@Override
public void populateDAG(DAG dag, Configuration conf) {
KafkaSinglePortInputOperator in =
dag.addOperator("kafkaIn", new KafkaSinglePortInputOperator());
in.setInitialOffset(AbstractKafkaInputOperator.InitialOffset.EARLIEST.name());
LineOutputOperator out = dag.addOperator("fileOut", new LineOutputOperator());
dag.addStream("data", in.outputPort, out.input);
}
@Override
public void populateDAG(DAG dag, Configuration configuration) {
NBTestInputOperator inp =
(NBTestInputOperator) dag.addOperator("vectorize", new NBTestInputOperator());
ClassifierOperator classifier =
(ClassifierOperator) dag.addOperator("classifier", new ClassifierOperator());
ConsoleOutputOperator out =
(ConsoleOutputOperator) dag.addOperator("console", new ConsoleOutputOperator());
dag.addStream("vectorize_classifier", inp.data, classifier.input).setLocality(this.locality);
dag.addStream("classifier_console", classifier.output, out.input).setLocality(this.locality);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
HBaseRowStringGenerator row = dag.addOperator("rand", new HBaseRowStringGenerator());
HBaseCsvMappingPutOperator csvMappingPutOperator =
dag.addOperator("HBaseoper", new HBaseCsvMappingPutOperator());
csvMappingPutOperator.getStore().setTableName("table1");
csvMappingPutOperator.getStore().setZookeeperQuorum("127.0.0.1");
csvMappingPutOperator.getStore().setZookeeperClientPort(2181);
csvMappingPutOperator.setMappingString("colfam0.street,colfam0.city,colfam0.state,row");
dag.addStream("hbasestream", row.outputPort, csvMappingPutOperator.input).setLocality(locality);
}
/**
* Test AbstractKafkaSinglePortInputOperator (i.e. an input adapter for Kafka, aka consumer). This
* module receives data from an outside test generator through Kafka message bus and feed that
* data into Malhar streaming platform.
*
* <p>[Generate message and send that to Kafka message bus] ==> [Receive that message through
* Kafka input adapter(i.e. consumer) and send using emitTuples() interface on output port during
* onMessage call]
*
* @throws Exception
*/
public void testKafkaInputOperator(
int sleepTime, final int totalCount, KafkaConsumer consumer, boolean isValid)
throws Exception {
// initial the latch for this test
latch = new CountDownLatch(1);
// Start producer
KafkaTestProducer p = new KafkaTestProducer(TEST_TOPIC);
p.setSendCount(totalCount);
new Thread(p).start();
// Create DAG for testing.
LocalMode lma = LocalMode.newInstance();
DAG dag = lma.getDAG();
// Create KafkaSinglePortStringInputOperator
KafkaSinglePortStringInputOperator node =
dag.addOperator("Kafka message consumer", KafkaSinglePortStringInputOperator.class);
consumer.setTopic(TEST_TOPIC);
if (isValid) {
Set<String> brokerSet = new HashSet<String>();
brokerSet.add("localhost:9092");
consumer.setBrokerSet(brokerSet);
}
node.setConsumer(consumer);
// Create Test tuple collector
CollectorModule<String> collector =
dag.addOperator("TestMessageCollector", new CollectorModule<String>());
// Connect ports
dag.addStream("Kafka message", node.outputPort, collector.inputPort)
.setLocality(Locality.CONTAINER_LOCAL);
// Create local cluster
final LocalMode.Controller lc = lma.getController();
lc.setHeartbeatMonitoringEnabled(false);
lc.runAsync();
// Wait 30s for consumer finish consuming all the messages
Assert.assertTrue("TIMEOUT: 30s ", latch.await(30000, TimeUnit.MILLISECONDS));
// Check results
Assert.assertEquals("Collections size", 1, collections.size());
Assert.assertEquals("Tuple count", totalCount, collections.get(collector.inputPort.id).size());
logger.debug(
String.format(
"Number of emitted tuples: %d", collections.get(collector.inputPort.id).size()));
