Struct faiss::index::scalar_quantizer::IVFScalarQuantizerIndexImpl

pub struct IVFScalarQuantizerIndexImpl<Q> { /* private fields */ }

Native implementation of a scalar quantizer index.

Implementations

impl<Q> IVFScalarQuantizerIndexImpl<Q>where Q: NativeIndex,

pub fn new_with_metric( quantizer: Q, d: u32, qt: QuantizerType, nlist: u32, metric: MetricType, encode_residual: Option<bool> ) -> Result<Self>

Create a new IVF scalar quantizer index with metric. The index owns the quantizer.

pub fn new_l2( quantizer: Q, d: u32, qt: QuantizerType, nlist: u32 ) -> Result<Self>

Create a new IVF scalar quantizer index with L2 metric. The index owns the quantizer.

pub fn new_ip( quantizer: Q, d: u32, qt: QuantizerType, nlist: u32 ) -> Result<Self>

Create a new IVF scalar quantizer index with IP metric. The index owns the quantizer.

pub fn nlist(&self) -> u32

Get number of possible key values

pub fn nprobe(&self) -> u32

Get number of probes at query time

pub fn set_nprobe(&mut self, value: u32)

Set number of probes at query time

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

impl<Q> IVFScalarQuantizerIndexImpl<Q>where Q: NativeIndex + CpuIndex,

pub fn to_gpu<'gpu, G>( &self, gpu_res: &'gpu G, device: i32 ) -> Result<GpuIndexImpl<'gpu, IVFScalarQuantizerIndexImpl<Q>>>where G: GpuResourcesProvider,

Build a GPU in from the given CPU native index, yielding two independent indices. The operation fails if the index does not provide GPU support.

pub fn into_gpu<'gpu, G>( self, gpu_res: &'gpu G, device: i32 ) -> Result<GpuIndexImpl<'gpu, IVFScalarQuantizerIndexImpl<Q>>>where G: GpuResourcesProvider,

Build a GPU index from the given CPU native index, discarding the CPU-backed index. The operation fails if the index does not provide GPU support.

pub fn to_gpu_multiple<'gpu, G>( &self, gpu_res: &'gpu [G], devices: &[i32] ) -> Result<GpuIndexImpl<'gpu, IVFScalarQuantizerIndexImpl<Q>>>where G: GpuResourcesProvider + 'gpu,

Build a GPU index from the given CPU native index.

Errors

The operation fails if the number of GPU resources and number of devices do not match, or the index does not provide GPU support.

pub fn into_gpu_multiple<'gpu, G>( self, gpu_res: &'gpu [G], devices: &[i32] ) -> Result<GpuIndexImpl<'gpu, IVFScalarQuantizerIndexImpl<Q>>>where G: GpuResourcesProvider + 'gpu,

Build a GPU index from the given CPU native index. The index residing in CPU memory is discarded in the process.

Errors

The operation fails if the number of GPU resources and number of devices do not match, or the index does not provide GPU support.

Trait Implementations

impl<Q> ConcurrentIndex for IVFScalarQuantizerIndexImpl<Q>where Q: ConcurrentIndex,

fn assign(&self, query: &[f32], k: usize) -> Result<AssignSearchResult>

Similar to search, but only provides the labels.

fn search(&self, query: &[f32], k: usize) -> Result<SearchResult>

Perform a search for the k closest vectors to the given query vectors.

fn range_search(&self, query: &[f32], radius: f32) -> Result<RangeSearchResult>

Perform a ranged search for the vectors closest to the given query vectors by the given radius.

impl<Q: Debug> Debug for IVFScalarQuantizerIndexImpl<Q>

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

Formats the value using the given formatter.

impl<Q> Drop for IVFScalarQuantizerIndexImpl<Q>

fn drop(&mut self)

Executes the destructor for this type.

impl FromInnerPtr for IVFScalarQuantizerIndexImpl<IndexImpl>

unsafe fn from_inner_ptr(inner_ptr: *mut FaissIndex) -> Self

Create an index using the given pointer to a native object.

impl<Q> Index for IVFScalarQuantizerIndexImpl<Q>

fn is_trained(&self) -> bool

Whether the Index does not require training, or if training is done already

fn ntotal(&self) -> u64

The total number of vectors indexed

fn d(&self) -> u32

The dimensionality of the indexed vectors

fn metric_type(&self) -> MetricType

The metric type assumed by the index

fn add(&mut self, x: &[f32]) -> Result<()>

Add new data vectors to the index. This assumes a C-contiguous memory slice of vectors, where the total number of vectors is x.len() / d.

fn add_with_ids(&mut self, x: &[f32], xids: &[Idx]) -> Result<()>

Add new data vectors to the index with IDs. This assumes a C-contiguous memory slice of vectors, where the total number of vectors is x.len() / d. Not all index types may support this operation.

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

Train the underlying index with the given data.

fn assign(&mut self, query: &[f32], k: usize) -> Result<AssignSearchResult>

Similar to search, but only provides the labels.

fn search(&mut self, query: &[f32], k: usize) -> Result<SearchResult>

Perform a search for the k closest vectors to the given query vectors.

fn range_search( &mut self, query: &[f32], radius: f32 ) -> Result<RangeSearchResult>

Perform a ranged search for the vectors closest to the given query vectors by the given radius.

fn reset(&mut self) -> Result<()>

Clear the entire index.

fn remove_ids(&mut self, sel: &IdSelector) -> Result<usize>

Remove data vectors represented by IDs.

fn verbose(&self) -> bool

Index verbosity level

fn set_verbose(&mut self, value: bool)

Set Index verbosity level

impl<Q> NativeIndex for IVFScalarQuantizerIndexImpl<Q>

fn inner_ptr(&self) -> *mut FaissIndex

Retrieve a pointer to the native index object.

impl<Q> TryClone for IVFScalarQuantizerIndexImpl<Q>

fn try_clone(&self) -> Result<Self>where Self: Sized,

Create an independent clone of this index.

impl<Q: CpuIndex> CpuIndex for IVFScalarQuantizerIndexImpl<Q>

impl<Q: Send> Send for IVFScalarQuantizerIndexImpl<Q>

impl<Q: Sync> Sync for IVFScalarQuantizerIndexImpl<Q>