K-1 Coloring

This section describes the K-1 Coloring algorithm in the Neo4j Graph Data Science library.

This algorithm is in the beta tier. For more information on algorithm tiers, see Algorithms.

1. Introduction

The K-1 Coloring algorithm assigns a color to every node in the graph, trying to optimize for two objectives:

To make sure that every neighbor of a given node has a different color than the node itself.
To use as few colors as possible.

Note that the graph coloring problem is proven to be NP-complete, which makes it intractable on anything but trivial graph sizes. For that reason the implemented algorithm is a greedy algorithm. Thus it is neither guaranteed that the result is an optimal solution, using as few colors as theoretically possible, nor does it always produce a correct result where no two neighboring nodes have different colors. However the precision of the latter can be controlled by the number of iterations this algorithm runs.

For more information on this algorithm, see:

Running this algorithm requires sufficient memory availability. Before running this algorithm, we recommend that you read Memory Estimation.

2. Syntax

Example 1. K-1 Coloring syntax per mode

The following describes the API for running the algorithm and stream results:

CALL gds.beta.k1coloring.stream(graphName: String, {
  // additional configuration
})
YIELD nodeId, color

Table 1. Parameters
Name	Type	Default	Optional	Description
graphName	String	`null`	yes	The name of an existing graph on which to run the algorithm. If no graph name is provided, the configuration map must contain configuration for creating a graph.
configuration	Map	`{}`	yes	Additional configuration, see below.

Table 2. Configuration
Name	Type	Default	Optional	Description
nodeProjection	String	null	yes	The projection of nodes to use when creating the implicit graph.
relationshipProjection	String	null	yes	The projection of relationships to use when creating the implicit graph.
concurrency	Integer	4	yes	The number of concurrent threads used for running the algorithm. Also provides the default value for 'readConcurrency' and 'writeConcurrency'. This is dependent on the Neo4j edition; for more information, see :/installation/index.adoc#system-requirements-cpu.
readConcurrency	Integer	value of 'concurrency'	yes	The number of concurrent threads used for reading the graph.
maxIterations	Integer	10	yes	The maximum number of iterations of K1 Coloring to run.

Table 3. Results
Name	Type	Description
nodeId	Integer	The ID of the Node
color	Integer	The color of the Node

The following describes the API for running the algorithm and returning the computation statistics:

CALL gds.beta.k1coloring.stats(
    graphName: String,
    configuration: Map
)
YIELD
    nodes,
    colorCount,
    ranIterations,
    didConverge,
    configuration,
    createMillis,
    computeMillis

Table 4. Parameters
Name	Type	Default	Optional	Description
graphNameOrConfig	String or Map	`null`	no	Either the name of a graph stored in the catalog or a Map configuring the graph creation and algorithm execution.
configuration	Map	`{}`	yes	Configuration for algorithm-specifics and/or graph filtering. Must be empty if graphNameOrConfig is a Map.

Table 5. Configuration
Name	Type	Default	Optional	Description
nodeProjection	String	null	yes	The projection of nodes to use when creating the implicit graph.
relationshipProjection	String	null	yes	The projection of relationships to use when creating the implicit graph.
concurrency	Integer	4	yes	The number of concurrent threads used for running the algorithm. Also provides the default value for 'readConcurrency' and 'writeConcurrency'. This is dependent on the Neo4j edition; for more information, see :/installation/index.adoc#system-requirements-cpu.
readConcurrency	Integer	value of 'concurrency'	yes	The number of concurrent threads used for reading the graph.
maxIterations	Integer	10	yes	The maximum number of iterations of K1 Coloring to run.

Table 6. Results
Name	Type	Description
nodeCount	Integer	The number of nodes considered.
ranIterations	Integer	The actual number of iterations the algorithm ran.
didConverge	Boolean	An indicator of whether the algorithm found a correct coloring.
colorCount	Integer	The number of colors used.
createMillis	Integer	Milliseconds for loading data.
computeMillis	Integer	Milliseconds for running the algorithm.
configuration	Map	The configuration used for running the algorithm.

The following describes the API for running the algorithm and mutating the in-memory graph:

CALL gds.beta.k1coloring.mutate(graphName: String, configuration: Map)
YIELD nodes, colorCount, ranIterations, didConverge, configuration, createMillis, computeMillis, mutateMillis

The configuration for the mutate mode is similar to the write mode. Instead of specifying a writeProperty, we need to specify a mutateProperty. Also, specifying writeConcurrency is not possible in mutate mode.

