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

Introduce the depth_first_search path resolution function

Browse Source
This commit is contained in:
Kerollmops 2020-09-28 13:54:29 +02:00
parent 51c237f9d8
commit a3821a0b33
No known key found for this signature in database
GPG Key ID: 92ADA4E935E71FA4
1 changed files with 40 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
src/search.rs
View File

@ -186,8 +186,6 @@ impl<'a> Search<'a> {
debug!("candidates: {:?}", candidates); debug!("candidates: {:?}", candidates);
let mut documents = Vec::new();
// If there is only one query word, no need to compute the best proximities. // If there is only one query word, no need to compute the best proximities.
if derived_words.len() == 1 || candidates.is_empty() { if derived_words.len() == 1 || candidates.is_empty() {
let found_words = derived_words.into_iter().flat_map(|(w, _)| w).map(|(w, _)| w).collect(); let found_words = derived_words.into_iter().flat_map(|(w, _)| w).map(|(w, _)| w).collect();
@ -211,31 +209,53 @@ impl<'a> Search<'a> {
pairs pairs
} }
let mut answer = RoaringBitmap::new(); fn depth_first_search(
for (i, words) in derived_words.windows(2).enumerate() { index: &Index,
let pairs = words_pair_combinations(&words[0].0, &words[1].0); rtxn: &heed::RoTxn,
eprintln!("found pairs {:?}", pairs); words: &[(HashMap<String, (u8, RoaringBitmap)>, RoaringBitmap)],
candidates: &RoaringBitmap,
parent_docids: Option<&RoaringBitmap>,
) -> anyhow::Result<Option<RoaringBitmap>>
{
let (words1, words2) = (&words[0].0, &words[1].0);
let pairs = words_pair_combinations(words1, words2);
let mut pairs_union = RoaringBitmap::new(); for proximity in 1..=8 {
'pairs: for (w1, w2) in pairs {
for prox in 1..=7 { let mut docids = RoaringBitmap::new();
let key = (w1, w2, prox); if proximity == 8 {
eprintln!("{:?}", key); docids = candidates.clone();
if let Some(docids) = index.word_pair_proximity_docids.get(rtxn, &key)? { } else {
pairs_union.union_with(&docids); for (w1, w2) in pairs.iter().cloned() {
continue 'pairs; let key = (w1, w2, proximity);
if let Some(di) = index.word_pair_proximity_docids.get(rtxn, &key)? {
docids.union_with(&di);
}
}
};
if let Some(parent_docids) = &parent_docids {
docids.intersect_with(parent_docids);
}
if !docids.is_empty() {
let words = &words[1..];
// We are the last word.
if words.len() < 2 { return Ok(Some(docids)) }
if let Some(di) = depth_first_search(index, rtxn, words, candidates, Some(&docids))? {
return Ok(Some(di))
} }
} }
} }
if i == 0 { Ok(None)
answer = pairs_union;
} else {
answer.intersect_with(&pairs_union);
}
} }
documents.push(answer); let mut documents = Vec::new();
let answer = depth_first_search(index, rtxn, &derived_words, &candidates, None)?;
if let Some(answer) = answer {
documents.push(answer);
}
let found_words = derived_words.into_iter().flat_map(|(w, _)| w).map(|(w, _)| w).collect(); let found_words = derived_words.into_iter().flat_map(|(w, _)| w).map(|(w, _)| w).collect();
let documents_ids = documents.into_iter().flatten().take(limit).collect(); let documents_ids = documents.into_iter().flatten().take(limit).collect();