Struct CCHOneToMany 

pub struct CCHOneToMany<'a> { /* private fields */ }

A reusable one-to-many query with a fixed (pinned) set of targets.

Pinning precomputes the search space of the targets once; afterwards each CCHOneToMany::distances_from call answers distances from a source to all pinned targets in a single elimination-tree sweep, which is much faster than running one point-to-point query per target. Use this for distance tables / matrices (one CCHOneToMany per row) or k-nearest style workloads.

Thread-safety: Send but not Sync; holds mutable per-query state.

Implementations

impl<'a> CCHOneToMany<'a>

pub fn new(metric: &'a CCHMetric<'a>, targets: &[u32]) -> Self

Create a one-to-many query bound to metric with a fixed set of pinned targets. Distances returned later are aligned with the order of targets (duplicates are allowed and answered per entry).

Panics if any target node id is out of range.

pub fn target_count(&self) -> usize

Number of pinned targets (= length of the distance vectors returned).

pub fn repin_targets(&mut self, targets: &[u32])

Replace the pinned target set, reusing the query’s internal O(n) label buffers (cheaper than constructing a new CCHOneToMany).

Panics if any target node id is out of range.

pub fn distances_from(&mut self, source: u32) -> Vec<Option<u32>>

Compute the shortest distances from source to every pinned target. Result is aligned with the pinned target order; None = unreachable.

pub fn distances_from_multi( &mut self, sources: &[(u32, u32)], ) -> Vec<Option<u32>>

Multi-source variant: for each pinned target t, returns min over the given (source, initial_dist) pairs of initial_dist + dist(source, t).

Panics if sources is empty or contains node ids out of range.