Merge pull request #4160 from meilisearch/diff-indexing-vector-points

Diff Indexing for the vector points

milli/src/update/index_documents/extract/extract_vector_points.rs

@@ -1,15 +1,25 @@
+use std::cmp::Ordering;
 use std::convert::TryFrom;
 use std::fs::File;
-use std::io::{self, BufReader};
+use std::io::{self, BufReader, BufWriter};
+use std::mem::size_of;
+use std::str::from_utf8;
 
 use bytemuck::cast_slice;
+use grenad::Writer;
+use itertools::EitherOrBoth;
+use ordered_float::OrderedFloat;
 use serde_json::{from_slice, Value};
 
 use super::helpers::{create_writer, writer_into_reader, GrenadParameters};
 use crate::error::UserError;
+use crate::update::del_add::{DelAdd, KvReaderDelAdd, KvWriterDelAdd};
 use crate::update::index_documents::helpers::try_split_at;
 use crate::{DocumentId, FieldId, InternalError, Result, VectorOrArrayOfVectors};
 
+/// The length of the elements that are always in the buffer when inserting new values.
+const TRUNCATE_SIZE: usize = size_of::<DocumentId>();
+
 /// Extracts the embedding vector contained in each document under the `_vectors` field.
 ///
 /// Returns the generated grenad reader containing the docid as key associated to the Vec<f32>
@@ -27,45 +37,112 @@ pub fn extract_vector_points<R: io::Read + io::Seek>(
         tempfile::tempfile()?,
     );
 
+    let mut key_buffer = Vec::new();
     let mut cursor = obkv_documents.into_cursor()?;
     while let Some((key, value)) = cursor.move_on_next()? {
         // this must always be serialized as (docid, external_docid);
         let (docid_bytes, external_id_bytes) =
             try_split_at(key, std::mem::size_of::<DocumentId>()).unwrap();
-        debug_assert!(std::str::from_utf8(external_id_bytes).is_ok());
+        debug_assert!(from_utf8(external_id_bytes).is_ok());
 
         let obkv = obkv::KvReader::new(value);
+        key_buffer.clear();
+        key_buffer.extend_from_slice(docid_bytes);
 
         // since we only needs the primary key when we throw an error we create this getter to
         // lazily get it when needed
-        let document_id = || -> Value { std::str::from_utf8(external_id_bytes).unwrap().into() };
+        let document_id = || -> Value { from_utf8(external_id_bytes).unwrap().into() };
 
         // first we retrieve the _vectors field
-        if let Some(vectors) = obkv.get(vectors_fid) {
+        if let Some(value) = obkv.get(vectors_fid) {
-            // extract the vectors
+            let vectors_obkv = KvReaderDelAdd::new(value);
-            let vectors = match from_slice(vectors) {
-                Ok(vectors) => VectorOrArrayOfVectors::into_array_of_vectors(vectors),
-                Err(_) => {
-                    return Err(UserError::InvalidVectorsType {
-                        document_id: document_id(),
-                        value: from_slice(vectors).map_err(InternalError::SerdeJson)?,
-                    }
-                    .into())
-                }
-            };
 
-            if let Some(vectors) = vectors {
+            // then we extract the values
-                for (i, vector) in vectors.into_iter().enumerate().take(u16::MAX as usize) {
+            let del_vectors = vectors_obkv
-                    let index = u16::try_from(i).unwrap();
+                .get(DelAdd::Deletion)
-                    let mut key = docid_bytes.to_vec();
+                .map(|vectors| extract_vectors(vectors, document_id))
-                    key.extend_from_slice(&index.to_be_bytes());
+                .transpose()?
-                    let bytes = cast_slice(&vector);
+                .flatten();
-                    writer.insert(key, bytes)?;
+            let add_vectors = vectors_obkv
+                .get(DelAdd::Addition)
+                .map(|vectors| extract_vectors(vectors, document_id))
+                .transpose()?
+                .flatten();
+
+            // and we finally push the unique vectors into the writer
+            push_vectors_diff(
+                &mut writer,
+                &mut key_buffer,
+                del_vectors.unwrap_or_default(),
+                add_vectors.unwrap_or_default(),
+            )?;
         }
     }
-        }
-        // else => the `_vectors` object was `null`, there is nothing to do
-    }
 
