use a bufreader everytime there is a grenad<file>

This commit is contained in:
Tamo
2023-09-28 16:26:01 +02:00
parent 8fe8ddea79
commit d772073dfa
22 changed files with 122 additions and 86 deletions

View File

@ -1,5 +1,6 @@
use std::borrow::Cow;
use std::fs::File;
use std::io::BufReader;
use grenad::CompressionType;
use heed::types::ByteSlice;
@ -30,7 +31,7 @@ pub struct FacetsUpdateBulk<'i> {
facet_type: FacetType,
field_ids: Vec<FieldId>,
// None if level 0 does not need to be updated
new_data: Option<grenad::Reader<File>>,
new_data: Option<grenad::Reader<BufReader<File>>>,
}
impl<'i> FacetsUpdateBulk<'i> {
@ -38,7 +39,7 @@ impl<'i> FacetsUpdateBulk<'i> {
index: &'i Index,
field_ids: Vec<FieldId>,
facet_type: FacetType,
new_data: grenad::Reader<File>,
new_data: grenad::Reader<BufReader<File>>,
group_size: u8,
min_level_size: u8,
) -> FacetsUpdateBulk<'i> {
@ -187,7 +188,7 @@ impl<R: std::io::Read + std::io::Seek> FacetsUpdateBulkInner<R> {
&self,
field_id: FieldId,
txn: &RoTxn,
) -> Result<(Vec<grenad::Reader<File>>, RoaringBitmap)> {
) -> Result<(Vec<grenad::Reader<BufReader<File>>>, RoaringBitmap)> {
let mut all_docids = RoaringBitmap::new();
let subwriters = self.compute_higher_levels(txn, field_id, 32, &mut |bitmaps, _| {
for bitmap in bitmaps {
@ -259,7 +260,7 @@ impl<R: std::io::Read + std::io::Seek> FacetsUpdateBulkInner<R> {
field_id: u16,
level: u8,
handle_group: &mut dyn FnMut(&[RoaringBitmap], &'t [u8]) -> Result<()>,
) -> Result<Vec<grenad::Reader<File>>> {
) -> Result<Vec<grenad::Reader<BufReader<File>>>> {
if level == 0 {
self.read_level_0(rtxn, field_id, handle_group)?;
// Level 0 is already in the database

View File

@ -1,5 +1,6 @@
use std::collections::HashMap;
use std::fs::File;
use std::io::BufReader;
use heed::types::{ByteSlice, DecodeIgnore};
use heed::{BytesDecode, Error, RoTxn, RwTxn};
@ -34,14 +35,14 @@ pub struct FacetsUpdateIncremental<'i> {
index: &'i Index,
inner: FacetsUpdateIncrementalInner,
facet_type: FacetType,
new_data: grenad::Reader<File>,
new_data: grenad::Reader<BufReader<File>>,
}
impl<'i> FacetsUpdateIncremental<'i> {
pub fn new(
index: &'i Index,
facet_type: FacetType,
new_data: grenad::Reader<File>,
new_data: grenad::Reader<BufReader<File>>,
group_size: u8,
min_level_size: u8,
max_group_size: u8,

View File

@ -78,6 +78,7 @@ pub const FACET_MIN_LEVEL_SIZE: u8 = 5;
use std::collections::BTreeSet;
use std::fs::File;
use std::io::BufReader;
use std::iter::FromIterator;
use charabia::normalizer::{Normalize, NormalizerOption};
@ -108,13 +109,17 @@ pub struct FacetsUpdate<'i> {
index: &'i Index,
database: heed::Database<FacetGroupKeyCodec<ByteSliceRefCodec>, FacetGroupValueCodec>,
facet_type: FacetType,
new_data: grenad::Reader<File>,
new_data: grenad::Reader<BufReader<File>>,
group_size: u8,
max_group_size: u8,
min_level_size: u8,
}
impl<'i> FacetsUpdate<'i> {
pub fn new(index: &'i Index, facet_type: FacetType, new_data: grenad::Reader<File>) -> Self {
pub fn new(
index: &'i Index,
facet_type: FacetType,
new_data: grenad::Reader<BufReader<File>>,
) -> Self {
let database = match facet_type {
FacetType::String => index
.facet_id_string_docids