snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2025-01-19 09:35:51 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Breaking change: remove vector for score details

Browse Source
This commit is contained in:
Louis Dureuil 2024-03-27 15:36:49 +01:00
parent 5509bafff8
commit 928e6e4c05
No known key found for this signature in database
3 changed files with 15 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
meilisearch/src/search.rs
View File

@ -633,10 +633,8 @@ pub fn perform_search(
let mut semantic_score = None; let mut semantic_score = None;
for details in &score { for details in &score {
if let ScoreDetails::Vector(score_details::Vector { if let ScoreDetails::Vector(score_details::Vector { similarity: Some(similarity) }) =
target_vector: _, details
value_similarity: Some((_matching_vector, similarity)),
}) = details
{ {
semantic_score = Some(*similarity); semantic_score = Some(*similarity);
break; break;

16
milli/src/score_details.rs
View File

@ -98,9 +98,9 @@ impl ScoreDetails {
ScoreDetails::ExactWords(e) => RankOrValue::Rank(e.rank()), ScoreDetails::ExactWords(e) => RankOrValue::Rank(e.rank()),
ScoreDetails::Sort(sort) => RankOrValue::Sort(sort), ScoreDetails::Sort(sort) => RankOrValue::Sort(sort),
ScoreDetails::GeoSort(geosort) => RankOrValue::GeoSort(geosort), ScoreDetails::GeoSort(geosort) => RankOrValue::GeoSort(geosort),
ScoreDetails::Vector(vector) => RankOrValue::Score( ScoreDetails::Vector(vector) => {
vector.value_similarity.as_ref().map(|(_, s)| *s as f64).unwrap_or(0.0f64), RankOrValue::Score(vector.similarity.as_ref().map(|s| *s as f64).unwrap_or(0.0f64))
), }
ScoreDetails::Skipped => RankOrValue::Rank(Rank { rank: 0, max_rank: 1 }), ScoreDetails::Skipped => RankOrValue::Rank(Rank { rank: 0, max_rank: 1 }),
} }
} }
@ -249,16 +249,13 @@ impl ScoreDetails {
order += 1; order += 1;
} }
ScoreDetails::Vector(s) => { ScoreDetails::Vector(s) => {
let vector = format!("vectorSort({:?})", s.target_vector); let similarity = s.similarity.as_ref();
let value = s.value_similarity.as_ref().map(|(v, _)| v);
let similarity = s.value_similarity.as_ref().map(|(_, s)| s);
let details = serde_json::json!({ let details = serde_json::json!({
"order": order, "order": order,
"value": value,
"similarity": similarity, "similarity": similarity,
}); });
details_map.insert(vector, details); details_map.insert("vectorSort".into(), details);
order += 1; order += 1;
} }
ScoreDetails::Skipped => { ScoreDetails::Skipped => {
@ -494,8 +491,7 @@ impl PartialOrd for GeoSort {
#[derive(Debug, Clone, PartialEq, PartialOrd)] #[derive(Debug, Clone, PartialEq, PartialOrd)]
pub struct Vector { pub struct Vector {
pub target_vector: Vec<f32>, pub similarity: Option<f32>,
pub value_similarity: Option<(Vec<f32>, f32)>,
} }
impl GeoSort { impl GeoSort {

28
milli/src/search/new/vector_sort.rs
View File

@ -12,7 +12,7 @@ pub struct VectorSort<Q: RankingRuleQueryTrait> {
query: Option<Q>, query: Option<Q>,
target: Vec<f32>, target: Vec<f32>,
vector_candidates: RoaringBitmap, vector_candidates: RoaringBitmap,
cached_sorted_docids: std::vec::IntoIter<(DocumentId, f32, Vec<f32>)>, cached_sorted_docids: std::vec::IntoIter<(DocumentId, f32)>,
limit: usize, limit: usize,
distribution_shift: Option<DistributionShift>, distribution_shift: Option<DistributionShift>,
embedder_index: u8, embedder_index: u8,
@ -70,14 +70,9 @@ impl<Q: RankingRuleQueryTrait> VectorSort<Q> {
for reader in readers.iter() { for reader in readers.iter() {
let nns_by_vector = let nns_by_vector =
reader.nns_by_vector(ctx.txn, target, self.limit, None, Some(vector_candidates))?; reader.nns_by_vector(ctx.txn, target, self.limit, None, Some(vector_candidates))?;
let vectors: std::result::Result<Vec<_>, _> = nns_by_vector results.extend(nns_by_vector.into_iter());
.iter()
.map(|(docid, _)| reader.item_vector(ctx.txn, *docid).transpose().unwrap())
.collect();
let vectors = vectors?;
results.extend(nns_by_vector.into_iter().zip(vectors).map(|((x, y), z)| (x, y, z)));
} }
results.sort_unstable_by_key(|(_, distance, _)| OrderedFloat(*distance)); results.sort_unstable_by_key(|(_, distance)| OrderedFloat(*distance));
self.cached_sorted_docids = results.into_iter(); self.cached_sorted_docids = results.into_iter();
Ok(()) Ok(())
@ -118,14 +113,11 @@ impl<'ctx, Q: RankingRuleQueryTrait> RankingRule<'ctx, Q> for VectorSort<Q> {
return Ok(Some(RankingRuleOutput { return Ok(Some(RankingRuleOutput {
query, query,
candidates: universe.clone(), candidates: universe.clone(),
score: ScoreDetails::Vector(score_details::Vector { score: ScoreDetails::Vector(score_details::Vector { similarity: None }),
target_vector: self.target.clone(),
value_similarity: None,
}),
})); }));
} }
for (docid, distance, vector) in self.cached_sorted_docids.by_ref() { for (docid, distance) in self.cached_sorted_docids.by_ref() {
if vector_candidates.contains(docid) { if vector_candidates.contains(docid) {
let score = 1.0 - distance; let score = 1.0 - distance;
let score = self let score = self
@ -135,10 +127,7 @@ impl<'ctx, Q: RankingRuleQueryTrait> RankingRule<'ctx, Q> for VectorSort<Q> {
return Ok(Some(RankingRuleOutput { return Ok(Some(RankingRuleOutput {
query, query,
candidates: RoaringBitmap::from_iter([docid]), candidates: RoaringBitmap::from_iter([docid]),
score: ScoreDetails::Vector(score_details::Vector { score: ScoreDetails::Vector(score_details::Vector { similarity: Some(score) }),
target_vector: self.target.clone(),
value_similarity: Some((vector, score)),
}),
})); }));
} }
} }
@ -154,10 +143,7 @@ impl<'ctx, Q: RankingRuleQueryTrait> RankingRule<'ctx, Q> for VectorSort<Q> {
return Ok(Some(RankingRuleOutput { return Ok(Some(RankingRuleOutput {
query, query,
candidates: universe.clone(), candidates: universe.clone(),
score: ScoreDetails::Vector(score_details::Vector { score: ScoreDetails::Vector(score_details::Vector { similarity: None }),
target_vector: self.target.clone(),
value_similarity: None,
}),
})); }));
} }