Struct Clustering 

pub struct Clustering { /* private fields */ }

The clustering algorithm.

Implementations

impl Clustering

pub fn new(d: u32, k: u32) -> Result<Self>

Obtain a new clustering object with the given dimensionality d and number of centroids k.

pub fn new_with_params( d: u32, k: u32, params: &ClusteringParameters, ) -> Result<Self>

Obtain a new clustering object, with the given clustering parameters.

pub fn train<I>(&mut self, x: &[f32], index: &mut I) -> Result<()>

where I: ?Sized + NativeIndex<Inner = FaissIndex>,

Perform the clustering algorithm with the given data and index. The index is used during the assignment stage.

pub fn centroids(&self) -> Result<Vec<&[f32]>>

Retrieve the centroids from the clustering process. Returns a vector of k slices of size d.

pub fn centroids_mut(&mut self) -> Result<Vec<&mut [f32]>>

Retrieve the centroids from the clustering process. Returns a vector of k slices of size d.

pub fn iteration_stats(&self) -> &[ClusteringIterationStats]

Retrieve the stats achieved from the clustering process. Returns as many values as the number of iterations made.

pub fn iteration_stats_mut(&mut self) -> &mut [ClusteringIterationStats]

Retrieve the stats. Returns as many values as the number of iterations made.

pub fn d(&self) -> u32

Getter for the clustering object’s vector dimensionality.

pub fn k(&self) -> u32

Getter for the number of centroids.

pub fn niter(&self) -> u32

Getter for the number of k-means iterations.

pub fn nredo(&self) -> u32

Getter for the nredo property of Clustering.

pub fn verbose(&self) -> bool

Getter for the verbose property of Clustering.

pub fn spherical(&self) -> bool

Getter for whether spherical clustering is intended.

pub fn int_centroids(&self) -> bool

Getter for the int_centroids property of Clustering. Round centroids coordinates to integer

pub fn update_index(&self) -> bool

Getter for the update_index property of Clustering.

pub fn frozen_centroids(&self) -> bool

Getter for the frozen_centroids property of Clustering.

pub fn seed(&self) -> u32

Getter for the seed property of Clustering.

pub fn decode_block_size(&self) -> usize

Getter for the decode_block_size property of Clustering. How many vectors at a time to decode

pub fn min_points_per_centroid(&self) -> u32

Getter for the minimum number of points per centroid.

pub fn max_points_per_centroid(&self) -> u32

Getter for the maximum number of points per centroid.

Trait Implementations

impl Drop for Clustering

fn drop(&mut self)

Executes the destructor for this type.

impl Send for Clustering

impl Sync for Clustering