searchengines-web/MeiliSearch
searchengines-web/MeiliSearch
1
0
Fork 0
mirror of https://github.com/meilisearch/MeiliSearch synced 2024-12-23 21:20:24 +01:00
Code Issues Releases Wiki Activity

take hard separator into account

Browse Source
This commit is contained in:
many 2021-03-02 11:14:10 +01:00 committed by Kerollmops
parent 6bf6b40495
commit 246286f0eb
No known key found for this signature in database
GPG Key ID: 92ADA4E935E71FA4
1 changed files with 34 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
milli/src/update/index_documents/store.rs
View File

@ -13,7 +13,7 @@ use grenad::{Reader, FileFuse, Writer, Sorter, CompressionType};
use heed::BytesEncode;
use linked_hash_map::LinkedHashMap;
use log::{debug, info};
use meilisearch_tokenizer::{Analyzer, AnalyzerConfig};
use meilisearch_tokenizer::{Analyzer, AnalyzerConfig, Token, TokenKind, token::SeparatorKind};
use ordered_float::OrderedFloat;
use roaring::RoaringBitmap;
use serde_json::Value;
@ -471,14 +471,11 @@ impl<'s, A: AsRef<[u8]>> Store<'s, A> {
};
let analyzed = self.analyzer.analyze(&content);
let tokens = analyzed
.tokens()
.filter(|t| t.is_word())
.map(|t| t.text().to_string());
let tokens = process_tokens(analyzed.tokens());
for (pos, word) in tokens.enumerate().take(MAX_POSITION) {
for (pos, token) in tokens.take_while(|(pos, _)| *pos < MAX_POSITION) {
let position = (attr as usize * MAX_POSITION + pos) as u32;
words_positions.entry(word).or_insert_with(SmallVec32::new).push(position);
words_positions.entry(token.text().to_string()).or_insert_with(SmallVec32::new).push(position);
}
}
}
@ -609,6 +606,36 @@ enum FacetValue {
Integer(i64),
}
/// take an iterator on tokens and compute their relative position depending on separator kinds
/// if it's an `Hard` separator we add an additional relative proximity of 8 between words,
/// else we keep the standart proximity of 1 between words.
fn process_tokens<'a>(tokens: impl Iterator<Item = Token<'a>>) -> impl Iterator<Item = (usize, Token<'a>)> {
tokens
.skip_while(|token| token.is_separator().is_some())
.scan((0, None), |(offset, prev_kind), token| {
match token.kind {
TokenKind::Word | TokenKind::StopWord | TokenKind::Unknown => {
*offset += match *prev_kind {
Some(TokenKind::Separator(SeparatorKind::Hard)) => 8,
Some(_) => 1,
None => 0,
};
*prev_kind = Some(token.kind)
}
TokenKind::Separator(SeparatorKind::Hard) => {
*prev_kind = Some(token.kind);
}
TokenKind::Separator(SeparatorKind::Soft)
if *prev_kind != Some(TokenKind::Separator(SeparatorKind::Hard)) => {
*prev_kind = Some(token.kind);
}
_ => (),
}
Some((*offset, token))
})
.filter(|(_, t)| t.is_word())
}
fn parse_facet_value(ftype: FacetType, value: &Value) -> anyhow::Result<SmallVec8<FacetValue>> {
use FacetValue::*;