Replace the hnsw crate by the instant-distance one

Cargo.lock

@@ -1197,12 +1197,6 @@ dependencies = [
  "winapi",
 ]
 
-[[package]]
-name = "doc-comment"
-version = "0.3.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fea41bba32d969b513997752735605054bc0dfa92b4c56bf1189f2e174be7a10"
-
 [[package]]
 name = "dump"
 version = "1.3.0"
@@ -1707,15 +1701,6 @@ dependencies = [
  "byteorder",
 ]
 
-[[package]]
-name = "hashbrown"
-version = "0.11.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
-dependencies = [
- "ahash 0.7.6",
-]
-
 [[package]]
 name = "hashbrown"
 version = "0.12.3"
@@ -1814,22 +1799,6 @@ dependencies = [
  "digest",
 ]
 
-[[package]]
-name = "hnsw"
-version = "0.11.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2b9740ebf8769ec4ad6762cc951ba18f39bba6dfbc2fbbe46285f7539af79752"
-dependencies = [
- "ahash 0.7.6",
- "hashbrown 0.11.2",
- "libm",
- "num-traits",
- "rand_core",
- "serde",
- "smallvec",
- "space",
-]
-
 [[package]]
 name = "http"
 version = "0.2.9"
@@ -2008,6 +1977,21 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "instant-distance"
+version = "0.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8c619cdaa30bb84088963968bee12a45ea5fbbf355f2c021bcd15589f5ca494a"
+dependencies = [
+ "num_cpus",
+ "ordered-float",
+ "parking_lot",
+ "rand",
+ "rayon",
+ "serde",
+ "serde-big-array",
+]
+
 [[package]]
 name = "io-lifetimes"
 version = "1.0.11"
@@ -2701,9 +2685,9 @@ dependencies = [
  "geoutils",
  "grenad",
  "heed",
- "hnsw",
  "indexmap 1.9.3",
  "insta",
+ "instant-distance",
  "itertools",
  "json-depth-checker",
  "levenshtein_automata",
@@ -2727,7 +2711,6 @@ dependencies = [
  "smallstr",
  "smallvec",
  "smartstring",
- "space",
  "tempfile",
  "thiserror",
  "time",
@@ -3607,6 +3590,15 @@ dependencies = [
  "serde_derive",
 ]
 
+[[package]]
+name = "serde-big-array"
+version = "0.5.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "11fc7cc2c76d73e0f27ee52abbd64eec84d46f370c88371120433196934e4b7f"
+dependencies = [
+ "serde",
+]
+
 [[package]]
 name = "serde-cs"
 version = "0.2.4"
@@ -3756,9 +3748,6 @@ name = "smallvec"
 version = "1.10.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a507befe795404456341dfab10cef66ead4c041f62b8b11bbb92bffe5d0953e0"
-dependencies = [
- "serde",
-]
 
 [[package]]
 name = "smartstring"
@@ -3781,16 +3770,6 @@ dependencies = [
  "winapi",
 ]
 
-[[package]]
-name = "space"
-version = "0.17.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c5ab9701ae895386d13db622abf411989deff7109b13b46b6173bb4ce5c1d123"
-dependencies = [
- "doc-comment",
- "num-traits",
-]
-
 [[package]]
 name = "spin"
 version = "0.5.2"

milli/Cargo.toml

@@ -33,8 +33,8 @@ heed = { git = "https://github.com/meilisearch/heed", tag = "v0.12.6", default-f
     "lmdb",
     "sync-read-txn",
 ] }
-hnsw = { version = "0.11.0", features = ["serde1"] }
 indexmap = { version = "1.9.3", features = ["serde"] }
+instant-distance = { version = "0.6.1", features = ["with-serde"] }
 json-depth-checker = { path = "../json-depth-checker" }
 levenshtein_automata = { version = "0.2.1", features = ["fst_automaton"] }
 memmap2 = "0.5.10"
@@ -48,7 +48,6 @@ rstar = { version = "0.10.0", features = ["serde"] }
 serde = { version = "1.0.160", features = ["derive"] }
 serde_json = { version = "1.0.95", features = ["preserve_order"] }
 slice-group-by = "0.3.0"
-space = "0.17.0"
 smallstr = { version = "0.3.0", features = ["serde"] }
 smallvec = "1.10.0"
 smartstring = "1.0.1"

milli/src/distance.rs

@@ -1,20 +1,36 @@
+use std::ops;
+
+use instant_distance::Point;
 use serde::{Deserialize, Serialize};
-use space::Metric;
 
