GDS 1.6 Preview

Release Date: 20 May 2021

GDS 1.6 is compatible with Neo4j 4.0, 4.1, and 4.2 but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Breaking changes

• Degree centrality has been promoted to the product tier
  • Added procedures:
    • gds.degree.stream.estimate
    • gds.degree.write.estimate
    • gds.degree.mutate
    • gds.degree.mutate.estimate
    • gds.degree.stats
    • gds.degree.stats.estimate
  • Removed alpha procedures:
    • gds.alpha.degree.stream
    • Gds.alpha.degree.write
• Article Rank has been promoted to the product tier
  • Added procedures:
    • gds.articleRank.stream
    • gds.articleRank.stream.estimate
    • gds.articleRank.write
    • gds.articleRank.write.estimate
    • gds.articleRank.mutate
    • gds.articleRank.mutate.estimate
    • gds.articleRank.stats
    • gds.articleRank.stats.estimate
  • Removed alpha procedures:
    • gds.alpha.articleRank.stream
    • gds.alpha.articleRank.write
• Eigenvector Centrality has been promoted to the product tier
  • Added procedures:
    • gds.eigenvector.stream
    • gds.eigenvector.stream.estimate
    • gds.eigenvector.write
    • gds.eigenvector.write.estimate
    • gds.eigenvector.mutate
    • gds.eigenvector.mutate.estimate
    • gds.eigenvector.stats
    • gds.eigenvector.stats.estimate
  • Removed alpha procedures:
    • gds.alpha.eigenvector.stream
    • Gds.alpha.eigenvector.write
• AStar has been promoted to the product tier
  • Added procedures:
    • gds.astar.stream
    • gds.astar.stream.estimate
    • gds.astar.write
    • gds.astar.write.estimate
    • gds.astar.mutate
    • gds.astar.mutate.estimate
  • Removed alpha procedures:
    • gds.beta.astar.stream
    • gds.beta.astar.stream.estimate
    • gds.beta.astar.write
    • gds.beta.astar.write.estimate
    • gds.beta.astar.mutate
    • gds.beta.astar.mutate.estimate
  • The parameter path was removed. The path computation is controlled by the YIELD.
• Yens K Shortest Paths has been promoted to the product tier:
  • Added procedures:
    • gds.yens.stream
    • gds.yens.stream.estimate
    • gds.yens.write
    • gds.yens.write.estimate
    • gds.yens.mutate
    • gds.yens.mutate.estimate
  • Removed alpha procedures:
    • gds.beta.yens.stream
    • gds.beta.yens.stream.estimate
    • gds.beta.yens.write
    • gds.beta.yens.write.estimate
    • gds.beta.yens.mutate
    • gds.beta.yens.mutate.estimate
  • The parameter path was removed. The path computation is controlled by the cypher YIELD sub-clause.
• Dijkstra Source-Target has been promoted to the product tier:
  • Added procedures:
    • gds.shortestPath.dijkstra.stream
    • gds.shortestPath.dijkstra.stream.estimate
    • gds.shortestPath.dijkstra.write
    • gds.shortestPath.dijkstra.write.estimate
    • gds.shortestPath.dijkstra.mutate
    • gds.shortestPath.dijkstra.mutate.estimate
  • Removed alpha procedures:
    • gds.beta.shortestPath.dijkstra.stream
    • gds.beta.shortestPath.dijkstra.stream.estimate
    • gds.beta.shortestPath.dijkstra.write
    • gds.beta.shortestPath.dijkstra.write.estimate
    • gds.beta.shortestPath.dijkstra.mutate
    • gds.beta.shortestPath.dijkstra.mutate.estimate
  • The parameter path was removed. The path computation is controlled by the cypher YIELD sub-clause.
• Dijkstra Single-Source has been promoted to the product tier:
  • Added procedures:
    • gds.allShortestPath.dijkstra.stream
    • gds.allShortestPath.dijkstra.stream.estimate
    • gds.allShortestPath.dijkstra.write
    • gds.allShortestPath.dijkstra.write.estimate
    • gds.allShortestPath.dijkstra.mutate
    • gds.allShortestPath.dijkstra.mutate.estimate
  • Removed alpha procedures:
    • gds.beta.allShortestPath.dijkstra.stream
    • gds.beta.allShortestPath.dijkstra.stream.estimate
    • gds.beta.allShortestPath.dijkstra.write
    • gds.beta.allShortestPath.dijkstra.write.estimate
    • gds.beta.allShortestPath.dijkstra.mutate
    • gds.beta.allShortestPath.dijkstra.mutate.estimate
  • The parameter path was removed. The path computation is controlled by the cypher YIELD sub-clause.
• Node2Vec has been promoted to the beta tier
  • Added procedures:
    • gds.beta.node2vec.stream
    • gds.beta.node2vec.stream.estimate
    • gds.beta.node2vec.write
    • gds.beta.node2vec.write.estimate
    • gds.beta.node2vec.mutate
    • gds.beta.node2vec.mutate.estimate
  • Removed alpha procedures:
    • gds.alpha.node2vec.stream
    • gds.alpha.node2vec.write
• The parameter centerSamplingFactor is renamed to positiveSamplingFactor
• The parameter contextSamplingExponent is renamed to negativeSamplingExponent

