Enum Engine 

#[non_exhaustive]
pub enum Engine {
    Auto,
    Dense,
    Sparse,
}

Which engine reduces a dense input.

A dense matrix can be reduced as it stands, or converted to the graph of its edges at the threshold and reduced by the sparse engine. The second is faster when few pairs are edges, because the sparse enumerator walks a neighbor list where the dense one scans every vertex. Auto picks between them from the edge density at the resolved threshold and from the memory the conversion would take; Dense and Sparse force one.

The diagram is identical under all three, bit for bit. Every simplex of the complex has diameter at most the threshold, and a diameter is the largest of the edge lengths, so every edge of every simplex survives the conversion. The conversion keeps only edges at or below the threshold. The vertex set carries over because the conversion passes the point count explicitly.

An infinite threshold reads as f64::MAX in both engines. An absent pair has distance +inf, so it enters neither complex, and the conversion keeps exactly the pairs the dense engine admits.

The rule and its constants were frozen on 2026-08-18 from disclosed engineering data. Performance assessment is WIP.

A sparse input is never routed, so this setting does not reach crate::rips_persistence_sparse. With RipsParams::collapse_edges set, the collapse already produces a graph and the sparse engine reduces it.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Auto

Reduce a low-density dense input with the sparse engine, and every other dense input with the dense engine. The rule takes no argument. The conversion has a memory budget: 32 MiB, or the bytes of the compact matrix, whichever is larger. A matrix of mostly absent pairs is low-density at any threshold, including an infinite one, so it routes too.

Dense

Always reduce the distance matrix as it stands.

Sparse

Always convert to the thresholded graph and reduce that. The conversion costs one pass over the matrix. This is an explicit request, so the Auto memory budget does not apply.

Trait Implementations

impl Clone for Engine

fn clone(&self) -> Engine

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for Engine

impl Debug for Engine

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

Formats the value using the given formatter.

impl Default for Engine

fn default() -> Engine

Returns the “default value” for a type.

impl Eq for Engine

impl PartialEq for Engine

fn eq(&self, other: &Engine) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for Engine