The following will run the algorithm and store the results in myGraph:

CALL gds.beta.k1coloring.mutate('myGraph', { mutateProperty: 'color' })

The following describes the API for running the algorithm and writing results back to Neo4j:

CALL gds.beta.k1coloring.write(graphName: String, configuration: Map)
YIELD nodes, colorCount, ranIterations, didConverge, configuration, createMillis, computeMillis, writeMillis

Table 7. Parameters
Name	Type	Default	Optional	Description
graphNameOrConfig	String or Map	`null`	no	Either the name of a graph stored in the catalog or a Map configuring the graph creation and algorithm execution.
configuration	Map	`{}`	yes	Configuration for algorithm-specifics and/or graph filtering. Must be empty if graphNameOrConfig is a Map.

Table 8. Configuration
Name	Type	Default	Optional	Description
nodeProjection	String	null	yes	The projection of nodes to use when creating the implicit graph.
relationshipProjection	String	null	yes	The projection of relationships to use when creating the implicit graph.
concurrency	Integer	4	yes	The number of concurrent threads used for running the algorithm. Also provides the default value for 'readConcurrency' and 'writeConcurrency'. This is dependent on the Neo4j edition; for more information, see :/installation/index.adoc#system-requirements-cpu.
readConcurrency	Integer	value of 'concurrency'	yes	The number of concurrent threads used for reading the graph.
writeConcurrency	Integer	value of 'concurrency'	yes	The number of concurrent threads used for writing the result.
maxIterations	Integer	10	yes	The maximum number of iterations of K1 Coloring to run.
writeProperty	String	n/a	no	The node property this procedure writes the color to.

Table 9. Results
Name	Type	Description
nodeCount	Integer	The number of nodes considered.
ranIterations	Integer	The actual number of iterations the algorithm ran.
didConverge	Boolean	An indicator of whether the algorithm found a correct coloring.
colorCount	Integer	The number of colors used.
write	Boolean	Specifies if the result was written back as a node property.
createMillis	Integer	Milliseconds for loading data.
computeMillis	Integer	Milliseconds for running the algorithm.
writeMillis	Integer	Milliseconds for writing result data back to Neo4j.
configuration	Map	The configuration used for running the algorithm.

3. Examples

Consider the graph created by the following Cypher statement:

CREATE (alice:User {name: 'Alice'}),
       (bridget:User {name: 'Bridget'}),
       (charles:User {name: 'Charles'}),
       (doug:User {name: 'Doug'}),

       (alice)-[:LINK]->(bridget),
       (alice)-[:LINK]->(charles),
       (alice)-[:LINK]->(doug),
       (bridget)-[:LINK]->(charles)

This graph has a super node with name "Alice" that connects to all other nodes. It should therefore not be possible for any other node to be assigned the same color as the Alice node.

CALL gds.graph.create(
    'myGraph',
    'User',
    {
        LINK : {
            orientation: 'UNDIRECTED'
        }
    }
)

We can now go ahead and create an in-memory graph with all the User nodes and the LINK relationships with UNDIRECTED orientation.

In the examples below we will use named graphs and standard projections as the norm. However, Cypher projection and anonymous graphs could also be used.

The following statement will create the graph and store it in the graph catalog.

CALL gds.graph.create('myGraph', 'Person', 'LIKES')

In the following examples we will demonstrate using the K-1 Coloring algorithm on this graph.

Running the K-1 Coloring algorithm in stream mode:

CALL gds.beta.k1coloring.stream('myGraph')
YIELD nodeId, color
RETURN gds.util.asNode(nodeId).name AS name, color
ORDER BY name

Table 10. Results
name	color
`"Alice"`	`0`
`"Bridget"`	`1`
`"Charles"`	`2`
`"Doug"`	`1`

It is also possible to write the assigned colors back to the database using the write mode.

Running the K-1 Coloring algorithm in write mode:

CALL gds.beta.k1coloring.write('myGraph', {writeProperty: 'color'})
YIELD nodeCount, colorCount, ranIterations, didConverge

Table 11. Results
nodeCount	colorCount	ranIterations	didConverge
`4`	`3`	`1`	`true`

When using write mode the procedure will return information about the algorithm execution. In this example we return the number of processed nodes, the number of colors used to color the graph, the number of iterations and information whether the algorithm converged.

To instead mutate the in-memory graph with the assigned colors, the mutate mode can be used as follows.

Running the K-1 Coloring algorithm in mutate mode:

CALL gds.beta.k1coloring.mutate('myGraph', {mutateProperty: 'color'})
YIELD nodeCount, colorCount, ranIterations, didConverge

Table 12. Results
nodeCount	colorCount	ranIterations	didConverge
`4`	`3`	`1`	`true`

Similar to the write mode, stats mode can run the algorithm and return only the execution statistics without persisting the results.

Running the K-1 Coloring algorithm in stats mode:

CALL gds.beta.k1coloring.stats('myGraph')
YIELD nodeCount, colorCount, ranIterations, didConverge

Table 13. Results
nodeCount	colorCount	ranIterations	didConverge
`4`	`3`	`1`	`true`

Was this page helpful?

Filter By Category

K-1 Coloring

1. Introduction

2. Syntax

3. Examples