@Test
public void shouldWriteToArbitraryRDD() throws Exception {
final Configuration configuration = new BaseConfiguration();
configuration.setProperty("spark.master", "local[4]");
configuration.setProperty("spark.serializer", GryoSerializer.class.getCanonicalName());
configuration.setProperty(Graph.GRAPH, HadoopGraph.class.getName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_INPUT_FORMAT, GryoInputFormat.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_SPARK_GRAPH_OUTPUT_RDD, ExampleOutputRDD.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_OUTPUT_LOCATION,
TestHelper.makeTestDataDirectory(this.getClass(), "shouldWriteToArbitraryRDD"));
configuration.setProperty(Constants.GREMLIN_HADOOP_JARS_IN_DISTRIBUTED_CACHE, false);
////////
Graph graph = GraphFactory.open(configuration);
graph
.compute(SparkGraphComputer.class)
.result(GraphComputer.ResultGraph.NEW)
.persist(GraphComputer.Persist.EDGES)
.program(
TraversalVertexProgram.build()
.traversal(
graph.traversal().withComputer(g -> g.compute(SparkGraphComputer.class)),
"gremlin-groovy",
"g.V()")
.create(graph))
.submit()
.get();
}
static {
try {
final List<String> groovyScriptResources = Arrays.asList("GremlinExecutorInit.groovy");
for (final String fileName : groovyScriptResources) {
PATHS.put(
fileName,
TestHelper.generateTempFileFromResource(GremlinExecutorTest.class, fileName, "")
.getAbsolutePath());
}
} catch (Exception e) {
e.printStackTrace();
}
}
@Override
public Map<String, Object> getBaseConfiguration(
final String graphName,
final Class<?> test,
final String testMethodName,
final LoadGraphWith.GraphData graphData) {
final String directory =
getWorkingDirectory()
+ File.separator
+ TestHelper.cleanPathSegment(this.getClass().getSimpleName())
+ File.separator
+ TestHelper.cleanPathSegment(graphName)
+ File.separator
+ cleanParameters(TestHelper.cleanPathSegment(testMethodName));
return new HashMap<String, Object>() {
{
put(Graph.GRAPH, Neo4jGraph.class.getName());
put(Neo4jGraph.CONFIG_DIRECTORY, directory);
put(Neo4jGraph.CONFIG_META_PROPERTIES, true);
put(Neo4jGraph.CONFIG_MULTI_PROPERTIES, true);
}
};
}
@Test
public void shouldNotHaveDanglingPersistedComputeRDDs() throws Exception {
Spark.create("local[4]");
final String rddName =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, UUID.randomUUID().toString());
final Configuration configuration = super.getBaseConfiguration();
configuration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, GryoInputFormat.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, GryoOutputFormat.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName);
configuration.setProperty(Constants.GREMLIN_SPARK_PERSIST_CONTEXT, true);
Graph graph = GraphFactory.open(configuration);
///
assertEquals(
6,
graph
.traversal()
.withComputer(Computer.compute(SparkGraphComputer.class))
.V()
.out()
.count()
.next()
.longValue());
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(0, Spark.getContext().getPersistentRDDs().size());
//
assertEquals(
2,
graph
.traversal()
.withComputer(Computer.compute(SparkGraphComputer.class))
.V()
.out()
.out()
.count()
.next()
.longValue());
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(0, Spark.getContext().getPersistentRDDs().size());
///////
Spark.close();
}
@Test
public void shouldPersistRDDBasedOnStorageLevel() throws Exception {
Spark.create("local[4]");
int counter = 0;
for (final String storageLevel :
Arrays.asList("MEMORY_ONLY", "DISK_ONLY", "MEMORY_ONLY_SER", "MEMORY_AND_DISK_SER")) {
assertEquals(counter, Spark.getRDDs().size());
assertEquals(counter, Spark.getContext().getPersistentRDDs().size());
counter++;
final String rddName =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, UUID.randomUUID().toString());
final Configuration configuration = super.getBaseConfiguration();
configuration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, GryoInputFormat.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, PersistedOutputRDD.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_SPARK_PERSIST_STORAGE_LEVEL, storageLevel);
configuration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName);
configuration.setProperty(Constants.GREMLIN_SPARK_PERSIST_CONTEXT, true);
Graph graph = GraphFactory.open(configuration);
