snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2025-07-26 00:01:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge #696

Browse Source 
696: Fix Facet Indexing bugs r=Kerollmops a=loiclec

1. Handle keys with variable length correctly

Closes (partially) https://github.com/meilisearch/meilisearch/issues/3042 
This issue is now easily reproducible with the updated fuzz tests, which now generate keys with variable lengths.

2. Prevent adding facets to the database if their encoded value does not satisfy `valid_lmdb_key`.

Closes (partially) https://github.com/meilisearch/meilisearch/issues/2743
This fixes an indexing failure when a document had a filterable attribute containing a value whose length is higher than ~500 bytes. For now, this fix is just meant to prevent crashes. Better handling of long values of filterable attributes will be handled in a separate PR.


Co-authored-by: Loïc Lecrenier <loic.lecrenier@me.com>
This commit is contained in:
bors[bot]
2022-11-17 11:56:16 +00:00
committed by GitHub
parent a651397afc 0caadedd3b
commit 8ceb199dca
7 changed files with 115 additions and 255 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
milli/src/update/facet/bulk.rs
View File

@ -12,7 +12,7 @@ use crate::heed_codec::facet::{
FacetGroupKey, FacetGroupKeyCodec, FacetGroupValue, FacetGroupValueCodec,
};
use crate::heed_codec::ByteSliceRefCodec;
use crate::update::index_documents::{create_writer, writer_into_reader};
use crate::update::index_documents::{create_writer, valid_lmdb_key, writer_into_reader};
use crate::{CboRoaringBitmapCodec, FieldId, Index, Result};
/// Algorithm to insert elememts into the `facet_id_(string/f64)_docids` databases
@ -142,6 +142,9 @@ impl<R: std::io::Read + std::io::Seek> FacetsUpdateBulkInner<R> {
let mut database = self.db.iter_mut(wtxn)?.remap_types::<ByteSlice, ByteSlice>();
let mut cursor = new_data.into_cursor()?;
while let Some((key, value)) = cursor.move_on_next()? {
if !valid_lmdb_key(key) {
continue;
}
buffer.clear();
// the group size for level 0
buffer.push(1);
@ -155,6 +158,9 @@ impl<R: std::io::Read + std::io::Seek> FacetsUpdateBulkInner<R> {
let mut cursor = new_data.into_cursor()?;
while let Some((key, value)) = cursor.move_on_next()? {
if !valid_lmdb_key(key) {
continue;
}
// the value is a CboRoaringBitmap, but I still need to prepend the
// group size for level 0 (= 1) to it
buffer.clear();

