Search in sources :

Example 6 with PeriodicQueryResultStorage

use of org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage in project incubator-rya by apache.

the class PeriodicBindingSetExporterFactory method build.

@Override
public Optional<IncrementalResultExporter> build(Context context) throws IncrementalExporterFactoryException, ConfigurationException {
    checkNotNull(context);
    // Wrap the context's parameters for parsing.
    final RyaExportParameters params = new RyaExportParameters(context.getObserverConfiguration().toMap());
    if (params.getUsePeriodicBindingSetExporter()) {
        // Setup Zookeeper connection info.
        final String accumuloInstance = params.getAccumuloInstanceName().get();
        final String zookeeperServers = params.getZookeeperServers().get().replaceAll(";", ",");
        final Instance inst = new ZooKeeperInstance(accumuloInstance, zookeeperServers);
        try {
            // Setup Accumulo connection info.
            final String exporterUsername = params.getExporterUsername().get();
            final String exporterPassword = params.getExporterPassword().get();
            final Connector accumuloConn = inst.getConnector(exporterUsername, new PasswordToken(exporterPassword));
            // Setup Rya PCJ Storage.
            final String ryaInstanceName = params.getRyaInstanceName().get();
            final PeriodicQueryResultStorage periodicStorage = new AccumuloPeriodicQueryResultStorage(accumuloConn, ryaInstanceName);
            // Make the exporter.
            final IncrementalBindingSetExporter exporter = new PeriodicBindingSetExporter(periodicStorage);
            return Optional.of(exporter);
        } catch (final AccumuloException | AccumuloSecurityException e) {
            throw new IncrementalExporterFactoryException("Could not initialize the Accumulo connector using the provided configuration.", e);
        }
    } else {
        return Optional.absent();
    }
}
Also used : Connector(org.apache.accumulo.core.client.Connector) AccumuloException(org.apache.accumulo.core.client.AccumuloException) Instance(org.apache.accumulo.core.client.Instance) ZooKeeperInstance(org.apache.accumulo.core.client.ZooKeeperInstance) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) IncrementalBindingSetExporter(org.apache.rya.indexing.pcj.fluo.app.export.IncrementalBindingSetExporter) ZooKeeperInstance(org.apache.accumulo.core.client.ZooKeeperInstance) PasswordToken(org.apache.accumulo.core.client.security.tokens.PasswordToken) AccumuloSecurityException(org.apache.accumulo.core.client.AccumuloSecurityException) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage)

Example 7 with PeriodicQueryResultStorage

use of org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage in project incubator-rya by apache.

the class PeriodicNotificationBinPrunerIT method periodicPrunerTest.

