snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2024-11-23 18:45:06 +08:00
Code Issues Actions 3 Packages Projects Releases Wiki Activity

feat: Made query handle synonyms via the Store

Browse Source
This commit is contained in:
Clément Renault 2019-06-13 16:38:37 +02:00
parent 18736bdcd0
commit 707d7b062b
No known key found for this signature in database
GPG Key ID: 0151CDAB43460DAE
1 changed files with 30 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

58
meilidb-core/src/query_builder.rs
View File

@ -8,40 +8,46 @@ use rayon::slice::ParallelSliceMut;
use slice_group_by::GroupByMut; use slice_group_by::GroupByMut;
use meilidb_tokenizer::{is_cjk, split_query_string}; use meilidb_tokenizer::{is_cjk, split_query_string};
use hashbrown::{HashMap, HashSet}; use hashbrown::{HashMap, HashSet};
use fst::Streamer; use fst::{Streamer, IntoStreamer};
use log::info; use log::info;
use crate::automaton::{self, DfaExt, AutomatonExt}; use crate::automaton::{self, DfaExt, AutomatonExt, build_dfa, build_prefix_dfa};
use crate::distinct_map::{DistinctMap, BufferedDistinctMap}; use crate::distinct_map::{DistinctMap, BufferedDistinctMap};
use crate::criterion::Criteria; use crate::criterion::Criteria;
use crate::raw_documents_from_matches; use crate::raw_documents_from_matches;
use crate::{Match, DocumentId, Store, RawDocument, Document}; use crate::{Match, DocumentId, Store, RawDocument, Document};
fn generate_automatons(query: &str, synonyms: &HashMap<&str, &[&str]>) -> Vec<(usize, DfaExt)> { fn generate_automatons<S: Store>(query: &str, store: &S) -> Result<Vec<(usize, DfaExt)>, S::Error> {
let has_end_whitespace = query.chars().last().map_or(false, char::is_whitespace); let has_end_whitespace = query.chars().last().map_or(false, char::is_whitespace);
let mut groups = split_query_string(query).map(str::to_lowercase).peekable(); let mut groups = split_query_string(query).map(str::to_lowercase).peekable();
let mut automatons = Vec::new(); let mut automatons = Vec::new();
let mut index = 0; let mut index = 0;
let synonyms = store.synonyms()?;
while let Some(word) = groups.next() { while let Some(word) = groups.next() {
let word = word.as_str(); let word = word.as_str();
let has_following_word = groups.peek().is_some(); let has_following_word = groups.peek().is_some();
let is_prefix_dfa = has_following_word || has_end_whitespace || word.chars().all(is_cjk); let not_prefix_dfa = has_following_word || has_end_whitespace || word.chars().all(is_cjk);
let words = synonyms.get(word).cloned().unwrap_or_default().iter().chain(Some(&word));
for word in words { let lev = if not_prefix_dfa { build_dfa(word) } else { build_prefix_dfa(word) };
let lev = if is_prefix_dfa { let mut stream = synonyms.search(&lev).into_stream();
automaton::build_dfa(word) while let Some(synonym) = stream.next() {
} else { if let Some(words) = store.alternatives_to(synonym)? {
automaton::build_prefix_dfa(word) let mut stream = words.into_stream();
}; while let Some(word) = stream.next() {
automatons.push((index, lev)); let word = std::str::from_utf8(word).unwrap();
let lev = if not_prefix_dfa { build_dfa(word) } else { build_prefix_dfa(word) };
automatons.push((index, lev));
}
}
} }
automatons.push((index, lev));
index += 1; index += 1;
} }
automatons Ok(automatons)
} }
pub struct QueryBuilder<'c, S, FI = fn(DocumentId) -> bool> { pub struct QueryBuilder<'c, S, FI = fn(DocumentId) -> bool> {
@ -91,17 +97,7 @@ impl<'c, S, FI> QueryBuilder<'c, S, FI>
where S: Store, where S: Store,
{ {
fn query_all(&self, query: &str) -> Result<Vec<RawDocument>, S::Error> { fn query_all(&self, query: &str) -> Result<Vec<RawDocument>, S::Error> {
let map = { let automatons = generate_automatons(query, &self.store)?;
let mut map = HashMap::new();
map.insert("hello", &["bonjour", "salut"][..]);
map.insert("bonjour", &["hello", "salut"]);
map.insert("salut", &["hello", "bonjour"]);
map
};
let automatons = generate_automatons(query, &map);
let words = self.store.words()?.as_fst(); let words = self.store.words()?.as_fst();
let mut stream = { let mut stream = {
@ -450,11 +446,13 @@ mod tests {
} }
#[test] #[test]
fn simple_synonymes() { fn simple_synonyms() {
let store = InMemorySetStore::from_iter(vec![ let mut store = InMemorySetStore::from_iter(vec![
(&b"hello"[..], &[doc_index(0, 0)][..]), (&b"hello"[..], &[doc_index(0, 0)][..]),
]); ]);
store.add_synonym("bonjour", SetBuf::from_dirty(vec!["hello"]));
let builder = QueryBuilder::new(&store); let builder = QueryBuilder::new(&store);
let results = builder.query("hello", 0..20).unwrap(); let results = builder.query("hello", 0..20).unwrap();
let mut iter = results.into_iter(); let mut iter = results.into_iter();
@ -481,13 +479,17 @@ mod tests {
} }
#[test] #[test]
fn harder_synonymes() { fn harder_synonyms() {
let store = InMemorySetStore::from_iter(vec![ let mut store = InMemorySetStore::from_iter(vec![
(&b"hello"[..], &[doc_index(0, 0)][..]), (&b"hello"[..], &[doc_index(0, 0)][..]),
(&b"bonjour"[..], &[doc_index(1, 3)]), (&b"bonjour"[..], &[doc_index(1, 3)]),
(&b"salut"[..], &[doc_index(2, 5)]), (&b"salut"[..], &[doc_index(2, 5)]),
]); ]);
store.add_synonym("hello", SetBuf::from_dirty(vec!["bonjour", "salut"]));
store.add_synonym("bonjour", SetBuf::from_dirty(vec!["hello", "salut"]));
store.add_synonym("salut", SetBuf::from_dirty(vec!["hello", "bonjour"]));
let builder = QueryBuilder::new(&store); let builder = QueryBuilder::new(&store);
let results = builder.query("hello", 0..20).unwrap(); let results = builder.query("hello", 0..20).unwrap();
let mut iter = results.into_iter(); let mut iter = results.into_iter();