The model catalog list feature no longer throws an error when a non-existent model name is given

Node Classification and Link Prediction”

• maxStreakCount configuration parameter is renamed to patience. It is used in the train modes of Node Classification and Link Prediction.
• maxIterations and minIterations configuration parameters are renamed to maxEpochs and minEpochs. It is used in the train modes of Node Classification and Link Prediction.
• windowSize configuration parameters is removed from the train modes of Node Classification and Link Prediction.

gds.alpha.ml.linkPrediction.train configuration parameter classRatio is renamed to negativeClassWeight. It is also mandatory now.

Removed degreeAsProperty configuration parameter from GraphSAGE

• The same effect can be achieved by using gds.degree.mutate and use the mutated property as feature for GraphSAGE training.
• Important: GraphSAGE models persisted with earlier versions of GDS are not compatible with this version.

New features

• New ScaleProperties procedures to transform and scale node properties. Available scalers: Min-max, Max, Mean, Log, Standard Score, L1 Norm, L2 Norm
  • gds.alpha.scaleProperties.stream
  • gds.alpha.scaleProperties.mutate
• Added ability to create new in-memory graphs by filtering existing named graphs based on node and relationship properties with new catalog procedure gds.beta.graph.create.subgraph
• Two new centrality algorithms for influence maximization were contributed by community member @xkitsios
  • gds.alpha.influenceMaximization.celf.stream
  • gds.alpha.influenceMaximization.greedy.stream
• Link Prediction:
  • Added support for storing, loading and publishing Link Prediction models.
  • Added progress logging for gds.alpha.ml.linkPrediction.train and gds.alpha.ml.linkPrediction.predict.
  • Added write and stream modes to gds.alpha.ml.linkPrediction.predict
    • gds.alpha.ml.linkPrediction.stream
    • gds.alpha.ml.linkPrediction.write
  • Added estimate mode for Link Prediction:
    • gds.alpha.ml.linkPrediction.train.estimate
    • gds.alpha.ml.lin...

1.5.2

Release Date: 11 May 2021

GDS 1.5 is compatible with Neo4j 4.0, 4.1, and 4.2 but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Bug fixes

• Fixed a bug in FastRPExtended concerning implementation internals, especially when propertyDimesion == embeddingDimension output contained NaNs.
• Fixed a bug where Alpha similarity algorithms in some cases could fail on division by 0 when writing results back.
• Fixed an issue where gds.graph.drop could take a long time when the graph contained node embeddings.
• Fixed a bug where gds.beta.graphSage.train was failing in the presence of array properties.

1.5.1

Release Date: 3 March, 2021

GDS 1.5 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Bug fixes

• Fixed a bug which caused gds.graph.list and gds.graph.drop to throw an error when specifying a graph with duplicate property keys by failing early.
• Fixed potential ArrayIndexOutOfBoundsException when running gds.triangleCount on a relationship-filtered graph.
• Fixed a bug that can lead to inconsistencies when writing or mutating new relationships created from a label-filtered graph.

Improvements

• Progress logging: Removed a "disabled" log message from the database startup when GDS was running in its default configuration. It is replaced with a more elaborate "enabled" message when the progress tracking feature is enabled.
• We now return the name of the current database in the error message if graph is not found.