266
milli/src/update/facet/incremental.rs
View File

@ -11,6 +11,7 @@ use crate::heed_codec::facet::{
};
use crate::heed_codec::ByteSliceRefCodec;
use crate::search::facet::get_highest_level;
use crate::update::index_documents::valid_lmdb_key;
use crate::{CboRoaringBitmapCodec, FieldId, Index, Result};
enum InsertionResult {
@ -70,6 +71,9 @@ impl<'i> FacetsUpdateIncremental<'i> {
let mut cursor = self.new_data.into_cursor()?;
while let Some((key, value)) = cursor.move_on_next()? {
if !valid_lmdb_key(key) {
continue;
}
let key = FacetGroupKeyCodec::<ByteSliceRefCodec>::bytes_decode(key)
.ok_or(heed::Error::Encoding)?;
let docids = CboRoaringBitmapCodec::bytes_decode(value).ok_or(heed::Error::Encoding)?;
@ -114,37 +118,18 @@ impl FacetsUpdateIncrementalInner {
txn: &RoTxn,
) -> Result<(FacetGroupKey<Vec<u8>>, FacetGroupValue)> {
assert!(level > 0);
let mut prefix = vec![];
prefix.extend_from_slice(&field_id.to_be_bytes());
prefix.push(level);
prefix.extend_from_slice(facet_value);
let mut prefix_iter = self
.db
.as_polymorph()
.prefix_iter::<_, ByteSlice, FacetGroupValueCodec>(txn, prefix.as_slice())?;
if let Some(e) = prefix_iter.next() {
let (key_bytes, value) = e?;
Ok((
FacetGroupKeyCodec::<ByteSliceRefCodec>::bytes_decode(key_bytes)
.ok_or(Error::Encoding)?
.into_owned(),
value,
))
} else {
let key = FacetGroupKey { field_id, level, left_bound: facet_value };
match self.db.get_lower_than(txn, &key)? {
match self.db.get_lower_than_or_equal_to(
txn,
&FacetGroupKey { field_id, level, left_bound: facet_value },
)? {
Some((key, value)) => {
if key.level != level {
let mut prefix = vec![];
prefix.extend_from_slice(&field_id.to_be_bytes());
prefix.push(level);
let mut iter = self
.db
.as_polymorph()
.prefix_iter::<_, ByteSlice, FacetGroupValueCodec>(
let mut iter =
self.db.as_polymorph().prefix_iter::<_, ByteSlice, FacetGroupValueCodec>(
txn,
prefix.as_slice(),
)?;
@ -159,7 +144,11 @@ impl FacetsUpdateIncrementalInner {
Ok((key.into_owned(), value))
}
}
None => panic!(),
None => {
// We checked that the level is > 0
// Since all keys of level 1 are greater than those of level 0,
// we are guaranteed that db.get_lower_than_or_equal_to(key) exists
panic!()
}
}
}
@ -1050,9 +1039,7 @@ ensures that:
2. its content is the same as a trivially correct implementation of the same database
*/
mod fuzz {
use std::borrow::Cow;
use std::collections::{BTreeMap, HashMap};
use std::convert::TryFrom;
use std::iter::FromIterator;
use std::rc::Rc;
@ -1060,49 +1047,30 @@ mod fuzz {
use fuzzcheck::mutators::integer_within_range::{U16WithinRangeMutator, U8WithinRangeMutator};
use fuzzcheck::mutators::vector::VecMutator;
use fuzzcheck::DefaultMutator;
use heed::BytesEncode;
use roaring::RoaringBitmap;
use tempfile::TempDir;
use super::*;
use crate::update::facet::tests::FacetIndex;
struct NEU16Codec;
impl<'a> BytesEncode<'a> for NEU16Codec {
type EItem = u16;
#[no_coverage]
fn bytes_encode(item: &'a Self::EItem) -> Option<std::borrow::Cow<'a, [u8]>> {
Some(Cow::Owned(item.to_be_bytes().to_vec()))
}
}
impl<'a> BytesDecode<'a> for NEU16Codec {
type DItem = u16;
#[no_coverage]
fn bytes_decode(bytes: &'a [u8]) -> Option<Self::DItem> {
let bytes = <[u8; 2]>::try_from(&bytes[0..=1]).unwrap();
Some(u16::from_be_bytes(bytes))
}
}
#[derive(Default)]
pub struct TrivialDatabase<T> {
pub elements: BTreeMap<u16, BTreeMap<T, RoaringBitmap>>,
}
impl<T> TrivialDatabase<T>
where
T: Ord + Clone + Copy + Eq + std::fmt::Debug,
T: Ord + Clone + Eq + std::fmt::Debug,
{
#[no_coverage]
pub fn insert(&mut self, field_id: u16, new_key: T, new_values: &RoaringBitmap) {
pub fn insert(&mut self, field_id: u16, new_key: &T, new_values: &RoaringBitmap) {
if new_values.is_empty() {
return;
}
let values_field_id = self.elements.entry(field_id).or_default();
let values = values_field_id.entry(new_key).or_default();
let values = values_field_id.entry(new_key.clone()).or_default();
*values |= new_values;
}
#[no_coverage]
pub fn delete(&mut self, field_id: u16, key: T, values_to_remove: &RoaringBitmap) {
pub fn delete(&mut self, field_id: u16, key: &T, values_to_remove: &RoaringBitmap) {
if let Some(values_field_id) = self.elements.get_mut(&field_id) {
if let Some(values) = values_field_id.get_mut(&key) {
*values -= values_to_remove;
@ -1117,8 +1085,9 @@ mod fuzz {
}
}
#[derive(Clone, DefaultMutator, serde::Serialize, serde::Deserialize)]
struct Operation<Key> {
key: Key,
struct Operation {
#[field_mutator(VecMutator<u8, U8Mutator> = { VecMutator::new(u8::default_mutator(), 0 ..= 5) })]
key: Vec<u8>,
#[field_mutator(U8WithinRangeMutator = { U8WithinRangeMutator::new(..32) })]
group_size: u8,
#[field_mutator(U8WithinRangeMutator = { U8WithinRangeMutator::new(..32) })]
@ -1142,13 +1111,12 @@ mod fuzz {
}
#[no_coverage]
