Struct Shells 

pub struct Shells;

Concentric L-infinity shells from a chosen origin.

Trait Implementations

impl Strategy for Shells

fn selects_from_shape(&self) -> bool

The shells are distances in the index space.

fn name(&self) -> StrategyName

The strategy’s name. Mirrors StrategyName.

fn accepts_input(&self, idx: Option<&IndexFn>) -> bool

V4 input-shape check (comprehension_forms.md §3.6). None represents an input with no closed-form index function; only Lex accepts that. Concrete IndexFn variants are accepted per the per-strategy rules in §3.6’s table.

fn has_closed_form_for(&self, idx: &IndexFn) -> bool

R2 push-down eligibility (§10.2 R2). true if this strategy has a closed-form multi-index rule over the given input; otherwise select orders the input’s positions as one axis.

fn select( &self, index_fn: &IndexFn, cardinality: u64, truncation: Option<u64>, _seed: Option<u64>, ) -> Selection

The positions this strategy emits over an input of cardinality tuples addressed by index_fn, cut to truncation, under the authored seed (comprehension_forms.md §3.6: a seeded strategy, Shuffle or Lhs, derives its state from the seed and the input’s structural identity, and from its fixed default when seed is None; every other strategy ignores it).

fn select_surviving( &self, index_fn: &IndexFn, cardinality: u64, truncation: Option<u64>, seed: Option<u64>, survivors: &[u64], ) -> Selection

The positions this strategy emits over an input of which only the positions in survivors (ascending) pass a filter (comprehension_forms.md §5 V5): the strategy selects from the input’s whole index space, keeps the survivors in the order it emits them, and applies its truncation to them. The positions are the survivors’ original positions in the input, so order(filter(c, p), halton, n) yields n survivors whenever at least n exist, and when every tuple survives the selection is select’s.

fn apply(&self, input: &EvaluatedInput, truncation: Option<u64>) -> Vec<Tuple>

Apply this strategy to the given input: its selection looked up against input.tuples.

fn apply_seeded( &self, input: &EvaluatedInput, truncation: Option<u64>, seed: Option<u64>, ) -> Vec<Tuple>

apply under an authored seed.