1.5.0

Release Date: 9 February, 2021

GDS 1.5 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Breaking changes

• Promote several shortest path algorithms to beta tier: Dijkstra, A*, and Yens k-shortest paths. The APIs have been standardized, and all include the ability to return source/target nodes, nodes traversed, and paths.
  • This adds procedures
    • gds.beta.shortestPath.dijkstra.mutate
    • gds.beta.shortestPath.dijkstra.mutate.estimate
    • gds.beta.shortestPath.dijkstra.stream
    • gds.beta.shortestPath.dijkstra.stream.estimate
    • gds.beta.shortestPath.dijkstra.write
    • gds.beta.shortestPath.dijkstra.write.estimate
    • gds.beta.shortestPath.astar.mutate
    • gds.beta.shortestPath.astar.mutate.estimate
    • gds.beta.shortestPath.astar.stream
    • gds.beta.shortestPath.astar.stream.estimate
    • gds.beta.shortestPath.astar.write
    • gds.beta.shortestPath.astar.write.estimate
    • gds.beta.shortestPath.yens.mutate
    • gds.beta.shortestPath.yens.mutate.estimate
    • gds.beta.shortestPath.yens.stream
    • gds.beta.shortestPath.yens.stream.estimate
    • gds.beta.shortestPath.yens.write
    • gds.beta.shortestPath.yens.write.estimate
    • gds.beta.allShortestPaths.dijkstra.mutate
    • gds.beta.allShortestPaths.dijkstra.mutate.estimate
    • gds.beta.allShortestPaths.dijkstra.stream
    • gds.beta.allShortestPaths.dijkstra.stream.estimate
    • gds.beta.allShortestPaths.dijkstra.write
    • gds.beta.allShortestPaths.dijkstra.write.estimate
  • And removes alpha procedures
    • gds.alpha.shortestPath.stream
    • gds.alpha.shortestPath.write
    • gds.alpha.shortestPath.astar.stream
    • gds.alpha.kShortestPaths.stream
    • gds.alpha.kShortestPaths.write
    • gds.alpha.shortestPaths.stream
    • gds.alpha.shortestPaths.write
• GDS will now throw an error when a user tries to use a mutate procedure on graphs not stored in the graph catalog (anonymous graphs)

New Features

• Introduced machine learning based multi-class node classification procedures:
  • Add gds.alpha.ml.nodeClassification.train to train a model to predict a node label
  • Add gds.alpha.ml.nodeClassification.predict.mutate to make predictions using a trained model
• Introduced machine learning based link prediction procedures:
  • Add gds.alpha.linkPrediction.train procedure for training Link Prediction models.
  • Added gds.alpha.linkPrediction.predict.mutate procedure for predicting relationships based on a trained Link Prediction model.
• Added support for list properties as features for
  • gds.alpha.nodeClassification
  • gds.beta.fastRPExtended
  • gds.beta.graphSage
• Added support for storing trained models on disk (Enterprise only)
  • gds.alpha.model.store
  • gds.alpha.model.load
  • gds.alpha.model.delete
• Added procedure for publishing trained models (Enterprise only)
  • gds.alpha.model.publish
• Added HITS algorithm to the alpha tier
  • gds.alpha.hits.mutate and gds.alpha.hits.mutate.estimate
  • gds.alpha.hits.stats and gds.alpha.hits.stats.estimate
  • gds.alpha.hits.stream and gds.alpha.hits.stream.estimate
  • gds.alpha.hits.write and gds.alpha.hits.write.estimate
• Added Speaker-Listener Label Propagation Algorithm (SLLPA) to the alpha tier
  • gds.alpha.sllpa.mutate and gds.alpha.sllpa.mutate.estimate
  • gds.alpha.sllpa.stats and gds.alpha.sllpa.stats.estimate
  • gds.alpha.sllpa.stream and gds.alpha.sllpa.stream.estimate
  • gds.alpha.sllpa.write and gds.alpha.sllpa.write.estimate
• Added CSV export capabilities with the gds.beta.graph.export.csv procedure to allow users to export their in-memory graph to CSV
• Add message reducer capability to Pregel framework to improve memory consumption and computation runtime.
• Added a progress logging procedure with gds.beta.listProgress, to return status of running algorithms. This is turned off by default, but can be enabled with gds.progress_tracking_enabled in the config.
• Add a new BitIdMap data structure to represent node id mappings (Enterprise only)
  • The data structure can lead to a significant reduction in required heap space for an in-memory graph.
  • The data structure is used for native graph projections and in some algorithms, e.g., Louvain.
  • The data structure is not used in Cypher projections.
  • The feature is enabled by default on GDS Enterprise Edition and can be disabled using the USE_BIT_ID_MAP feature toggle.

Bug fixes

• Adding projection parameters as additional configuration in gds.graph.create and gds.graph.create.cypher will throw an exception if improperly configured, instead of being silently ignored.
• Fixed a bug in gds.alpha.articleRank where centrality scores were not normalized correctly
• Fixed a bug in path stream procedures where the path object (path: true) used incorrect node identifiers.
• Fixed a bug in path write procedures where the relationship property nodeIds contained incorrect node identifiers.
• Fixed a race condition that could cause exceptions thrown by scheduled tasks to be supressed.

Improvements

• Improved progress logging to write progress per individual node label in gds.graph.writeNodeProperties.
• When a named graph does not exist, the graph catalog will display similarly named stored graphs.
• When a saved model does not exist, the model catalog will display similarly named stored graphs.
• Added centralityDistribution to the return fields for the write mode of the alpha centrality algorithms.
• gds.beta.graph.generate using relationshipDistribution: 'POWER_LAW' applies the distribution to the native orientation.
• Added centralityDistribution as a return field in gds.betweenness.[write/mutate/stats]
• Added getNeighbours and isMultiGraph to the Pregel-API.
• Added new message queue implementations for the Pregel framework, which
  • replace the previously used JCTools queue and work with primitive double arrays instead of boxed values.
  • lead to 3x to 5x faster runtimes for Pregel based algorithms.
  • reduce GC pressure due to less object allocations which leads to more predictable runtimes.
  • support synchronous and asynchronous Pregel computations.