p.close();
lc.shutdown();
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
// Sample DAG with 2 operators
// Replace this code with the DAG you want to build
RandomNumberGenerator randomGenerator =
dag.addOperator("randomGenerator", RandomNumberGenerator.class);
randomGenerator.setNumTuples(500);
TupleReceiver tupleReceiver = dag.addOperator("console", new TupleReceiver());
dag.addStream("randomData", randomGenerator.out, tupleReceiver.input);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
int maxValue = 1000;
RandomEventGenerator rand = dag.addOperator("rand", new RandomEventGenerator());
rand.setMinvalue(0);
rand.setMaxvalue(maxValue);
rand.setTuplesBlast(200);
CouchBaseOutputOperator couchbaseOutput =
dag.addOperator("couchbaseOutput", new CouchBaseOutputOperator());
// couchbaseOutput.getStore().setBucket("default");
// couchbaseOutput.getStore().setPassword("");
dag.addStream("ss", rand.integer_data, couchbaseOutput.input).setLocality(locality);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
// Sample DAG with 2 operators
// Replace this code with the DAG you want to build
SeedEventGenerator seedGen = dag.addOperator("seedGen", SeedEventGenerator.class);
seedGen.setSeedstart(1);
seedGen.setSeedend(10);
seedGen.addKeyData("x", 0, 10);
seedGen.addKeyData("y", 0, 100);
ConsoleOutputOperator cons = dag.addOperator("console", new ConsoleOutputOperator());
cons.setStringFormat("hello: %s");
dag.addStream("seeddata", seedGen.val_list, cons.input).setLocality(Locality.CONTAINER_LOCAL);
}
@Override
@SuppressWarnings("unchecked")
public void populateDAG(DAG dag, Configuration conf) {
RandomWordGenerator randomWordGenerator = new RandomWordGenerator();
randomWordGenerator.setTuplesPerWindow(2);
dag.addOperator("random", randomWordGenerator);
if (maxLength != null) {
fsWriter.setMaxLength(maxLength);
}
fsWriter.setFilePath(testDir.getPath());
dag.addOperator("fswriter", fsWriter);
dag.addStream("fswriterstream", randomWordGenerator.output, fsWriter.input);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
// Create ActiveMQStringSinglePortOutputOperator
SpillableTestInputOperator input = new SpillableTestInputOperator();
input.batchSize = 100;
input.sleepBetweenBatch = 0;
input = dag.addOperator("input", input);
SpillableTestOperator testOperator = new SpillableTestOperator();
testOperator.store = createStore(conf);
testOperator.shutdownCount = -1;
testOperator = dag.addOperator("test", testOperator);
// Connect ports
dag.addStream("stream", input.output, testOperator.input)
.setLocality(DAG.Locality.CONTAINER_LOCAL);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
int maxValue = 30000;
RandomEventGenerator rand = dag.addOperator("rand", new RandomEventGenerator());
rand.setMinvalue(0);
rand.setMaxvalue(maxValue);
RoundRobinHashMap<String, Object> rrhm =
dag.addOperator("rrhm", new RoundRobinHashMap<String, Object>());
rrhm.setKeys(new String[] {"x", "y"});
JavaScriptOperator calc = dag.addOperator("picalc", new JavaScriptOperator());
calc.setPassThru(false);
calc.put("i", 0);
calc.put("count", 0);
calc.addSetupScript(
"function pi() { if (x*x+y*y <= "
+ maxValue * maxValue
+ ") { i++; } count++; return i / count * 4; }");
calc.setInvoke("pi");
dag.addStream("rand_rrhm", rand.integer_data, rrhm.data);
dag.addStream("rrhm_calc", rrhm.map, calc.inBindings);
ConsoleOutputOperator console = dag.addOperator("console", new ConsoleOutputOperator());
dag.addStream("rand_console", calc.result, console.input);
}
/**
* Test KafkaOutputOperator (i.e. an output adapter for Kafka, aka producer). This module sends
* data into an ActiveMQ message bus.
