Feature/multi dump file

benchmark.ttl

@@ -22,6 +22,7 @@
   hobbit:hasParameter
     :seed,
     :numberOfNodes,
+    :numberOfGraphs,
     :averageTriplesPerNode,
     :blankNodesRatio,
     :averageNodeGraphDegree,
@@ -80,6 +81,7 @@
   hobbit:hasParameter
     :seed,
     :numberOfNodes,
+    :numberOfGraphs,
     :averageTriplesPerNode,
     :blankNodesRatio,
     :averageNodeGraphDegree,
@@ -136,6 +138,13 @@
   rdfs:domain hobbit:Experiment, hobbit:Challenge;
   rdfs:range xsd:integer;
   hobbit:defaultValue "10"^^xsd:integer .
+
+:numberOfGraphs a hobbit:Parameter, hobbit:ConfigurableParameter;
+  rdfs:label "Number of graphs"@en;
+  rdfs:comment "The number of graphs a node can have."@en;
+  rdfs:domain hobbit:Experiment, hobbit:Challenge;
+  rdfs:range xsd:integer;
+  hobbit:defaultValue "1"^^xsd:integer .
 
 :averageTriplesPerNode a hobbit:Parameter, hobbit:ConfigurableParameter;
   rdfs:label "RDF dataset size"@en;

docs/README.md

@@ -41,6 +41,7 @@ For the full license text, see [LICENSE](../LICENSE).
 | Parameter | Description | Ontology resources
 | --- | --- | ---
 | Number of nodes | The number of nodes in the synthetic graph. | [orca:numberOfNodes](http://w3id.org/dice-research/orca/ontology#numberOfNodes)
+| Number of graphs | The number of graphs a node can have. | [orca:numberOfGraphs](http://w3id.org/dice-research/orca/ontology#numberOfGraphs)
 | Average node degree | The average degree of the nodes in the generated graph. | [orca:averageNodeGraphDegree](http://w3id.org/dice-research/orca/ontology#averageNodeGraphDegree)
 | Ratio of blank nodes | The ration of blank nodes in the generated graph. | [orca:blankNodesRatio](http://w3id.org/dice-research/orca/ontology#blankNodesRatio)
 | RDF dataset size | Average number of triples of the generated RDF graphs. | [orca:averageTriplesPerNode](http://w3id.org/dice-research/orca/ontology#averageTriplesPerNode)

ldcbench.api/src/main/java/org/dice_research/ldcbench/ApiConstants.java

@@ -14,6 +14,7 @@ public class ApiConstants {
     public static final String ENV_CRAWL_DELAY_KEY = "LDCBENCH_CRAWL_DELAY";
     public static final String ENV_DISALLOWED_RATIO_KEY = "LDCBENCH_DISALLOWED_RATIO";
     public static final String ENV_COMPRESSED_RATIO_KEY = "LDCBENCH_COMPRESSED_RATIO";
+    public static final String ENV_NUMBER_OF_GRAPHS_KEY = "LDCBENCH_DATAGENERATOR_NUMBER_OF_GRAPHS";
 
     public static final String ENV_SEED_KEY = "LDCBENCH_DATAGENERATOR_SEED";
     public static final String ENV_COMPONENT_COUNT_KEY = "LDCBENCH_COMPONENT_COUNT";

ldcbench.api/src/main/java/org/dice_research/ldcbench/graph/Graph.java

@@ -117,6 +117,8 @@ public default int incomingEdgeCount(int nodeId) {
      *         the ID of the external graph the node belongs to.
      */
     public int getGraphId(int nodeId);
+
+    public int getGraphId();
 
     /**
      * Returns the ID of the node in the external graph if this node is an external

ldcbench.api/src/main/java/org/dice_research/ldcbench/graph/GraphBuilder.java

@@ -3,7 +3,7 @@
 /**
  * An extension of the {@link Graph} interface offering methods for the addition
  * of nodes and edges.
- * 
+ *
  * @author Michael Röder (michael.roeder@uni-paderborn.de)
  *
  */
@@ -13,7 +13,7 @@ public interface GraphBuilder extends Graph {
      * Returns a {@link Graph} instance based on the graph data that the builder
      * collected until now. This instance might offer a faster implementation than
      * the builder class.
-     * 
+     *
      * @return a {@link Graph} instance containing the built graph.
      */
     public default Graph build() {
@@ -23,7 +23,7 @@ public default Graph build() {
     /**
      * Adds a new edge with the given information if possible. Returns {@code true}
      * if the edge could be added.
-     * 
+     *
      * @param sourceId
      *            the id of the source node
      * @param targetId
@@ -37,15 +37,15 @@ public default Graph build() {
 