Other Changes

• The PageRank configuration parameter cacheWeights has been deprecated. The parameter had no effect.
• Deprecated minimumScore, maximumScore, scoreSum return fields in gds.betweenness.[write/mutate/stats]

GDS 1.5 Preview

Release date: 29 January, 2021

Warning: This is a preview release and not intended for production use. If you have any feedback, please let us know: https://github.com/neo4j/graph-data-science/issues

GDS 1.5 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Breaking changes

• Promote several shortest path algorithms to beta tier: Dijkstra, A*, and Yens k-shortest paths. The APIs have been standardized, and all include the ability to return source/target nodes, nodes traversed, and paths.
  • This adds procedures
    • gds.beta.shortestPath.dijkstra.mutate
    • gds.beta.shortestPath.dijkstra.mutate.estimate
    • gds.beta.shortestPath.dijkstra.stream
    • gds.beta.shortestPath.dijkstra.stream.estimate
    • gds.beta.shortestPath.dijkstra.write
    • gds.beta.shortestPath.dijkstra.write.estimate
    • gds.beta.shortestPath.astar.mutate
    • gds.beta.shortestPath.astar.mutate.estimate
    • gds.beta.shortestPath.astar.stream
    • gds.beta.shortestPath.astar.stream.estimate
    • gds.beta.shortestPath.astar.write
    • gds.beta.shortestPath.astar.write.estimate
    • gds.beta.shortestPath.yens.mutate
    • gds.beta.shortestPath.yens.mutate.estimate
    • gds.beta.shortestPath.yens.stream
    • gds.beta.shortestPath.yens.stream.estimate
    • gds.beta.shortestPath.yens.write
    • gds.beta.shortestPath.yens.write.estimate
    • gds.beta.allShortestPaths.dijkstra.mutate
    • gds.beta.allShortestPaths.dijkstra.mutate.estimate
    • gds.beta.allShortestPaths.dijkstra.stream
    • gds.beta.allShortestPaths.dijkstra.stream.estimate
    • gds.beta.allShortestPaths.dijkstra.write
    • gds.beta.allShortestPaths.dijkstra.write.estimate
  • And removes alpha procedures
    • gds.alpha.shortestPath.stream
    • gds.alpha.shortestPath.write
    • gds.alpha.shortestPath.astar.stream
    • gds.alpha.kShortestPaths.stream
    • gds.alpha.kShortestPaths.write
    • gds.alpha.shortestPaths.stream
    • gds.alpha.shortestPaths.write
• GDS will now throw an error when a user tries to use a mutate procedure on graphs not stored in the graph catalog (anonymous graphs)

New Features

• Introduced machine learning based multi-class node classification procedures:
  • Add gds.alpha.ml.nodeClassification.train to train a model to predict a node label
  • Add gds.alpha.ml.nodeClassification.predict.mutate to make predictions using a trained model
• Introduced machine learning based link prediction procedures:
  • Add gds.alpha.linkPrediction.train procedure for training Link Prediction models.
  • Added gds.alpha.linkPrediction.predict.mutate procedure for predicting relationships based on a trained Link Prediction model.
• Added support for list properties as features for
  • gds.alpha.nodeClassification
  • gds.beta.fastRPExtended
  • gds.beta.graphSage
• Added support for storing trained models on disk (Enterprise only)
  • gds.alpha.model.store
  • gds.alpha.model.load
  • gds.alpha.model.delete
• Added procedure for publishing trained models (Enterprise only)
  • Gds.alpha.model.publish
• Added HITS algorithm to the alpha tier
  • gds.alpha.hits.mutate and gds.alpha.hits.mutate.estimate
  • gds.alpha.hits.stats and gds.alpha.hits.stats.estimate
  • gds.alpha.hits.stream and gds.alpha.hits.stream.estimate
  • gds.alpha.hits.write and gds.alpha.hits.write.estimate
• Added Speaker-Listener Label Propagation Algorithm (SLLPA) to the alpha tier
  • gds.alpha.sllpa.mutate and gds.alpha.sllpa.mutate.estimate
  • gds.alpha.sllpa.stats and gds.alpha.sllpa.stats.estimate
  • gds.alpha.sllpa.stream and gds.alpha.sllpa.stream.estimate
  • gds.alpha.sllpa.write and gds.alpha.sllpa.write.estimate
• Added CSV export capabilities with the gds.beta.graph.export.csv procedure to allow users to export their in-memory graph to CSV
• Added a progress logging procedure with gds.beta.listProgress, to return status of running algorithms. This is turned off by default, but can be enabled with gds.progress_tracking_enabled in the config.
• Add message reducer capability to Pregel framework to improve memory consumption and computation runtime.
• Add a new BitIdMap data structure to represent node id mappings (Enterprise only)
  • The data structure can lead to a significant reduction in required heap space for an in-memory graph.
  • The data structure is used for native graph projections and in some algorithms, e.g., Louvain.
  • The data structure is not used in Cypher projections.
  • The feature is enabled by default on GDS Enterprise Edition and can be disabled using the USE_BIT_ID_MAP feature toggle.

