rust/tantivy
1
0
Fork 0
mirror of https://github.com/quickwit-oss/tantivy.git synced 2026-01-06 09:12:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

Renaming intersection / exclude

Browse Source
This commit is contained in:
Paul Masurel
2018-02-16 11:55:56 +09:00
parent 930d3db2f7
commit 31e5580bfa
6 changed files with 54 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
src/postings/intersection.rs
View File

@@ -1,20 +1,22 @@
use postings::DocSet;
use postings::SkipResult;
use query::Scorer;
use DocId;
use Score;
/// Creates a `DocSet` that iterator through the intersection of two `DocSet`s.
pub struct IntersectionDocSet<TDocSet: DocSet> {
pub struct Intersection<TDocSet: DocSet> {
docsets: Vec<TDocSet>,
finished: bool,
doc: DocId,
}
impl<TDocSet: DocSet> From<Vec<TDocSet>> for IntersectionDocSet<TDocSet> {
fn from(mut docsets: Vec<TDocSet>) -> IntersectionDocSet<TDocSet> {
impl<TDocSet: DocSet> From<Vec<TDocSet>> for Intersection<TDocSet> {
fn from(mut docsets: Vec<TDocSet>) -> Intersection<TDocSet> {
assert!(docsets.len() >= 2);
docsets.sort_by_key(|docset| docset.size_hint());
IntersectionDocSet {
Intersection {
docsets,
finished: false,
doc: 0u32,
@@ -22,7 +24,7 @@ impl<TDocSet: DocSet> From<Vec<TDocSet>> for IntersectionDocSet<TDocSet> {
}
}
impl<TDocSet: DocSet> IntersectionDocSet<TDocSet> {
impl<TDocSet: DocSet> Intersection<TDocSet> {
/// Returns an array to the underlying `DocSet`s of the intersection.
/// These `DocSet` are in the same position as the `IntersectionDocSet`,
/// so that user can access their `docfreq` and `positions`.
@@ -31,7 +33,7 @@ impl<TDocSet: DocSet> IntersectionDocSet<TDocSet> {
}
}
impl<TDocSet: DocSet> DocSet for IntersectionDocSet<TDocSet> {
impl<TDocSet: DocSet> DocSet for Intersection<TDocSet> {
#[allow(never_loop)]
fn advance(&mut self) -> bool {
if self.finished {
@@ -114,10 +116,10 @@ impl<TDocSet: DocSet> DocSet for IntersectionDocSet<TDocSet> {
}
fn doc(&self) -> DocId {
self.doc
}
fn size_hint(&self) -> u32 {
self.docsets
.iter()
@@ -127,10 +129,21 @@ impl<TDocSet: DocSet> DocSet for IntersectionDocSet<TDocSet> {
}
}
impl<TScorer> Scorer for Intersection<TScorer>
where TScorer: Scorer {
fn score(&mut self) -> Score {
self.docsets
.iter_mut()
.map(Scorer::score)
.sum()
}
}
#[cfg(test)]
mod tests {
use postings::SkipResult;
use postings::{DocSet, IntersectionDocSet, VecPostings};
use postings::{DocSet, Intersection, VecPostings};
use postings::tests::test_skip_against_unoptimized;
#[test]
@@ -138,7 +151,7 @@ mod tests {
{
let left = VecPostings::from(vec![1, 3, 9]);
let right = VecPostings::from(vec![3, 4, 9, 18]);
let mut intersection = IntersectionDocSet::from(vec![left, right]);
let mut intersection = Intersection::from(vec![left, right]);
assert!(intersection.advance());
assert_eq!(intersection.doc(), 3);
assert!(intersection.advance());
@@ -149,7 +162,7 @@ mod tests {
let a = VecPostings::from(vec![1, 3, 9]);
let b = VecPostings::from(vec![3, 4, 9, 18]);
let c = VecPostings::from(vec![1, 5, 9, 111]);
let mut intersection = IntersectionDocSet::from(vec![a, b, c]);
let mut intersection = Intersection::from(vec![a, b, c]);
assert!(intersection.advance());
assert_eq!(intersection.doc(), 9);
assert!(!intersection.advance());
@@ -160,7 +173,7 @@ mod tests {
fn test_intersection_zero() {
let left = VecPostings::from(vec![0]);
let right = VecPostings::from(vec![0]);
let mut intersection = IntersectionDocSet::from(vec![left, right]);
let mut intersection = Intersection::from(vec![left, right]);
assert!(intersection.advance());
assert_eq!(intersection.doc(), 0);
}
@@ -170,7 +183,7 @@ mod tests {
fn test_intersection_skip() {
let left = VecPostings::from(vec![0, 1, 2, 4]);
let right = VecPostings::from(vec![2, 5]);
let mut intersection = IntersectionDocSet::from(vec![left, right]);
let mut intersection = Intersection::from(vec![left, right]);
assert_eq!(intersection.skip_next(2), SkipResult::Reached);
assert_eq!(intersection.doc(), 2);
}
@@ -181,17 +194,17 @@ mod tests {
test_skip_against_unoptimized(|| {
let left = VecPostings::from(vec![4]);
let right = VecPostings::from(vec![2, 5]);
box IntersectionDocSet::from(vec![left, right])
box Intersection::from(vec![left, right])
}, vec![0,2,4,5,6]);
test_skip_against_unoptimized(|| {
let mut left = VecPostings::from(vec![1, 4, 5, 6]);
let mut right = VecPostings::from(vec![2, 5, 10]);
left.advance();
right.advance();
box IntersectionDocSet::from(vec![left, right])
box Intersection::from(vec![left, right])
}, vec![0,1,2,3,4,5,6,7,10,11]);
test_skip_against_unoptimized(|| {
box IntersectionDocSet::from(vec![
box Intersection::from(vec![
VecPostings::from(vec![1, 4, 5, 6]),
VecPostings::from(vec![1, 2, 5, 6]),
VecPostings::from(vec![1, 4, 5, 6]),
@@ -206,7 +219,7 @@ mod tests {
let a = VecPostings::from(vec![1, 3]);
let b = VecPostings::from(vec![1, 4]);
let c = VecPostings::from(vec![3, 9]);
let mut intersection = IntersectionDocSet::from(vec![a, b, c]);
let mut intersection = Intersection::from(vec![a, b, c]);
assert!(!intersection.advance());
}
}