-#[derive(Debug, Default, Clone, Copy, Serialize, Deserialize)]
+use crate::normalize_vector;
-pub struct DotProduct;
 
-impl Metric<Vec<f32>> for DotProduct {
+#[derive(Debug, Default, Clone, Serialize, Deserialize)]
-    type Unit = u32;
+pub struct NDotProductPoint(Vec<f32>);
 
-    // Following <https://docs.rs/space/0.17.0/space/trait.Metric.html>.
+impl NDotProductPoint {
-    //
+    pub fn new(point: Vec<f32>) -> Self {
-    // Here is a playground that validate the ordering of the bit representation of floats in range 0.0..=1.0:
+        NDotProductPoint(normalize_vector(point))
-    // <https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=6c59e31a3cc5036b32edf51e8937b56e>
+    }
-    fn distance(&self, a: &Vec<f32>, b: &Vec<f32>) -> Self::Unit {
+
-        let dist = 1.0 - dot_product_similarity(a, b);
+    pub fn into_inner(self) -> Vec<f32> {
+        self.0
+    }
+}
+
+impl ops::Deref for NDotProductPoint {
+    type Target = [f32];
+
+    fn deref(&self) -> &Self::Target {
+        self.0.as_slice()
+    }
+}
+
+impl Point for NDotProductPoint {
+    fn distance(&self, other: &Self) -> f32 {
+        let dist = 1.0 - dot_product_similarity(&self.0, &other.0);
         debug_assert!(!dist.is_nan());
-        dist.to_bits()
+        dist
     }
 }
 

milli/src/index.rs

@@ -8,12 +8,11 @@ use charabia::{Language, Script};
 use heed::flags::Flags;
 use heed::types::*;
 use heed::{CompactionOption, Database, PolyDatabase, RoTxn, RwTxn};
-use rand_pcg::Pcg32;
 use roaring::RoaringBitmap;
 use rstar::RTree;
 use time::OffsetDateTime;
 
-use crate::distance::DotProduct;
+use crate::distance::NDotProductPoint;
 use crate::error::{InternalError, UserError};
 use crate::facet::FacetType;
 use crate::fields_ids_map::FieldsIdsMap;
@@ -31,7 +30,7 @@ use crate::{
 };
 
 /// The HNSW data-structure that we serialize, fill and search in.
-pub type Hnsw = hnsw::Hnsw<DotProduct, Vec<f32>, Pcg32, 12, 24>;
+pub type Hnsw = instant_distance::Hnsw<NDotProductPoint>;
 
 pub const DEFAULT_MIN_WORD_LEN_ONE_TYPO: u8 = 5;
 pub const DEFAULT_MIN_WORD_LEN_TWO_TYPOS: u8 = 9;

milli/src/search/new/mod.rs

@@ -28,7 +28,7 @@ use db_cache::DatabaseCache;
 use exact_attribute::ExactAttribute;
 use graph_based_ranking_rule::{Exactness, Fid, Position, Proximity, Typo};
 use heed::RoTxn;
-use hnsw::Searcher;
+use instant_distance::Search;
 use interner::{DedupInterner, Interner};
 pub use logger::visual::VisualSearchLogger;
 pub use logger::{DefaultSearchLogger, SearchLogger};
@@ -40,19 +40,18 @@ use ranking_rules::{
 use resolve_query_graph::{compute_query_graph_docids, PhraseDocIdsCache};
 use roaring::RoaringBitmap;
 use sort::Sort;
-use space::Neighbor;
 
 use self::distinct::facet_string_values;
 use self::geo_sort::GeoSort;
 pub use self::geo_sort::Strategy as GeoSortStrategy;
 use self::graph_based_ranking_rule::Words;
 use self::interner::Interned;
+use crate::distance::NDotProductPoint;
 use crate::error::FieldIdMapMissingEntry;
 use crate::score_details::{ScoreDetails, ScoringStrategy};
 use crate::search::new::distinct::apply_distinct_rule;
 use crate::{
-    normalize_vector, AscDesc, DocumentId, Filter, Index, Member, Result, TermsMatchingStrategy,
+    AscDesc, DocumentId, Filter, Index, Member, Result, TermsMatchingStrategy, UserError, BEU32,
-    UserError, BEU32,
 };
 
