0
0
Fork 0
bleve/search/collector/topn.go

279 lines
7.6 KiB
Go
Raw Normal View History

2016-08-10 10:13:38 +02:00
// Copyright (c) 2014 Couchbase, Inc.
2016-10-02 16:13:14 +02:00
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package collector
2016-08-09 15:18:53 +02:00
import (
2016-08-10 10:13:38 +02:00
"time"
2016-08-09 15:18:53 +02:00
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/search"
"golang.org/x/net/context"
)
// PreAllocSizeSkipCap will cap preallocation to this amount when
// size+skip exceeds this value
var PreAllocSizeSkipCap = 1000
type collectorCompare func(i, j *search.DocumentMatch) int
type collectorFixup func(d *search.DocumentMatch) error
// TopNCollector collects the top N hits, optionally skipping some results
type TopNCollector struct {
2016-08-09 15:18:53 +02:00
size int
skip int
total uint64
maxScore float64
2016-08-09 15:18:53 +02:00
took time.Duration
sort search.SortOrder
results search.DocumentMatchCollection
2016-08-09 15:18:53 +02:00
facetsBuilder *search.FacetsBuilder
switch collector store impl from slice to heap Additional testing has shown that the heap collector performs significantly better when larger numbers of hits are requested. The heap is also faster (though very close) when fewer (10) hits are requested. Here are the numbers from my laptop: slice: go test -run=xxx -bench=. -benchmem BenchmarkTop10of10000Scores-4 5000 396943 ns/op 2472 B/op 26 allocs/op BenchmarkTop100of10000Scores-4 2000 630894 ns/op 18848 B/op 116 allocs/op BenchmarkTop1000of10000Scores-4 100 14996445 ns/op 176552 B/op 1016 allocs/op BenchmarkTop10000of100000Scores-4 1 1878796320 ns/op 1857768 B/op 19023 allocs/op BenchmarkTop10of100000Scores-4 500 3858309 ns/op 2480 B/op 26 allocs/op BenchmarkTop100of100000Scores-4 300 4270086 ns/op 19000 B/op 116 allocs/op BenchmarkTop1000of100000Scores-4 50 30163705 ns/op 178024 B/op 1016 allocs/op BenchmarkTop10000of1000000Scores-4 1 3429557237 ns/op 1882008 B/op 19023 allocs/op PASS ok github.com/blevesearch/bleve/search/collector 16.316s heap: go test -run=xxx -bench=. -benchmem BenchmarkTop10of10000Scores-4 5000 341064 ns/op 2552 B/op 27 allocs/op BenchmarkTop100of10000Scores-4 3000 501922 ns/op 19744 B/op 117 allocs/op BenchmarkTop1000of10000Scores-4 1000 1759088 ns/op 184744 B/op 1017 allocs/op BenchmarkTop10000of100000Scores-4 50 25954696 ns/op 1939608 B/op 19024 allocs/op BenchmarkTop10of100000Scores-4 500 3814933 ns/op 2560 B/op 27 allocs/op BenchmarkTop100of100000Scores-4 300 4009369 ns/op 19896 B/op 117 allocs/op BenchmarkTop1000of100000Scores-4 200 6397276 ns/op 186184 B/op 1017 allocs/op BenchmarkTop10000of1000000Scores-4 20 81815315 ns/op 1963912 B/op 19024 allocs/op PASS ok github.com/blevesearch/bleve/search/collector 14.980s
2017-03-24 17:38:06 +01:00
store *collectStoreHeap
needDocIds bool
neededFields []string
cachedScoring []bool
cachedDesc []bool
lowestMatchOutsideResults *search.DocumentMatch
2016-08-09 15:18:53 +02:00
}
// CheckDoneEvery controls how frequently we check the context deadline
const CheckDoneEvery = uint64(1024)
2016-08-10 10:13:38 +02:00
// NewTopNCollector builds a collector to find the top 'size' hits
// skipping over the first 'skip' hits
// ordering hits by the provided sort order
func NewTopNCollector(size int, skip int, sort search.SortOrder) *TopNCollector {
hc := &TopNCollector{size: size, skip: skip, sort: sort}
// pre-allocate space on the store to avoid reslicing
// unless the size + skip is too large, then cap it
// everything should still work, just reslices as necessary
backingSize := size + skip + 1
if size+skip > PreAllocSizeSkipCap {
backingSize = PreAllocSizeSkipCap + 1
}