Bug fixes

• Adding projection parameters as additional configuration in gds.graph.create and gds.graph.create.cypher will throw an exception if improperly configured, instead of being silently ignored.
• Fixed a bug in gds.alpha.articleRank where centrality scores were not normalized correctly

Improvements

• Improved progress logging to write progress per individual node label in gds.graph.writeNodeProperties.
• When a named graph does not exist, the graph catalog will display similarly named stored graphs.
• When a saved model does not exist, the model catalog will display similarly named stored graphs.
• Add centralityDistribution to the return fields for the write mode of the alpha centrality algorithms.
• gds.beta.graph.generate using relationshipDistribution: 'POWER_LAW' applies the distribution to the native orientation.
• Add getNeighbours and isMultiGraph to the Pregel-API.
• Add centralityDistribution as a return field in gds.betweenness.[write/mutate/stats]

Other Changes

• The PageRank configuration parameter cacheWeights has been deprecated. The parameter had no effect.
• Deprecate minimumScore, maximumScore, scoreSum return fields in gds.betweenness.[write/mutate/stats]

GDS 1.4.1

Release date: 7 December, 2020

GDS 1.4.1 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Bug fixes

• Fixed a bug in progress logging for gds.graph.writeNodeProperties() and gds.graph.writeRelationships() where some percentages were missed, or others reported multiple times.
• Fixed a bug where gds.graph.writeNodeProperties() and gds.alpha.shortestPathDeltaStepping.write() were single threaded by default
• Fixed a bug where gds.alpha.node2vec ignored relationships for graphs with multiple projected relationship types.
• Fixed a bug where gds.pagerank.*.estimate would fail for very large node counts.
• Fixed a bug where using float array node properties (e.g. after running gds.fastRP.mutate) would fail in some situations.
• Fixed a bug where a graph with multiple labels and all nodes sharing at least one label could lead to either an exception or a wrongly mapped Neo4j id.

Improvements

• gds.pageRank will now select batches more dynamically to properly respect the requested concurrency.

1.3.5

Release date: 23 November, 2020

GDS 1.3.5 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x or 4.2. For a 3.5 compatible release, please see GDS 1.1.6. For a 4.2 compatible release, please see GDS 1.4.0

See also 1.3.0 release notes, 1.3.1 release notes, 1.3.2 release notes, 1.3.3 release notes, and 1.3.4 release notes,

Bug fixes

• Fixed a bug in gds.graph.export where at most one relationship property per relationship type would be exported.
• Fixed a bug in Louvain where changes to maxIterations were ignored.
• Fixed a bug where gds.alpha.node2vec would ignore relationships for graphs with multiple projected relationship types.

GDS 1.4.0

Release date: 5 November, 2020

GDS 1.4.0 is compatible with Neo4j 4.0 and 4.1, but not Neo4j 3.5.x. For a 3.5 compatible release, please see GDS 1.1.6.

