fix bug with minimum_should_match and AllScorer (#2774)

This commit is contained in:
trinity-1686a
2025-12-14 10:10:45 +01:00
committed by GitHub
parent e9020d17d4
commit d0e1600135
3 changed files with 54 additions and 7 deletions

View File

@@ -23,7 +23,11 @@ pub struct AllWeight;
impl Weight for AllWeight {
fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
let all_scorer = AllScorer::new(reader.max_doc());
if boost != 1.0 {
Ok(Box::new(BoostScorer::new(all_scorer, boost)))
} else {
Ok(Box::new(all_scorer))
}
}
fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {

View File

@@ -193,18 +193,18 @@ impl<TScoreCombiner: ScoreCombiner> BooleanWeight<TScoreCombiner> {
return Ok(SpecializedScorer::Other(Box::new(EmptyScorer)));
}
let minimum_number_should_match = self
let effective_minimum_number_should_match = self
.minimum_number_should_match
.saturating_sub(should_special_scorer_counts.num_all_scorers);
let should_scorers: ShouldScorersCombinationMethod = {
let num_of_should_scorers = should_scorers.len();
if minimum_number_should_match > num_of_should_scorers {
if effective_minimum_number_should_match > num_of_should_scorers {
// We don't have enough scorers to satisfy the minimum number of should matches.
// The request will match no documents.
return Ok(SpecializedScorer::Other(Box::new(EmptyScorer)));
}
match minimum_number_should_match {
match effective_minimum_number_should_match {
0 if num_of_should_scorers == 0 => ShouldScorersCombinationMethod::Ignored,
0 => ShouldScorersCombinationMethod::Optional(scorer_union(
should_scorers,
@@ -226,7 +226,7 @@ impl<TScoreCombiner: ScoreCombiner> BooleanWeight<TScoreCombiner> {
scorer_disjunction(
should_scorers,
score_combiner_fn(),
self.minimum_number_should_match,
effective_minimum_number_should_match,
),
)),
}

View File

@@ -9,12 +9,14 @@ pub use self::boolean_weight::BooleanWeight;
#[cfg(test)]
mod tests {
use std::ops::Bound;
use super::*;
use crate::collector::tests::TEST_COLLECTOR_WITH_SCORE;
use crate::collector::TopDocs;
use crate::collector::{Count, TopDocs};
use crate::query::term_query::TermScorer;
use crate::query::{
AllScorer, EmptyScorer, EnableScoring, Intersection, Occur, Query, QueryParser,
AllScorer, EmptyScorer, EnableScoring, Intersection, Occur, Query, QueryParser, RangeQuery,
RequiredOptionalScorer, Scorer, SumCombiner, TermQuery,
};
use crate::schema::*;
@@ -374,4 +376,45 @@ mod tests {
}
Ok(())
}
#[test]
pub fn test_min_should_match_with_all_query() -> crate::Result<()> {
let mut schema_builder = Schema::builder();
let text_field = schema_builder.add_text_field("text", TEXT);
let num_field =
schema_builder.add_i64_field("num", NumericOptions::default().set_fast().set_indexed());
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut index_writer: IndexWriter = index.writer_for_tests()?;
index_writer.add_document(doc!(text_field => "apple", num_field => 10i64))?;
index_writer.add_document(doc!(text_field => "banana", num_field => 20i64))?;
index_writer.commit()?;
let searcher = index.reader()?.searcher();
let effective_all_match_query: Box<dyn Query> = Box::new(RangeQuery::new(
Bound::Excluded(Term::from_field_i64(num_field, 0)),
Bound::Unbounded,
));
let term_query: Box<dyn Query> = Box::new(TermQuery::new(
Term::from_field_text(text_field, "apple"),
IndexRecordOption::Basic,
));
// in some previous version, we would remove the 2 all_match, but then say we need *4*
// matches out of the 3 term queries, which matches nothing.
let mut bool_query = BooleanQuery::new(vec![
(Occur::Should, effective_all_match_query.box_clone()),
(Occur::Should, effective_all_match_query.box_clone()),
(Occur::Should, term_query.box_clone()),
(Occur::Should, term_query.box_clone()),
(Occur::Should, term_query.box_clone()),
]);
bool_query.set_minimum_number_should_match(4);
let count = searcher.search(&bool_query, &Count)?;
assert_eq!(count, 1);
Ok(())
}
}