switch collector store impl from slice to heap Additional testing has shown that the heap collector performs significantly better when larger numbers of hits are requested. The heap is also faster (though very close) when fewer (10) hits are requested. Here are the numbers from my laptop: slice: go test -run=xxx -bench=. -benchmem BenchmarkTop10of10000Scores-4 5000 396943 ns/op 2472 B/op 26 allocs/op BenchmarkTop100of10000Scores-4 2000 630894 ns/op 18848 B/op 116 allocs/op BenchmarkTop1000of10000Scores-4 100 14996445 ns/op 176552 B/op 1016 allocs/op BenchmarkTop10000of100000Scores-4 1 1878796320 ns/op 1857768 B/op 19023 allocs/op BenchmarkTop10of100000Scores-4 500 3858309 ns/op 2480 B/op 26 allocs/op BenchmarkTop100of100000Scores-4 300 4270086 ns/op 19000 B/op 116 allocs/op BenchmarkTop1000of100000Scores-4 50 30163705 ns/op 178024 B/op 1016 allocs/op BenchmarkTop10000of1000000Scores-4 1 3429557237 ns/op 1882008 B/op 19023 allocs/op PASS ok github.com/blevesearch/bleve/search/collector 16.316s heap: go test -run=xxx -bench=. -benchmem BenchmarkTop10of10000Scores-4 5000 341064 ns/op 2552 B/op 27 allocs/op BenchmarkTop100of10000Scores-4 3000 501922 ns/op 19744 B/op 117 allocs/op BenchmarkTop1000of10000Scores-4 1000 1759088 ns/op 184744 B/op 1017 allocs/op BenchmarkTop10000of100000Scores-4 50 25954696 ns/op 1939608 B/op 19024 allocs/op BenchmarkTop10of100000Scores-4 500 3814933 ns/op 2560 B/op 27 allocs/op BenchmarkTop100of100000Scores-4 300 4009369 ns/op 19896 B/op 117 allocs/op BenchmarkTop1000of100000Scores-4 200 6397276 ns/op 186184 B/op 1017 allocs/op BenchmarkTop10000of1000000Scores-4 20 81815315 ns/op 1963912 B/op 19024 allocs/op PASS ok github.com/blevesearch/bleve/search/collector 14.980s
2017-03-24 17:38:06 +01:00
hc.store = newStoreHeap(backingSize, func(i, j *search.DocumentMatch) int {
return hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, i, j)
})
// these lookups traverse an interface, so do once up-front
if sort.RequiresDocID() {
hc.needDocIds = true
}
hc.neededFields = sort.RequiredFields()
hc.cachedScoring = sort.CacheIsScore()
hc.cachedDesc = sort.CacheDescending()
2016-08-09 15:18:53 +02:00
return hc
}
// Collect goes to the index to find the matching documents
func (hc *TopNCollector) Collect(ctx context.Context, searcher search.Searcher, reader index.IndexReader) error {
2016-08-09 15:18:53 +02:00
startTime := time.Now()
var err error
var next *search.DocumentMatch
// pre-allocate enough space in the DocumentMatchPool
// unless the size + skip is too large, then cap it
// everything should still work, just allocates DocumentMatches on demand
backingSize := hc.size + hc.skip + 1
if hc.size+hc.skip > PreAllocSizeSkipCap {
backingSize = PreAllocSizeSkipCap + 1
}
searchContext := &search.SearchContext{
DocumentMatchPool: search.NewDocumentMatchPool(backingSize+searcher.DocumentMatchPoolSize(), len(hc.sort)),
}
2016-08-09 15:18:53 +02:00
select {
case <-ctx.Done():
return ctx.Err()
default:
next, err = searcher.Next(searchContext)
2016-08-09 15:18:53 +02:00
}
for err == nil && next != nil {
if hc.total%CheckDoneEvery == 0 {
2016-08-09 15:18:53 +02:00
select {
case <-ctx.Done():
return ctx.Err()
default:
}
}
err = hc.collectSingle(searchContext, reader, next)
if err != nil {
break
}
next, err = searcher.Next(searchContext)
2016-08-09 15:18:53 +02:00
}
// compute search duration
hc.took = time.Since(startTime)
if err != nil {
return err
}
// finalize actual results
err = hc.finalizeResults(reader)
if err != nil {
return err
}
2016-08-09 15:18:53 +02:00
return nil
}
var sortByScoreOpt = []string{"_score"}
func (hc *TopNCollector) collectSingle(ctx *search.SearchContext, reader index.IndexReader, d *search.DocumentMatch) error {
var err error
// visit field terms for features that require it (sort, facets)
if len(hc.neededFields) > 0 {
err = hc.visitFieldTerms(reader, d)
if err != nil {