4
src/postings/mod.rs
View File

@@ -30,7 +30,7 @@ pub use self::postings::Postings;
pub use self::vec_postings::VecPostings;
pub use self::segment_postings::{BlockSegmentPostings, SegmentPostings};
pub use self::intersection::IntersectionDocSet;
pub use self::intersection::Intersection;
pub use self::union::UnionDocSet;
pub use common::HasLen;
@@ -551,7 +551,7 @@ pub mod tests {
.inverted_index(TERM_D.field())
.read_postings(&*TERM_D, IndexRecordOption::Basic)
.unwrap();
let mut intersection = IntersectionDocSet::from(vec![
let mut intersection = Intersection::from(vec![
segment_postings_a,
segment_postings_b,
segment_postings_c,

8
src/query/boolean_query/boolean_weight.rs
View File

@@ -1,10 +1,10 @@
use query::Weight;
use core::SegmentReader;
use postings::{IntersectionDocSet, UnionDocSet};
use postings::{Intersection, UnionDocSet};
use std::collections::HashMap;
use query::EmptyScorer;
use query::Scorer;
use query::ExcludeScorer;
use query::Exclude;
use super::BooleanScorer;
use query::OccurFilter;
use query::ConstScorer;
@@ -60,7 +60,7 @@ impl BooleanWeight {
let must_scorer_opt: Option<Box<Scorer + 'a>> = per_occur_scorers
.remove(&Occur::Must)
.map(|scorers| {
let scorer: Box<Scorer> = box ConstScorer::new(IntersectionDocSet::from(scorers));
let scorer: Box<Scorer> = box ConstScorer::new(Intersection::from(scorers));
scorer
});
@@ -77,7 +77,7 @@ impl BooleanWeight {
};
if let Some(exclude_scorer) = exclude_scorer_opt {
Ok(box ExcludeScorer::new(positive_scorer, exclude_scorer))
Ok(box Exclude::new(positive_scorer, exclude_scorer))
} else {
Ok(positive_scorer)
}

