefix + ">\n"
+				+ "PREFIX geof: <" + GEOF.NAMESPACE + ">\n"
+				+ "PREFIX geo: <" + CoreDatatype.GEO.NAMESPACE + ">\n"
+				+ "PREFIX uom: <http://www.opengis.net/def/uom/OGC/1.0/>\n";
+		Model data = new TreeModel();
+
+		IRI cp = vf.createIRI(prefix + "cp");
+		IRI bm = vf.createIRI(prefix + "bm");
+		IRI nkv = vf.createIRI(prefix + "nkv");
+
+		data.add(cp, GEO.AS_WKT, vf.createLiteral("Point(4.38436 45.44917)", CoreDatatype.GEO.WKT_LITERAL));
+		data.add(bm, GEO.AS_WKT, vf.createLiteral("Point(4.38311 45.45423)", CoreDatatype.GEO.WKT_LITERAL));
+		data.add(nkv, GEO.AS_WKT, vf.createLiteral("Point(4.87306 45.77903)", CoreDatatype.GEO.WKT_LITERAL));
+		data.add(vf.createIRI(prefix + "arp"), GEO.AS_WKT,
+				vf.createLiteral("Point(2.89271 42.69848)", CoreDatatype.GEO.WKT_LITERAL));
+
+		String polyVill = "POLYGON((4.864712 45.784405, 4.883165 45.787756, 4.889946 45.785781, 4.904881 45.767403, 4.900761 45.765487, 4.872093 45.770995, 4.86454 45.770457, 4.858789 45.770277, 4.859905 45.784644, 4.864712 45.784405))";
+		String polySain = "POLYGON((4.380627 45.463983, 4.400539 45.462177, 4.428349 45.436286, 4.399509 45.411346, 4.374447 45.426528, 4.370499 45.450618, 4.380627 45.463983))";
+
+		SailRepository m1 = new SailRepository(new MemoryStore());
+		LuceneSail lc = new LuceneSail();
+		lc.setBaseSail(new MemoryStore());
+		lc.setParameter(LuceneSail.WKT_FIELDS, GEO.AS_WKT.toString());
+		lc.setLuceneIndex(index);
+		lc.setEvaluationMode(TupleFunctionEvaluationMode.NATIVE);
+		SailRepository m2 = new SailRepository(lc);
+
+		// add test data
+		Repositories.consume(m1, conn -> conn.add(data));
+		Repositories.consume(m2, conn -> conn.add(data));
+
+		lc.reindex();
+
+		Function<TupleQueryResult, Set<Value>> toval = (res) -> {
+			Set<Value> list = new HashSet<>();
+			while (res.hasNext()) {
+				BindingSet next = res.next();
+				list.add(next.getValue("v"));
+			}
+			return list;
+		};
+
+		// test queries
+
+		String q0 = prefixes
+				+ "SELECT * {\n"
+				+ "  ?v geo:asWKT ?loc .\n"
+				+ "  FILTER(geof:distance(\"Point(4.386914 45.440637)\"^^geo:wktLiteral, ?loc, uom:metre) < 10000) \n"
+				+ "}\n";
+		Set<Value> q0ex = Set.of(bm, cp);
+
+		String q1 = prefixes
+				+ "SELECT * {\n"
+				+ "  ?v geo:asWKT ?loc .\n"
+				+ "  FILTER(geof:ehContains(\"" + polySain + "\"^^geo:wktLiteral, ?loc)) \n"
+				+ "}\n";
+		Set<Value> q1ex = Set.of(bm, cp);
+
+		String q2 = prefixes
+				+ "SELECT * {\n"
+				+ "  ?v geo:asWKT ?loc .\n"
+				+ "  FILTER(geof:ehContains(\"" + polyVill + "\"^^geo:wktLiteral, ?loc)) \n"
+				+ "}\n";
+		Set<Value> q2ex = Set.of(nkv);
+
+		Set<Value> nlcq0 = Repositories.tupleQuery(m1, q0, toval);
+		Set<Value> nlcq1 = Repositories.tupleQuery(m1, q1, toval);
+		Set<Value> nlcq2 = Repositories.tupleQuery(m1, q2, toval);
+
+		assertEquals(q0ex, nlcq0);
+		assertEquals(q1ex, nlcq1);
+		assertEquals(q2ex, nlcq2);
+
+		assertEquals(nlcq0, Repositories.tupleQuery(m2, q0, toval));
+		assertEquals(nlcq1, Repositories.tupleQuery(m2, q1, toval));
+		assertEquals(nlcq2, Repositories.tupleQuery(m2, q2, toval));
+	}
 }