return err
}
}
2016-08-09 15:18:53 +02:00
// increment total hits
hc.total++
d.HitNumber = hc.total
// update max score
if d.Score > hc.maxScore {
hc.maxScore = d.Score
}
// see if we need to load ID (at this early stage, for example to sort on it)
if hc.needDocIds {
d.ID, err = reader.ExternalID(d.IndexInternalID)
2016-08-09 15:18:53 +02:00
if err != nil {
return err
}
}
// compute this hits sort value
if len(hc.sort) == 1 && hc.cachedScoring[0] {
d.Sort = sortByScoreOpt
} else {
hc.sort.Value(d)
}
// optimization, we track lowest sorting hit already removed from heap
2016-10-02 18:11:15 +02:00
// with this one comparison, we can avoid all heap operations if
// this hit would have been added and then immediately removed
if hc.lowestMatchOutsideResults != nil {
cmp := hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, d, hc.lowestMatchOutsideResults)
if cmp >= 0 {
// this hit can't possibly be in the result set, so avoid heap ops
ctx.DocumentMatchPool.Put(d)
return nil
}
}
hc.store.Add(d)
if hc.store.Len() > hc.size+hc.skip {
removed := hc.store.RemoveLast()
if hc.lowestMatchOutsideResults == nil {
hc.lowestMatchOutsideResults = removed
} else {
cmp := hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, removed, hc.lowestMatchOutsideResults)
if cmp < 0 {
tmp := hc.lowestMatchOutsideResults
hc.lowestMatchOutsideResults = removed
ctx.DocumentMatchPool.Put(tmp)
}
}
2016-08-10 10:13:38 +02:00
}
2016-08-09 15:18:53 +02:00
return nil
}
// visitFieldTerms is responsible for visiting the field terms of the
// search hit, and passing visited terms to the sort and facet builder
func (hc *TopNCollector) visitFieldTerms(reader index.IndexReader, d *search.DocumentMatch) error {
if hc.facetsBuilder != nil {
hc.facetsBuilder.StartDoc()
}
err := reader.DocumentVisitFieldTerms(d.IndexInternalID, hc.neededFields, func(field string, term []byte) {
if hc.facetsBuilder != nil {
hc.facetsBuilder.UpdateVisitor(field, term)
}
hc.sort.UpdateVisitor(field, term)
})
if hc.facetsBuilder != nil {
hc.facetsBuilder.EndDoc()
}
return err
}
// SetFacetsBuilder registers a facet builder for this collector
func (hc *TopNCollector) SetFacetsBuilder(facetsBuilder *search.FacetsBuilder) {
2016-08-09 15:18:53 +02:00
hc.facetsBuilder = facetsBuilder
hc.neededFields = append(hc.neededFields, hc.facetsBuilder.RequiredFields()...)
2016-08-09 15:18:53 +02:00
}
// finalizeResults starts with the heap containing the final top size+skip
// it now throws away the results to be skipped
// and does final doc id lookup (if necessary)
func (hc *TopNCollector) finalizeResults(r index.IndexReader) error {
var err error
hc.results, err = hc.store.Final(hc.skip, func(doc *search.DocumentMatch) error {
if doc.ID == "" {
// look up the id since we need it for lookup
var err error
doc.ID, err = r.ExternalID(doc.IndexInternalID)
if err != nil {
return err
}
2016-08-10 10:13:38 +02:00
}
return nil
})
return err
}
// Results returns the collected hits
func (hc *TopNCollector) Results() search.DocumentMatchCollection {
return hc.results
2016-08-09 15:18:53 +02:00
}
// Total returns the total number of hits
func (hc *TopNCollector) Total() uint64 {
2016-08-09 15:18:53 +02:00
return hc.total
}
// MaxScore returns the maximum score seen across all the hits
func (hc *TopNCollector) MaxScore() float64 {
return hc.maxScore
2016-08-10 10:13:38 +02:00
}
// Took returns the time spent collecting hits
func (hc *TopNCollector) Took() time.Duration {
2016-08-10 10:13:38 +02:00
return hc.took
2016-08-09 15:18:53 +02:00
}
// FacetResults returns the computed facets results
func (hc *TopNCollector) FacetResults() search.FacetResults {
2016-08-09 15:18:53 +02:00
if hc.facetsBuilder != nil {
return hc.facetsBuilder.Results()
}
return search.FacetResults{}
}