@Test
public void periodicPrunerTest() throws Exception {
    String sparql = // n
    "prefix function: <http://org.apache.rya/function#> " + // n
    "prefix time: <http://www.w3.org/2006/time#> " + // n
    "select ?id (count(?obs) as ?total) where {" + // n
    "Filter(function:periodic(?time, 2, .5, time:hours)) " + // n
    "?obs <uri:hasTime> ?time. " + // n
    "?obs <uri:hasId> ?id } group by ?id";
    FluoClient fluo = new FluoClientImpl(super.getFluoConfiguration());
    // initialize resources and create pcj
    PeriodicQueryResultStorage periodicStorage = new AccumuloPeriodicQueryResultStorage(super.getAccumuloConnector(), getRyaInstanceName());
    CreatePeriodicQuery createPeriodicQuery = new CreatePeriodicQuery(fluo, periodicStorage);
    String queryId = FluoQueryUtils.convertFluoQueryIdToPcjId(createPeriodicQuery.createPeriodicQuery(sparql).getQueryId());
    // create statements to ingest into Fluo
    final ValueFactory vf = new ValueFactoryImpl();
    final DatatypeFactory dtf = DatatypeFactory.newInstance();
    ZonedDateTime time = ZonedDateTime.now();
    long currentTime = time.toInstant().toEpochMilli();
    ZonedDateTime zTime1 = time.minusMinutes(30);
    String time1 = zTime1.format(DateTimeFormatter.ISO_INSTANT);
    ZonedDateTime zTime2 = zTime1.minusMinutes(30);
    String time2 = zTime2.format(DateTimeFormatter.ISO_INSTANT);
    ZonedDateTime zTime3 = zTime2.minusMinutes(30);
    String time3 = zTime3.format(DateTimeFormatter.ISO_INSTANT);
    ZonedDateTime zTime4 = zTime3.minusMinutes(30);
    String time4 = zTime4.format(DateTimeFormatter.ISO_INSTANT);
    final Collection<Statement> statements = Sets.newHashSet(vf.createStatement(vf.createURI("urn:obs_1"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time1))), vf.createStatement(vf.createURI("urn:obs_1"), vf.createURI("uri:hasId"), vf.createLiteral("id_1")), vf.createStatement(vf.createURI("urn:obs_2"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time2))), vf.createStatement(vf.createURI("urn:obs_2"), vf.createURI("uri:hasId"), vf.createLiteral("id_2")), vf.createStatement(vf.createURI("urn:obs_3"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time3))), vf.createStatement(vf.createURI("urn:obs_3"), vf.createURI("uri:hasId"), vf.createLiteral("id_3")), vf.createStatement(vf.createURI("urn:obs_4"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time4))), vf.createStatement(vf.createURI("urn:obs_4"), vf.createURI("uri:hasId"), vf.createLiteral("id_4")), vf.createStatement(vf.createURI("urn:obs_1"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time4))), vf.createStatement(vf.createURI("urn:obs_1"), vf.createURI("uri:hasId"), vf.createLiteral("id_1")), vf.createStatement(vf.createURI("urn:obs_2"), vf.createURI("uri:hasTime"), vf.createLiteral(dtf.newXMLGregorianCalendar(time3))), vf.createStatement(vf.createURI("urn:obs_2"), vf.createURI("uri:hasId"), vf.createLiteral("id_2")));
    // add statements to Fluo
    InsertTriples inserter = new InsertTriples();
    statements.forEach(x -> inserter.insert(fluo, RdfToRyaConversions.convertStatement(x)));
    super.getMiniFluo().waitForObservers();
    // FluoITHelper.printFluoTable(fluo);
    // Create the expected results of the SPARQL query once the PCJ has been
    // computed.
    final Set<BindingSet> expected1 = new HashSet<>();
    final Set<BindingSet> expected2 = new HashSet<>();
    final Set<BindingSet> expected3 = new HashSet<>();
    final Set<BindingSet> expected4 = new HashSet<>();
    long period = 1800000;
    long binId = (currentTime / period) * period;
    long bin1 = binId;
    long bin2 = binId + period;
    long bin3 = binId + 2 * period;
    long bin4 = binId + 3 * period;
    MapBindingSet bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("2", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_1", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin1));
    expected1.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("2", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_2", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin1));
    expected1.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_3", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin1));
    expected1.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_4", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin1));
    expected1.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_1", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin2));
    expected2.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("2", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_2", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin2));
    expected2.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_3", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin2));
    expected2.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_1", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin3));
    expected3.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_2", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin3));
    expected3.add(bs);
    bs = new MapBindingSet();
    bs.addBinding("total", vf.createLiteral("1", XMLSchema.INTEGER));
    bs.addBinding("id", vf.createLiteral("id_1", XMLSchema.STRING));
    bs.addBinding("periodicBinId", vf.createLiteral(bin4));
    expected4.add(bs);
    // make sure that expected and actual results align after ingest
    compareResults(periodicStorage, queryId, bin1, expected1);
    compareResults(periodicStorage, queryId, bin2, expected2);
    compareResults(periodicStorage, queryId, bin3, expected3);
    compareResults(periodicStorage, queryId, bin4, expected4);
    BlockingQueue<NodeBin> bins = new LinkedBlockingQueue<>();
    PeriodicQueryPrunerExecutor pruner = new PeriodicQueryPrunerExecutor(periodicStorage, fluo, 1, bins);
    pruner.start();
    bins.add(new NodeBin(queryId, bin1));
    bins.add(new NodeBin(queryId, bin2));
    bins.add(new NodeBin(queryId, bin3));
    bins.add(new NodeBin(queryId, bin4));
    Thread.sleep(10000);
    compareResults(periodicStorage, queryId, bin1, new HashSet<>());
    compareResults(periodicStorage, queryId, bin2, new HashSet<>());
    compareResults(periodicStorage, queryId, bin3, new HashSet<>());
    compareResults(periodicStorage, queryId, bin4, new HashSet<>());
    compareFluoCounts(fluo, queryId, bin1);
    compareFluoCounts(fluo, queryId, bin2);
    compareFluoCounts(fluo, queryId, bin3);
    compareFluoCounts(fluo, queryId, bin4);
    pruner.stop();
}
Also used : MapBindingSet(org.openrdf.query.impl.MapBindingSet) QueryBindingSet(org.openrdf.query.algebra.evaluation.QueryBindingSet) BindingSet(org.openrdf.query.BindingSet) FluoClient(org.apache.fluo.api.client.FluoClient) FluoClientImpl(org.apache.fluo.core.client.FluoClientImpl) DatatypeFactory(javax.xml.datatype.DatatypeFactory) InsertTriples(org.apache.rya.indexing.pcj.fluo.api.InsertTriples) NodeBin(org.apache.rya.periodic.notification.api.NodeBin) Statement(org.openrdf.model.Statement) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) ValueFactoryImpl(org.openrdf.model.impl.ValueFactoryImpl) ValueFactory(org.openrdf.model.ValueFactory) LinkedBlockingQueue(java.util.concurrent.LinkedBlockingQueue) ZonedDateTime(java.time.ZonedDateTime) MapBindingSet(org.openrdf.query.impl.MapBindingSet) CreatePeriodicQuery(org.apache.rya.indexing.pcj.fluo.api.CreatePeriodicQuery) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage) HashSet(java.util.HashSet) Test(org.junit.Test)

