Struct bigs::Sampler

pub struct Sampler { /* private fields */ }

A sampler for regular bipartite graph.

This is the main struct of this crate. The builder method is used to build a sampler and the sample_with method is used to sample a random Graph.

A sampler is specified by 4 parameters: - variable’s degree: the same for all variables, - constraint’s degree: the same for all contraints, - number of variables, - number of contraints.

To make sure that the graphs are regular, the number of variables times their degree need to equal the number of contraints times their degree. If this is not satified, the builder will panic during construction.

Example

This can be used to generate a random graph with 5 variables of degree 3 and 3 constraints of degree 5.

use rand::thread_rng;

let sampler = Sampler::builder()
    .number_of_variables(5)
    .variable_degree(3)
    .number_of_constraints(3)
    .constraint_degree(5)
    .build()
    .unwrap();

let graph = sampler.sample_with(&mut thread_rng());

assert_eq!(graph.number_of_variables(), 5);
for variable in graph.variables() {
    assert_eq!(variable.degree(), 3);
}

assert_eq!(graph.number_of_constraints(), 3);
for constraint in graph.constraints() {
    assert_eq!(constraint.degree(), 5);
    }

However, this will return an error since the parameters do not define a regular graph.

let sampler = Sampler::builder()
    .number_of_variables(5)
    .variable_degree(3)
    .number_of_constraints(3)
    .constraint_degree(3)
    .build();

assert!(sampler.is_err());

Reproductibility

If you use the same random number generator, the graphs should be the same.

use rand::SeedableRng;
use rand::rngs::SmallRng; // require the small_rng feature

let sampler = Sampler::builder()
    .number_of_variables(5)
    .variable_degree(3)
    .number_of_constraints(3)
    .constraint_degree(5)
    .build()
    .unwrap();

// Create two rngs with the same seed.
let mut rng = SmallRng::seed_from_u64(123);
let mut other_rng = SmallRng::seed_from_u64(123);

let graph = sampler.sample_with(&mut rng);
let other_graph = sampler.sample_with(&mut other_rng);

assert_eq!(graph, other_graph);

Implementations

impl Sampler

pub fn builder() -> Builder

Instanciates a builder for samplers.

pub fn sample_with<R: Rng>(&self, rng: &mut R) -> Graph

Samples a random graph with the sampler parameters.

pub fn number_of_variables(&self) -> usize

Returns the number of variables in the graphs that will be generated by the sampler.

pub fn number_of_constraints(&self) -> usize

Returns the number of constraints in the graphs that will be generated by the sampler.

pub fn number_of_edges(&self) -> usize

Returns the number of edges in the graphs that will be generated by the sampler.

pub fn variable_degree(&self) -> usize

Returns the variable’s degree in the graphs that will be generated by the sampler.

pub fn constraint_degree(&self) -> usize

Returns the contraint’s degree in the graphs that will be generated by the sampler.

Trait Implementations

impl Debug for Sampler

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.