Skip to content

Commit

Permalink
[#874] Restructure pattern matching in LogicalGraph (#889)
Browse files Browse the repository at this point in the history 
fixes #874
  • Loading branch information
@ChrizZz110
Rascat authored and ChrizZz110 committed Jul 24, 2018
1 parent a644ab2 commit 68efc3d
Show file tree
Hide file tree
Showing 5 changed files with 168 additions and 47 deletions.
14 changes: 12 additions & 2 deletions gradoop-examples/src/main/java/org/gradoop/examples/io/EdgeListExample.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,10 @@
import org.gradoop.flink.io.impl.edgelist.EdgeListDataSource;
import org.gradoop.flink.model.api.epgm.GraphCollection;
import org.gradoop.flink.model.api.epgm.LogicalGraph;
import org.gradoop.flink.model.impl.operators.matching.common.MatchStrategy;
import org.gradoop.flink.model.impl.operators.matching.common.query.DFSTraverser;
import org.gradoop.flink.model.impl.operators.matching.single.preserving.explorative.ExplorativePatternMatching;
import org.gradoop.flink.model.impl.operators.matching.single.preserving.explorative.traverser.TraverserStrategy;
import org.gradoop.flink.util.GradoopFlinkConfig;

/**
Expand Down Expand Up @@ -83,8 +87,14 @@ public static void main(String[] args) throws Exception {
});

// do some analytics (e.g. match two-node cycles)
GraphCollection matches = logicalGraph
.match("(a:Node)-[:link]->(b:Node)-[:link]->(a)");
String query = "(a:Node)-[:link]->(b:Node)-[:link]->(a)";
ExplorativePatternMatching patternMatchingOperator = new ExplorativePatternMatching.Builder()
.setQuery(query)
.setAttachData(true)
.setMatchStrategy(MatchStrategy.ISOMORPHISM)
.setTraverserStrategy(TraverserStrategy.SET_PAIR_BULK_ITERATION)
.setTraverser(new DFSTraverser()).build();
GraphCollection matches = logicalGraph.callForCollection(patternMatchingOperator);

// print number of matching subgraphs
System.out.println(matches.getGraphHeads().count());
Expand Down
2 changes: 1 addition & 1 deletion gradoop-examples/src/main/java/org/gradoop/examples/patternmatching/CypherExample.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -61,7 +61,7 @@ public static void main(String[] args) throws Exception {

// run a Cypher query (vertex homomorphism, edge isomorphism)
// the result is a graph collection containing all matching subgraphs
GraphCollection matches = socialNetwork.cypher(
GraphCollection matches = socialNetwork.query(
"MATCH (u1:Person)<-[:hasModerator]-(f:Forum)" +
"(u2:Person)<-[:hasMember]-(f)" +
"WHERE u1.name = \"Alice\"", statistics);
Expand Down
59 changes: 42 additions & 17 deletions gradoop-flink/src/main/java/org/gradoop/flink/model/api/epgm/LogicalGraph.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -47,11 +47,8 @@
import org.gradoop.flink.model.impl.operators.grouping.GroupingStrategy;
import org.gradoop.flink.model.impl.operators.grouping.functions.aggregation.PropertyValueAggregator;
import org.gradoop.flink.model.impl.operators.matching.common.MatchStrategy;
import org.gradoop.flink.model.impl.operators.matching.common.query.DFSTraverser;
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.matching.single.preserving.explorative.ExplorativePatternMatching;
import org.gradoop.flink.model.impl.operators.matching.single.preserving.explorative.traverser.TraverserStrategy;
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 @@ -174,6 +171,7 @@ public DataSet<Edge> getIncomingEdges(GradoopId vertexID) {
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query) {
return cypher(query, new GraphStatistics(1, 1, 1, 1));
}
Expand All @@ -182,6 +180,7 @@ public GraphCollection cypher(String query) {
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query, String constructionPattern) {
return cypher(query, constructionPattern, new GraphStatistics(1, 1, 1, 1));
}
Expand All @@ -190,6 +189,7 @@ public GraphCollection cypher(String query, String constructionPattern) {
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query, GraphStatistics graphStatistics) {
return cypher(query, true,
MatchStrategy.HOMOMORPHISM, MatchStrategy.ISOMORPHISM, graphStatistics);
Expand All @@ -199,6 +199,7 @@ public GraphCollection cypher(String query, GraphStatistics graphStatistics) {
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query, String constructionPattern,
GraphStatistics graphStatistics) {
return cypher(query, constructionPattern, true,
Expand All @@ -210,6 +211,7 @@ public GraphCollection cypher(String query, String constructionPattern,
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query, boolean attachData, MatchStrategy vertexStrategy,
MatchStrategy edgeStrategy, GraphStatistics graphStatistics) {
return cypher(query, null, attachData, vertexStrategy, edgeStrategy, graphStatistics);
Expand All @@ -219,6 +221,7 @@ public GraphCollection cypher(String query, boolean attachData, MatchStrategy ve
* {@inheritDoc}
*/
@Override
@Deprecated
public GraphCollection cypher(String query, String constructionPattern, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy, GraphStatistics graphStatistics) {
return callForCollection(new CypherPatternMatching(query, constructionPattern, attachData,
Expand All @@ -229,34 +232,56 @@ public GraphCollection cypher(String query, String constructionPattern, boolean
* {@inheritDoc}
*/
@Override
public GraphCollection match(String pattern) {
return match(pattern, true);
public GraphCollection query(String query) {
return query(query, new GraphStatistics(1, 1, 1, 1));
}