 /// A structure used throughout the execution of a search query.
@@ -445,29 +444,31 @@ pub fn execute_search(
     check_sort_criteria(ctx, sort_criteria.as_ref())?;
 
     if let Some(vector) = vector {
-        let mut searcher = Searcher::new();
+        let mut search = Search::default();
-        let hnsw = ctx.index.vector_hnsw(ctx.txn)?.unwrap_or_default();
+        let docids = match ctx.index.vector_hnsw(ctx.txn)? {
-        let ef = hnsw.len().min(100);
+            Some(hnsw) => {
-        let mut dest = vec![Neighbor { index: 0, distance: 0 }; ef];
+                let vector = NDotProductPoint::new(vector.clone());
-        let vector = normalize_vector(vector.clone());
+                let neighbors = hnsw.search(&vector, &mut search);
-        let neighbors = hnsw.nearest(&vector, ef, &mut searcher, &mut dest[..]);
 
-        let mut docids = Vec::new();
+                let mut docids = Vec::new();
-        let mut uniq_docids = RoaringBitmap::new();
+                let mut uniq_docids = RoaringBitmap::new();
-        for Neighbor { index, distance: _ } in neighbors.iter() {
+                for instant_distance::Item { distance: _, pid, point: _ } in neighbors {
-            let index = BEU32::new(*index as u32);
+                    let index = BEU32::new(pid.into_inner());
-            let docid = ctx.index.vector_id_docid.get(ctx.txn, &index)?.unwrap().get();
+                    let docid = ctx.index.vector_id_docid.get(ctx.txn, &index)?.unwrap().get();
-            if universe.contains(docid) && uniq_docids.insert(docid) {
+                    if universe.contains(docid) && uniq_docids.insert(docid) {
-                docids.push(docid);
+                        docids.push(docid);
-                if docids.len() == (from + length) {
+                        if docids.len() == (from + length) {
-                    break;
+                            break;
+                        }
+                    }
                 }
-            }
-        }
 
-        // return the nearest documents that are also part of the candidates
+                // return the nearest documents that are also part of the candidates
-        // along with a dummy list of scores that are useless in this context.
+                // along with a dummy list of scores that are useless in this context.
-        let docids: Vec<_> = docids.into_iter().skip(from).take(length).collect();
+                docids.into_iter().skip(from).take(length).collect()
+            }
+            None => Vec::new(),
+        };
 
         return Ok(PartialSearchResult {
             candidates: universe,

milli/src/update/delete_documents.rs

@@ -4,10 +4,9 @@ use std::collections::{BTreeSet, HashMap, HashSet};
 use fst::IntoStreamer;
 use heed::types::{ByteSlice, DecodeIgnore, Str, UnalignedSlice};
 use heed::{BytesDecode, BytesEncode, Database, RwIter};
-use hnsw::Searcher;
+use instant_distance::PointId;
 use roaring::RoaringBitmap;
 use serde::{Deserialize, Serialize};
-use space::KnnPoints;
 use time::OffsetDateTime;
 
 use super::facet::delete::FacetsDelete;
@@ -436,24 +435,24 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
 
         // An ugly and slow way to remove the vectors from the HNSW
         // It basically reconstructs the HNSW from scratch without editing the current one.
-        let current_hnsw = self.index.vector_hnsw(self.wtxn)?.unwrap_or_default();
+        if let Some(current_hnsw) = self.index.vector_hnsw(self.wtxn)? {
-        if !current_hnsw.is_empty() {
+            let mut points = Vec::new();
-            let mut new_hnsw = Hnsw::default();
+            let mut docids = Vec::new();
-            let mut searcher = Searcher::new();
-            let mut new_vector_id_docids = Vec::new();
-
             for result in vector_id_docid.iter(self.wtxn)? {
                 let (vector_id, docid) = result?;
                 if !self.to_delete_docids.contains(docid.get()) {
-                    let vector = current_hnsw.get_point(vector_id.get() as usize).clone();
+                    let pid = PointId::from(vector_id.get());
-                    let vector_id = new_hnsw.insert(vector, &mut searcher);
+                    let vector = current_hnsw[pid].clone();
-                    new_vector_id_docids.push((vector_id as u32, docid));
+                    points.push(vector);
+                    docids.push(docid);
                 }
             }
 
