0
0
Fork 0
bleve/search/search.go

233 lines
6.1 KiB
Go
Raw Normal View History

2014-04-17 22:55:53 +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.
2014-04-17 22:55:53 +02:00
package search
import (
"fmt"
"reflect"
"github.com/blevesearch/bleve/document"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/size"
)
major refactor of index/search API index id's are now opaque (until finally returned to top-level user) - the TermFieldDoc's returned by TermFieldReader no longer contain doc id - instead they return an opaque IndexInternalID - items returned are still in the "natural index order" - but that is no longer guaranteed to be "doc id order" - correct behavior requires that they all follow the same order - but not any particular order - new API FinalizeDocID which converts index internal ID's to public string ID - APIs used internally which previously took doc id now take IndexInternalID - that is DocumentFieldTerms() and DocumentFieldTermsForFields() - however, APIs that are used externally do not reflect this change - that is Document() - DocumentIDReader follows the same changes, but this is less obvious - behavior clarified, used to iterate doc ids, BUT NOT in doc id order - method STILL available to iterate doc ids in range - but again, you won't get them in any meaningful order - new method to iterate actual doc ids from list of possible ids - this was introduced to make the DocIDSearcher continue working searchers now work with the new opaque index internal doc ids - they return new DocumentMatchInternal (which does not have string ID) scorerers also work with these opaque index internal doc ids - they return DocumentMatchInternal (which does not have string ID) collectors now also perform a final step of converting the final result - they STILL return traditional DocumentMatch (with string ID) - but they now also require an IndexReader (so that they can do the conversion)
2016-07-31 19:46:18 +02:00
var reflectStaticSizeDocumentMatch int
var reflectStaticSizeSearchContext int
var reflectStaticSizeLocation int
func init() {
var dm DocumentMatch
reflectStaticSizeDocumentMatch = int(reflect.TypeOf(dm).Size())
var sc SearchContext
reflectStaticSizeSearchContext = int(reflect.TypeOf(sc).Size())
var l Location
reflectStaticSizeLocation = int(reflect.TypeOf(l).Size())
}
type ArrayPositions []uint64
2014-04-17 22:55:53 +02:00
func (ap ArrayPositions) Equals(other ArrayPositions) bool {
if len(ap) != len(other) {
return false
}
for i := range ap {
if ap[i] != other[i] {
return false
}
}
return true
}
type Location struct {
// Pos is the position of the term within the field, starting at 1
Pos uint64 `json:"pos"`
2017-05-25 17:32:10 +02:00
// Start and End are the byte offsets of the term in the field
Start uint64 `json:"start"`
End uint64 `json:"end"`
2017-05-25 17:32:10 +02:00
// ArrayPositions contains the positions of the term within any elements.
ArrayPositions ArrayPositions `json:"array_positions"`
}
func (l *Location) Size() int {
return reflectStaticSizeLocation + size.SizeOfPtr +
len(l.ArrayPositions)*size.SizeOfUint64
}
2014-04-17 22:55:53 +02:00
type Locations []*Location
type TermLocationMap map[string]Locations
func (t TermLocationMap) AddLocation(term string, location *Location) {
2016-10-11 21:14:48 +02:00
t[term] = append(t[term], location)
}
2014-04-17 22:55:53 +02:00
type FieldTermLocationMap map[string]TermLocationMap
type FieldFragmentMap map[string][]string
2014-04-17 22:55:53 +02:00
type DocumentMatch struct {
Index string `json:"index,omitempty"`
ID string `json:"id"`
IndexInternalID index.IndexInternalID `json:"-"`
Score float64 `json:"score"`
Expl *Explanation `json:"explanation,omitempty"`
Locations FieldTermLocationMap `json:"locations,omitempty"`
Fragments FieldFragmentMap `json:"fragments,omitempty"`
improved implementation to address perf regressions primary change is going back to sort values be []string and not []interface{}, this avoid allocatiosn converting into the interface{} that sounds obvious, so why didn't we just do that first? because a common (default) sort is score, which is naturally a number, not a string (like terms). converting into the number was also expensive, and the common case. so, this solution also makes the change to NOT put the score into the sort value list. instead you see the dummy value "_score". this is just a placeholder, the actual sort impl knows that field of the sort is the score, and will sort using the actual score. also, several other aspets of the benchmark were cleaned up so that unnecessary allocations do not pollute the cpu profiles Here are the updated benchmarks: $ go test -run=xxx -bench=. -benchmem -cpuprofile=cpu.out BenchmarkTop10of100000Scores-4 3000 465809 ns/op 2548 B/op 33 allocs/op BenchmarkTop100of100000Scores-4 2000 626488 ns/op 21484 B/op 213 allocs/op BenchmarkTop10of1000000Scores-4 300 5107658 ns/op 2560 B/op 33 allocs/op BenchmarkTop100of1000000Scores-4 300 5275403 ns/op 21624 B/op 213 allocs/op PASS ok github.com/blevesearch/bleve/search/collectors 7.188s Prior to this PR, master reported: $ go test -run=xxx -bench=. -benchmem BenchmarkTop10of100000Scores-4 3000 453269 ns/op 360161 B/op 42 allocs/op BenchmarkTop100of100000Scores-4 2000 519131 ns/op 388275 B/op 219 allocs/op BenchmarkTop10of1000000Scores-4 200 7459004 ns/op 4628236 B/op 52 allocs/op BenchmarkTop100of1000000Scores-4 200 8064864 ns/op 4656596 B/op 232 allocs/op PASS ok github.com/blevesearch/bleve/search/collectors 7.385s So, we're pretty close on the smaller datasets, and we scale better on the larger datasets. We also show fewer allocations and bytes in all cases (some of this is artificial due to test cleanup).