graph
.compute(SparkGraphComputer.class)
.result(GraphComputer.ResultGraph.NEW)
.persist(GraphComputer.Persist.EDGES)
.program(
TraversalVertexProgram.build()
.traversal(
graph.traversal().withComputer(SparkGraphComputer.class),
"gremlin-groovy",
"g.V().groupCount('m').by('name').out()")
.create(graph))
.submit()
.get();
////////
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(
StorageLevel.fromString(storageLevel),
Spark.getRDD(Constants.getGraphLocation(rddName)).getStorageLevel());
assertEquals(counter, Spark.getRDDs().size());
assertEquals(counter, Spark.getContext().getPersistentRDDs().size());
}
Spark.close();
}
@Test
public void shouldNotPersistRDDAcrossJobs() throws Exception {
Spark.create("local[4]");
final String rddName =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, UUID.randomUUID().toString());
final Configuration configuration = super.getBaseConfiguration();
configuration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, GryoInputFormat.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, PersistedOutputRDD.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName);
configuration.setProperty(
Constants.GREMLIN_SPARK_PERSIST_CONTEXT,
false); // because the spark context is NOT persisted, neither is the RDD
Graph graph = GraphFactory.open(configuration);
graph
.compute(SparkGraphComputer.class)
.result(GraphComputer.ResultGraph.NEW)
.persist(GraphComputer.Persist.EDGES)
.program(
TraversalVertexProgram.build()
.traversal(
graph.traversal().withComputer(SparkGraphComputer.class),
"gremlin-groovy",
"g.V()")
.create(graph))
.submit()
.get();
////////
Spark.create("local[4]");
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(0, Spark.getContext().getPersistentRDDs().size());
Spark.close();
}
public static String convertToRelative(final Class clazz, final File f) {
final File root = TestHelper.getRootOfBuildDirectory(clazz).getParentFile().getAbsoluteFile();
return root.toURI().relativize(f.getAbsoluteFile().toURI()).toString();
}
@Test
public void testComplexChain() throws Exception {
Spark.create("local[4]");
final String rddName =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, "testComplexChain", "graphRDD");
final String rddName2 =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, "testComplexChain", "graphRDD2");
final Configuration configuration = super.getBaseConfiguration();
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, GryoInputFormat.class.getCanonicalName());
configuration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, PersistedOutputRDD.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName);
configuration.setProperty(Constants.GREMLIN_SPARK_PERSIST_CONTEXT, true);
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(0, Spark.getContext().getPersistentRDDs().size());
Graph graph = GraphFactory.open(configuration);
graph =
graph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.EDGES)
.program(PageRankVertexProgram.build().iterations(2).create(graph))
.submit()
.get()
.graph();
GraphTraversalSource g = graph.traversal();
assertEquals(6l, g.V().count().next().longValue());
assertEquals(6l, g.E().count().next().longValue());
assertEquals(6l, g.V().values(PageRankVertexProgram.PAGE_RANK).count().next().longValue());
////
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(1, Spark.getContext().getPersistentRDDs().size());
////
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, PersistedInputRDD.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_HADOOP_INPUT_LOCATION, rddName);
configuration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, PersistedOutputRDD.class.getCanonicalName());
configuration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName2);
////
graph = GraphFactory.open(configuration);
graph =
graph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.EDGES)
.mapReduce(PageRankMapReduce.build().create())
.program(PageRankVertexProgram.build().iterations(2).create(graph))
.submit()
.get()
.graph();
g = graph.traversal();
assertEquals(6l, g.V().count().next().longValue());
assertEquals(6l, g.E().count().next().longValue());
assertEquals(6l, g.V().values(PageRankVertexProgram.PAGE_RANK).count().next().longValue());
////
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName2)));