fn compare_with_trivial_database(tempdir: Rc<TempDir>, operations: &[Operation<u16>]) {
let index = FacetIndex::<NEU16Codec>::open_from_tempdir(tempdir, 4, 8, 5); // dummy params, they'll be overwritten
// let mut txn = index.env.write_txn().unwrap();
fn compare_with_trivial_database(tempdir: Rc<TempDir>, operations: &[Operation]) {
let index = FacetIndex::<ByteSliceRefCodec>::open_from_tempdir(tempdir, 4, 8, 5); // dummy params, they'll be overwritten
let mut txn = index.env.write_txn().unwrap();
let mut trivial_db = TrivialDatabase::<u16>::default();
let mut value_to_keys = HashMap::<u8, Vec<u16>>::new();
let mut trivial_db = TrivialDatabase::<Vec<u8>>::default();
let mut value_to_keys = HashMap::<u8, Vec<Vec<u8>>>::new();
for Operation { key, group_size, max_group_size, min_level_size, field_id, kind } in
operations
{
@ -1160,10 +1128,10 @@ mod fuzz {
let mut bitmap = RoaringBitmap::new();
for value in values {
bitmap.insert(*value as u32);
value_to_keys.entry(*value).or_default().push(*key);
value_to_keys.entry(*value).or_default().push(key.clone());
}
index.insert(&mut txn, *field_id, key, &bitmap);
trivial_db.insert(*field_id, *key, &bitmap);
index.insert(&mut txn, *field_id, &key.as_slice(), &bitmap);
trivial_db.insert(*field_id, &key, &bitmap);
}
OperationKind::Delete(values) => {
let values = RoaringBitmap::from_iter(values.iter().copied().map(|x| x as u32));
@ -1179,8 +1147,8 @@ mod fuzz {
}
}
for (key, values) in values_per_key {
index.delete(&mut txn, *field_id, &key, &values);
trivial_db.delete(*field_id, *key, &values);
index.delete(&mut txn, *field_id, &key.as_slice(), &values);
trivial_db.delete(*field_id, &key, &values);
}
}
}
@ -1198,7 +1166,8 @@ mod fuzz {
for ((key, values), group) in values_field_id.iter().zip(level0iter) {
let (group_key, group_values) = group.unwrap();
let group_key = FacetGroupKeyCodec::<NEU16Codec>::bytes_decode(group_key).unwrap();
let group_key =
FacetGroupKeyCodec::<ByteSliceRefCodec>::bytes_decode(group_key).unwrap();
assert_eq!(key, &group_key.left_bound);
assert_eq!(values, &group_values.bitmap);
}
@ -1213,7 +1182,8 @@ mod fuzz {
for ((key, values), group) in values_field_id.iter().zip(level0iter) {
let (group_key, group_values) = group.unwrap();
let group_key = FacetGroupKeyCodec::<NEU16Codec>::bytes_decode(group_key).unwrap();
let group_key =
FacetGroupKeyCodec::<ByteSliceRefCodec>::bytes_decode(group_key).unwrap();
assert_eq!(key, &group_key.left_bound);
assert_eq!(values, &group_values.bitmap);
}
@ -1227,7 +1197,7 @@ mod fuzz {
fn fuzz() {
let tempdir = Rc::new(TempDir::new().unwrap());
let tempdir_cloned = tempdir.clone();
let result = fuzzcheck::fuzz_test(move |operations: &[Operation<u16>]| {
let result = fuzzcheck::fuzz_test(move |operations: &[Operation]| {
compare_with_trivial_database(tempdir_cloned.clone(), operations)
})
.default_mutator()
@ -1243,168 +1213,4 @@ mod fuzz {
.launch();
assert!(!result.found_test_failure);
}
#[test]
#[no_coverage]
fn reproduce_bug1() {
let operations = r#"
[
{"key":0, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[109]}},
{"key":143, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[243]}},
{"key":90, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[217]}},
{"key":172, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[94]}},
{"key":27, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[4]}},
{"key":124, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[0]}},
{"key":123, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[0]}},
{"key":67, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[109]}},
{"key":13, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[0]}},
{"key":162, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[213]}},
{"key":235, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[67]}},
{"key":251, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[50]}},
{"key":218, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[164]}},
{"key":166, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[67]}},
{"key":64, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[61]}},
{"key":183, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[210]}},
{"key":250, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Delete":[50]}}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(operations).unwrap();
let tempdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tempdir), &operations);
}
#[test]
#[no_coverage]
fn reproduce_bug2() {
let operations = r#"
[
{"key":102, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[122]}},
{"key":73, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[132]}},
{"key":20, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[215]}},
{"key":39, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[152]}},
{"key":151, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[226]}},