     /**
      * Adds a new node to the graph and returns its ID.
-     * 
+     *
      * @return the ID of the newly created node
      */
     public int addNode();
 
     /**
      * Adds the given number of nodes to the graph and returns the range of node
      * IDs.
-     * 
+     *
      * @param nodeCount
      *            the number of nodes that should be added
      * @return the range of IDs from the first added node ([0], inclusive) to the
@@ -56,7 +56,7 @@ public default Graph build() {
     /**
      * Sets the IDs of the entrance nodes (i.e., nodes from which all other nodes
      * can be reached).
-     * 
+     *
      * @param entranceNodes
      *            the IDs of the entrance nodes
      */
@@ -68,7 +68,7 @@ public default Graph build() {
      * graph. However, to mark external nodes (i.e., nodes that are not part of this
      * graph but are linked from one of the nodes of this graph), this method can be
      * used to store their graph ID as well as their ID in the other graph.
-     * 
+     *
      * @param nodeId
      *            the ID of the node in this graph
      * @param graphId
@@ -80,7 +80,8 @@ public default Graph build() {
      *            graphID != {@link #INTERNAL_NODE_GRAPH_ID})
      */
     public void setGraphIdOfNode(int nodeId, int graphId, int externalId);
-
+
+    public void setGraphId(int i);
     /**
      * Sets the index of the first blank node of the graph.
      *

ldcbench.api/src/main/java/org/dice_research/ldcbench/graph/GraphMetadata.java

@@ -8,4 +8,5 @@ public class GraphMetadata implements Serializable {
 
     public int numberOfNodes;
     public int[] entranceNodes;
+    public int graphId;
 }

ldcbench.api/src/main/java/org/dice_research/ldcbench/graph/GrphBasedGraph.java

@@ -39,6 +39,8 @@ public class GrphBasedGraph implements GraphBuilder {
      */
     protected int blankNodesIndex = Integer.MAX_VALUE;
 
+    protected int actualGraphId;
+
     /**
      * Constructor for an empty GraphBuilder.
      */
@@ -112,7 +114,7 @@ public int[] outgoingEdgeTargets(int nodeId) {
     public int[] outgoingEdgeTypes(int nodeId) {
         return orderedOutgoingEdges(nodeId).map(this::getEdgeType).toArray();
     }
-    
+
     @Override
     public int outgoingEdgeCount(int nodeId) {
         return graph.getOutEdgeDegree(nodeId);
@@ -125,7 +127,7 @@ public int[] incomingEdgeSources(int nodeId) {
     public int[] incomingEdgeTypes(int nodeId) {
         return orderedIncomingEdges(nodeId).map(this::getEdgeType).toArray();
     }
-    
+
     @Override
     public int incomingEdgeCount(int nodeId) {
         return graph.getInEdgeDegree(nodeId);
@@ -263,6 +265,14 @@ public String toString() {
         return graph.toString();
     }
 
+    public int getGraphId() {
+    	return actualGraphId;
+    }
+
+    public void setGraphId(int id) {
+        actualGraphId = id;
+    }
+
     @Override
     public boolean isBlankNode(int nodeId) {
         return nodeId >= getBlankNodesIndex();

ldcbench.api/src/main/java/org/dice_research/ldcbench/vocab/LDCBench.java

@@ -28,6 +28,7 @@ public class LDCBench {
     public static final Property httpDumpNodeCompressedRatio = property("httpDumpNodeCompressedRatio");
     public static final Property httpDumpNodeWeight = property("httpDumpNodeWeight");
     public static final Property numberOfNodes = property("numberOfNodes");
+    public static final Property numberOfGraphs = property("numberOfGraphs");
     public static final Property blankNodesRatio = property("blankNodesRatio");
     public static final Property macroRecall = property("macroRecall");
     public static final Property microRecall = property("microRecall");

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/BenchmarkController.java

@@ -229,6 +229,8 @@ public void init() throws Exception {
                 .getInt();
         int averageRdfGraphDegree = RdfHelper.getLiteral(benchmarkParamModel, null, LDCBench.averageRdfGraphDegree)
                 .getInt();
+        int numberOfGraphs = RdfHelper.getLiteral(benchmarkParamModel, null, LDCBench.numberOfGraphs)
+                .getInt();
         double httpDumpNodeCompressedRatio = RdfHelper
                 .getLiteral(benchmarkParamModel, null, LDCBench.httpDumpNodeCompressedRatio).getDouble();
 
