Skip to content

Commit

Permalink
[#1366] Move generified operators to BaseGraphOperators. (#1420)
Browse files Browse the repository at this point in the history 
fixes #1366
  • Loading branch information
@timo95
timo95 authored and Kevin Gómez committed Oct 17, 2019
1 parent 5057ebe commit 50b81ab
Show file tree
Hide file tree
Showing 11 changed files with 175 additions and 320 deletions.
18 changes: 18 additions & 0 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/BaseGraph.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -15,10 +15,14 @@
*/
package org.gradoop.flink.model.api.epgm;

import org.apache.flink.api.java.DataSet;
import org.gradoop.common.model.api.entities.GraphHead;
import org.gradoop.common.model.api.entities.Edge;
import org.gradoop.common.model.api.entities.Vertex;
import org.gradoop.flink.model.api.layouts.LogicalGraphLayout;
import org.gradoop.flink.model.impl.functions.bool.Not;
import org.gradoop.flink.model.impl.functions.bool.Or;
import org.gradoop.flink.model.impl.functions.bool.True;
import org.gradoop.flink.util.GradoopFlinkConfig;

/**
Expand Down Expand Up @@ -58,4 +62,18 @@ public interface BaseGraph<
* @return a factory that can be used to create a {@link GC} instance.
*/
BaseGraphCollectionFactory<G, V, E, LG, GC> getCollectionFactory();

//----------------------------------------------------------------------------
// Utility methods
//----------------------------------------------------------------------------

@Override
default DataSet<Boolean> isEmpty() {
return getVertices()
.map(new True<>())
.distinct()
.union(getConfig().getExecutionEnvironment().fromElements(false))
.reduce(new Or())
.map(new Not());
}
}
17 changes: 17 additions & 0 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/BaseGraphCollection.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
import org.gradoop.common.model.impl.id.GradoopId;
import org.gradoop.common.model.impl.id.GradoopIdSet;
import org.gradoop.flink.model.api.layouts.GraphCollectionLayout;
import org.gradoop.flink.model.impl.functions.bool.Not;
import org.gradoop.flink.model.impl.functions.bool.Or;
import org.gradoop.flink.model.impl.functions.bool.True;
import org.gradoop.flink.model.impl.functions.epgm.BySameId;
import org.gradoop.flink.model.impl.functions.graphcontainment.InAnyGraph;
import org.gradoop.flink.model.impl.functions.graphcontainment.InGraph;
Expand Down Expand Up @@ -98,4 +101,18 @@ default GC getGraphs(final GradoopIdSet identifiers) {

return getFactory().fromDataSets(newGraphHeads, vertices, edges);
}

//----------------------------------------------------------------------------
// Utility methods
//----------------------------------------------------------------------------

@Override
default DataSet<Boolean> isEmpty() {
return getGraphHeads()
.map(new True<>())
.distinct()
.union(getConfig().getExecutionEnvironment().fromElements(false))
.reduce(new Or())
.map(new Not());
}
}
10 changes: 10 additions & 0 deletions ...op-flink/src/main/java/org/gradoop/flink/model/api/epgm/BaseGraphCollectionOperators.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -299,6 +299,16 @@ default DataSet<Boolean> equalsByGraphData(GC otherCollection) {
// Utility methods
//----------------------------------------------------------------------------

/**
* Returns a 1-element dataset containing a {@code boolean} value which indicates if the collection
* is empty.
*
* A collection is considered empty, if it contains no logical graphs.
*
* @return 1-element dataset containing {@code true}, if the collection is empty or {@code false} if not
*/
DataSet<Boolean> isEmpty();

/**
* Returns a distinct collection of base graphs. Graph equality is based on graph identifiers.
*
Expand Down
130 changes: 130 additions & 0 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/BaseGraphOperators.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,7 @@
package org.gradoop.flink.model.api.epgm;

import org.apache.flink.api.common.functions.FilterFunction;
import org.apache.flink.api.java.DataSet;
import org.gradoop.common.model.api.entities.Edge;
import org.gradoop.common.model.api.entities.GraphHead;
import org.gradoop.common.model.api.entities.Vertex;
Expand All @@ -34,6 +35,9 @@
import org.gradoop.flink.model.impl.operators.exclusion.Exclusion;
import org.gradoop.flink.model.impl.operators.grouping.Grouping;
import org.gradoop.flink.model.impl.operators.grouping.GroupingStrategy;
import org.gradoop.flink.model.impl.operators.matching.common.MatchStrategy;
import org.gradoop.flink.model.impl.operators.matching.common.statistics.GraphStatistics;
import org.gradoop.flink.model.impl.operators.matching.single.cypher.CypherPatternMatching;
import org.gradoop.flink.model.impl.operators.neighborhood.Neighborhood;
import org.gradoop.flink.model.impl.operators.neighborhood.ReduceEdgeNeighborhood;
import org.gradoop.flink.model.impl.operators.neighborhood.ReduceVertexNeighborhood;
Expand Down Expand Up @@ -66,6 +70,118 @@ public interface BaseGraphOperators<
// Unary Operators
//----------------------------------------------------------------------------