*
* <p>[Generate tuple] ==> [send tuple through Kafka output adapter(i.e. producer) into Kafka
* message bus] ==> [receive data in outside Kaka listener (i.e consumer)]
*
* @throws Exception
*/
@Test
@SuppressWarnings({"SleepWhileInLoop", "empty-statement"})
public void testKafkaOutputOperator() throws Exception {
// Setup a message listener to receive the message
KafkaConsumer listener = new KafkaConsumer("topic1");
new Thread(listener).start();
// Malhar module to send message
// Create DAG for testing.
LocalMode lma = LocalMode.newInstance();
DAG dag = lma.getDAG();
// Create ActiveMQStringSinglePortOutputOperator
StringGeneratorInputOperator generator =
dag.addOperator("TestStringGenerator", StringGeneratorInputOperator.class);
KafkaStringSinglePortOutputOperator node =
dag.addOperator("Kafka message producer", KafkaStringSinglePortOutputOperator.class);
// Set configuration parameters for Kafka
node.setTopic("topic1");
// Connect ports
dag.addStream("Kafka message", generator.outputPort, node.inputPort)
.setLocality(Locality.CONTAINER_LOCAL);
// Create local cluster
final LocalMode.Controller lc = lma.getController();
lc.runAsync();
Thread.sleep(2000);
lc.shutdown();
// Check values send vs received
Assert.assertEquals("Number of emitted tuples", tupleCount, listener.holdingBuffer.size());
logger.debug(String.format("Number of emitted tuples: %d", listener.holdingBuffer.size()));
Assert.assertEquals(
"First tuple", "testString 1", listener.getMessage(listener.holdingBuffer.peek()));
listener.close();
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
// Setup the operator to get the data from twitter sample stream injected into the system.
TwitterSampleInput twitterFeed = new TwitterSampleInput();
twitterFeed = dag.addOperator("TweetSampler", twitterFeed);
// Setup a node to count the unique Hashtags within a window.
UniqueCounter<String> uniqueCounter =
dag.addOperator("UniqueHashtagCounter", new UniqueCounter<String>());
// Get the aggregated Hashtag counts and count them over last 5 mins.
WindowedTopCounter<String> topCounts =
dag.addOperator("TopCounter", new WindowedTopCounter<String>());
topCounts.setTopCount(10);
topCounts.setSlidingWindowWidth(600);
topCounts.setDagWindowWidth(1);
dag.addStream("TwittedHashtags", twitterFeed.hashtag, uniqueCounter.data).setLocality(locality);
// Count unique Hashtags
dag.addStream("UniqueHashtagCounts", uniqueCounter.count, topCounts.input);
TwitterTopCounterApplication.consoleOutput(
dag, "topHashtags", topCounts.output, SNAPSHOT_SCHEMA, "hashtag");
}
@SuppressWarnings("unchecked")
@Override
public void populateDAG(DAG dag, Configuration conf) {
KeyGen keyGen = dag.addOperator("KeyGenerator", new KeyGen());
UniqueValueCount<Integer> valCount =
dag.addOperator("ValueCounter", new UniqueValueCount<Integer>());
IntegerUniqueValueCountAppender uniqueUnifier =
dag.addOperator("Unique", new IntegerUniqueValueCountAppender());
VerifyTable verifyTable = dag.addOperator("VerifyTable", new VerifyTable());
@SuppressWarnings("rawtypes")
DefaultOutputPort valOut = valCount.output;
@SuppressWarnings("rawtypes")
DefaultOutputPort uniqueOut = uniqueUnifier.output;
dag.addStream("DataIn", keyGen.output, valCount.input);
dag.addStream("UnifyWindows", valOut, uniqueUnifier.input);
