Struct SparseFeatures 

pub struct SparseFeatures { /* private fields */ }

Sparse feature vector: sorted (FeatureId, value) pairs.

Requirements

• Sorted by FeatureId in ascending order.
• No duplicate FeatureIds.
• Zero values may be omitted.
• Not a HashMap<String, f64>.

Examples

use rill_ml::sparse::SparseFeatures;

let features = SparseFeatures::from_sorted(vec![
    (1, 0.5),
    (3, 2.0),
    (7, -1.0),
]).unwrap();
assert_eq!(features.len(), 3);
assert_eq!(features.get(3), Some(2.0));

Implementations

impl SparseFeatures

pub fn new() -> Self

Create an empty sparse feature vector.

pub fn from_sorted(values: Vec<(FeatureId, f64)>) -> Result<Self, RillError>

Build from a Vec that is already sorted by FeatureId and contains no duplicates.

Returns an error if the input is unsorted, has duplicates, or contains non-finite values.

pub fn from_unsorted(values: Vec<(FeatureId, f64)>) -> Result<Self, RillError>

Build from an unsorted Vec, sorting and merging duplicates by summing their values.

Non-finite values are rejected after merging.

pub fn push(&mut self, id: FeatureId, value: f64) -> Result<(), RillError>

Append a single feature. The id must be strictly greater than the last inserted id.

pub fn values(&self) -> &[(FeatureId, f64)]

Access the internal (FeatureId, value) slice.

pub fn len(&self) -> usize

Number of non-zero (explicitly stored) features.

pub fn is_empty(&self) -> bool

Whether no features are stored.

pub fn get(&self, id: FeatureId) -> Option<f64>

Look up the value for a given FeatureId via binary search.

pub fn iter(&self) -> impl Iterator<Item = (FeatureId, f64)> + '_

Iterate over (FeatureId, &value) pairs.

pub fn validate(&self) -> Result<(), RillError>

Validate sorting, uniqueness, and finiteness.

Trait Implementations

impl Clone for SparseFeatures

fn clone(&self) -> SparseFeatures

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for SparseFeatures

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

Formats the value using the given formatter.

impl Default for SparseFeatures

fn default() -> SparseFeatures

Returns the “default value” for a type.