+            let (new_hnsw, pids) = Hnsw::builder().build_hnsw(points);
+
             vector_id_docid.clear(self.wtxn)?;
-            for (vector_id, docid) in new_vector_id_docids {
+            for (pid, docid) in pids.into_iter().zip(docids) {
-                vector_id_docid.put(self.wtxn, &BEU32::new(vector_id), &docid)?;
+                vector_id_docid.put(self.wtxn, &BEU32::new(pid.into_inner()), &docid)?;
             }
             self.index.put_vector_hnsw(self.wtxn, &new_hnsw)?;
         }

milli/src/update/index_documents/typed_chunk.rs

@@ -9,22 +9,19 @@ use charabia::{Language, Script};
 use grenad::MergerBuilder;
 use heed::types::ByteSlice;
 use heed::RwTxn;
-use hnsw::Searcher;
 use roaring::RoaringBitmap;
-use space::KnnPoints;
 
 use super::helpers::{
     self, merge_ignore_values, serialize_roaring_bitmap, valid_lmdb_key, CursorClonableMmap,
 };
 use super::{ClonableMmap, MergeFn};
+use crate::distance::NDotProductPoint;
 use crate::error::UserError;
 use crate::facet::FacetType;
+use crate::index::Hnsw;
 use crate::update::facet::FacetsUpdate;
 use crate::update::index_documents::helpers::{as_cloneable_grenad, try_split_array_at};
-use crate::{
+use crate::{lat_lng_to_xyz, CboRoaringBitmapCodec, DocumentId, GeoPoint, Index, Result, BEU32};
-    lat_lng_to_xyz, normalize_vector, CboRoaringBitmapCodec, DocumentId, GeoPoint, Index, Result,
-    BEU32,
-};
 
 pub(crate) enum TypedChunk {
     FieldIdDocidFacetStrings(grenad::Reader<CursorClonableMmap>),
@@ -230,17 +227,20 @@ pub(crate) fn write_typed_chunk_into_index(
             index.put_geo_faceted_documents_ids(wtxn, &geo_faceted_docids)?;
         }
         TypedChunk::VectorPoints(vector_points) => {
-            let mut hnsw = index.vector_hnsw(wtxn)?.unwrap_or_default();
+            let (pids, mut points): (Vec<_>, Vec<_>) = match index.vector_hnsw(wtxn)? {
-            let mut searcher = Searcher::new();
+                Some(hnsw) => hnsw.iter().map(|(pid, point)| (pid, point.clone())).unzip(),
-
+                None => Default::default(),
-            let mut expected_dimensions = match index.vector_id_docid.iter(wtxn)?.next() {
-                Some(result) => {
-                    let (vector_id, _) = result?;
-                    Some(hnsw.get_point(vector_id.get() as usize).len())
-                }
-                None => None,
             };
 
+            // Convert the PointIds into DocumentIds
+            let mut docids = Vec::new();
+            for pid in pids {
+                let docid =
+                    index.vector_id_docid.get(wtxn, &BEU32::new(pid.into_inner()))?.unwrap();
+                docids.push(docid.get());
+            }
+
+            let mut expected_dimensions = points.get(0).map(|p| p.len());
             let mut cursor = vector_points.into_cursor()?;
             while let Some((key, value)) = cursor.move_on_next()? {
                 // convert the key back to a u32 (4 bytes)
@@ -256,12 +256,26 @@ pub(crate) fn write_typed_chunk_into_index(
                     return Err(UserError::InvalidVectorDimensions { expected, found })?;
                 }
 
-                let vector = normalize_vector(vector);
+                points.push(NDotProductPoint::new(vector));
-                let vector_id = hnsw.insert(vector, &mut searcher) as u32;
+                docids.push(docid);
-                index.vector_id_docid.put(wtxn, &BEU32::new(vector_id), &BEU32::new(docid))?;
             }
-            log::debug!("There are {} entries in the HNSW so far", hnsw.len());
+
-            index.put_vector_hnsw(wtxn, &hnsw)?;
+            assert_eq!(docids.len(), points.len());
+
+            let hnsw_length = points.len();
+            let (new_hnsw, pids) = Hnsw::builder().build_hnsw(points);
+
+            index.vector_id_docid.clear(wtxn)?;
+            for (docid, pid) in docids.into_iter().zip(pids) {
+                index.vector_id_docid.put(
+                    wtxn,
+                    &BEU32::new(pid.into_inner()),
+                    &BEU32::new(docid),
+                )?;
+            }
+
+            log::debug!("There are {} entries in the HNSW so far", hnsw_length);
+            index.put_vector_hnsw(wtxn, &new_hnsw)?;
         }
         TypedChunk::ScriptLanguageDocids(hash_pair) => {
             let mut buffer = Vec::new();