Example 8 with PeriodicQueryResultStorage

use of org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage in project incubator-rya by apache.

the class PeriodicNotificationApplicationFactory method getPeriodicApplication.

/**
 * Create a PeriodicNotificationApplication.
 * @param conf - Configuration object that specifies the parameters needed to create the application
 * @return PeriodicNotificationApplication to periodically poll Rya Fluo for new results
 * @throws PeriodicApplicationException
 */
public static PeriodicNotificationApplication getPeriodicApplication(final PeriodicNotificationApplicationConfiguration conf) throws PeriodicApplicationException {
    final Properties kafkaConsumerProps = getKafkaConsumerProperties(conf);
    final Properties kafkaProducerProps = getKafkaProducerProperties(conf);
    final BlockingQueue<TimestampedNotification> notifications = new LinkedBlockingQueue<>();
    final BlockingQueue<NodeBin> bins = new LinkedBlockingQueue<>();
    final BlockingQueue<BindingSetRecord> bindingSets = new LinkedBlockingQueue<>();
    FluoClient fluo = null;
    try {
        final PeriodicQueryResultStorage storage = getPeriodicQueryResultStorage(conf);
        fluo = FluoClientFactory.getFluoClient(conf.getFluoAppName(), Optional.of(conf.getFluoTableName()), conf);
        final NotificationCoordinatorExecutor coordinator = getCoordinator(conf.getCoordinatorThreads(), notifications);
        addRegisteredNotices(coordinator, fluo.newSnapshot());
        final KafkaExporterExecutor exporter = getExporter(conf.getExporterThreads(), kafkaProducerProps, bindingSets);
        final PeriodicQueryPrunerExecutor pruner = getPruner(storage, fluo, conf.getPrunerThreads(), bins);
        final NotificationProcessorExecutor processor = getProcessor(storage, notifications, bins, bindingSets, conf.getProcessorThreads());
        final KafkaNotificationProvider provider = getProvider(conf.getProducerThreads(), conf.getNotificationTopic(), coordinator, kafkaConsumerProps);
        return PeriodicNotificationApplication.builder().setCoordinator(coordinator).setProvider(provider).setExporter(exporter).setProcessor(processor).setPruner(pruner).build();
    } catch (AccumuloException | AccumuloSecurityException e) {
        throw new PeriodicApplicationException(e.getMessage());
    }
}
Also used : KafkaNotificationProvider(org.apache.rya.periodic.notification.registration.kafka.KafkaNotificationProvider) AccumuloException(org.apache.accumulo.core.client.AccumuloException) FluoClient(org.apache.fluo.api.client.FluoClient) NodeBin(org.apache.rya.periodic.notification.api.NodeBin) NotificationCoordinatorExecutor(org.apache.rya.periodic.notification.api.NotificationCoordinatorExecutor) PeriodicNotificationCoordinatorExecutor(org.apache.rya.periodic.notification.coordinator.PeriodicNotificationCoordinatorExecutor) BindingSetRecord(org.apache.rya.periodic.notification.api.BindingSetRecord) NotificationProcessorExecutor(org.apache.rya.periodic.notification.processor.NotificationProcessorExecutor) TimestampedNotification(org.apache.rya.periodic.notification.notification.TimestampedNotification) Properties(java.util.Properties) LinkedBlockingQueue(java.util.concurrent.LinkedBlockingQueue) KafkaExporterExecutor(org.apache.rya.periodic.notification.exporter.KafkaExporterExecutor) PeriodicQueryPrunerExecutor(org.apache.rya.periodic.notification.pruner.PeriodicQueryPrunerExecutor) AccumuloSecurityException(org.apache.accumulo.core.client.AccumuloSecurityException) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage)

