fixes

Author: hmottestad

core/sail/base/src/main/java/org/eclipse/rdf4j/sail/base/SketchBasedJoinEstimator.java

@@ -1934,8 +1934,6 @@ public void startBackgroundRefresh(int stalenessThreshold) {
 					}
 					continue;
 				}
-//				Staleness staleness = staleness();
-//				System.out.println(staleness.toString());
 
 				try {
 					rebuild();
@@ -3259,11 +3257,6 @@ private static double enforceCardinalityLowerBound(ArrayOfDoublesSketch sketch,
 		if (sketch == null || !Double.isFinite(cardinality)) {
 			return cardinality;
 		}
-//		if(cardinality < sketch.getEstimate()){
-//			System.out.println("Cardinality " + cardinality + " is lower than sketch estimate " + sketch.getEstimate());
-//		}else if(cardinality > sketch.getEstimate()) {
-//			System.out.println("Cardinality " + cardinality + " is higher than sketch estimate " + sketch.getEstimate());
-//		}
 
 		// Cardinality cannot be lower than the estimated number of distinct join bindings.
 //		return Math.max(cardinality, sketch.getEstimate());
@@ -3547,10 +3540,9 @@ private static final class State {
 
 		State(int k, int subjectBuckets, int predicateBuckets, int objectBuckets, int contextBuckets,
 				boolean contextPairSketchesEnabled) {
-			System.out.println("Initializing state: k=" + k + ", subjectBuckets=" + subjectBuckets
-					+ ", predicateBuckets="
-					+ predicateBuckets + ", objectBuckets=" + objectBuckets + ", contextBuckets=" + contextBuckets
-					+ ", contextPairSketchesEnabled=" + contextPairSketchesEnabled);
+			logger.info(
+					"Initializing state: k={}, subjectBuckets={}, predicateBuckets={}, objectBuckets={}, contextBuckets={}, contextPairSketchesEnabled={}",
+					k, subjectBuckets, predicateBuckets, objectBuckets, contextBuckets, contextPairSketchesEnabled);
 			this.k = k;
 			this.subjectBuckets = subjectBuckets;
 			this.predicateBuckets = predicateBuckets;
@@ -8207,7 +8199,13 @@ private long samplingInterval() {
 	 */
 	public boolean persistIfDirty() {
 		flushPendingIncremental();
-		SketchEstimatorPersistenceStore store = persistenceStore;
+		SketchEstimatorPersistenceStore store;
+		try {
+			store = ensurePersistenceStore();
+		} catch (Throwable t) {
+			logger.warn("Failed to open join estimator persistence store at {}", persistenceFile, t);
+			return false;
+		}
 		if (!persistenceEnabled || persistenceFile == null || store == null) {
 			return false;
 		}
@@ -8935,7 +8933,7 @@ private boolean appendDirtySketchIfResident(int entryId, SketchAddress address,
 
 	private void appendNativeSketchPayload(State state, int entryId, SketchAddress address, byte[] payload)
 			throws IOException {
-		SketchEstimatorPersistenceStore store = persistenceStore;
+		SketchEstimatorPersistenceStore store = ensurePersistenceStore();
 		if (store != null) {
 			byte slot = slotByte(slotOf(state));
 			int slotBytes = shouldPersistCompactly(address) ? payload.length : state.maxSketchBytes;
@@ -8960,6 +8958,25 @@ private void appendNativeSketchPayload(State state, int entryId, SketchAddress a
 		throw new IOException("No directory-backed sketch store configured");
 	}
 
+	private SketchEstimatorPersistenceStore ensurePersistenceStore() throws IOException {
+		if (!persistenceEnabled || persistenceFile == null) {
+			return null;
+		}
+		SketchEstimatorPersistenceStore store = persistenceStore;
+		if (store != null) {
+			return store;
+		}
+		synchronized (persistLock) {
+			store = persistenceStore;
+			if (store == null && persistenceEnabled && persistenceFile != null) {
+				logger.info("RdfJoinEstimator: Opening default directory-backed sketch store at {}", persistenceFile);
+				store = SketchEstimatorPersistenceStore.open(persistenceFile, logger, sketchFileChunks());
+				persistenceStore = store;
+			}
+			return store;
+		}
+	}
+
 	private void clearDirtyMappedSketch(int entryId, byte slot) {
 		synchronized (sketchCacheLock) {
 			cacheDirectory.clearDirty(entryId, slot);

core/sail/lmdb/src/test/java/org/eclipse/rdf4j/sail/lmdb/LmdbSailStoreEstimatorPersistenceTest.java

@@ -159,6 +159,35 @@ void closeWritesEstimatorSnapshotAndReopensLazy(@TempDir File dataDir) throws Ex
 		}
 	}
 
+	@Test
+	void persistIfDirtyReopensDefaultDirectoryBackedSketchStore(@TempDir File dataDir) throws Exception {
+		var vf = SimpleValueFactory.getInstance();
+		var s = vf.createIRI("urn:default-store:s");
+		var p = vf.createIRI("urn:default-store:p");
+		var o = vf.createIRI("urn:default-store:o");
+
+		LmdbStore store = new LmdbStore(dataDir, new LmdbStoreConfig("spoc"));
+		store.init();
+		try {
+			SketchBasedJoinEstimator estimator = store.getBackingStore().getSketchBasedJoinEstimator();
+			estimator.stop();
+			estimator.rebuild();
+
+			AutoCloseable persistenceStore = (AutoCloseable) objectField(estimator, "persistenceStore");
+			persistenceStore.close();
+			setObjectField(estimator, "persistenceStore", null);
+
+			estimator.addStatement(vf.createStatement(s, p, o));
+
+			assertTrue(estimator.persistIfDirty(),
+					"Estimator persistence should reopen the default directory-backed sketch store");
+		} finally {
+			store.shutDown();
+		}
+
+		assertTrue(estimatorMetadata(dataDir).isFile(), "Expected estimator metadata in the default store directory");
+	}
+
 	@Test
 	void closeWritesFilterSelectivitySidecarAndReloadsItAfterRestart(@TempDir File dataDir) throws Exception {
 		Filter learnedFilter = firstFilter(LEARNED_FILTER_QUERY);
@@ -662,6 +691,12 @@ private static Object objectField(Object target, String name) throws Exception {
 		return field.get(target);
 	}
 
+	private static void setObjectField(Object target, String name, Object value) throws Exception {
+		Field field = target.getClass().getDeclaredField(name);
+		field.setAccessible(true);
+		field.set(target, value);
+	}
+
 	private static int pendingIncrementalStatementQueueSize(SketchBasedJoinEstimator estimator) throws Exception {
 		Object[] queues = (Object[]) objectField(estimator, "incrementalStatementQueues");
 		int pending = 0;

core/sail/lmdb/src/test/java/org/eclipse/rdf4j/sail/lmdb/benchmark/DatagovLoadIsolationBenchmark.java

@@ -63,7 +63,7 @@ public class DatagovLoadIsolationBenchmark {
 
 	private static final String DATA_FILE = "benchmarkFiles/datagovbe-valid.ttl.gz";
 
-	@Param({ "NONE", "READ_COMMITTED" })
+	@Param({ "NONE", "READ_COMMITTED", "SNAPSHOT_READ", "SNAPSHOT", "SERIALIZABLE" })
 	public IsolationLevels isolationLevel;
 
 	@Param({ "256" })