{"key":17, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[101]}},
{"key":74, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[210]}},
{"key":2, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[130]}},
{"key":64, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[180]}},
{"key":83, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[250]}},
{"key":80, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[210]}},
{"key":113, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[63]}},
{"key":201, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[210]}},
{"key":200, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[5]}},
{"key":93, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[98]}},
{"key":162, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Insert":[5]}},
{"key":80, "field_id": 0, "group_size":4, "max_group_size":8, "min_level_size":5, "kind":{"Delete":[210]}}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(operations).unwrap();
let tempdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tempdir), &operations);
}
#[test]
#[no_coverage]
fn reproduce_bug3() {
let operations = r#"
[
{"key":27488, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[206]}},
{"key":64716, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[216]}},
{"key":60886, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[206]}},
{"key":59509, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[187,231]}},
{"key":55057, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[37]}},
{"key":45200, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[206]}},
{"key":55056, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[37]}},
{"key":63679, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[206]}},
{"key":52155, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[74]}},
{"key":20648, "field_id": 0, "group_size":0, "max_group_size":7, "min_level_size":0, "kind":{"Insert":[47,138,157]}}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(operations).unwrap();
let tempdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tempdir), &operations);
}
#[test]
#[no_coverage]
fn reproduce_bug4() {
let operations = r#"[
{"key":63499, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[87]}},
{"key":25374, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[14]}},
{"key":64481, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Delete":[87]}},
{"key":23038, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[173]}},
{"key":14862, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[8]}},
{"key":13145, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[5,64]}},
{"key":23446, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[86,59]}},
{"key":17972, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[58,137]}},
{"key":21273, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[121,132,81,147]}},
{"key":28264, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[36]}},
{"key":46659, "field_id": 0, "group_size":2, "max_group_size":1, "min_level_size":0, "kind":{"Insert":[]}}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(operations).unwrap();
let tempdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tempdir), &operations);
}
#[test]
#[no_coverage]
fn reproduce_bug5() {
let input = r#"
[
{
"key":3438,
"group_size":11,
"max_group_size":0,
"min_level_size":17,
"field_id":3,
"kind":{"Insert":[198]}
},
{
"key":47098,
"group_size":0,
"max_group_size":8,
"min_level_size":0,
"field_id":3,
"kind":{"Insert":[11]}
},
{
"key":22453,
"group_size":0,
"max_group_size":0,
"min_level_size":0,
"field_id":3,
"kind":{"Insert":[145]}
},
{
"key":14105,
"group_size":14,
"max_group_size":4,
"min_level_size":25,
"field_id":3,
"kind":{"Delete":[11]}
}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(input).unwrap();
let tmpdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tmpdir), &operations);
}
#[test]
#[no_coverage]
fn reproduce_bug6() {
let input = r#"
[
{"key":45720,"group_size":1,"max_group_size":4,"min_level_size":0,"field_id":0,"kind":{"Insert":[120]}},
{"key":37463,"group_size":1,"max_group_size":4,"min_level_size":0,"field_id":0,"kind":{"Insert":[187]}},
{"key":21512,"group_size":23,"max_group_size":20,"min_level_size":23,"field_id":0,"kind":{"Insert":[181]}},
{"key":21511,"group_size":23,"max_group_size":20,"min_level_size":23,"field_id":0,"kind":{"Insert":[181]}},
{"key":37737,"group_size":12,"max_group_size":0,"min_level_size":6,"field_id":0,"kind":{"Insert":[181]}},
{"key":53042,"group_size":23,"max_group_size":20,"min_level_size":23,"field_id":0,"kind":{"Insert":[181]}}
]
"#;
let operations: Vec<Operation<u16>> = serde_json::from_str(input).unwrap();
let tmpdir = TempDir::new().unwrap();
compare_with_trivial_database(Rc::new(tmpdir), &operations);
}
}

