Available on crate feature
quantized only.Expand description
Quantized graph index (QG Index)
Defining the properties of a new QG index:
use ngt::qg::{QgProperties, QgDistance};
// Defaut properties with vectors of dimension 3
let prop = QgProperties::<f32>::dimension(3)?;
// Or customize values (here are the defaults)
let prop = QgProperties::<f32>::dimension(3)?
.creation_edge_size(10)?
.search_edge_size(40)?
.distance_type(QgDistance::L2)?;
Creating/Opening a QG index and using it:
use ngt::NgtIndex;
use ngt::qg::{QgDistance, QgIndex, QgProperties, QgQuantizationParams, QgQuery};
// Create a new quantizable NGT index
let prop = QgProperties::dimension(3)?;
let mut index: NgtIndex<f32> =
NgtIndex::create("target/path/to/qg_index/dir", prop.try_into()?)?;
// Insert two vectors and get their id
let vec1 = vec![1.0, 2.0, 3.0];
let vec2 = vec![4.0, 5.0, 6.0];
let id1 = index.insert(vec1)?;
let id2 = index.insert(vec2)?;
// Add enough dummy vectors to build an index
for i in 0..64 {
index.insert(vec![100. + i as f32; 3])?;
}
// Build the index in RAM and persist it on disk
index.build(1)?;
index.persist()?;
// Quantize the NGT index
let params = QgQuantizationParams {
dimension_of_subvector: 1.,
max_number_of_edges: 50,
};
let index = QgIndex::quantize(index, params)?;
// Open an existing QG index
let index = QgIndex::open("target/path/to/qg_index/dir")?;
// Perform a vector search (with 1 result)
let query = vec![1.1, 2.1, 3.1];
let res = index.search(QgQuery::new(&query).size(1))?;
assert_eq!(res[0].id, id1);
assert_eq!(index.get_vec(id1)?, vec![1.0, 2.0, 3.0]);