2016-08-25 21:47:07 +02:00
Sort []string `json:"sort,omitempty"`
// Fields contains the values for document fields listed in
// SearchRequest.Fields. Text fields are returned as strings, numeric
// fields as float64s and date fields as time.RFC3339 formatted strings.
Fields map[string]interface{} `json:"fields,omitempty"`
// if we load the document for this hit, remember it so we dont load again
Document *document.Document `json:"-"`
// used to maintain natural index order
HitNumber uint64 `json:"-"`
}
func (dm *DocumentMatch) AddFieldValue(name string, value interface{}) {
if dm.Fields == nil {
dm.Fields = make(map[string]interface{})
}
existingVal, ok := dm.Fields[name]
if !ok {
dm.Fields[name] = value
return
}
valSlice, ok := existingVal.([]interface{})
if ok {
// already a slice, append to it
valSlice = append(valSlice, value)
} else {
// create a slice
valSlice = []interface{}{existingVal, value}
}
dm.Fields[name] = valSlice
2014-04-17 22:55:53 +02:00
}
// Reset allows an already allocated DocumentMatch to be reused
func (dm *DocumentMatch) Reset() *DocumentMatch {
// remember the []byte used for the IndexInternalID
indexInternalID := dm.IndexInternalID
// remember the []interface{} used for sort
sort := dm.Sort
// idiom to copy over from empty DocumentMatch (0 allocations)
*dm = DocumentMatch{}
// reuse the []byte already allocated (and reset len to 0)
dm.IndexInternalID = indexInternalID[:0]
// reuse the []interface{} already allocated (and reset len to 0)
dm.Sort = sort[:0]
return dm
}
func (dm *DocumentMatch) Size() int {
sizeInBytes := reflectStaticSizeDocumentMatch + size.SizeOfPtr +
len(dm.Index) +
len(dm.ID) +
len(dm.IndexInternalID)
if dm.Expl != nil {
sizeInBytes += dm.Expl.Size()
}
for k, v := range dm.Locations {
sizeInBytes += size.SizeOfString + len(k)
for k1, v1 := range v {
sizeInBytes += size.SizeOfString + len(k1) +
size.SizeOfSlice
for _, entry := range v1 {
sizeInBytes += entry.Size()
}
}
}
for k, v := range dm.Fragments {
sizeInBytes += size.SizeOfString + len(k) +
size.SizeOfSlice
for _, entry := range v {
sizeInBytes += size.SizeOfString + len(entry)
}
}
for _, entry := range dm.Sort {
sizeInBytes += size.SizeOfString + len(entry)
}
for k, _ := range dm.Fields {
sizeInBytes += size.SizeOfString + len(k) +
size.SizeOfPtr
}
if dm.Document != nil {
sizeInBytes += dm.Document.Size()
}
return sizeInBytes
}
func (dm *DocumentMatch) String() string {
return fmt.Sprintf("[%s-%f]", string(dm.IndexInternalID), dm.Score)
}
2014-04-17 22:55:53 +02:00
type DocumentMatchCollection []*DocumentMatch
func (c DocumentMatchCollection) Len() int { return len(c) }
func (c DocumentMatchCollection) Swap(i, j int) { c[i], c[j] = c[j], c[i] }
func (c DocumentMatchCollection) Less(i, j int) bool { return c[i].Score > c[j].Score }
2014-04-17 22:55:53 +02:00
type Searcher interface {
Next(ctx *SearchContext) (*DocumentMatch, error)
Advance(ctx *SearchContext, ID index.IndexInternalID) (*DocumentMatch, error)
Close() error
2014-04-17 22:55:53 +02:00
Weight() float64
SetQueryNorm(float64)
Count() uint64
Min() int
Size() int
DocumentMatchPoolSize() int
}
type SearcherOptions struct {
Explain bool
IncludeTermVectors bool
}
// SearchContext represents the context around a single search
type SearchContext struct {
DocumentMatchPool *DocumentMatchPool
2014-04-17 22:55:53 +02:00
}
func (sc *SearchContext) Size() int {
sizeInBytes := reflectStaticSizeSearchContext + size.SizeOfPtr +
reflectStaticSizeDocumentMatchPool + size.SizeOfPtr
if sc.DocumentMatchPool != nil {
for _, entry := range sc.DocumentMatchPool.avail {
if entry != nil {
sizeInBytes += entry.Size()
}
}
}
return sizeInBytes
}