assertTrue(
Spark.hasRDD(Constants.getMemoryLocation(rddName2, PageRankMapReduce.DEFAULT_MEMORY_KEY)));
assertEquals(3, Spark.getContext().getPersistentRDDs().size());
////
graph = GraphFactory.open(configuration);
graph =
graph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.VERTEX_PROPERTIES)
.program(PageRankVertexProgram.build().iterations(2).create(graph))
.submit()
.get()
.graph();
g = graph.traversal();
assertEquals(6l, g.V().count().next().longValue());
assertEquals(0l, g.E().count().next().longValue());
assertEquals(6l, g.V().values(PageRankVertexProgram.PAGE_RANK).count().next().longValue());
////
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName2)));
assertFalse(
Spark.hasRDD(Constants.getMemoryLocation(rddName2, PageRankMapReduce.DEFAULT_MEMORY_KEY)));
assertEquals(2, Spark.getContext().getPersistentRDDs().size());
////
graph = GraphFactory.open(configuration);
graph =
graph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.NOTHING)
.program(PageRankVertexProgram.build().iterations(2).create(graph))
.submit()
.get()
.graph();
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName2)));
g = graph.traversal();
assertEquals(0l, g.V().count().next().longValue());
assertEquals(0l, g.E().count().next().longValue());
assertEquals(0l, g.V().values(PageRankVertexProgram.PAGE_RANK).count().next().longValue());
////
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertFalse(Spark.hasRDD(Constants.getGraphLocation(rddName2)));
assertFalse(
Spark.hasRDD(Constants.getMemoryLocation(rddName2, PageRankMapReduce.DEFAULT_MEMORY_KEY)));
assertEquals(1, Spark.getContext().getPersistentRDDs().size());
Spark.close();
}
@Test
public void testBulkLoaderVertexProgramChainWithInputOutputHelperMapping() throws Exception {
Spark.create("local[4]");
final String rddName =
TestHelper.makeTestDataDirectory(
PersistedInputOutputRDDIntegrateTest.class, UUID.randomUUID().toString());
final Configuration readConfiguration = super.getBaseConfiguration();
readConfiguration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_READER, GryoInputFormat.class.getCanonicalName());
readConfiguration.setProperty(
Constants.GREMLIN_HADOOP_INPUT_LOCATION,
SparkHadoopGraphProvider.PATHS.get("tinkerpop-modern.kryo"));
readConfiguration.setProperty(
Constants.GREMLIN_HADOOP_GRAPH_WRITER, PersistedOutputRDD.class.getCanonicalName());
readConfiguration.setProperty(Constants.GREMLIN_HADOOP_OUTPUT_LOCATION, rddName);
readConfiguration.setProperty(Constants.GREMLIN_SPARK_PERSIST_CONTEXT, true);
Graph pageRankGraph = GraphFactory.open(readConfiguration);
///////////////
final Configuration writeConfiguration = new BaseConfiguration();
writeConfiguration.setProperty(Graph.GRAPH, TinkerGraph.class.getCanonicalName());
writeConfiguration.setProperty(TinkerGraph.GREMLIN_TINKERGRAPH_GRAPH_FORMAT, "gryo");
writeConfiguration.setProperty(
TinkerGraph.GREMLIN_TINKERGRAPH_GRAPH_LOCATION,
TestHelper.makeTestDataDirectory(PersistedInputOutputRDDIntegrateTest.class)
+ "testBulkLoaderVertexProgramChainWithInputOutputHelperMapping.kryo");
final Graph bulkLoaderGraph =
pageRankGraph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.EDGES)
.program(PageRankVertexProgram.build().create(pageRankGraph))
.submit()
.get()
.graph();
bulkLoaderGraph
.compute(SparkGraphComputer.class)
.persist(GraphComputer.Persist.NOTHING)
.workers(1)
.program(
BulkLoaderVertexProgram.build()
.userSuppliedIds(true)
.writeGraph(writeConfiguration)
.create(bulkLoaderGraph))
.submit()
.get();
////
Spark.create(readConfiguration);
assertTrue(Spark.hasRDD(Constants.getGraphLocation(rddName)));
assertEquals(1, Spark.getContext().getPersistentRDDs().size());
////
final Graph graph = TinkerGraph.open();
final GraphTraversalSource g = graph.traversal();
graph
.io(IoCore.gryo())
.readGraph(
TestHelper.makeTestDataDirectory(PersistedInputOutputRDDIntegrateTest.class)
+ "testBulkLoaderVertexProgramChainWithInputOutputHelperMapping.kryo");
assertEquals(6l, g.V().count().next().longValue());
assertEquals(6l, g.E().count().next().longValue());
assertEquals("marko", g.V().has("name", "marko").values("name").next());
assertEquals(6l, g.V().values(PageRankVertexProgram.PAGE_RANK).count().next().longValue());
////
Spark.close();
}