Breaking changes

• License key configuration was renamed from licenseFile to license_file for consistency with Bloom
• Removed sparsity parameter from gds.alpha.randomProjection.*
• Renamed gds.alpha.randomProjection to gds.fastRP due to productization.
• Renamed embeddingSize parameter to embeddingDimension for fastRP, GraphSAGE and Node2Vec.
• Renamed projectedFeatureSize to projectedFeatureDimension for GraphSAGE
• Renamed nodePropertyNames has been renamed to featureProperties in gds.beta.fastRPExtended and gds.beta.graphSage.train
• Renamed gds.alpha.randomProjection to gds.fastRP due to productization.
• Default parameters for gds.fastRP have changed on the following configuration parameters:
  • iterationWeights now has default [0.0, 1.0, 1.0]
  • normalizeL2 has been removed and its effect is always applied
• Removed alpha procedures for GraphSage (replaced with beta tier, see New Features section)
  • gds.alpha.graphSage.stream
  • gds.alpha.graphSage.write
• GraphSage no longer directly calculates embeddings, instead it has been split into train (to generate a named model) and write, mutate, and stream to apply the model predictions to your data.
• Due to the creation of a train mode for graph sage, the following configuration parameters were moved:
  • embeddingSize - moved as configuration parameter of gds.beta.graphSage.train
  • aggregator - moved as configuration parameter of gds.beta.graphSage.train
  • activationFunction - moved as configuration parameter of gds.beta.graphSage.train
  • sampleSizes - moved as configuration parameter of gds.beta.graphSage.train
  • nodePropertyNames - moved as configuration parameter of gds.beta.graphSage.train
  • tolerance - moved as configuration parameter of gds.beta.graphSage.train
  • learningRate - moved as configuration parameter of gds.beta.graphSage.train
  • epochs - moved as configuration parameter of gds.beta.graphSage.train
  • maxIterations - moved as configuration parameter of gds.beta.graphSage.train
  • searchDepth - moved as configuration parameter of gds.beta.graphSage.train
  • negativeSampleWeight - moved as configuration parameter of gds.beta.graphSage.train
  • degreeAsProperty - moved as configuration parameter of gds.beta.graphSage.train
• gds.beta.graphSage.stream procedure now requires modelName configuration parameter.
• gds.beta.graphSage.write procedure requires modelName configuration parameter.
• Removed startLoss and epochLosses from the result columns of gds.beta.graphSage.write.
• Added the graph create config as a return field to the train procedure, affecting gds.beta.graphSage.train
• Fixed result column name embeddings to embedding in GraphSAGE, to align with the other embeddings.
• Removed configuration parameter maxCost from gds.alpha.bfs/dfs.
• Unlocking the Enterprise Edition of the Graph Data Science library requires a license key. The previous config setting has been removed.
• Removed degreeDistribution from gds.graph.drop return columns.
• gds.pageRank now respects the concurrency setting. It will not run if there is insufficient memory for the given concurrency setting.
• Alpha similarity algorithms no longer accept graph name as a parameter. The algorithm never used the named graph, and now the possibility to specify one is removed.

New features

• Promote GraphSage to beta tier and added support for inductive models with the train mode
  • This adds procedures
    • gds.beta.graphSage.mutate
    • gds.beta.graphSage.mutate.estimate
    • gds.beta.graphSage.stream
    • gds.beta.graphSage.stream.estimate
    • gds.beta.graphSage.train
    • gds.beta.graphSage.train.estimate
    • gds.beta.graphSage.write
    • gds.beta.graphSage.write.estimate
  • And removes alpha procedures
    • gds.alpha.graphSage.stream
    • gds.alpha.graphSage.write
• GraphSage supports relationship weights, driven by relationshipWeightProperty
• GraphSage supports node labels via projectedFeatureSize
• Introduced the model catalog to manage trained models, including:
  • gds.beta.model.exists - a procedure to check if a model exists in the catalog
  • Gds.beta.model.list- list all available models
  • gds.beta.model.drop - removes a model from the catalog
• The Random Projection algorithm has been promoted to the product tier and we have added:
  • gds.fastRP.stats
  • gds.fastRP.mutate
  • gds.fastRP.estimate
  • Added procedures for stats and mutate mode, as well as, estimates for all modes.
• FastRP has been extended to support relationship weights and directions
• FastRP supports integer configuration for iteration weights.
• We’ve added support for node property features for FastRP in the beta namespace with FastRPExtended:
  • gds.beta.fastRPExtended.mutate
  • gds.beta.fastRPExtended.stream
  • gds.beta.fastRPExtended.stats
  • gds.beta.fastRPExtended.write
  • gds.beta.fastRPExtended.mutate.estimate
  • gds.beta.fastRPExtended.stream.estimate
  • gds.beta.fastRPExtended.stats.estimate
  • gds.beta.fastRPExtended.write.estimate
• We’ve added the K-Nearest Neighbors (KNN) algorithm to the beta tier
• gds.beta.knn.mutate and gds.beta.knn.mutate.estimate
• gds.beta.knn.stats and gds.beta.knn.stats.estimate
• gds.beta.knn.stream and gds.beta.knn.stream.estimate
• gds.beta.knn.write and gds.beta.knn.write.estimate
• The in memory graph can now support list properties, enabling embedding results to be stored in memory, or loading embeddings from nodes for KNN or similarity calculations.
• Pregel framework
  • Added Pregel annotation processor to generate GDS procedures for custom Pregel algorithms.
  • Pregel now supports long and double array node values.
  • Add support for composite node state to allow complex data types on nodes.
  • Reduced memory consumption.
  • Improved memory estimation.
  • Simplified message iteration in compute methods.
  • Split context into Init- and ComputeContext and simplified API.
  • Removed K1ColoringExample standalone project.
  • Added pregel-bootstrap standalone project.
  • Added pregel-examples module.
• Licensing: GDS Enterprise edition now requires license keys issued by Neo4j to unlock enterprise features
• Added density property to the output of graph in graph.list.
• Added a failIfMissing flag to gds.graph.drop