/**
* {@inheritDoc}
*/
@Override
public GraphCollection match(String pattern, boolean attachData) {
return match(pattern, attachData, MatchStrategy.ISOMORPHISM,
TraverserStrategy.SET_PAIR_BULK_ITERATION);
public GraphCollection query(String query, String constructionPattern) {
return query(query, constructionPattern, new GraphStatistics(1, 1, 1, 1));
}

/**
* {@inheritDoc}
*/
@Override
public GraphCollection match(String pattern, boolean attachData,
MatchStrategy matchStrategy, TraverserStrategy traverserStrategy) {
public GraphCollection query(String query, GraphStatistics graphStatistics) {
return query(query, true,
MatchStrategy.HOMOMORPHISM, MatchStrategy.ISOMORPHISM, graphStatistics);
}

/**
* {@inheritDoc}
*/
@Override
public GraphCollection query(String query, String constructionPattern,
GraphStatistics graphStatistics) {
return query(query, constructionPattern, true,
MatchStrategy.HOMOMORPHISM, MatchStrategy.ISOMORPHISM, graphStatistics);
}


ExplorativePatternMatching op = new ExplorativePatternMatching.Builder()
.setQuery(pattern)
.setAttachData(attachData)
.setMatchStrategy(matchStrategy)
.setTraverserStrategy(traverserStrategy)
.setTraverser(new DFSTraverser()).build();
/**
* {@inheritDoc}
*/
@Override
public GraphCollection query(String query, boolean attachData, MatchStrategy vertexStrategy,
MatchStrategy edgeStrategy, GraphStatistics graphStatistics) {
return query(query, null, attachData, vertexStrategy, edgeStrategy, graphStatistics);
}

return callForCollection(op);
/**
* {@inheritDoc}
*/
@Override
public GraphCollection query(String query, String constructionPattern, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy,
GraphStatistics graphStatistics) {
return callForCollection(new CypherPatternMatching(query, constructionPattern, attachData,
vertexStrategy, edgeStrategy, graphStatistics));
}

/**
Expand Down
134 changes: 110 additions & 24 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 @@ -34,7 +34,6 @@
import org.gradoop.flink.model.impl.operators.grouping.functions.aggregation.PropertyValueAggregator;
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.preserving.explorative.traverser.TraverserStrategy;
import org.gradoop.flink.model.impl.operators.neighborhood.Neighborhood;
import org.gradoop.flink.model.impl.operators.subgraph.Subgraph;

Expand All @@ -61,7 +60,10 @@ public interface LogicalGraphOperators extends GraphBaseOperators {
*
* @param query Cypher query
* @return graph collection containing matching subgraphs
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String)} instead.
*/
@Deprecated
GraphCollection cypher(String query);

