Update bleve dependency to latest master revision (#6100)

* update bleve to master b17287a86f6cac923a5d886e10618df994eeb54b6724eac2e3b8dde89cfbe3a2

* remove unused pkg from dep file

* change bleve from master to recent revision
This commit is contained in:
Lunny Xiao 2019-02-18 08:50:26 +08:00 committed by techknowlogick
parent 11e316654e
commit a380cfd8e0
161 changed files with 9911 additions and 4233 deletions

View file

@ -25,9 +25,9 @@ type collectStoreHeap struct {
compare collectorCompare
}
func newStoreHeap(cap int, compare collectorCompare) *collectStoreHeap {
func newStoreHeap(capacity int, compare collectorCompare) *collectStoreHeap {
rv := &collectStoreHeap{
heap: make(search.DocumentMatchCollection, 0, cap),
heap: make(search.DocumentMatchCollection, 0, capacity),
compare: compare,
}
heap.Init(rv)

View file

@ -25,7 +25,7 @@ type collectStoreList struct {
compare collectorCompare
}
func newStoreList(cap int, compare collectorCompare) *collectStoreList {
func newStoreList(capacity int, compare collectorCompare) *collectStoreList {
rv := &collectStoreList{
results: list.New(),
compare: compare,
@ -34,8 +34,7 @@ func newStoreList(cap int, compare collectorCompare) *collectStoreList {
return rv
}
func (c *collectStoreList) AddNotExceedingSize(doc *search.DocumentMatch,
size int) *search.DocumentMatch {
func (c *collectStoreList) AddNotExceedingSize(doc *search.DocumentMatch, size int) *search.DocumentMatch {
c.add(doc)
if c.len() > size {
return c.removeLast()

View file

@ -21,9 +21,9 @@ type collectStoreSlice struct {
compare collectorCompare
}
func newStoreSlice(cap int, compare collectorCompare) *collectStoreSlice {
func newStoreSlice(capacity int, compare collectorCompare) *collectStoreSlice {
rv := &collectStoreSlice{
slice: make(search.DocumentMatchCollection, 0, cap),
slice: make(search.DocumentMatchCollection, 0, capacity),
compare: compare,
}
return rv

