use of co.cask.cdap.test.SparkManager in project cdap by caskdata.
the class SparkStreamIntegrationTestRun method testSparkWithStream.
@Test
public void testSparkWithStream() throws Exception {
ApplicationManager applicationManager = deployApplication(TestSparkStreamIntegrationApp.class);
StreamManager streamManager = getStreamManager("testStream");
for (int i = 0; i < 50; i++) {
streamManager.send(String.valueOf(i));
}
SparkManager sparkManager = applicationManager.getSparkManager("SparkStreamProgram").start();
sparkManager.waitForRun(ProgramRunStatus.COMPLETED, 120, TimeUnit.SECONDS);
// The Spark job simply turns every stream event body into key/value pairs, with key==value.
DataSetManager<KeyValueTable> datasetManager = getDataset("result");
verifyDatasetResult(datasetManager);
}
use of co.cask.cdap.test.SparkManager in project cdap by caskdata.
the class Spark2Test method testScalaSparkWithObjectStore.
@Test
public void testScalaSparkWithObjectStore() throws Exception {
ApplicationManager applicationManager = deploy(NamespaceId.DEFAULT, SparkAppUsingObjectStore.class);
DataSetManager<ObjectStore<String>> keysManager = getDataset("keys");
prepareInputData(keysManager);
SparkManager sparkManager = applicationManager.getSparkManager(ScalaCharCountProgram.class.getSimpleName()).start();
sparkManager.waitForRun(ProgramRunStatus.COMPLETED, 1, TimeUnit.MINUTES);
DataSetManager<KeyValueTable> countManager = getDataset("count");
checkOutputData(countManager);
}
use of co.cask.cdap.test.SparkManager in project cdap by caskdata.
the class DecisionTreeRegressionAppTest method test.
@Test
public void test() throws Exception {
// Deploy the Application
ApplicationManager appManager = deployApplication(DecisionTreeRegressionApp.class);
// Start the Service
ServiceManager serviceManager = appManager.getServiceManager(ModelDataService.SERVICE_NAME).start();
serviceManager.waitForStatus(true, 30, 1);
URL serviceURL = serviceManager.getServiceURL(15, TimeUnit.SECONDS);
URL addDataURL = new URL(serviceURL, "labels");
HttpRequest request = HttpRequest.builder(HttpMethod.PUT, addDataURL).withBody(new InputSupplier<InputStream>() {
@Override
public InputStream getInput() throws IOException {
return getClass().getClassLoader().getResourceAsStream("sample_libsvm_data.txt");
}
}).build();
HttpResponse response = HttpRequests.execute(request);
Assert.assertEquals(200, response.getResponseCode());
// Start a Spark Program
SparkManager sparkManager = appManager.getSparkManager(ModelTrainer.NAME).start();
sparkManager.waitForRun(ProgramRunStatus.COMPLETED, 60, TimeUnit.SECONDS);
// Check that there is a new model
URL listModelsURL = new URL(serviceURL, "models");
request = HttpRequest.builder(HttpMethod.GET, listModelsURL).build();
response = HttpRequests.execute(request);
Assert.assertEquals(200, response.getResponseCode());
List<String> models = GSON.fromJson(response.getResponseBodyAsString(), new TypeToken<List<String>>() {
}.getType());
Assert.assertEquals(1, models.size());
// Check that there is some model metadata
String modelId = models.get(0);
URL modelMetaURL = new URL(serviceURL, "models/" + modelId);
request = HttpRequest.builder(HttpMethod.GET, modelMetaURL).build();
response = HttpRequests.execute(request);
Assert.assertEquals(200, response.getResponseCode());
ModelMeta meta = GSON.fromJson(response.getResponseBodyAsString(), ModelMeta.class);
Assert.assertNotNull(meta);
Assert.assertEquals(0.7, meta.getTrainingPercentage(), 0.000001);
Assert.assertEquals(692, meta.getNumFeatures());
// Check that the corresponding model file exists
DataSetManager<FileSet> modelFiles = getDataset(DecisionTreeRegressionApp.MODEL_DATASET);
Assert.assertTrue(modelFiles.get().getBaseLocation().append(modelId).exists());
}
use of co.cask.cdap.test.SparkManager in project cdap by caskdata.
the class LogAnalysisAppTest method test.
