Upgrade blevesearch dependency to v2.0.1 (#14346)

* Upgrade blevesearch dependency to v2.0.1

* Update rupture to v1.0.0

* Fix test

vendor/github.com/blevesearch/bleve/v2/search/collector/heap.go

@@ -0,0 +1,95 @@
+//  Copyright (c) 2014 Couchbase, Inc.
+//
+// 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
+
+import (
+	"container/heap"
+
+	"github.com/blevesearch/bleve/v2/search"
+)
+
+type collectStoreHeap struct {
+	heap    search.DocumentMatchCollection
+	compare collectorCompare
+}
+
+func newStoreHeap(capacity int, compare collectorCompare) *collectStoreHeap {
+	rv := &collectStoreHeap{
+		heap:    make(search.DocumentMatchCollection, 0, capacity),
+		compare: compare,
+	}
+	heap.Init(rv)
+	return rv
+}
+
+func (c *collectStoreHeap) AddNotExceedingSize(doc *search.DocumentMatch,
+	size int) *search.DocumentMatch {
+	c.add(doc)
+	if c.Len() > size {
+		return c.removeLast()
+	}
+	return nil
+}
+
+func (c *collectStoreHeap) add(doc *search.DocumentMatch) {
+	heap.Push(c, doc)
+}
+
+func (c *collectStoreHeap) removeLast() *search.DocumentMatch {
+	return heap.Pop(c).(*search.DocumentMatch)
+}
+
+func (c *collectStoreHeap) Final(skip int, fixup collectorFixup) (search.DocumentMatchCollection, error) {
+	count := c.Len()
+	size := count - skip
+	if size <= 0 {
+		return make(search.DocumentMatchCollection, 0), nil
+	}
+	rv := make(search.DocumentMatchCollection, size)
+	for i := size - 1; i >= 0; i-- {
+		doc := heap.Pop(c).(*search.DocumentMatch)
+		rv[i] = doc
+		err := fixup(doc)
+		if err != nil {
+			return nil, err
+		}
+	}
+	return rv, nil
+}
+
+// heap interface implementation
+
+func (c *collectStoreHeap) Len() int {
+	return len(c.heap)
+}
+
+func (c *collectStoreHeap) Less(i, j int) bool {
+	so := c.compare(c.heap[i], c.heap[j])
+	return -so < 0
+}
+
+func (c *collectStoreHeap) Swap(i, j int) {
+	c.heap[i], c.heap[j] = c.heap[j], c.heap[i]
+}
+
+func (c *collectStoreHeap) Push(x interface{}) {
+	c.heap = append(c.heap, x.(*search.DocumentMatch))
+}
+
+func (c *collectStoreHeap) Pop() interface{} {
+	var rv *search.DocumentMatch
+	rv, c.heap = c.heap[len(c.heap)-1], c.heap[:len(c.heap)-1]
+	return rv
+}

vendor/github.com/blevesearch/bleve/v2/search/collector/list.go

@@ -0,0 +1,86 @@
+//  Copyright (c) 2014 Couchbase, Inc.
+//
+// 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
+
+import (
+	"container/list"
+
+	"github.com/blevesearch/bleve/v2/search"
+)
+
+type collectStoreList struct {
+	results *list.List
+	compare collectorCompare
+}
+
+func newStoreList(capacity int, compare collectorCompare) *collectStoreList {
+	rv := &collectStoreList{
+		results: list.New(),
+		compare: compare,
+	}
+
+	return rv
+}
+
+func (c *collectStoreList) AddNotExceedingSize(doc *search.DocumentMatch, size int) *search.DocumentMatch {
+	c.add(doc)
+	if c.len() > size {
+		return c.removeLast()
+	}
+	return nil
+}
+
+func (c *collectStoreList) add(doc *search.DocumentMatch) {
+	for e := c.results.Front(); e != nil; e = e.Next() {
+		curr := e.Value.(*search.DocumentMatch)
+		if c.compare(doc, curr) >= 0 {
+			c.results.InsertBefore(doc, e)
+			return
+		}
+	}
+	// if we got to the end, we still have to add it
+	c.results.PushBack(doc)
+}
+
+func (c *collectStoreList) removeLast() *search.DocumentMatch {
+	return c.results.Remove(c.results.Front()).(*search.DocumentMatch)
+}
+
+func (c *collectStoreList) Final(skip int, fixup collectorFixup) (search.DocumentMatchCollection, error) {
+	if c.results.Len()-skip > 0 {
+		rv := make(search.DocumentMatchCollection, c.results.Len()-skip)
+		i := 0
+		skipped := 0
+		for e := c.results.Back(); e != nil; e = e.Prev() {
+			if skipped < skip {
+				skipped++
+				continue
+			}
+
+			rv[i] = e.Value.(*search.DocumentMatch)
+			err := fixup(rv[i])
+			if err != nil {
+				return nil, err
+			}
+			i++
+		}
+		return rv, nil
+	}
+	return search.DocumentMatchCollection{}, nil
+}
+
+func (c *collectStoreList) len() int {
+	return c.results.Len()
+}