26
src/query/exclude.rs
View File

@@ -10,24 +10,28 @@ enum State {
Finished
}
pub struct ExcludeScorer<TDocSet, TDocSetExclude> {
/// Filters a given `DocSet` by removing the docs from a given `DocSet`.
///
/// The excluding docset has no impact on scoring.
pub struct Exclude<TDocSet, TDocSetExclude> {
underlying_docset: TDocSet,
excluding_docset: TDocSetExclude,
excluding_state: State,
}
impl<TDocSet, TDocSetExclude> ExcludeScorer<TDocSet, TDocSetExclude>
impl<TDocSet, TDocSetExclude> Exclude<TDocSet, TDocSetExclude>
where TDocSetExclude: DocSet {
pub fn new(underlying_docset: TDocSet, mut excluding_docset: TDocSetExclude) -> ExcludeScorer<TDocSet, TDocSetExclude> {
/// Creates a new `ExcludeScorer`
pub fn new(underlying_docset: TDocSet, mut excluding_docset: TDocSetExclude) -> Exclude<TDocSet, TDocSetExclude> {
let state =
if excluding_docset.advance() {
State::ExcludeOne(excluding_docset.doc())
} else {
State::Finished
};
ExcludeScorer {
Exclude {
underlying_docset,
excluding_docset,
excluding_state: state,
@@ -35,7 +39,7 @@ impl<TDocSet, TDocSetExclude> ExcludeScorer<TDocSet, TDocSetExclude>
}
}
impl<TDocSet, TDocSetExclude> ExcludeScorer<TDocSet, TDocSetExclude>
impl<TDocSet, TDocSetExclude> Exclude<TDocSet, TDocSetExclude>
where TDocSet: DocSet, TDocSetExclude: DocSet {
/// Returns true iff the doc is not removed.
@@ -73,7 +77,7 @@ impl<TDocSet, TDocSetExclude> ExcludeScorer<TDocSet, TDocSetExclude>
}
}
impl<TDocSet, TDocSetExclude> DocSet for ExcludeScorer<TDocSet, TDocSetExclude>
impl<TDocSet, TDocSetExclude> DocSet for Exclude<TDocSet, TDocSetExclude>
where TDocSet: DocSet, TDocSetExclude: DocSet {
fn advance(&mut self) -> bool {
@@ -113,8 +117,8 @@ impl<TDocSet, TDocSetExclude> DocSet for ExcludeScorer<TDocSet, TDocSetExclude>
}
impl<TDocSet, TDocSetExclude> Scorer for ExcludeScorer<TDocSet, TDocSetExclude>
where TDocSet: Scorer, TDocSetExclude: Scorer {
impl<TScorer, TDocSetExclude> Scorer for Exclude<TScorer, TDocSetExclude>
where TScorer: Scorer, TDocSetExclude: DocSet {
fn score(&mut self) -> Score {
self.underlying_docset.score()
}
@@ -130,7 +134,7 @@ mod tests {
#[test]
fn test_exclude() {
let mut exclude_scorer = ExcludeScorer::new(
let mut exclude_scorer = Exclude::new(
VecPostings::from(vec![1,2,5,8,10,15,24]),
VecPostings::from(vec![1,2,3,10,16,24])
);
@@ -144,7 +148,7 @@ mod tests {
#[test]
fn test_exclude_skip() {
test_skip_against_unoptimized(
|| box ExcludeScorer::new(
|| box Exclude::new(
VecPostings::from(vec![1, 2, 5, 8, 10, 15, 24]),
VecPostings::from(vec![1, 2, 3, 10, 16, 24])
),
@@ -158,7 +162,7 @@ mod tests {
let sample_exclude = sample_with_seed(10_000, 0.05, 2);
let sample_skip = sample_with_seed(10_000, 0.005, 3);
test_skip_against_unoptimized(
|| box ExcludeScorer::new(
|| box Exclude::new(
VecPostings::from(sample_include.clone()),
VecPostings::from(sample_exclude.clone())
),

2
src/query/mod.rs
View File

@@ -18,7 +18,7 @@ mod exclude;
mod reqopt_scorer;
pub use self::reqopt_scorer::RequiredOptionalScorer;
pub use self::exclude::ExcludeScorer;
pub use self::exclude::Exclude;
pub use self::bitset::BitSetDocSet;
pub use self::boolean_query::BooleanQuery;
pub use self::occur_filter::OccurFilter;

6
src/query/phrase_query/phrase_scorer.rs
View File

@@ -1,6 +1,6 @@
use query::Scorer;
use DocId;
use postings::{DocSet, IntersectionDocSet, Postings, SegmentPostings, SkipResult};
use postings::{DocSet, Intersection, Postings, SegmentPostings, SkipResult};
struct PostingsWithOffset {
offset: u32,
@@ -45,7 +45,7 @@ impl DocSet for PostingsWithOffset {
}
pub struct PhraseScorer {
intersection_docset: IntersectionDocSet<PostingsWithOffset>,
intersection_docset: Intersection<PostingsWithOffset>,
}
impl PhraseScorer {
@@ -56,7 +56,7 @@ impl PhraseScorer {
.map(|(offset, postings)| PostingsWithOffset::new(postings, offset as u32))
.collect();
PhraseScorer {
intersection_docset: IntersectionDocSet::from(postings_with_offsets),
intersection_docset: Intersection::from(postings_with_offsets),
}
}