feat: expose hnsw indices (#1595)

PR closes #1522

---------

Co-authored-by: Will Jones <willjones127@gmail.com>
This commit is contained in:
Gagan Bhullar
2024-09-10 12:08:13 -06:00
committed by GitHub
parent 2bde5401eb
commit 205fc530cf
6 changed files with 291 additions and 3 deletions

View File

@@ -16,7 +16,11 @@ use std::sync::Mutex;
use lancedb::index::scalar::FtsIndexBuilder;
use lancedb::{
index::{scalar::BTreeIndexBuilder, vector::IvfPqIndexBuilder, Index as LanceDbIndex},
index::{
scalar::BTreeIndexBuilder,
vector::{IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder},
Index as LanceDbIndex,
},
DistanceType,
};
use pyo3::{
@@ -24,6 +28,8 @@ use pyo3::{
pyclass, pymethods, PyResult,
};
use crate::util::parse_distance_type;
#[pyclass]
pub struct Index {
inner: Mutex<Option<LanceDbIndex>>,
@@ -110,6 +116,78 @@ impl Index {
inner: Mutex::new(Some(LanceDbIndex::FTS(opts))),
}
}
#[staticmethod]
pub fn hnsw_pq(
distance_type: Option<String>,
num_partitions: Option<u32>,
num_sub_vectors: Option<u32>,
max_iterations: Option<u32>,
sample_rate: Option<u32>,
m: Option<u32>,
ef_construction: Option<u32>,
) -> PyResult<Self> {
let mut hnsw_pq_builder = IvfHnswPqIndexBuilder::default();
if let Some(distance_type) = distance_type {
let distance_type = parse_distance_type(distance_type)?;
hnsw_pq_builder = hnsw_pq_builder.distance_type(distance_type);
}
if let Some(num_partitions) = num_partitions {
hnsw_pq_builder = hnsw_pq_builder.num_partitions(num_partitions);
}
if let Some(num_sub_vectors) = num_sub_vectors {
hnsw_pq_builder = hnsw_pq_builder.num_sub_vectors(num_sub_vectors);
}
if let Some(max_iterations) = max_iterations {
hnsw_pq_builder = hnsw_pq_builder.max_iterations(max_iterations);
}
if let Some(sample_rate) = sample_rate {
hnsw_pq_builder = hnsw_pq_builder.sample_rate(sample_rate);
}
if let Some(m) = m {
hnsw_pq_builder = hnsw_pq_builder.num_edges(m);
}
if let Some(ef_construction) = ef_construction {
hnsw_pq_builder = hnsw_pq_builder.ef_construction(ef_construction);
}
Ok(Self {
inner: Mutex::new(Some(LanceDbIndex::IvfHnswPq(hnsw_pq_builder))),
})
}
#[staticmethod]
pub fn hnsw_sq(
distance_type: Option<String>,
num_partitions: Option<u32>,
max_iterations: Option<u32>,
sample_rate: Option<u32>,
m: Option<u32>,
ef_construction: Option<u32>,
) -> PyResult<Self> {
let mut hnsw_sq_builder = IvfHnswSqIndexBuilder::default();
if let Some(distance_type) = distance_type {
let distance_type = parse_distance_type(distance_type)?;
hnsw_sq_builder = hnsw_sq_builder.distance_type(distance_type);
}
if let Some(num_partitions) = num_partitions {
hnsw_sq_builder = hnsw_sq_builder.num_partitions(num_partitions);
}
if let Some(max_iterations) = max_iterations {
hnsw_sq_builder = hnsw_sq_builder.max_iterations(max_iterations);
}
if let Some(sample_rate) = sample_rate {
hnsw_sq_builder = hnsw_sq_builder.sample_rate(sample_rate);
}
if let Some(m) = m {
hnsw_sq_builder = hnsw_sq_builder.num_edges(m);
}
if let Some(ef_construction) = ef_construction {
hnsw_sq_builder = hnsw_sq_builder.ef_construction(ef_construction);
}
Ok(Self {
inner: Mutex::new(Some(LanceDbIndex::IvfHnswSq(hnsw_sq_builder))),
})
}
}
#[pyclass(get_all)]