vendor/github.com/blevesearch/bleve/v2/search/collector/slice.go

@@ -0,0 +1,77 @@
+//  Copyright (c) 2014 Couchbase, Inc.
+//
+// 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
+
+import "github.com/blevesearch/bleve/v2/search"
+
+type collectStoreSlice struct {
+	slice   search.DocumentMatchCollection
+	compare collectorCompare
+}
+
+func newStoreSlice(capacity int, compare collectorCompare) *collectStoreSlice {
+	rv := &collectStoreSlice{
+		slice:   make(search.DocumentMatchCollection, 0, capacity),
+		compare: compare,
+	}
+	return rv
+}
+
+func (c *collectStoreSlice) AddNotExceedingSize(doc *search.DocumentMatch,
+	size int) *search.DocumentMatch {
+	c.add(doc)
+	if c.len() > size {
+		return c.removeLast()
+	}
+	return nil
+}
+
+func (c *collectStoreSlice) add(doc *search.DocumentMatch) {
+	// find where to insert, starting at end (lowest)
+	i := len(c.slice)
+	for ; i > 0; i-- {
+		cmp := c.compare(doc, c.slice[i-1])
+		if cmp >= 0 {
+			break
+		}
+	}
+	// insert at i
+	c.slice = append(c.slice, nil)
+	copy(c.slice[i+1:], c.slice[i:])
+	c.slice[i] = doc
+}
+
+func (c *collectStoreSlice) removeLast() *search.DocumentMatch {
+	var rv *search.DocumentMatch
+	rv, c.slice = c.slice[len(c.slice)-1], c.slice[:len(c.slice)-1]
+	return rv
+}
+
+func (c *collectStoreSlice) Final(skip int, fixup collectorFixup) (search.DocumentMatchCollection, error) {
+	for i := skip; i < len(c.slice); i++ {
+		err := fixup(c.slice[i])
+		if err != nil {
+			return nil, err
+		}
+	}
+	if skip <= len(c.slice) {
+		return c.slice[skip:], nil
+	}
+	return search.DocumentMatchCollection{}, nil
+}
+
+func (c *collectStoreSlice) len() int {
+	return len(c.slice)
+}