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism strategies,
* which is vertex homomorphism and edge isomorphism. The vertex and edge data of the data graph elements
* is attached to the resulting vertices.
* <p>
* Note, that this method used no statistics about the data graph which may result in bad runtime
* performance. Use {@link #query(String, GraphStatistics)} to provide statistics for the query planner.
*
* @param query Cypher query
* @return graph collection containing matching subgraphs
*/
default GC query(String query) {
return query(query, new GraphStatistics(1, 1, 1, 1));
}

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism strategies,
* which is vertex homomorphism and edge isomorphism. The vertex and edge data of the data graph elements
* is attached to the resulting vertices.
* <p>
* Note, that this method used no statistics about the data graph which may result in bad runtime
* performance. Use {@link #query(String, String, GraphStatistics)} to provide statistics for the query planner.
* <p>
* In addition, the operator can be supplied with a construction pattern allowing the creation of new graph
* elements based on variable bindings of the match pattern. Consider the following example:
* <br>
* {@code graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")}
* <p>
* The query pattern is looking for pairs of authors that worked on the same paper.
* The construction pattern defines a new edge of type CO_AUTHOR between the two entities.
*
* @param query Cypher query string
* @param constructionPattern Construction pattern
* @return graph collection containing the output of the construct pattern
*/
default GC query(String query, String constructionPattern) {
return query(query, constructionPattern, new GraphStatistics(1, 1, 1, 1));
}

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism strategies,
* which is vertex homomorphism and edge isomorphism. The vertex and edge data of the data graph elements
* is attached to the resulting vertices.
*
* @param query Cypher query
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
default GC query(String query, GraphStatistics graphStatistics) {
return query(query, true, MatchStrategy.HOMOMORPHISM, MatchStrategy.ISOMORPHISM,
graphStatistics);
}

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism strategies,
* which is vertex homomorphism and edge isomorphism. The vertex and edge data of the data graph elements
* is attached to the resulting vertices.
* <p>
* In addition, the operator can be supplied with a construction pattern allowing the creation of new graph
* elements based on variable bindings of the match pattern. Consider the following example:
* <br>
* {@code graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")}
* <p>
* The query pattern is looking for pairs of authors that worked on the same paper.
* The construction pattern defines a new edge of type CO_AUTHOR between the two entities.
*
* @param query Cypher query
* @param constructionPattern Construction pattern
* @param graphStatistics statistics about the data graph
* @return graph collection containing the output of the construct pattern
*/
default GC query(String query, String constructionPattern, GraphStatistics graphStatistics) {
return query(query, constructionPattern, true, MatchStrategy.HOMOMORPHISM,
MatchStrategy.ISOMORPHISM, graphStatistics);
}

/**
* Evaluates the given query using the Cypher query engine.
*
* @param query Cypher query
* @param attachData attach original vertex and edge data to the result
* @param vertexStrategy morphism setting for vertex mapping
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
default GC query(String query, boolean attachData, MatchStrategy vertexStrategy,
MatchStrategy edgeStrategy, GraphStatistics graphStatistics) {
return query(query, null, attachData, vertexStrategy, edgeStrategy, graphStatistics);
}

/**
* Evaluates the given query using the Cypher query engine.
*
* @param query Cypher query
* @param constructionPattern Construction pattern
* @param attachData attach original vertex and edge data to the result
* @param vertexStrategy morphism setting for vertex mapping
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
default GC query(String query, String constructionPattern, boolean attachData, MatchStrategy vertexStrategy,
MatchStrategy edgeStrategy, GraphStatistics graphStatistics) {
return callForCollection(new CypherPatternMatching<>(query, constructionPattern, attachData,
vertexStrategy, edgeStrategy, graphStatistics));
}

/**
* Creates a copy of the base graph.
* <p>
Expand Down Expand Up @@ -376,6 +492,20 @@ default LG exclude(LG otherGraph) {
return callForGraph(new Exclusion<>(), otherGraph);
}

//----------------------------------------------------------------------------
// Utility methods
//----------------------------------------------------------------------------

/**
* Returns a 1-element dataset containing a {@code boolean} value which indicates if the graph is empty.
*
* A graph is considered empty, if it contains no vertices.
*
* @return 1-element dataset containing {@code true}, if the collection is
* empty or {@code false} if not
*/
DataSet<Boolean> isEmpty();

//----------------------------------------------------------------------------
// Call for Operators
//----------------------------------------------------------------------------
Expand Down
12 changes: 0 additions & 12 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/GraphCollectionOperators.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,6 @@
*/
package org.gradoop.flink.model.api.epgm;

import org.apache.flink.api.java.DataSet;
import org.gradoop.common.model.impl.pojo.EPGMEdge;
import org.gradoop.common.model.impl.pojo.EPGMGraphHead;
import org.gradoop.common.model.impl.pojo.EPGMVertex;
Expand All @@ -35,17 +34,6 @@ public interface GraphCollectionOperators
// Auxiliary operators
//----------------------------------------------------------------------------