Bug fixes

• Pregel:
  • Fixed a bug in Pregel that could lead to incorrect results when running in parallel.
  • Fix cast exception when returning array node properties in generated Pregel procedures.
• Fixed a bug in a multi-source BFS traversal strategy that could affect the following procedures:
  • gds.alpha.closeness
  • gds.alpha.closeness.harmonic
  • gds.alpha.allShortestPaths
• Fixed a bug in gds.alpha.shortestPath.deltaStepping where large relationship weights led to incorrect results
• Weakly connected components:
  • Fixed a bug in WCC where componentCount would be negative when the graph is empty.
  • Fixed a regression where WCC could run more slowly with increased concurrency.
• Fixed bugs in Louvain:
  • communityCount is no longer negative when the graph is empty.
  • changes to maxIterations are no longer ignored.
• Fixed a bug in LabelPropagation where communityCount would be negative when the graph is empty.
• Fixed a bug in KNN where it failed when run on graphs with filtere...

GDS 1.4 Preview

Breaking changes

• Removed sparsity parameter from gds.alpha.randomProjection.*
• Renamed gds.alpha.randomProjection to gds.fastRP due to productization.
• Renamed embeddingSize parameter to embeddingDimension for fastRP, GraphSAGE and Node2Vec.
• Renamed gds.alpha.randomProjection to gds.fastRP due to productization.
• Default parameters for gds.fastRP have changed on the following configuration parameters:
  • iterationWeights now has default [0.0, 1.0, 1.0]
  • normalizeL2 has been removed and its effect is always applied
• Removed alpha procedures for GraphSage (replaced with beta tier, see New Features section)
  • gds.alpha.graphSage.stream
  • gds.alpha.graphSage.write
• GraphSage no longer directly calculates embeddings, instead it has been split into train (to generate a named model) and write, mutate, and stream to apply the model predictions to your data.
• Due to the creation of a train mode for graph sage, the following configuration parameters were moved:
  • embeddingSize - moved as configuration parameter of gds.beta.graphSage.train
  • aggregator - moved as configuration parameter of gds.beta.graphSage.train
  • activationFunction - moved as configuration parameter of gds.beta.graphSage.train
  • sampleSizes - moved as configuration parameter of gds.beta.graphSage.train
  • nodePropertyNames - moved as configuration parameter of gds.beta.graphSage.train
  • tolerance - moved as configuration parameter of gds.beta.graphSage.train
  • learningRate - moved as configuration parameter of gds.beta.graphSage.train
  • epochs - moved as configuration parameter of gds.beta.graphSage.train
  • maxIterations - moved as configuration parameter of gds.beta.graphSage.train
  • searchDepth - moved as configuration parameter of gds.beta.graphSage.train
  • negativeSampleWeight - moved as configuration parameter of gds.beta.graphSage.train
  • degreeAsProperty - moved as configuration parameter of gds.beta.graphSage.train
• gds.beta.graphSage.stream procedure now requires modelName configuration parameter.
• gds.beta.graphSage.write procedure requires modelName configuration parameter.
• Removed startLoss and epochLosses from the result columns of gds.beta.graphSage.write.
• Added the graph create config as a return field to the train procedure, affecting gds.beta.graphSage.train
• Fixed result column name embeddings to embedding in GraphSAGE, to align with the other embeddings.
• Removed configuration parameter maxCost from gds.alpha.bfs/dfs.
• Unlocking the Enterprise Edition of the Graph Data Science library requires a license key. The previous config setting has been removed.
• Removed degreeDistribution from gds.graph.drop return columns.
• gds.pageRank now respects the concurrency setting. It will not run if there is insufficient memory for the given concurrency setting.
• Alpha similarity algorithms no longer accept graph name as a parameter. The algorithm never used the named graph, and now the possibility to specify one is removed.

