private static JavaStreamingContext createContext(String input, String checkpointDirectory) {
System.out.println("Creating new context");
// final File outputFile = new File("/flume_recover");
// if (outputFile.exists()) {
// outputFile.delete();
// }
SparkConf conf =
new SparkConf()
.setMaster("local[2]")
.setAppName("Stream File")
.set("spark.driver.allowMultipleContexts", "true");
conf.set("spark.serializer", KryoSerializer.class.getName());
conf.set("es.index.auto.create", "true");
conf.set("es.nodes", "10.26.1.134:9200");
conf.set("es.resource", "flume/test");
conf.set("es.input.json", "true");
JavaStreamingContext jssc = new JavaStreamingContext(conf, new Duration(3000));
jssc.checkpoint(checkpointDirectory);
JavaDStream<String> textFile = jssc.textFileStream(input);
JavaDStream<String> jsonStr =
textFile.map(
new Function<String, String>() {
public String call(String arg0) throws Exception {
Matcher m = log.matcher(arg0);
if (m.find()) {
return transferJson(m);
}
return "";
}
});
jsonStr.print();
jsonStr.foreach(
new Function<JavaRDD<String>, Void>() {
public Void call(JavaRDD<String> arg0) throws Exception {
if (!arg0.isEmpty() && arg0 != null) {
JavaEsSpark.saveToEs(arg0, "flume/test");
}
return null;
}
});
return jssc;
}
public static void main(String[] args) {
// Create a Spark Context.
SparkConf conf = new SparkConf().setAppName("Activity").set("spark.eventLog.enabled", "true");
;
JavaSparkContext sc = new JavaSparkContext(conf);
JavaStreamingContext jssc = new JavaStreamingContext(sc, STREAM_INTERVAL);
String TOPIC = "activityevent";
String zkQuorum = "localhost:2181";
String group = "1";
Map<String, Integer> topicMap = new HashMap<String, Integer>();
topicMap.put(TOPIC, 1);
JavaPairReceiverInputDStream<String, String> messages =
KafkaUtils.createStream(jssc, zkQuorum, group, topicMap);
// messages.print();
JavaDStream<String> activitydatastream =
messages.map(
new Function<Tuple2<String, String>, String>() {
@Override
public String call(Tuple2<String, String> tuple2) {
return tuple2._2();
}
});
final Long teamWindowDurationMs = Durations.minutes(1).milliseconds();
JavaDStream<Activity> ActivityEntryDStream = activitydatastream.map(Activity::parseFromLine);
JavaPairDStream<WithTimestamp<String>, Double> ActivityWindowDStream =
ActivityEntryDStream.mapToPair(
windows ->
new Tuple2<>(
WithTimestamp.create(
windows.getActivity(),
// Apply Fixed Window by rounding the timestamp down to the nearest
// multiple of the window size
(convertMillsecs(windows.getTimestamp()) / teamWindowDurationMs)
* teamWindowDurationMs),
windows.getXaxis()))
.reduceByKey(SUM_REDUCER);
ActivityWindowDStream.print();
jssc.start();
jssc.awaitTermination();
// jssc.close();
sc.stop();
sc.close();
}
public JavaStreamingContext createContext() {
CassandraConnectionContext connectionContext = connectToCassandra();
JavaDStream<String> rdd = connectionContext.getRDD();
JavaStreamingContext sparkContext = connectionContext.getStreamingContext();
final CassandraConnector connector = connectionContext.getCassandraConnector();
final SimpleDateFormat format = new SimpleDateFormat("yyyy-MM-dd");
JavaDStream<Tuple2<String, Transaction>> transactionList =
rdd.map(
new Function<String, Tuple2<String, Transaction>>() {
@Override
public Tuple2<String, Transaction> call(String line) throws Exception {
String[] columns = line.split(",");
String taxId = columns[0];
String name = columns[1];
String merchant = columns[2];
BigDecimal amount = new BigDecimal(columns[3]);
Date transactionDate;
try {
transactionDate = format.parse(columns[4]);
} catch (ParseException e) {
e.printStackTrace();
throw new RuntimeException(e);
}
String tranId = columns[5];
System.out.println(line);
Tuple2<String, Transaction> transaction =
new Tuple2<>(
taxId, new Transaction(name, merchant, amount, transactionDate, tranId));
return transaction;
}
});
transactionList.cache();
final String warningsTableName = "warnings";
try (Session session = connector.openSession()) {
session.execute(
String.format("DROP TABLE IF EXISTS %s.%s", getKeySpaceName(), warningsTableName));
session.execute(
String.format(
"CREATE TABLE IF NOT EXISTS %s.%s (ssn text, "
+ "id uuid, amount decimal, rule text, PRIMARY KEY(ssn, id))",
getKeySpaceName(), warningsTableName));
}
// setup warning on more than certain number of transactions by user in a 60 second window,
// every 10 seconds
JavaPairDStream<String, BigDecimal> warnings =
transactionList
.window(new Duration(60000), new Duration(10000))
.mapToPair(
new PairFunction<Tuple2<String, Transaction>, String, BigDecimal>() {
@Override
public Tuple2<String, BigDecimal> call(Tuple2<String, Transaction> transaction)
throws Exception {
String taxId = transaction._1();
BigDecimal amount = transaction._2().getAmount();
return new Tuple2<>(taxId, amount);
}
})
.reduceByKey(
new Function2<BigDecimal, BigDecimal, BigDecimal>() {
@Override
public BigDecimal call(
BigDecimal transactionAmount1, BigDecimal transactionAmount2)
throws Exception {
return transactionAmount1.add(transactionAmount2);
}
})
.filter(
new Function<Tuple2<String, BigDecimal>, Boolean>() {
@Override
public Boolean call(Tuple2<String, BigDecimal> transactionSumByTaxId)
throws Exception {
// returns true if total is greater than 999
Boolean result = transactionSumByTaxId._2().compareTo(new BigDecimal(999)) == 1;
System.out.println(
"tran "
+ transactionSumByTaxId._1()
+ " with amount "
+ transactionSumByTaxId._2());
if (result) {
System.out.println(
"warning "
+ transactionSumByTaxId._1()
+ " has value "
+ transactionSumByTaxId._2()
+ " is greater than 999");
}
return result;
}
});
JavaDStream<Warning> mappedWarnings =
warnings.map(
new Function<Tuple2<String, BigDecimal>, Warning>() {
@Override
public Warning call(Tuple2<String, BigDecimal> warnings) throws Exception {
Warning warning = new Warning();
warning.setSsn(warnings._1());
warning.setId(UUIDs.timeBased());
warning.setAmount(warnings._2());
warning.setRule("OVER_DOLLAR_AMOUNT");
return warning;
}
});
javaFunctions(mappedWarnings)
.writerBuilder(getKeySpaceName(), warningsTableName, mapToRow(Warning.class))
.saveToCassandra();
return sparkContext;
}