     writer_into_reader(writer)
 }
+
+/// Computes the diff between both Del and Add numbers and
+/// only inserts the parts that differ in the sorter.
+fn push_vectors_diff(
+    writer: &mut Writer<BufWriter<File>>,
+    key_buffer: &mut Vec<u8>,
+    mut del_vectors: Vec<Vec<f32>>,
+    mut add_vectors: Vec<Vec<f32>>,
+) -> Result<()> {
+    // We sort and dedup the vectors
+    del_vectors.sort_unstable_by(|a, b| compare_vectors(a, b));
+    add_vectors.sort_unstable_by(|a, b| compare_vectors(a, b));
+    del_vectors.dedup_by(|a, b| compare_vectors(a, b).is_eq());
+    add_vectors.dedup_by(|a, b| compare_vectors(a, b).is_eq());
+
+    let merged_vectors_iter =
+        itertools::merge_join_by(del_vectors, add_vectors, |del, add| compare_vectors(del, add));
+
+    // insert vectors into the writer
+    for (i, eob) in merged_vectors_iter.into_iter().enumerate().take(u16::MAX as usize) {
+        // Generate the key by extending the unique index to it.
+        key_buffer.truncate(TRUNCATE_SIZE);
+        let index = u16::try_from(i).unwrap();
+        key_buffer.extend_from_slice(&index.to_be_bytes());
+
+        match eob {
+            EitherOrBoth::Both(_, _) => (), // no need to touch anything
+            EitherOrBoth::Left(vector) => {
+                // We insert only the Del part of the Obkv to inform
+                // that we only want to remove all those vectors.
+                let mut obkv = KvWriterDelAdd::memory();
+                obkv.insert(DelAdd::Deletion, cast_slice(&vector))?;
+                let bytes = obkv.into_inner()?;
+                writer.insert(&key_buffer, bytes)?;
+            }
+            EitherOrBoth::Right(vector) => {
+                // We insert only the Add part of the Obkv to inform
+                // that we only want to remove all those vectors.
+                let mut obkv = KvWriterDelAdd::memory();
+                obkv.insert(DelAdd::Addition, cast_slice(&vector))?;
+                let bytes = obkv.into_inner()?;
+                writer.insert(&key_buffer, bytes)?;
+            }
+        }
+    }
+
+    Ok(())
+}
+
+/// Compares two vectors by using the OrderingFloat helper.
+fn compare_vectors(a: &[f32], b: &[f32]) -> Ordering {
+    a.iter().copied().map(OrderedFloat).cmp(b.iter().copied().map(OrderedFloat))
+}
+
+/// Extracts the vectors from a JSON value.
+fn extract_vectors(value: &[u8], document_id: impl Fn() -> Value) -> Result<Option<Vec<Vec<f32>>>> {
+    match from_slice(value) {
+        Ok(vectors) => Ok(VectorOrArrayOfVectors::into_array_of_vectors(vectors)),
+        Err(_) => Err(UserError::InvalidVectorsType {
+            document_id: document_id(),
+            value: from_slice(value).map_err(InternalError::SerdeJson)?,
+        }
+        .into()),
+    }
+}

milli/src/update/index_documents/typed_chunk.rs

@@ -1,4 +1,4 @@
-use std::collections::HashMap;
+use std::collections::{HashMap, HashSet};
 use std::convert::TryInto;
 use std::fs::File;
 use std::io::{self, BufReader};
@@ -8,7 +8,9 @@ use charabia::{Language, Script};
 use grenad::MergerBuilder;
 use heed::types::ByteSlice;
 use heed::RwTxn;
+use log::error;
 use obkv::{KvReader, KvWriter};
+use ordered_float::OrderedFloat;
 use roaring::RoaringBitmap;
 
 use super::helpers::{self, merge_ignore_values, valid_lmdb_key, CursorClonableMmap};
@@ -22,10 +24,9 @@ use crate::index::Hnsw;
 use crate::update::del_add::{DelAdd, KvReaderDelAdd};
 use crate::update::facet::FacetsUpdate;
 use crate::update::index_documents::helpers::{as_cloneable_grenad, try_split_array_at};
-use crate::update::index_documents::validate_document_id_value;
 use crate::{
-    lat_lng_to_xyz, CboRoaringBitmapCodec, DocumentId, FieldId, GeoPoint, Index, InternalError,
+    lat_lng_to_xyz, CboRoaringBitmapCodec, DocumentId, FieldId, GeoPoint, Index, Result,
-    Result, SerializationError, BEU32,
+    SerializationError, BEU32,
 };
 
