Skip to content

Commit

Permalink
Use the algorithm runner in the algorithm facade
Browse files Browse the repository at this point in the history 
Co-authored-by: Ioannis Panagiotas <ioannis.panagiotas@neotechnology.com>
  • Loading branch information
@vnickolov @IoannisPanagiotas
vnickolov and IoannisPanagiotas committed Nov 15, 2023
1 parent 84e5a11 commit d82176a
Show file tree
Hide file tree
Showing 12 changed files with 155 additions and 378 deletions.
122 changes: 18 additions & 104 deletions algo/src/main/java/org/neo4j/gds/algorithms/community/CommunityAlgorithmsFacade.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,14 +19,8 @@
*/
package org.neo4j.gds.algorithms.community;

import org.neo4j.gds.Algorithm;
import org.neo4j.gds.GraphAlgorithmFactory;
import org.neo4j.gds.PreconditionsProvider;
import org.neo4j.gds.algorithms.AlgorithmComputationResult;
import org.neo4j.gds.algorithms.AlgorithmMemoryEstimation;
import org.neo4j.gds.algorithms.AlgorithmMemoryValidationService;
import org.neo4j.gds.api.DatabaseId;
import org.neo4j.gds.api.GraphName;
import org.neo4j.gds.api.User;
import org.neo4j.gds.approxmaxkcut.ApproxMaxKCutAlgorithmFactory;
import org.neo4j.gds.approxmaxkcut.ApproxMaxKCutResult;
Expand All @@ -35,12 +29,7 @@
import org.neo4j.gds.conductance.ConductanceAlgorithmFactory;
import org.neo4j.gds.conductance.ConductanceBaseConfig;
import org.neo4j.gds.conductance.ConductanceResult;
import org.neo4j.gds.config.AlgoBaseConfig;
import org.neo4j.gds.core.GraphDimensions;
import org.neo4j.gds.core.loading.GraphStoreCatalogService;
import org.neo4j.gds.core.utils.paged.dss.DisjointSetStruct;
import org.neo4j.gds.core.utils.progress.TaskRegistryFactory;
import org.neo4j.gds.core.utils.warnings.UserLogRegistryFactory;
import org.neo4j.gds.k1coloring.K1ColoringAlgorithmFactory;
import org.neo4j.gds.k1coloring.K1ColoringBaseConfig;
import org.neo4j.gds.k1coloring.K1ColoringResult;
Expand All @@ -56,7 +45,6 @@
import org.neo4j.gds.leiden.LeidenAlgorithmFactory;
import org.neo4j.gds.leiden.LeidenBaseConfig;
import org.neo4j.gds.leiden.LeidenResult;
import org.neo4j.gds.logging.Log;
import org.neo4j.gds.louvain.LouvainAlgorithmFactory;
import org.neo4j.gds.louvain.LouvainBaseConfig;
import org.neo4j.gds.louvain.LouvainResult;
Expand All @@ -80,24 +68,11 @@
import java.util.Optional;

public class CommunityAlgorithmsFacade {
private final GraphStoreCatalogService graphStoreCatalogService;
private final TaskRegistryFactory taskRegistryFactory;
private final UserLogRegistryFactory userLogRegistryFactory;
private final AlgorithmMemoryValidationService memoryUsageValidator;
private final Log log;

public CommunityAlgorithmsFacade(
GraphStoreCatalogService graphStoreCatalogService,
TaskRegistryFactory taskRegistryFactory,
UserLogRegistryFactory userLogRegistryFactory,
AlgorithmMemoryValidationService memoryUsageValidator,
Log log
) {
this.graphStoreCatalogService = graphStoreCatalogService;
this.taskRegistryFactory = taskRegistryFactory;
this.userLogRegistryFactory = userLogRegistryFactory;
this.memoryUsageValidator = memoryUsageValidator;
this.log = log;
private final BasicAlgorithmRunner algorithmRunner;

public CommunityAlgorithmsFacade(BasicAlgorithmRunner algorithmRunner) {
this.algorithmRunner = algorithmRunner;
}