Example 9 with PeriodicQueryResultStorage

use of org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage in project incubator-rya by apache.

the class QueryIT method runTest.

public void runTest(final String sparql, final Collection<Statement> statements, final Collection<BindingSet> expectedResults, final ExportStrategy strategy) throws Exception {
    requireNonNull(sparql);
    requireNonNull(statements);
    requireNonNull(expectedResults);
    // Register the PCJ with Rya.
    final Connector accumuloConn = super.getAccumuloConnector();
    final RyaClient ryaClient = AccumuloRyaClientFactory.build(createConnectionDetails(), accumuloConn);
    switch(strategy) {
        case RYA:
            ryaClient.getCreatePCJ().createPCJ(getRyaInstanceName(), sparql);
            addStatementsAndWait(statements);
            // Fetch the value that is stored within the PCJ table.
            try (final PrecomputedJoinStorage pcjStorage = new AccumuloPcjStorage(accumuloConn, getRyaInstanceName())) {
                final String pcjId = pcjStorage.listPcjs().get(0);
                final Set<BindingSet> results = Sets.newHashSet(pcjStorage.listResults(pcjId));
                // Ensure the result of the query matches the expected result.
                assertEquals(expectedResults, results);
            }
            break;
        case PERIODIC:
            final PeriodicQueryResultStorage periodicStorage = new AccumuloPeriodicQueryResultStorage(accumuloConn, getRyaInstanceName());
            final String periodicId = periodicStorage.createPeriodicQuery(sparql);
            try (FluoClient fluo = new FluoClientImpl(super.getFluoConfiguration())) {
                new CreateFluoPcj().createPcj(periodicId, sparql, Sets.newHashSet(ExportStrategy.PERIODIC), fluo);
            }
            addStatementsAndWait(statements);
            final Set<BindingSet> results = Sets.newHashSet();
            try (CloseableIterator<BindingSet> resultIter = periodicStorage.listResults(periodicId, Optional.empty())) {
                while (resultIter.hasNext()) {
                    results.add(resultIter.next());
                }
            }
            assertEquals(expectedResults, results);
            break;
        default:
            throw new RuntimeException("Invalid export option");
    }
}
Also used : Connector(org.apache.accumulo.core.client.Connector) MapBindingSet(org.openrdf.query.impl.MapBindingSet) BindingSet(org.openrdf.query.BindingSet) FluoClient(org.apache.fluo.api.client.FluoClient) FluoClientImpl(org.apache.fluo.core.client.FluoClientImpl) AccumuloPcjStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPcjStorage) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) CreateFluoPcj(org.apache.rya.indexing.pcj.fluo.api.CreateFluoPcj) RyaClient(org.apache.rya.api.client.RyaClient) PrecomputedJoinStorage(org.apache.rya.indexing.pcj.storage.PrecomputedJoinStorage) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage)

