Struct HyperBallBuilder

pub struct HyperBallBuilder<'a, G1: RandomAccessGraph + Sync, G2: RandomAccessGraph + Sync, D: Succ<Input = usize, Output = usize>, L: MergeEstimationLogic<Item = G1::Label>, A: EstimatorArrayMut<L>> { /* private fields */ }

A builder for HyperBall.

After creating a builder with HyperBallBuilder::new you can configure it using setters such as HyperBallBuilder its methods, then call HyperBallBuilder::build on it to create a HyperBall instance.

Implementations

impl<'a, G1: RandomAccessGraph + Sync, G2: RandomAccessGraph + Sync, D: Succ<Input = usize, Output = usize>> HyperBallBuilder<'a, G1, G2, D, HyperLogLog<G1::Label, BuildHasherDefault<DefaultHasher>, usize>, SliceEstimatorArray<HyperLogLog<G1::Label, BuildHasherDefault<DefaultHasher>, usize>, usize, Box<[usize]>>>

pub fn with_hyper_log_log( graph: &'a G1, transposed: Option<&'a G2>, cumul_outdeg: &'a D, log2m: usize, weights: Option<&'a [usize]>, ) -> Result<Self>

A builder for HyperBall using a specified EstimationLogic.

Arguments

• graph: the graph to analyze.
• transpose: optionally, the transpose of graph. If None, no systolic iterations will be performed by the resulting HyperBall.
• cumul_outdeg: the outdegree cumulative function of the graph.
• log2m: the base-2 logarithm of the number m of register per HyperLogLog cardinality estimator.
• weights: the weights to use. If None every node is assumed to be of weight equal to 1.

impl<'a, D: Succ<Input = usize, Output = usize>, G: RandomAccessGraph + Sync, L: MergeEstimationLogic<Item = G::Label> + PartialEq, A: EstimatorArrayMut<L>> HyperBallBuilder<'a, G, G, D, L, A>

pub fn new(graph: &'a G, cumul_outdeg: &'a D, array_0: A, array_1: A) -> Self

Creates a new builder with default parameters.

Arguments

• graph: the graph to analyze.
• cumul_outdeg: the outdegree cumulative function of the graph.
• array_0: a first array of estimators.
• array_1: A second array of estimators of the same length and with the same logic of array_0.

impl<'a, G1: RandomAccessGraph + Sync, G2: RandomAccessGraph + Sync, D: Succ<Input = usize, Output = usize>, L: MergeEstimationLogic<Item = G1::Label>, A: EstimatorArrayMut<L>> HyperBallBuilder<'a, G1, G2, D, L, A>

pub fn with_transpose( graph: &'a G1, transpose: &'a G2, cumul_outdeg: &'a D, array_0: A, array_1: A, ) -> Self

Creates a new builder with default parameters using also the transpose.

• graph: the graph to analyze.
• transpose: optionally, the transpose of graph. If None, no systolic iterations will be performed by the resulting HyperBall.
• cumul_outdeg: the outdegree cumulative function of the graph.
• array_0: a first array of estimators.
• array_1: A second array of estimators of the same length and with the same logic of array_0.

pub fn sum_of_distances(self, do_sum_of_distances: bool) -> Self

Sets whether to compute the sum of distances.

pub fn sum_of_inverse_distances(self, do_sum_of_inverse_distances: bool) -> Self

Sets whether to compute the sum of inverse distances.

pub fn granularity(self, granularity: Granularity) -> Self

Sets the base granularity used in the parallel phases of the iterations.

pub fn weights(self, weights: Option<&'a [usize]>) -> Self

Sets optional weights for the nodes of the graph.

Arguments

• weights: weights to use for the nodes. If None, every node is assumed to be of weight equal to 1.

pub fn discount_function( self, discount_function: impl Fn(usize) -> f64 + Sync + 'a, ) -> Self

Adds a new discount function whose sum over all spheres should be computed.

pub fn no_discount_function(self) -> Self

Removes all custom discount functions.

impl<'a, G1: RandomAccessGraph + Sync, G2: RandomAccessGraph + Sync, D: Succ<Input = usize, Output = usize>, L: MergeEstimationLogic<Item = G1::Label> + Sync + Display, A: EstimatorArrayMut<L>> HyperBallBuilder<'a, G1, G2, D, L, A>

pub fn build( self, pl: &mut impl ConcurrentProgressLog, ) -> HyperBall<'a, G1, G2, D, L, A>

Builds a HyperBall instance.

Arguments

• pl: A progress logger.