HNSW

This is a wrapper for hnswlib, to load a CAGRA index as an immutable HNSW index. The loaded HNSW index is only compatible in cuVS, and can be searched using wrapper functions.

#include <raft/neighbors/hnsw.h>

Index search parameters

typedef struct cuvsHnswSearchParams *cuvsHnswSearchParams_t

cuvsError_t cuvsHnswSearchParamsCreate(cuvsHnswSearchParams_t *params)

Allocate HNSW search params, and populate with default values.

Parameters:

params – [in] cuvsHnswSearchParams_t to allocate

Returns:

cuvsError_t

cuvsError_t cuvsHnswSearchParamsDestroy(cuvsHnswSearchParams_t params)

De-allocate HNSW search params.

Parameters:

params – [in] cuvsHnswSearchParams_t to de-allocate

Returns:

cuvsError_t

struct cuvsHnswSearchParams

#include <hnsw.h>

Index

typedef cuvsHnswIndex *cuvsHnswIndex_t

cuvsError_t cuvsHnswIndexCreate(cuvsHnswIndex_t *index)

Allocate HNSW index.

Parameters:

index – [in] cuvsHnswIndex_t to allocate

Returns:

HnswError_t

cuvsError_t cuvsHnswIndexDestroy(cuvsHnswIndex_t index)

De-allocate HNSW index.

Parameters:

index – [in] cuvsHnswIndex_t to de-allocate

struct cuvsHnswIndex

#include <hnsw.h>

Struct to hold address of cuvs::neighbors::Hnsw::index and its active trained dtype.

Index search

cuvsError_t cuvsCagraSearch(cuvsResources_t res, cuvsCagraSearchParams_t params, cuvsCagraIndex_t index, DLManagedTensor *queries, DLManagedTensor *neighbors, DLManagedTensor *distances)

Search a CAGRA index with a DLManagedTensor which has underlying DLDeviceType equal to kDLCUDA, kDLCUDAHost, kDLCUDAManaged. It is also important to note that the CAGRA Index must have been built with the same type of queries, such that index.dtype.code == queries.dl_tensor.dtype.code Types for input are:

1. queries: a. kDLDataType.code == kDLFloat and kDLDataType.bits = 32 b. kDLDataType.code == kDLInt and kDLDataType.bits = 8 c. kDLDataType.code == kDLUInt and kDLDataType.bits = 8

2. neighbors: kDLDataType.code == kDLUInt and kDLDataType.bits = 32

3. distances: kDLDataType.code == kDLFloat and kDLDataType.bits = 32

#include <cuvs/core/c_api.h>
#include <cuvs/neighbors/cagra.h>

// Create cuvsResources_t
cuvsResources_t res;
cuvsError_t res_create_status = cuvsResourcesCreate(&res);

// Assume a populated `DLManagedTensor` type here
DLManagedTensor dataset;
DLManagedTensor queries;
DLManagedTensor neighbors;

// Create default search params
cuvsCagraSearchParams_t params;
cuvsError_t params_create_status = cuvsCagraSearchParamsCreate(&params);

// Search the `index` built using `cuvsCagraBuild`
cuvsError_t search_status = cuvsCagraSearch(res, params, index, &queries, &neighbors,
&distances);

// de-allocate `params` and `res`
cuvsError_t params_destroy_status = cuvsCagraSearchParamsDestroy(params);
cuvsError_t res_destroy_status = cuvsResourcesDestroy(res);

Parameters:

• res – [in] cuvsResources_t opaque C handle

• params – [in] cuvsCagraSearchParams_t used to search CAGRA index

• index – [in] cuvsCagraIndex which has been returned by cuvsCagraBuild

• queries – [in] DLManagedTensor* queries dataset to search

• neighbors – [out] DLManagedTensor* output k neighbors for queries

• distances – [out] DLManagedTensor* output k distances for queries

Index serialize

Warning

doxygengroup: Cannot find group “hnsw_c_index_serialize” in doxygen xml output for project “cuvs” from directory: ../../cpp/doxygen/_xml/