dag.addStream("ResultsOut", uniqueOut, verifyTable.input);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
dag.setAttribute(DAG.APPLICATION_NAME, "HDHTBenchmarkApplication");
Generator gen = dag.addOperator("Generator", new Generator());
gen.setTupleBlast(1000);
gen.setSleepms(0);
dag.getOperatorMeta("Generator")
.getAttributes()
.put(Context.OperatorContext.APPLICATION_WINDOW_COUNT, 1);
HDSOperator hdsOut = dag.addOperator("Store", new HDSOperator());
TFileImpl.DTFileImpl hdsFile = new TFileImpl.DTFileImpl();
hdsFile.setBasePath("WALBenchMarkDir");
hdsOut.setFileStore(hdsFile);
dag.getOperatorMeta("Store")
.getAttributes()
.put(Context.OperatorContext.COUNTERS_AGGREGATOR, new HDHTWriter.BucketIOStatAggregator());
dag.addStream("s1", gen.out, hdsOut.input).setLocality(DAG.Locality.THREAD_LOCAL);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
String appName = conf.get("appName");
if (appName == null) {
appName = "VisualDataDemo";
}
dag.setAttribute(DAG.APPLICATION_NAME, appName);
int maxValue = 30000;
RandomEventGenerator rand = dag.addOperator("random", new RandomEventGenerator());
rand.setMinvalue(0);
rand.setMaxvalue(maxValue);
DemoValueGenerator demo = dag.addOperator("chartValue", new DemoValueGenerator());
demo.setRandomIncrement(5);
demo.setRandomIncrement2(20);
PiCalculateOperator calc = dag.addOperator("picalc", new PiCalculateOperator());
calc.setBase(maxValue * maxValue);
dag.addStream("rand_calc", rand.integer_data, calc.input).setLocality(locality);
WidgetOutputOperator woo =
dag.addOperator("widget output operator", new WidgetOutputOperator());
WidgetOutputOperator wooa =
dag.addOperator("widget output operator2", new WidgetOutputOperator());
// wire to simple input gadget
dag.addStream(
"ws_pi_data", calc.output, woo.simpleInput.setTopic("app." + appName + ".piValue"))
.setLocality(locality);
// wire to time series chart gadget
dag.addStream(
"ws_chart_data",
demo.simpleOutput,
woo.timeSeriesInput.setTopic("app." + appName + ".chartValue").setMin(0).setMax(100))
.setLocality(locality);
// wire to another time series chart gadget
dag.addStream(
"ws_chart_data2",
demo.simpleOutput2,
wooa.timeSeriesInput.setTopic("app." + appName + ".chartValue2"))
.setLocality(locality);
// wire to percentage chart gadget
dag.addStream(
"ws_percentage_data",
demo.percentageOutput,
woo.percentageInput.setTopic("app." + appName + ".percentage"))
.setLocality(locality);
// wire to top N chart gadget
dag.addStream(
"ws_topn_data",
demo.top10Output,
woo.topNInput.setN(10).setTopic("app." + appName + ".topn"))
.setLocality(locality);
// wire to progress bar chart gadget
dag.addStream(
"ws_progress_data",
demo.progressOutput,
wooa.percentageInput.setTopic("app." + appName + ".progress"))
.setLocality(locality);
// wire to piechart gadget
dag.addStream(
"ws_piechart_data",
demo.pieChartOutput,
wooa.pieChartInput.setTopic("app." + appName + ".piechart"))
.setLocality(locality);
}
@Override
public void populateDAG(DAG dag, Configuration conf) {
String lPhoneRange = conf.get(PHONE_RANGE_PROP, null);
if (lPhoneRange != null) {
String[] tokens = lPhoneRange.split("-");
if (tokens.length != 2) {
throw new IllegalArgumentException("Invalid range: " + lPhoneRange);
}