AlgorithmComputationResult<DisjointSetStruct> wcc(
Expand All @@ -106,7 +81,7 @@ AlgorithmComputationResult<DisjointSetStruct> wcc(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -122,7 +97,7 @@ AlgorithmComputationResult<TriangleCountResult> triangleCount(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -138,7 +113,7 @@ AlgorithmComputationResult<KCoreDecompositionResult> kCore(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -154,7 +129,7 @@ AlgorithmComputationResult<LouvainResult> louvain(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -170,7 +145,7 @@ AlgorithmComputationResult<LeidenResult> leiden(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -186,7 +161,7 @@ AlgorithmComputationResult<LabelPropagationResult> labelPropagation(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
configuration,
configuration.relationshipWeightProperty(),
Expand All @@ -202,7 +177,7 @@ AlgorithmComputationResult<HugeLongArray> scc(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -218,7 +193,7 @@ AlgorithmComputationResult<ModularityResult> modularity(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -234,7 +209,7 @@ AlgorithmComputationResult<KmeansResult> kmeans(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -250,7 +225,7 @@ public AlgorithmComputationResult<LocalClusteringCoefficientResult> localCluster
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -266,7 +241,7 @@ AlgorithmComputationResult<K1ColoringResult> k1Coloring(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
Optional.empty(),
Expand All @@ -282,7 +257,7 @@ AlgorithmComputationResult<ConductanceResult> conductance(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -298,7 +273,7 @@ AlgorithmComputationResult<ApproxMaxKCutResult> approxMaxKCut(
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -315,7 +290,7 @@ public AlgorithmComputationResult<ModularityOptimizationResult> modularityOptimi
User user,
DatabaseId databaseId
) {
return run(
return algorithmRunner.run(
graphName,
config,
config.relationshipWeightProperty(),
Expand All @@ -326,65 +301,4 @@ public AlgorithmComputationResult<ModularityOptimizationResult> modularityOptimi
}


private <A extends Algorithm<R>, R, C extends AlgoBaseConfig> AlgorithmComputationResult<R> run(
String graphName,
C config,
Optional<String> relationshipProperty,
GraphAlgorithmFactory<A, C> algorithmFactory,
User user,
DatabaseId databaseId
) {
// TODO: Is this the best place to check for preconditions???
PreconditionsProvider.preconditions().check();

// Go get the graph and graph store from the catalog
var graphWithGraphStore = graphStoreCatalogService.getGraphWithGraphStore(
GraphName.parse(graphName),
config,
relationshipProperty,
user,
databaseId
);

var graph = graphWithGraphStore.getLeft();
var graphStore = graphWithGraphStore.getRight();

// No algorithm execution when the graph is empty
if (graph.isEmpty()) {
return AlgorithmComputationResult.withoutAlgorithmResult(graph, graphStore);
}

// create the algorithm
var algorithmEstimator = new AlgorithmMemoryEstimation<>(
GraphDimensions.of(
graph.nodeCount(),
graph.relationshipCount()
),
algorithmFactory
);

memoryUsageValidator.validateAlgorithmCanRunWithTheAvailableMemory(
config,
algorithmEstimator::memoryEstimation,
graphStoreCatalogService.graphStoreCount()
);
var algorithm = algorithmFactory.build(
graph,
config,
(org.neo4j.logging.Log) log.getNeo4jLog(),
taskRegistryFactory,
userLogRegistryFactory
);

// run the algorithm
try {
var algorithmResult = algorithm.compute();

return AlgorithmComputationResult.of(algorithmResult, graph, graphStore, algorithm.getTerminationFlag());
} catch (Exception e) {
log.warn("Computation failed", e);
algorithm.getProgressTracker().endSubTaskWithFailure();
throw e;
}
}
}
Loading

0 comments on commit d82176a

Please sign in to comment.