/**
* Returns a 1-element dataset containing a {@code boolean} value which
* indicates if the collection is empty.
*
* A collection is considered empty, if it contains no logical graphs.
*
* @return 1-element dataset containing {@code true}, if the collection is
* empty or {@code false} if not
*/
DataSet<Boolean> isEmpty();

/**
* Writes the graph collection to the given data sink.
*
Expand Down
112 changes: 0 additions & 112 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/LogicalGraphOperators.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,8 +26,6 @@
import org.gradoop.flink.model.impl.epgm.GraphCollection;
import org.gradoop.flink.model.impl.epgm.LogicalGraph;
import org.gradoop.flink.model.impl.operators.cypher.capf.result.CAPFQueryResult;
import org.gradoop.flink.model.impl.operators.matching.common.MatchStrategy;
import org.gradoop.flink.model.impl.operators.matching.common.statistics.GraphStatistics;
import org.gradoop.flink.model.impl.operators.sampling.SamplingAlgorithm;

import java.io.IOException;
Expand Down Expand Up @@ -71,106 +69,6 @@ public interface LogicalGraphOperators
*/
CAPFQueryResult cypher(String query) throws Exception;

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism
* strategies, which is vertex homomorphism and edge isomorphism. The vertex and edge data of
* the data graph elements is attached to the resulting vertices.
* <p>
* Note, that this method used no statistics about the data graph which may result in bad
* runtime performance. Use {@link LogicalGraphOperators#query(String, GraphStatistics)} to
* provide statistics for the query planner.
*
* @param query Cypher query
* @return graph collection containing matching subgraphs
*/
GraphCollection query(String query);

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism
* strategies, which is vertex homomorphism and edge isomorphism. The vertex and edge data of
* the data graph elements is attached to the resulting vertices.
* <p>
* Note, that this method used no statistics about the data graph which may result in bad
* runtime performance. Use {@link LogicalGraphOperators#query(String, GraphStatistics)} to
* provide statistics for the query planner.
* <p>
* In addition, the operator can be supplied with a construction pattern allowing the creation
* of new graph elements based on variable bindings of the match pattern. Consider the following example:
* <p>
* {@code graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")}
* <p>
* The query pattern is looking for pairs of authors that worked on the same paper.
* The construction pattern defines a new edge of type CO_AUTHOR between the two entities.
*
* @param query Cypher query string
* @param constructionPattern Construction pattern
* @return graph collection containing the output of the construct pattern
*/
GraphCollection query(String query, String constructionPattern);

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism
* strategies, which is vertex homomorphism and edge isomorphism. The vertex and edge data of
* the data graph elements is attached to the resulting vertices.
*
* @param query Cypher query
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
GraphCollection query(String query, GraphStatistics graphStatistics);

/**
* Evaluates the given query using the Cypher query engine. The engine uses default morphism
* strategies, which is vertex homomorphism and edge isomorphism. The vertex and edge data of
* the data graph elements is attached to the resulting vertices.
* <p>
* In addition, the operator can be supplied with a construction pattern allowing the creation
* of new graph elements based on variable bindings of the match pattern. Consider the following example:
* <p>
* {@code graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")}
* <p>
* The query pattern is looking for pairs of authors that worked on the same paper.
* The construction pattern defines a new edge of type CO_AUTHOR between the two entities.
*
* @param query Cypher query
* @param constructionPattern Construction pattern
* @param graphStatistics statistics about the data graph
* @return graph collection containing the output of the construct pattern
*/
GraphCollection query(String query, String constructionPattern, GraphStatistics graphStatistics);

/**
* Evaluates the given query using the Cypher query engine.
*
* @param query Cypher query
* @param attachData attach original vertex and edge data to the result
* @param vertexStrategy morphism setting for vertex mapping
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
GraphCollection query(String query, boolean attachData, MatchStrategy vertexStrategy,
MatchStrategy edgeStrategy, GraphStatistics graphStatistics);

/**
* Evaluates the given query using the Cypher query engine.
*
* @param query Cypher query
* @param constructionPattern Construction pattern
* @param attachData attach original vertex and edge data to the result
* @param vertexStrategy morphism setting for vertex mapping
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
*/
GraphCollection query(String query, String constructionPattern, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy,
GraphStatistics graphStatistics);

/**
* Creates a new graph from a randomly chosen subset of nodes and their associated edges.
*
Expand Down Expand Up @@ -260,16 +158,6 @@ GraphCollection groupEdgesByRollUp(
*/
LogicalGraph callForGraph(GraphsToGraphOperator operator, LogicalGraph... otherGraphs);

/**
* Returns a 1-element dataset containing a {@code boolean} value which indicates if the graph is empty.
*
* A graph is considered empty, if it contains no vertices.
*
* @return 1-element dataset containing {@code true}, if the collection is
* empty or {@code false} if not
*/
DataSet<Boolean> isEmpty();

/**
* Writes the graph to given data sink.
*
Expand Down
Loading

0 comments on commit 50b81ab

Please sign in to comment.