/**
Expand Down Expand Up @@ -90,7 +92,10 @@ public interface LogicalGraphOperators extends GraphBaseOperators {
* @param query Cypher query string
* @param constructionPattern Construction pattern
* @return graph collection containing the output of the construct pattern
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String, String)} instead.
*/
@Deprecated
GraphCollection cypher(String query, String constructionPattern);

/**
Expand All @@ -101,7 +106,10 @@ public interface LogicalGraphOperators extends GraphBaseOperators {
* @param query Cypher query
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String, GraphStatistics)} instead.
*/
@Deprecated
GraphCollection cypher(String query, GraphStatistics graphStatistics);

/**
Expand All @@ -127,7 +135,10 @@ public interface LogicalGraphOperators extends GraphBaseOperators {
* @param constructionPattern Construction pattern
* @param graphStatistics statistics about the data graph
* @return graph collection containing the output of the construct pattern
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String, String, GraphStatistics)} instead.
*/
@Deprecated
GraphCollection cypher(String query, String constructionPattern, GraphStatistics graphStatistics);

/**
Expand All @@ -139,7 +150,10 @@ public interface LogicalGraphOperators extends GraphBaseOperators {
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String, boolean, MatchStrategy, MatchStrategy, GraphStatistics)} instead.
*/
@Deprecated
GraphCollection cypher(String query, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy, GraphStatistics graphStatistics);

Expand All @@ -153,47 +167,119 @@ GraphCollection cypher(String query, boolean attachData,
* @param edgeStrategy morphism setting for edge mapping
* @param graphStatistics statistics about the data graph
* @return graph collection containing matching subgraphs
* @deprecated because of API restructuring.
* Please use {@link LogicalGraph#query(String, String, boolean, MatchStrategy, MatchStrategy, GraphStatistics)} instead.
*/
@Deprecated
GraphCollection cypher(String query, String constructionPattern, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy, GraphStatistics graphStatistics);

/**
* Evaluates the given GDL query using the Traverser query engine.
* 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 pattern GDL graph pattern
* 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.
*
* @return subgraphs of the input graph that match the given graph pattern
* @param query Cypher query
* @return graph collection containing matching subgraphs
*/
GraphCollection match(String pattern);
GraphCollection query(String query);

/**
* Evaluates the given GDL query using the Traverser query engine.
* 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.
*
* This method allows to control if the original vertex and edge data
* (labels and properties) shall be attached to the resulting subgraphs.
* Note that this requires additional JOIN operations.
* 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 pattern GDL graph pattern
* @param attachData attach original vertex and edge data to the result
* @return subgraphs of the input graph that match the given graph pattern
* 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:
*
* <pre>
* <code>graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")
* </code>
* </pre>
*
* 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 match(String pattern, boolean attachData);
GraphCollection query(String query, String constructionPattern);

/**
* Evaluates the given GDL query using the Traverser query engine.
* 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.
*
* This method allows to control the match strategy. This influences mostly
* if vertices and edges can be matched to multiple vertices/edges in the
* query.
* @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.
*
* @param pattern GDL graph pattern
* @param attachData attach original vertex and edge data to the result
* @param matchStrategy strategy for vertex and edge mappings
* @param iterationStrategy strategy for internal iteration
* @return subgraphs of the input graph that match the given graph pattern
* 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:
*
* <pre>
* <code>graph.query(
* "MATCH (a:Author)-[:WROTE]->(:Paper)<-[:WROTE]-(b:Author) WHERE a <> b",
* "(a)-[:CO_AUTHOR]->(b)")
* </code>
* </pre>
*
* 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 match(String pattern, boolean attachData,
MatchStrategy matchStrategy, TraverserStrategy iterationStrategy);
GraphCollection query(String query, String constructionPattern, boolean attachData,
MatchStrategy vertexStrategy, MatchStrategy edgeStrategy, GraphStatistics graphStatistics);

/**
* Creates a copy of the logical graph.
Expand Down
Loading

0 comments on commit 68efc3d

Please sign in to comment.