View file

@ -16,12 +16,21 @@ package collector
import (
"context"
"reflect"
"time"
"github.com/blevesearch/bleve/index"
"github.com/blevesearch/bleve/search"
"github.com/blevesearch/bleve/size"
)
var reflectStaticSizeTopNCollector int
func init() {
var coll TopNCollector
reflectStaticSizeTopNCollector = int(reflect.TypeOf(coll).Size())
}
type collectorStore interface {
// Add the document, and if the new store size exceeds the provided size
// the last element is removed and returned. If the size has not been
@ -58,6 +67,8 @@ type TopNCollector struct {
cachedDesc []bool
lowestMatchOutsideResults *search.DocumentMatch
updateFieldVisitor index.DocumentFieldTermVisitor
dvReader index.DocValueReader
}
// CheckDoneEvery controls how frequently we check the context deadline
@ -98,6 +109,22 @@ func NewTopNCollector(size int, skip int, sort search.SortOrder) *TopNCollector
return hc
}
func (hc *TopNCollector) Size() int {
sizeInBytes := reflectStaticSizeTopNCollector + size.SizeOfPtr
if hc.facetsBuilder != nil {
sizeInBytes += hc.facetsBuilder.Size()
}
for _, entry := range hc.neededFields {
sizeInBytes += len(entry) + size.SizeOfString
}
sizeInBytes += len(hc.cachedScoring) + len(hc.cachedDesc)
return sizeInBytes
}
// Collect goes to the index to find the matching documents
func (hc *TopNCollector) Collect(ctx context.Context, searcher search.Searcher, reader index.IndexReader) error {
startTime := time.Now()
@ -113,8 +140,34 @@ func (hc *TopNCollector) Collect(ctx context.Context, searcher search.Searcher,
}
searchContext := &search.SearchContext{
DocumentMatchPool: search.NewDocumentMatchPool(backingSize+searcher.DocumentMatchPoolSize(), len(hc.sort)),
Collector: hc,
}
hc.dvReader, err = reader.DocValueReader(hc.neededFields)
if err != nil {
return err
}
hc.updateFieldVisitor = func(field string, term []byte) {
if hc.facetsBuilder != nil {
hc.facetsBuilder.UpdateVisitor(field, term)
}
hc.sort.UpdateVisitor(field, term)
}
dmHandlerMaker := MakeTopNDocumentMatchHandler
if cv := ctx.Value(search.MakeDocumentMatchHandlerKey); cv != nil {
dmHandlerMaker = cv.(search.MakeDocumentMatchHandler)
}
// use the application given builder for making the custom document match
// handler and perform callbacks/invocations on the newly made handler.
dmHandler, loadID, err := dmHandlerMaker(searchContext)
if err != nil {
return err
}
hc.needDocIds = hc.needDocIds || loadID
select {
case <-ctx.Done():
return ctx.Err()
@ -130,13 +183,26 @@ func (hc *TopNCollector) Collect(ctx context.Context, searcher search.Searcher,
}
}
err = hc.collectSingle(searchContext, reader, next)
err = hc.prepareDocumentMatch(searchContext, reader, next)
if err != nil {
break
}
err = dmHandler(next)
if err != nil {
break
}
next, err = searcher.Next(searchContext)
}
// help finalize/flush the results in case
// of custom document match handlers.
err = dmHandler(nil)
if err != nil {
return err
}
// compute search duration
hc.took = time.Since(startTime)
if err != nil {
@ -152,8 +218,8 @@ func (hc *TopNCollector) Collect(ctx context.Context, searcher search.Searcher,
var sortByScoreOpt = []string{"_score"}
func (hc *TopNCollector) collectSingle(ctx *search.SearchContext, reader index.IndexReader, d *search.DocumentMatch) error {
var err error
func (hc *TopNCollector) prepareDocumentMatch(ctx *search.SearchContext,
reader index.IndexReader, d *search.DocumentMatch) (err error) {
// visit field terms for features that require it (sort, facets)
if len(hc.neededFields) > 0 {
@ -187,35 +253,51 @@ func (hc *TopNCollector) collectSingle(ctx *search.SearchContext, reader index.I
hc.sort.Value(d)
}
// optimization, we track lowest sorting hit already removed from heap
// 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
}
}
removed := hc.store.AddNotExceedingSize(d, hc.size+hc.skip)
if removed != nil {
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)
}
}
}
return nil
}
func MakeTopNDocumentMatchHandler(
ctx *search.SearchContext) (search.DocumentMatchHandler, bool, error) {
var hc *TopNCollector
var ok bool
if hc, ok = ctx.Collector.(*TopNCollector); ok {
return func(d *search.DocumentMatch) error {
if d == nil {
return nil
}
// optimization, we track lowest sorting hit already removed from heap
// 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
}
}
removed := hc.store.AddNotExceedingSize(d, hc.size+hc.skip)
if removed != nil {
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)
}
}
}
return nil
}, false, nil
}
return nil, false, 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 {
@ -223,13 +305,7 @@ func (hc *TopNCollector) visitFieldTerms(reader index.IndexReader, d *search.Doc
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)
})
err := hc.dvReader.VisitDocValues(d.IndexInternalID, hc.updateFieldVisitor)
if hc.facetsBuilder != nil {
hc.facetsBuilder.EndDoc()
}
@ -257,6 +333,7 @@ func (hc *TopNCollector) finalizeResults(r index.IndexReader) error {
return err
}
}
doc.Complete(nil)
return nil
})
@ -288,5 +365,5 @@ func (hc *TopNCollector) FacetResults() search.FacetResults {
if hc.facetsBuilder != nil {
return hc.facetsBuilder.Results()
}
return search.FacetResults{}
return nil
}