Struct nblast::NblastArena

pub struct NblastArena<N>
where
    N: TargetNeuron,
{ /* private fields */ }

Struct for caching a number of neurons for multiple comparable NBLAST queries.

Implementations

impl<N> NblastArena<N>

where N: TargetNeuron + Sync,

pub fn new(score_calc: ScoreCalc, use_alpha: bool) -> Self

By default, runs in serial. See NblastArena.with_threads if the “parallel” feature is enabled.

pub fn size_of(&self, idx: NeuronIdx) -> Option<usize>

pub fn add_neuron(&mut self, neuron: N) -> NeuronIdx

Returns an index which is then used to make queries.

pub fn query_target( &self, query_idx: NeuronIdx, target_idx: NeuronIdx, normalize: bool, symmetry: &Option<Symmetry> ) -> Option<Precision>

Make a single query using the given indexes. normalize divides the result by the self-hit score of the query neuron. symmetry, if Some, also calculates the reverse score (normalizing it if necessary), and then applies a function to ensure that the query is symmetric/ commutative.

pub fn queries_targets( &self, query_idxs: &[NeuronIdx], target_idxs: &[NeuronIdx], normalize: bool, symmetry: &Option<Symmetry>, max_centroid_dist: Option<Precision> ) -> HashMap<(NeuronIdx, NeuronIdx), Precision>

Make many queries using the Cartesian product of the query and target indices.

See query_target for details on normalize and symmetry. Neurons whose centroids are greater than max_centroid_dist apart will return NaN. Indices which do not exist will be silently ignored.

pub fn query_target_pairs( &self, query_target_idxs: &[(NeuronIdx, NeuronIdx)], normalize: bool, symmetry: &Option<Symmetry>, max_centroid_dist: Option<Precision> ) -> HashMap<(NeuronIdx, NeuronIdx), Precision>

See query_target for details on normalize and symmetry. Neurons whose centroids are greater than max_centroid_dist apart will return NaN. Indices which do not exist will be silently ignored.

pub fn centroids_within_distance( &self, query_idx: NeuronIdx, target_idx: NeuronIdx, max_centroid_dist: Precision ) -> Option<bool>

pub fn self_hit(&self, idx: NeuronIdx) -> Option<Precision>

pub fn all_v_all( &self, normalize: bool, symmetry: &Option<Symmetry>, max_centroid_dist: Option<Precision> ) -> HashMap<(NeuronIdx, NeuronIdx), Precision>

Query every neuron against every other neuron. See queries_targets for more details.

pub fn is_empty(&self) -> bool

pub fn len(&self) -> usize

Number of neurons in the arena.

pub fn points( &self, idx: NeuronIdx ) -> Option<impl Iterator<Item = Point3> + '_>

pub fn tangents( &self, idx: NeuronIdx ) -> Option<impl Iterator<Item = Normal3> + '_>

pub fn alphas( &self, idx: NeuronIdx ) -> Option<impl Iterator<Item = Precision> + '_>