vendor/github.com/blevesearch/bleve/v2/search/collector/topn.go

@@ -0,0 +1,412 @@
+//  Copyright (c) 2014 Couchbase, Inc.
+//
+// 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
+
+import (
+	"context"
+	"reflect"
+	"strconv"
+	"time"
+
+	"github.com/blevesearch/bleve/v2/search"
+	"github.com/blevesearch/bleve/v2/size"
+	index "github.com/blevesearch/bleve_index_api"
+)
+
+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
+	// exceeded, nil is returned.
+	AddNotExceedingSize(doc *search.DocumentMatch, size int) *search.DocumentMatch
+
+	Final(skip int, fixup collectorFixup) (search.DocumentMatchCollection, error)
+}
+
+// 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 {
+	size          int
+	skip          int
+	total         uint64
+	maxScore      float64
+	took          time.Duration
+	sort          search.SortOrder
+	results       search.DocumentMatchCollection
+	facetsBuilder *search.FacetsBuilder
+
+	store collectorStore
+
+	needDocIds    bool
+	neededFields  []string
+	cachedScoring []bool
+	cachedDesc    []bool
+
+	lowestMatchOutsideResults *search.DocumentMatch
+	updateFieldVisitor        index.DocValueVisitor
+	dvReader                  index.DocValueReader
+	searchAfter               *search.DocumentMatch
+}
+
+// CheckDoneEvery controls how frequently we check the context deadline
+const CheckDoneEvery = uint64(1024)
+
+// 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 {
+	return newTopNCollector(size, skip, sort)
+}
+
+// NewTopNCollector builds a collector to find the top 'size' hits
+// skipping over the first 'skip' hits
+// ordering hits by the provided sort order
+func NewTopNCollectorAfter(size int, sort search.SortOrder, after []string) *TopNCollector {
+	rv := newTopNCollector(size, 0, sort)
+	rv.searchAfter = &search.DocumentMatch{
+		Sort: after,
+	}
+
+	for pos, ss := range sort {
+		if ss.RequiresDocID() {
+			rv.searchAfter.ID = after[pos]
+		}
+		if ss.RequiresScoring() {
+			if score, err := strconv.ParseFloat(after[pos], 64); err == nil {
+				rv.searchAfter.Score = score
+			}
+		}
+	}
+
+	return rv
+}
+
+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
+	}
+
+	if size+skip > 10 {
+		hc.store = newStoreHeap(backingSize, func(i, j *search.DocumentMatch) int {
+			return hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, i, j)
+		})
+	} else {
+		hc.store = newStoreSlice(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()
+
+	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()
+	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)),
+		Collector:         hc,
+		IndexReader:       reader,
+	}
+
+	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()
+	default:
+		next, err = searcher.Next(searchContext)
+	}
+	for err == nil && next != nil {
+		if hc.total%CheckDoneEvery == 0 {
+			select {
+			case <-ctx.Done():
+				return ctx.Err()
+			default:
+			}
+		}
+
+		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 {
+		return err
+	}
+	// finalize actual results
+	err = hc.finalizeResults(reader)
+	if err != nil {
+		return err
+	}
+	return nil
+}
+
+var sortByScoreOpt = []string{"_score"}
+
+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 {
+		err = hc.visitFieldTerms(reader, d)
+		if err != nil {
+			return err
+		}
+	}
+
+	// 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)
+		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)
+	}
+
+	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
+			}
+
+			// support search after based pagination,
+			// if this hit is <= the search after sort key
+			// we should skip it
+			if hc.searchAfter != nil {
+				// exact sort order matches use hit number to break tie
+				// but we want to allow for exact match, so we pretend
+				hc.searchAfter.HitNumber = d.HitNumber
+				if hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, d, hc.searchAfter) <= 0 {
+					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 {
+	if hc.facetsBuilder != nil {
+		hc.facetsBuilder.StartDoc()
+	}
+
+	err := hc.dvReader.VisitDocValues(d.IndexInternalID, hc.updateFieldVisitor)
+	if hc.facetsBuilder != nil {
+		hc.facetsBuilder.EndDoc()
+	}
+
+	return err
+}
+
+// SetFacetsBuilder registers a facet builder for this collector
+func (hc *TopNCollector) SetFacetsBuilder(facetsBuilder *search.FacetsBuilder) {
+	hc.facetsBuilder = facetsBuilder
+	hc.neededFields = append(hc.neededFields, hc.facetsBuilder.RequiredFields()...)
+}
+
+// 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
+			}
+		}
+		doc.Complete(nil)
+		return nil
+	})
+
+	return err
+}
+
+// Results returns the collected hits
+func (hc *TopNCollector) Results() search.DocumentMatchCollection {
+	return hc.results
+}
+
+// Total returns the total number of hits
+func (hc *TopNCollector) Total() uint64 {
+	return hc.total
+}
+
+// MaxScore returns the maximum score seen across all the hits
+func (hc *TopNCollector) MaxScore() float64 {
+	return hc.maxScore
+}
+
+// Took returns the time spent collecting hits
+func (hc *TopNCollector) Took() time.Duration {
+	return hc.took
+}
+
+// FacetResults returns the computed facets results
+func (hc *TopNCollector) FacetResults() search.FacetResults {
+	if hc.facetsBuilder != nil {
+		return hc.facetsBuilder.Results()
+	}
+	return nil
+}