1
milli/src/update/index_documents/extract/extract_facet_number_docids.rs
View File

@ -38,7 +38,6 @@ pub fn extract_facet_number_docids<R: io::Read + io::Seek>(
let key = FacetGroupKey { field_id, level: 0, left_bound: number };
let key_bytes = FacetGroupKeyCodec::<OrderedF64Codec>::bytes_encode(&key).unwrap();
facet_number_docids_sorter.insert(key_bytes, document_id.to_ne_bytes())?;
}

14
milli/src/update/index_documents/extract/extract_facet_string_docids.rs
View File

@ -6,6 +6,7 @@ use heed::BytesEncode;
use super::helpers::{create_sorter, sorter_into_reader, try_split_array_at, GrenadParameters};
use crate::heed_codec::facet::{FacetGroupKey, FacetGroupKeyCodec};
use crate::heed_codec::StrRefCodec;
use crate::update::index_documents::helpers::MAX_FACET_VALUE_LENGTH;
use crate::update::index_documents::merge_cbo_roaring_bitmaps;
use crate::{FieldId, Result};
@ -38,10 +39,19 @@ pub fn extract_facet_string_docids<R: io::Read + io::Seek>(
try_split_array_at::<_, 4>(bytes).unwrap();
let document_id = u32::from_be_bytes(document_id_bytes);
let normalised_value = std::str::from_utf8(normalized_value_bytes)?;
let mut normalised_value = std::str::from_utf8(normalized_value_bytes)?;
let normalised_truncated_value: String;
if normalised_value.len() > MAX_FACET_VALUE_LENGTH {
normalised_truncated_value = normalised_value
.char_indices()
.take_while(|(idx, _)| idx + 4 < MAX_FACET_VALUE_LENGTH)
.map(|(_, c)| c)
.collect();
normalised_value = normalised_truncated_value.as_str();
}
let key = FacetGroupKey { field_id, level: 0, left_bound: normalised_value };
let key_bytes = FacetGroupKeyCodec::<StrRefCodec>::bytes_encode(&key).unwrap();
// document id is encoded in native-endian because of the CBO roaring bitmap codec
facet_string_docids_sorter.insert(&key_bytes, document_id.to_ne_bytes())?;
}

9
milli/src/update/index_documents/extract/extract_fid_docid_facet_values.rs
View File

@ -12,6 +12,7 @@ use serde_json::Value;
use super::helpers::{create_sorter, keep_first, sorter_into_reader, GrenadParameters};
use crate::error::InternalError;
use crate::facet::value_encoding::f64_into_bytes;
use crate::update::index_documents::helpers::MAX_FACET_VALUE_LENGTH;
use crate::update::index_documents::{create_writer, writer_into_reader};
use crate::{CboRoaringBitmapCodec, DocumentId, FieldId, Result, BEU32};
@ -87,8 +88,14 @@ pub fn extract_fid_docid_facet_values<R: io::Read + io::Seek>(
// insert normalized and original facet string in sorter
for (normalized, original) in strings.into_iter().filter(|(n, _)| !n.is_empty()) {
let normalised_truncated_value: String = normalized
.char_indices()
.take_while(|(idx, _)| idx + 4 < MAX_FACET_VALUE_LENGTH)
.map(|(_, c)| c)
.collect();
key_buffer.truncate(size_of::<FieldId>() + size_of::<DocumentId>());
key_buffer.extend_from_slice(normalized.as_bytes());
key_buffer.extend_from_slice(normalised_truncated_value.as_bytes());
fid_docid_facet_strings_sorter.insert(&key_buffer, original.as_bytes())?;
}
}

14
milli/src/update/index_documents/helpers/mod.rs
View File

@ -18,8 +18,20 @@ pub use merge_functions::{
serialize_roaring_bitmap, MergeFn,
};
/// The maximum length a LMDB key can be.
///
/// Note that the actual allowed length is a little bit higher, but
/// we keep a margin of safety.
const MAX_LMDB_KEY_LENGTH: usize = 500;
/// The maximum length a field value can be when inserted in an LMDB key.
///
/// This number is determined by the keys of the different facet databases
/// and adding a margin of safety.
pub const MAX_FACET_VALUE_LENGTH: usize = MAX_LMDB_KEY_LENGTH - 20;
/// The maximum length a word can be
pub const MAX_WORD_LENGTH: usize = 250;
pub const MAX_WORD_LENGTH: usize = MAX_LMDB_KEY_LENGTH / 2;
pub fn valid_lmdb_key(key: impl AsRef<[u8]>) -> bool {
key.as_ref().len() <= MAX_WORD_LENGTH * 2 && !key.as_ref().is_empty()

20
milli/src/update/index_documents/mod.rs
View File

@ -1821,4 +1821,24 @@ mod tests {
let words_fst = index.words_fst(&rtxn).unwrap();
assert!(!words_fst.contains(&long_word));
}
#[test]
fn long_facet_values_must_not_crash() {
let index = TempIndex::new();
// this is obviousy too long
let long_word = "lol".repeat(1000);
let doc1 = documents! {[{
"id": "1",
"title": long_word,
}]};
index
.update_settings(|settings| {
settings.set_filterable_fields(hashset! { S("title") });
})
.unwrap();
index.add_documents(doc1).unwrap();
}
}