Neighbors Cagra Module
Rust module: cuvs::neighbors::cagra
Source: rust/cuvs/src/neighbors/cagra/mod.rs
CAGRA: a graph-based approximate nearest neighbors algorithm with state-of-the-art query throughput for both small and large batch sizes.
Build an [Index] from a dataset, then search it with
device-resident queries and output buffers. Tensors are passed through the
AsDlTensor / AsDlTensorMut traits; see the dlpack
module for the tensor model and examples/cagra.rs for a complete, runnable
example.
Parameter types ([IndexParams], [SearchParams], …) use the [bon]
builder pattern: every setter is optional and unset values keep the cuVS C
library defaults. Values are validated when the builder’s build() runs,
returning [CagraError::Validation] for out-of-range inputs.
crate::neighbors::filters::{Bitset, Filter}
Source: rust/cuvs/src/neighbors/cagra/mod.rs:23
index::Index
Source: rust/cuvs/src/neighbors/cagra/mod.rs:24
params::{CompressionParams, IndexParams, SearchParams}
Source: rust/cuvs/src/neighbors/cagra/mod.rs:25
GraphBuildAlgo
Algorithm for building the internal k-NN graph.
Source: rust/cuvs/src/neighbors/cagra/mod.rs:33
SearchAlgo
Search kernel implementation.
Source: rust/cuvs/src/neighbors/cagra/mod.rs:73
HashMode
Hash-table mode used during search.
Source: rust/cuvs/src/neighbors/cagra/mod.rs:109
CagraError
Error type for CAGRA operations.
Source: rust/cuvs/src/neighbors/cagra/mod.rs:141