New features

• Promote GraphSage to beta tier and added support for inductive models with the train mode
  • This adds procedures
    • gds.beta.graphSage.mutate
    • gds.beta.graphSage.mutate.estimate
    • gds.beta.graphSage.stream
    • gds.beta.graphSage.stream.estimate
    • gds.beta.graphSage.train
    • gds.beta.graphSage.train.estimate
    • gds.beta.graphSage.write
    • gds.beta.graphSage.write.estimate
  • And removes alpha procedures
    • gds.alpha.graphSage.stream
    • gds.alpha.graphSage.write
• GraphSage supports relationship weights, driven by relationshipWeightProperty
• GraphSage supports node labels via projectedFeatureSize
• Introduced the model catalog to manage trained models, including:
  • gds.beta.model.exists - a procedure to check if a model exists in the catalog
  • Gds.beta.model.list- list all available models
  • gds.beta.model.drop - removes a model from the catalog
• The Random Projection algorithm has been promoted to the product tier and we have added:
  • gds.fastRP.stats
  • gds.fastRP.mutate
  • gds.fastRP.estimate
  • Added procedures for stats and mutate mode, as well as, estimates for all modes.
• FastRP has been extended to support relationship weights and directions
• FastRP supports integer configuration for iteration weights.
• We’ve added support for node property features for FastRP in the beta namespace with FastRPExtended:
  • gds.beta.fastRPExtended.mutate
  • gds.beta.fastRPExtended.stream
  • gds.beta.fastRPExtended.stats
  • gds.beta.fastRPExtended.write
  • gds.beta.fastRPExtended.mutate.estimate
  • gds.beta.fastRPExtended.stream.estimate
  • gds.beta.fastRPExtended.stats.estimate
  • gds.beta.fastRPExtended.write.estimate
• We’ve added the K-Nearest Neighbors (KNN) algorithm to the beta tier
• gds.beta.knn.mutate and gds.beta.knn.mutate.estimate
• gds.beta.knn.stats and gds.beta.knn.stats.estimate
• gds.beta.knn.stream and gds.beta.knn.stream.estimate
• gds.beta.knn.write and gds.beta.knn.write.estimate
• The in memory graph can now support list properties, enabling embedding results to be stored in memory, or loading embeddings from nodes for KNN or similarity calculations.
• Pregel framework
  • Added Pregel annotation processor to generate GDS procedures for custom Pregel algorithms.
  • Pregel now supports long and double array node values.
  • Add support for composite node state to allow complex data types on nodes.
  • Reduced memory consumption.
  • Improved memory estimation.
  • Simplified message iteration in compute methods.
  • Split context into Init- and ComputeContext and simplified API.
  • Removed K1ColoringExample standalone project.
  • Added pregel-bootstrap standalone project.
  • Added pregel-examples module.
• Licensing: GDS Enterprise edition now requires license keys issued by Neo4j to unlock enterprise features
• Added density property to the output of graph in graph.list.
• Added a failIfMissing flag to gds.graph.drop

Bug fixes

• Pregel:
  • Fixed a bug in Pregel that could lead to incorrect results when running in parallel.
  • Fix cast exception when returning array node properties in generated Pregel procedures.
• Fixed a bug in a multi-source BFS traversal strategy that could affect the following procedures:
  • gds.alpha.closeness
  • gds.alpha.closeness.harmonic
  • gds.alpha.allShortestPaths
• Weakly connected components:
  • Fixed a bug in WCC where componentCount would be negative when the graph is empty.
  • Fixed a regression where WCC could run more slowly with increased concurrency.
• Fixed bugs in Louvain:
  • communityCount is no longer negative when the graph is empty.
  • changes to maxIterations are no longer ignored.
• Fixed a bug in LabelPropagation where communityCount would be negative when the graph is empty.
• Fixed a bug in gds.graph.export where at most one relationship property per relationship type would be exported.
• Graph loading:
  • Fixed a bug where using node label projections including properties on large graphs and high concurrency could lead to loss of some properties.
  • Fixed bug in graph creation which could cause an AIOOB exception during node loading.
  • The readConcurrency config parameter can no longer be overwritten by the concurrency param when it is explicitly set in an implicit graph creation config
• Fixed a bug in memory estimation of large anonymous fictitious graphs.
• Fixed bug in gds.alpha.dfs/bfs, where the algorithm did not terminate for graphs containing loops.
• Fixed result column name embeddings to embedding in GraphSAGE, to align with the other embeddings.
• Fixed a bug in Node2Vec where many disconnected nodes would cause a StackOverflowError
• Fixed a bug in RandomProjection each iteration weight was multiplied all previous iteration weights.
• Similarity algorithms:
  • Fixed a bug where Alpha Similarity algorithms would load a graph even though it was not needed
  • Fixed a bug where similarity algorithms would not remove the placeholder graph if config validation fails on invalid user input.
• Fixed a bug where community statistic computation could overflow for large community ids.
• Fixed a bug where DegreeCentrality returned incorrect values when concurrency > 1.
• Fixed a bug where ClosenessCentrality was using a slightly incorrect formula for Wasserman-Faust algorithm.
• Fixed a bug that affected gds.triangleCount() and gds.alpha.triangles() where not all triangles would be counted under certain conditions.
• Parallel edges in a graph no longer lead to incorrect Local Clustering Coefficient and Triangle Count results.

Improvements

• gds.fastRP now accepts integer iterationWeights
• If graphSage.train is run on a graph without relationships, GDS now fails gracefully with an appropriate error message
• Added validation that properties used by GraphSage exist on graph
• Added validation for <code>embeddingSize</code>>=1
• Added a failIfExists flag to graph creation to enable a user to specify that if a graph already exists, it should be overwritten without failing.
• Progress logging:
  • We now log progress in equally spaced percentages. This is 0-100% either in steps of 1, or in ...

1.4.0-alpha06

Tagging for 1.4.0-alpha06