@@ -443,7 +445,7 @@ public void handleNodeGraph(int senderId, Graph g) {
                         DataGenerator.ENV_ACCESS_URI_TEMPLATES_KEY + "=" + serializedAccessUris,
                         DataGenerator.ENV_RESOURCE_URI_TEMPLATES_KEY + "=" + serializedResourceUris, },
                         nodeManagers.get(i).getDataGeneratorEnvironment(averageRdfGraphDegree,
-                                nodeSizeDeterminer.getNodeSize()));
+                                nodeSizeDeterminer.getNodeSize(), numberOfGraphs));
                 createDataGenerator(nodeManagers.get(i), envVariables);
                 // FIXME: HOBBIT SDK workaround (setting environment for "containers")
                 if (sdk) {

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/CkanNodeManager.java

@@ -16,10 +16,11 @@ public boolean canBeHub() {
     }
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + 0,
             DataGenerator.ENV_NUMBER_OF_NODES_KEY + "=" + 1,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/DereferencingHttpNodeManager.java

@@ -32,10 +32,11 @@ public String[] getNodeEnvironment() {
     };
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + averageRdfGraphDegree,
             DataGenerator.ENV_NUMBER_OF_EDGES_KEY + "=" + triplesPerNode,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/HttpDumpNodeManager.java

@@ -23,10 +23,11 @@ public String[] getNodeEnvironment() {
     };
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + averageRdfGraphDegree,
             DataGenerator.ENV_NUMBER_OF_EDGES_KEY + "=" + triplesPerNode,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + numberOfGraphs,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/JsonLdNodeManager.java

@@ -29,10 +29,11 @@ public String[] getNodeEnvironment() {
     };
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + averageRdfGraphDegree,
             DataGenerator.ENV_NUMBER_OF_EDGES_KEY + "=" + triplesPerNode,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/MicrodataNodeManager.java

@@ -29,10 +29,11 @@ public String[] getNodeEnvironment() {
     };
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + 0,
             DataGenerator.ENV_NUMBER_OF_NODES_KEY + "=" + 1,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/NodeManager.java

@@ -6,7 +6,7 @@
  * Interface of a manager class that gives the benchmark controller the
  * necessary information to create a node of a certain type and supply it with
  * the given data and meta data.
- * 
+ *
  * @author Michael Röder (michael.roeder@uni-paderborn.de)
  *
  */
@@ -15,62 +15,62 @@ public interface NodeManager {
     /**
      * Flag indicating whether the node created by this node manager can be used as
      * a hub
-     * 
+     *
      * @return flag indicating whether the node created by this node manager can be
      *         used as a hub
      */
     public boolean canBeHub();
 
     /**
      * Returns the Docker image name of the node.
-     * 
+     *
      * @return the Docker image name of the node.
      */
     public String getNodeImageName();
 
     /**
      * Returns the environmental values for the generated node.
-     * 
+     *
      * @return the environmental values for the generated node.
      */
     public String[] getNodeEnvironment();
 
     /**
      * Returns the Docker image name of the node's data generator.
-     * 
+     *
      * @return the Docker image name of the node's data generator.
      */
     public String getDataGeneratorImageName();
 
     /**
      * Returns the environmental values for the generated node's data generator
      * based on the given data.
-     * 
+     *
      * @param averageRdfGraphDegree the average degree of a node within the
      *                              generated RDF graph
      * @param triplesPerNode        the average number of triples per node
      * @return the environmental values for the generated node's data generator.
      */
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode);
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs);
 
     /**
      * Returns the label of the node type.
-     * 
+     *
      * @return the label of the node type.
      */
     public abstract String getLabel();
 
     /**
      * Returns the metadata of this node.
-     * 
+     *
      * @return the metadata of this node.
      */
     public NodeMetadata getMetadata();
 
     /**
      * Returns the weight for a link from the given node manager instance to this
      * node manager instance.
-     * 
+     *
      * @param nodeManager the node manager instance that could create the outgoing link
      * @return the weight of such a link in the range of [0,1]
      */

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/RDFaNodeManager.java

@@ -29,10 +29,11 @@ public String[] getNodeEnvironment() {
     };
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + 0,
             DataGenerator.ENV_NUMBER_OF_NODES_KEY + "=" + 1,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }

ldcbench.controller/src/main/java/org/dice_research/ldcbench/benchmark/cloud/SparqlNodeManager.java

@@ -16,10 +16,11 @@ public boolean canBeHub() {
     }
 
     @Override
-    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode) {
+    public String[] getDataGeneratorEnvironment(long averageRdfGraphDegree, long triplesPerNode, long numberOfGraphs) {
         String[] env = new String[]{
             DataGenerator.ENV_AVERAGE_DEGREE_KEY + "=" + averageRdfGraphDegree,
             DataGenerator.ENV_NUMBER_OF_EDGES_KEY + "=" + triplesPerNode,
+            DataGenerator.ENV_NUMBER_OF_GRAPHS_KEY + "=" + 1,
         };
         return env;
     }