searchengines-web/MeiliSearch
searchengines-web/MeiliSearch
1
0
Fork 0
mirror of https://github.com/meilisearch/MeiliSearch synced 2024-11-30 00:34:26 +01:00
Code Issues Releases Wiki Activity

Fix the typo criterion

Browse Source
This commit is contained in:
Kerollmops 2021-02-23 17:33:20 +01:00
parent a273c46559
commit c5a32fd4fa
No known key found for this signature in database
GPG Key ID: 92ADA4E935E71FA4
1 changed files with 60 additions and 55 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
milli/src/search/criteria/typo.rs
View File

@ -3,17 +3,13 @@ use std::{borrow::Cow, collections::HashMap, mem::take};
use anyhow::bail; use anyhow::bail;
use roaring::RoaringBitmap; use roaring::RoaringBitmap;
use crate::search::query_tree::{Operation, Query, QueryKind}; use crate::search::query_tree::{maximum_typo, Operation, Query, QueryKind};
use crate::search::word_derivations; use crate::search::word_derivations;
use super::{Candidates, Criterion, CriterionResult, Context, query_docids, query_pair_proximity_docids}; use super::{Candidates, Criterion, CriterionResult, Context, query_docids, query_pair_proximity_docids};
// FIXME we must stop when the number of typos is equal to
// the maximum number of typos for this query tree.
const MAX_NUM_TYPOS: u8 = 8;
pub struct Typo<'t> { pub struct Typo<'t> {
ctx: &'t dyn Context, ctx: &'t dyn Context,
query_tree: Option<Operation>, query_tree: Option<(usize, Operation)>,
number_typos: u8, number_typos: u8,
candidates: Candidates, candidates: Candidates,
bucket_candidates: Option<RoaringBitmap>, bucket_candidates: Option<RoaringBitmap>,
@ -31,7 +27,7 @@ impl<'t> Typo<'t> {
{ {
Ok(Typo { Ok(Typo {
ctx, ctx,
query_tree, query_tree: query_tree.map(|op| (maximum_typo(&op), op)),
number_typos: 0, number_typos: 0,
candidates: candidates.map_or_else(Candidates::default, Candidates::Allowed), candidates: candidates.map_or_else(Candidates::default, Candidates::Allowed),
bucket_candidates: None, bucket_candidates: None,
@ -62,13 +58,17 @@ impl<'t> Typo<'t> {
impl<'t> Criterion for Typo<'t> { impl<'t> Criterion for Typo<'t> {
fn next(&mut self) -> anyhow::Result<Option<CriterionResult>> { fn next(&mut self) -> anyhow::Result<Option<CriterionResult>> {
use Candidates::{Allowed, Forbidden}; use Candidates::{Allowed, Forbidden};
while self.number_typos < MAX_NUM_TYPOS { loop {
match (&mut self.query_tree, &mut self.candidates) { match (&mut self.query_tree, &mut self.candidates) {
(_, Allowed(candidates)) if candidates.is_empty() => { (_, Allowed(candidates)) if candidates.is_empty() => {
self.query_tree = None; self.query_tree = None;
self.candidates = Candidates::default(); self.candidates = Candidates::default();
}, },
(Some(query_tree), Allowed(candidates)) => { (Some((max_typos, query_tree)), Allowed(candidates)) => {
if self.number_typos as usize > *max_typos {
self.query_tree = None;
self.candidates = Candidates::default();
} else {
let fst = self.ctx.words_fst(); let fst = self.ctx.words_fst();
let new_query_tree = if self.number_typos < 2 { let new_query_tree = if self.number_typos < 2 {
alterate_query_tree(&fst, query_tree.clone(), self.number_typos, &mut self.typo_cache)? alterate_query_tree(&fst, query_tree.clone(), self.number_typos, &mut self.typo_cache)?
@ -94,8 +94,13 @@ impl<'t> Criterion for Typo<'t> {
candidates: new_candidates, candidates: new_candidates,
bucket_candidates, bucket_candidates,
})); }));
}
}, },
(Some(query_tree), Forbidden(candidates)) => { (Some((max_typos, query_tree)), Forbidden(candidates)) => {
if self.number_typos as usize > *max_typos {
self.query_tree = None;
self.candidates = Candidates::default();
} else {
let fst = self.ctx.words_fst(); let fst = self.ctx.words_fst();
let new_query_tree = if self.number_typos < 2 { let new_query_tree = if self.number_typos < 2 {
alterate_query_tree(&fst, query_tree.clone(), self.number_typos, &mut self.typo_cache)? alterate_query_tree(&fst, query_tree.clone(), self.number_typos, &mut self.typo_cache)?
@ -121,6 +126,7 @@ impl<'t> Criterion for Typo<'t> {
candidates: new_candidates, candidates: new_candidates,
bucket_candidates, bucket_candidates,
})); }));
}
}, },
(None, Allowed(_)) => { (None, Allowed(_)) => {
let candidates = take(&mut self.candidates).into_inner(); let candidates = take(&mut self.candidates).into_inner();
@ -135,7 +141,8 @@ impl<'t> Criterion for Typo<'t> {
Some(parent) => { Some(parent) => {
match parent.next()? { match parent.next()? {
Some(CriterionResult { query_tree, candidates, bucket_candidates }) => { Some(CriterionResult { query_tree, candidates, bucket_candidates }) => {
self.query_tree = query_tree; self.query_tree = query_tree.map(|op| (maximum_typo(&op), op));
self.number_typos = 0;
self.candidates = Candidates::Allowed(candidates); self.candidates = Candidates::Allowed(candidates);
self.bucket_candidates = bucket_candidates; self.bucket_candidates = bucket_candidates;
}, },
@ -147,8 +154,6 @@ impl<'t> Criterion for Typo<'t> {
}, },
} }
} }
Ok(None)
} }
} }