 pub(crate) enum TypedChunk {
@@ -366,44 +367,70 @@ pub(crate) fn write_typed_chunk_into_index(
             index.put_geo_faceted_documents_ids(wtxn, &geo_faceted_docids)?;
         }
         TypedChunk::VectorPoints(vector_points) => {
-            let (pids, mut points): (Vec<_>, Vec<_>) = match index.vector_hnsw(wtxn)? {
+            let mut vectors_set = HashSet::new();
-                Some(hnsw) => hnsw.iter().map(|(pid, point)| (pid, point.clone())).unzip(),
+            // We extract and store the previous vectors
-                None => Default::default(),
+            if let Some(hnsw) = index.vector_hnsw(wtxn)? {
-            };
+                for (pid, point) in hnsw.iter() {
-
+                    let pid_key = BEU32::new(pid.into_inner());
-            // Convert the PointIds into DocumentIds
+                    let docid = index.vector_id_docid.get(wtxn, &pid_key)?.unwrap().get();
-            let mut docids = Vec::new();
+                    let vector: Vec<_> = point.iter().copied().map(OrderedFloat).collect();
-            for pid in pids {
+                    vectors_set.insert((docid, vector));
-                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)
                 let (left, _index) = try_split_array_at(key).unwrap();
                 let docid = DocumentId::from_be_bytes(left);
-                // convert the vector back to a Vec<f32>
-                let vector: Vec<f32> = pod_collect_to_vec(value);
 
-                // TODO Inform the user about the document that has a wrong `_vectors`
+                let vector_deladd_obkv = KvReaderDelAdd::new(value);
-                let found = vector.len();
+                if let Some(value) = vector_deladd_obkv.get(DelAdd::Deletion) {
-                let expected = *expected_dimensions.get_or_insert(found);
+                    // convert the vector back to a Vec<f32>
-                if expected != found {
+                    let vector = pod_collect_to_vec(value).into_iter().map(OrderedFloat).collect();
-                    return Err(UserError::InvalidVectorDimensions { expected, found }.into());
+                    let key = (docid, vector);
+                    if !vectors_set.remove(&key) {
+                        error!("Unable to delete the vector: {:?}", key.1);
+                    }
+                }
+                if let Some(value) = vector_deladd_obkv.get(DelAdd::Addition) {
+                    // convert the vector back to a Vec<f32>
+                    let vector = pod_collect_to_vec(value).into_iter().map(OrderedFloat).collect();
+                    vectors_set.insert((docid, vector));
+                }
             }
 
+            // Extract the most common vector dimension
+            let expected_dimension_size = {
+                let mut dims = HashMap::new();
+                vectors_set.iter().for_each(|(_, v)| *dims.entry(v.len()).or_insert(0) += 1);
+                dims.into_iter().max_by_key(|(_, count)| *count).map(|(len, _)| len)
+            };
+
+            // Ensure that the vector lenghts are correct and
+            // prepare the vectors before inserting them in the HNSW.
+            let mut points = Vec::new();
+            let mut docids = Vec::new();
+            for (docid, vector) in vectors_set {
+                if expected_dimension_size.map_or(false, |expected| expected != vector.len()) {
+                    return Err(UserError::InvalidVectorDimensions {
+                        expected: expected_dimension_size.unwrap_or(vector.len()),
+                        found: vector.len(),
+                    }
+                    .into());
+                } else {
+                    let vector = vector.into_iter().map(OrderedFloat::into_inner).collect();
                     points.push(NDotProductPoint::new(vector));
                     docids.push(docid);
                 }
-
+            }
-            assert_eq!(docids.len(), points.len());
 
             let hnsw_length = points.len();
             let (new_hnsw, pids) = Hnsw::builder().build_hnsw(points);
 
+            assert_eq!(docids.len(), pids.len());
+
+            // Store the vectors in the point-docid relation database
             index.vector_id_docid.clear(wtxn)?;
             for (docid, pid) in docids.into_iter().zip(pids) {
                 index.vector_id_docid.put(