@Test
public void test() throws Exception {
// Deploy the App
ApplicationManager appManager = deployApplication(LogAnalysisApp.class);
// Send a stream events to the Stream
StreamManager streamManager = getStreamManager(LogAnalysisApp.LOG_STREAM);
streamManager.send(LOG_1);
streamManager.send(LOG_2);
streamManager.send(LOG_3);
// run the spark program
SparkManager sparkManager = appManager.getSparkManager(LogAnalysisApp.ResponseCounterSpark.class.getSimpleName()).start();
sparkManager.waitForRun(ProgramRunStatus.COMPLETED, 60, TimeUnit.SECONDS);
// run the mapreduce job
MapReduceManager mapReduceManager = appManager.getMapReduceManager(HitCounterProgram.class.getSimpleName()).start();
mapReduceManager.waitForRun(ProgramRunStatus.COMPLETED, 3, TimeUnit.MINUTES);
// start and wait for services
ServiceManager hitCounterServiceManager = getServiceManager(appManager, LogAnalysisApp.HIT_COUNTER_SERVICE);
ServiceManager responseCounterServiceManager = getServiceManager(appManager, LogAnalysisApp.RESPONSE_COUNTER_SERVICE);
ServiceManager requestCounterServiceManager = getServiceManager(appManager, LogAnalysisApp.REQUEST_COUNTER_SERVICE);
// Query for hit counts and verify it
URL totalHitsURL = new URL(hitCounterServiceManager.getServiceURL(15, TimeUnit.SECONDS), LogAnalysisApp.HitCounterServiceHandler.HIT_COUNTER_SERVICE_PATH);
HttpResponse response = HttpRequests.execute(HttpRequest.post(totalHitsURL).withBody("{\"url\":\"" + "/home.html" + "\"}").build());
Assert.assertEquals(HttpURLConnection.HTTP_OK, response.getResponseCode());
Assert.assertEquals(TOTAL_HITS_VALUE, response.getResponseBodyAsString());
// query for total responses for a response code and verify it
URL responseCodeURL = new URL(responseCounterServiceManager.getServiceURL(15, TimeUnit.SECONDS), LogAnalysisApp.ResponseCounterHandler.RESPONSE_COUNT_PATH + "/" + RESPONSE_CODE);
HttpRequest request = HttpRequest.get(responseCodeURL).build();
response = HttpRequests.execute(request);
Assert.assertEquals(TOTAL_RESPONSE_VALUE, response.getResponseBodyAsString());
// query to get partitions in the request count tpfs
URL requestCountFilsetsURL = new URL(requestCounterServiceManager.getServiceURL(15, TimeUnit.SECONDS), LogAnalysisApp.RequestCounterHandler.REQUEST_COUNTER_PARTITIONS_PATH);
request = HttpRequest.get(requestCountFilsetsURL).build();
response = HttpRequests.execute(request);
TreeSet<String> partitions = GSON.fromJson(response.getResponseBodyAsString(), new TypeToken<TreeSet<String>>() {
}.getType());
Assert.assertEquals(1, partitions.size());
String partition = partitions.iterator().next();
// Query for the contents of the files in this partition and verify
URL requestFilesetContentURL = new URL(requestCounterServiceManager.getServiceURL(15, TimeUnit.SECONDS), LogAnalysisApp.RequestCounterHandler.REQUEST_FILE_CONTENT_PATH);
response = HttpRequests.execute(HttpRequest.post(requestFilesetContentURL).withBody("{\"" + LogAnalysisApp.RequestCounterHandler.REQUEST_FILE_PATH_HANDLER_KEY + "\":\"" + partition + "\"}").build());
Assert.assertEquals(HttpURLConnection.HTTP_OK, response.getResponseCode());
Map<String, Integer> responseMap = GSON.fromJson(response.getResponseBodyAsString(), new TypeToken<Map<String, Integer>>() {
}.getType());
Assert.assertTrue(responseMap.equals(TPFS_RESULT));
}
use of co.cask.cdap.test.SparkManager in project cdap by caskdata.
the class SparkPageRankAppTest method test.
@Test
public void test() throws Exception {
// Deploy the SparkPageRankApp
ApplicationManager appManager = deployApplication(SparkPageRankApp.class);
// Send a stream events to the Stream
StreamManager streamManager = getStreamManager(SparkPageRankApp.BACKLINK_URL_STREAM);
streamManager.send(Joiner.on(" ").join(URL_1, URL_2));
streamManager.send(Joiner.on(" ").join(URL_1, URL_3));
streamManager.send(Joiner.on(" ").join(URL_2, URL_1));
streamManager.send(Joiner.on(" ").join(URL_3, URL_1));
// Start service
ServiceManager serviceManager = appManager.getServiceManager(SparkPageRankApp.SERVICE_HANDLERS).start();
// Wait for service to start since the Spark program needs it
serviceManager.waitForStatus(true);
// Start the SparkPageRankProgram
SparkManager sparkManager = appManager.getSparkManager(SparkPageRankApp.PageRankSpark.class.getSimpleName()).start();
sparkManager.waitForRun(ProgramRunStatus.COMPLETED, 60, TimeUnit.SECONDS);
// Run RanksCounter which will count the number of pages for a pr
MapReduceManager mapReduceManager = appManager.getMapReduceManager(SparkPageRankApp.RanksCounter.class.getSimpleName()).start();
mapReduceManager.waitForRun(ProgramRunStatus.COMPLETED, 3, TimeUnit.MINUTES);
// Query for rank
URL url = new URL(serviceManager.getServiceURL(15, TimeUnit.SECONDS), SparkPageRankApp.SparkPageRankServiceHandler.RANKS_PATH);
HttpRequest request = HttpRequest.post(url).withBody(("{\"" + SparkPageRankApp.SparkPageRankServiceHandler.URL_KEY + "\":\"" + URL_1 + "\"}")).build();
HttpResponse response = HttpRequests.execute(request);
Assert.assertEquals(HttpURLConnection.HTTP_OK, response.getResponseCode());
Assert.assertEquals(RANK, response.getResponseBodyAsString());
// Request total pages for a page rank and verify it
url = new URL(serviceManager.getServiceURL(15, TimeUnit.SECONDS), SparkPageRankApp.SparkPageRankServiceHandler.TOTAL_PAGES_PATH + "/" + RANK);
response = HttpRequests.execute(HttpRequest.get(url).build());
Assert.assertEquals(TOTAL_PAGES, response.getResponseBodyAsString());
}
Aggregations