this.phoneRange = Range.between(Integer.parseInt(tokens[0]), Integer.parseInt(tokens[1]));
}
LOG.debug("Phone range {}", this.phoneRange);
RandomEventGenerator phones = dag.addOperator("Receiver", RandomEventGenerator.class);
phones.setMinvalue(this.phoneRange.getMinimum());
phones.setMaxvalue(this.phoneRange.getMaximum());
PhoneMovementGenerator movementGen =
dag.addOperator("LocationFinder", PhoneMovementGenerator.class);
dag.setAttribute(
movementGen,
OperatorContext.COUNTERS_AGGREGATOR,
new BasicCounters.LongAggregator<MutableLong>());
StatelessThroughputBasedPartitioner<PhoneMovementGenerator> partitioner =
new StatelessThroughputBasedPartitioner<PhoneMovementGenerator>();
partitioner.setCooldownMillis(conf.getLong(COOL_DOWN_MILLIS, 45000));
partitioner.setMaximumEvents(conf.getLong(MAX_THROUGHPUT, 30000));
partitioner.setMinimumEvents(conf.getLong(MIN_THROUGHPUT, 10000));
dag.setAttribute(
movementGen,
OperatorContext.STATS_LISTENERS,
Arrays.asList(new StatsListener[] {partitioner}));
dag.setAttribute(movementGen, OperatorContext.PARTITIONER, partitioner);
// generate seed numbers
Random random = new Random();
int maxPhone = phoneRange.getMaximum() - phoneRange.getMinimum();
int phonesToDisplay = conf.getInt(TOTAL_SEED_NOS, 10);
for (int i = phonesToDisplay; i-- > 0; ) {
int phoneNo = phoneRange.getMinimum() + random.nextInt(maxPhone + 1);
LOG.info("seed no: " + phoneNo);
movementGen.phoneRegister.add(phoneNo);
}
// done generating data
LOG.info("Finished generating seed data.");
String gatewayAddress = dag.getValue(DAG.GATEWAY_CONNECT_ADDRESS);
URI uri = URI.create("ws://" + gatewayAddress + "/pubsub");
PubSubWebSocketOutputOperator<Object> wsOut =
dag.addOperator("LocationResults", new PubSubWebSocketOutputOperator<Object>());
wsOut.setUri(uri);
PubSubWebSocketInputOperator<Map<String, String>> wsIn =
dag.addOperator("QueryLocation", new PubSubWebSocketInputOperator<Map<String, String>>());
wsIn.setUri(uri);
// default partitioning: first connected stream to movementGen will be partitioned
dag.addStream("Phone-Data", phones.integer_data, movementGen.data);
dag.addStream("Results", movementGen.locationQueryResult, wsOut.input);
dag.addStream("Query", wsIn.outputPort, movementGen.phoneQuery);
}
public void populateDAG(DAG dag, Configuration conf) {
/*
* Config for k-fold cross validation of the Naive Bayes Model
*/
numFolds = Integer.parseInt(conf.get("dt.ml.classification.nb.numFolds"));
numAttributes = Integer.parseInt(conf.get("dt.ml.classification.nb.numAttributes"));
numClasses = Integer.parseInt(conf.get("dt.ml.classification.nb.numClasses"));
inputDataFilePath = conf.get("dt.ml.classification.nb.inputDataFilePath");
modelDir = conf.get("dt.ml.classification.nb.modelDir");
modelFileName = conf.get("dt.ml.classification.nb.modelFileName");
resultDir = conf.get("dt.ml.classification.nb.resultDir");
resultFileName = conf.get("dt.ml.classification.nb.resultFileName");
isKFold = conf.getBoolean("dt.ml.classification.nb.isKFold", false);
isTrain = conf.getBoolean("dt.ml.classification.nb.isTrain", false);
isEvaluate = conf.getBoolean("dt.ml.classification.nb.isEvaluate", false);
System.out.println("NumFolds = " + numFolds);
System.out.println("isKFold = " + isKFold);
System.out.println("isTrain = " + isTrain);