Example 10 with PeriodicQueryResultStorage

use of org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage in project incubator-rya by apache.

the class AccumuloCreatePeriodicPCJ method updateFluoAppAndRegisterWithKafka.

private String updateFluoAppAndRegisterWithKafka(final String ryaInstance, final String fluoAppName, String sparql, String periodicTopic, String bootStrapServers) throws RepositoryException, MalformedQueryException, SailException, QueryEvaluationException, PcjException, RyaDAOException, UnsupportedQueryException, PeriodicQueryCreationException {
    requireNonNull(sparql);
    requireNonNull(periodicTopic);
    requireNonNull(bootStrapServers);
    final PeriodicQueryResultStorage periodicStorage = new AccumuloPeriodicQueryResultStorage(getConnector(), ryaInstance);
    // Connect to the Fluo application that is updating this instance's PCJs.
    final AccumuloConnectionDetails cd = super.getAccumuloConnectionDetails();
    try (final FluoClient fluoClient = new FluoClientFactory().connect(cd.getUsername(), new String(cd.getUserPass()), cd.getInstanceName(), cd.getZookeepers(), fluoAppName)) {
        // Initialize the PCJ within the Fluo application.
        final CreatePeriodicQuery periodicPcj = new CreatePeriodicQuery(fluoClient, periodicStorage);
        PeriodicNotificationClient periodicClient = new KafkaNotificationRegistrationClient(periodicTopic, createProducer(bootStrapServers));
        return periodicPcj.withRyaIntegration(sparql, periodicClient, getConnector(), ryaInstance).getQueryId();
    }
}
Also used : FluoClient(org.apache.fluo.api.client.FluoClient) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicNotificationClient(org.apache.rya.periodic.notification.api.PeriodicNotificationClient) CreatePeriodicQuery(org.apache.rya.indexing.pcj.fluo.api.CreatePeriodicQuery) AccumuloPeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage) PeriodicQueryResultStorage(org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage) KafkaNotificationRegistrationClient(org.apache.rya.periodic.notification.registration.KafkaNotificationRegistrationClient)

Aggregations

PeriodicQueryResultStorage (org.apache.rya.indexing.pcj.storage.PeriodicQueryResultStorage)11 AccumuloPeriodicQueryResultStorage (org.apache.rya.indexing.pcj.storage.accumulo.AccumuloPeriodicQueryResultStorage)11 FluoClient (org.apache.fluo.api.client.FluoClient)9 CreatePeriodicQuery (org.apache.rya.indexing.pcj.fluo.api.CreatePeriodicQuery)6 BindingSet (org.openrdf.query.BindingSet)6 HashSet (java.util.HashSet)5 Connector (org.apache.accumulo.core.client.Connector)5 Test (org.junit.Test)5 QueryBindingSet (org.openrdf.query.algebra.evaluation.QueryBindingSet)5 ZonedDateTime (java.time.ZonedDateTime)4 DatatypeFactory (javax.xml.datatype.DatatypeFactory)4 Statement (org.openrdf.model.Statement)4 ValueFactory (org.openrdf.model.ValueFactory)4 ValueFactoryImpl (org.openrdf.model.impl.ValueFactoryImpl)4 ArrayList (java.util.ArrayList)3 LinkedBlockingQueue (java.util.concurrent.LinkedBlockingQueue)3 KafkaConsumer (org.apache.kafka.clients.consumer.KafkaConsumer)3 StringDeserializer (org.apache.kafka.common.serialization.StringDeserializer)3 NodeBin (org.apache.rya.periodic.notification.api.NodeBin)3 BindingSetSerDe (org.apache.rya.periodic.notification.serialization.BindingSetSerDe)3