Package org.apache.tinkerpop.gremlin.algorithm.generator

Interface Distribution

All Known Implementing Classes:
CopyDistribution, NormalDistribution, PowerLawDistribution
public interface Distribution
Interface for a distribution over discrete values.

Used, for instance, by DistributionGenerator to define the in- and out-degree distributions and by CommunityGenerator to define the community size distribution.

Author:
Matthias Broecheler (me@matthiasb.com)

  • Method Summary

    Modifier and Type
    Method
    Description
    Distribution
    initialize(int invocations, int expectedTotal)
    Initializes the distribution such that expectedTotal is equal to the expected sum of generated values after the given number of invocatiosn.
    int
    nextConditionalValue(Random random, int otherValue)
    Returns the next value conditional on another given value.
    int
    nextValue(Random random)
    Returns the next value.

  • Method Details

    • initialize

      Distribution initialize(int invocations, int expectedTotal)
      Initializes the distribution such that expectedTotal is equal to the expected sum of generated values after the given number of invocatiosn.

      Since most distributions have an element of randomness, these values are the expected values.

      Returns:
      A new distribution configured to match the expected total for the number of invocations.

    • nextValue

      int nextValue(Random random)
      Returns the next value. If this value is randomly generated, the randomness must be drawn from the provided random generator.

      DO NOT use your own internal random generator as this makes the generated values non-reproducible and leads to faulty behavior.

      Parameters:
      random - random generator to use for randomness
      Returns:
      next value

    • nextConditionalValue

      int nextConditionalValue(Random random, int otherValue)
      Returns the next value conditional on another given value.

      This can be used, for instance, to define conditional degree distributions where the in-degree is conditional on the out-degree.

      If this value is randomly generated, the randomness must be drawn from the provided random generator. DO NOT use your own internal random generator as this makes the generated values non-reproducible and leads to faulty behavior.

      Parameters:
      random - random generator to use for randomness
      otherValue - The prior value
      Returns:
      next value