System.out.println("isEvaluate = " + isEvaluate);
System.out.println("Input file path = " + inputDataFilePath);
if (!isKFold && !isTrain && !isEvaluate) {
System.out.println("Invalid Params. K-Fold evaluation, Training or Testing must be selected");
return;
}
if (isKFold && numFolds <= 1) {
System.out.println("Invalid Params. Number of folds should be > 1");
return;
}
NBConfig nbc =
new NBConfig(
isKFold, // K-Fold Validation
isTrain, // Only Train
isEvaluate, // Only Evaluate
numFolds, // Number of folds
numAttributes, // Number of attributes
numClasses, // Number of classes
inputDataFilePath, // Input Data File
modelDir, // Model Dir
modelFileName, // Model File Name Base
resultDir, // Result Dir
resultFileName // Result file name
);
// NBConfig nbc = new NBConfig(
// true, // K-Fold Validation
// true, // Only Train
// false, // Only Evaluate
// 5, // Number of folds
// 28, //Number of attributes
// 2, //Number of classes
// "/input/HIGGS_TRAIN", // Input Data File
// "/pmmloutput", // Model Dir
// "PMML_HIGGS_CATEGORICAL.xml", // Model File Name Base
// "/testOutput", // Result Dir
// "output" // Result file name
// );
/*
* Define Operators
*/
// File Input Operator
NBLineReader opInput = dag.addOperator("File_Input", new NBLineReader(nbc));
// Input Reader
NBInputReader opNBInputReader = dag.addOperator("Parser", new NBInputReader(nbc));
// NB Counter
NBCounter opNBCounter = dag.addOperator("Counter", new NBCounter(nbc));
// NB Aggregator
NBModelAggregator<NBModelStorage> opNBAggregator =
dag.addOperator("Model_Updater", new NBModelAggregator<NBModelStorage>(nbc));
// NB Evaluator
NBEvaluator opNBEvaluator = dag.addOperator("Evaluator", new NBEvaluator(nbc));
// File Output Operator
NBOutputPerWindowOperator opNBOutput =
dag.addOperator("Model_Writer", new NBOutputPerWindowOperator(nbc));
/*
* Define Streams
*/
dag.addStream("To Parser", opInput.lineOutputPort, opNBInputReader.input)
.setLocality(Locality.THREAD_LOCAL);
dag.addStream("Control_Parser", opInput.controlOut, opNBInputReader.controlIn)
.setLocality(Locality.THREAD_LOCAL);
dag.addStream("To Counter", opNBInputReader.outForTraining, opNBCounter.inTraining);
dag.addStream("To Model Updater", opNBCounter.outTraining, opNBAggregator.inTraining);
dag.addStream("To Model Writer", opNBAggregator.outTraining, opNBOutput.inMultiWriter);
dag.addStream("To Evaluator", opNBInputReader.outForEvaluation, opNBEvaluator.inForEvaluation);
dag.addStream("To Result Writer", opNBEvaluator.outToWriter, opNBOutput.inStringWriter);
dag.addStream("Control_Counter", opNBInputReader.controlOut, opNBCounter.controlIn);
dag.addStream("Control_Updater", opNBCounter.controlOut, opNBAggregator.controlIn);
dag.addStream("Control_Writer", opNBAggregator.controlOut, opNBOutput.controlIn);
dag.addStream(
"To Evaluator (K-fold models)", opNBAggregator.outEvaluator, opNBEvaluator.inKFoldModels);
}
@Override
protected StringGeneratorInputOperator addGenerateOperator(DAG dag) {
return dag.addOperator("TestStringGenerator", StringGeneratorInputOperator.class);
// StringGeneratorInputOperator generator =
}
@Override
protected KinesisStringOutputOperator addTestingOperator(DAG dag) {
return dag.addOperator("KinesisMessageProducer", KinesisStringOutputOperator.class);
}