mirror of
https://github.com/go-gitea/gitea.git
synced 2024-12-04 14:46:57 -05:00
233 lines
5.5 KiB
Go
233 lines
5.5 KiB
Go
// 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 searcher
|
|
|
|
import (
|
|
"math"
|
|
"sort"
|
|
|
|
"github.com/blevesearch/bleve/index"
|
|
"github.com/blevesearch/bleve/search"
|
|
"github.com/blevesearch/bleve/search/scorer"
|
|
)
|
|
|
|
type ConjunctionSearcher struct {
|
|
indexReader index.IndexReader
|
|
searchers OrderedSearcherList
|
|
queryNorm float64
|
|
currs []*search.DocumentMatch
|
|
maxIDIdx int
|
|
scorer *scorer.ConjunctionQueryScorer
|
|
initialized bool
|
|
explain bool
|
|
}
|
|
|
|
func NewConjunctionSearcher(indexReader index.IndexReader, qsearchers []search.Searcher, explain bool) (*ConjunctionSearcher, error) {
|
|
// build the downstream searchers
|
|
searchers := make(OrderedSearcherList, len(qsearchers))
|
|
for i, searcher := range qsearchers {
|
|
searchers[i] = searcher
|
|
}
|
|
// sort the searchers
|
|
sort.Sort(searchers)
|
|
// build our searcher
|
|
rv := ConjunctionSearcher{
|
|
indexReader: indexReader,
|
|
explain: explain,
|
|
searchers: searchers,
|
|
currs: make([]*search.DocumentMatch, len(searchers)),
|
|
scorer: scorer.NewConjunctionQueryScorer(explain),
|
|
}
|
|
rv.computeQueryNorm()
|
|
return &rv, nil
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) computeQueryNorm() {
|
|
// first calculate sum of squared weights
|
|
sumOfSquaredWeights := 0.0
|
|
for _, termSearcher := range s.searchers {
|
|
sumOfSquaredWeights += termSearcher.Weight()
|
|
}
|
|
// now compute query norm from this
|
|
s.queryNorm = 1.0 / math.Sqrt(sumOfSquaredWeights)
|
|
// finally tell all the downstream searchers the norm
|
|
for _, termSearcher := range s.searchers {
|
|
termSearcher.SetQueryNorm(s.queryNorm)
|
|
}
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) initSearchers(ctx *search.SearchContext) error {
|
|
var err error
|
|
// get all searchers pointing at their first match
|
|
for i, termSearcher := range s.searchers {
|
|
if s.currs[i] != nil {
|
|
ctx.DocumentMatchPool.Put(s.currs[i])
|
|
}
|
|
s.currs[i], err = termSearcher.Next(ctx)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
}
|
|
s.initialized = true
|
|
return nil
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Weight() float64 {
|
|
var rv float64
|
|
for _, searcher := range s.searchers {
|
|
rv += searcher.Weight()
|
|
}
|
|
return rv
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) SetQueryNorm(qnorm float64) {
|
|
for _, searcher := range s.searchers {
|
|
searcher.SetQueryNorm(qnorm)
|
|
}
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Next(ctx *search.SearchContext) (*search.DocumentMatch, error) {
|
|
if !s.initialized {
|
|
err := s.initSearchers(ctx)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
var rv *search.DocumentMatch
|
|
var err error
|
|
OUTER:
|
|
for s.currs[s.maxIDIdx] != nil {
|
|
maxID := s.currs[s.maxIDIdx].IndexInternalID
|
|
|
|
i := 0
|
|
for i < len(s.currs) {
|
|
if s.currs[i] == nil {
|
|
return nil, nil
|
|
}
|
|
|
|
if i == s.maxIDIdx {
|
|
i++
|
|
continue
|
|
}
|
|
|
|
cmp := maxID.Compare(s.currs[i].IndexInternalID)
|
|
if cmp == 0 {
|
|
i++
|
|
continue
|
|
}
|
|
|
|
if cmp < 0 {
|
|
// maxID < currs[i], so we found a new maxIDIdx
|
|
s.maxIDIdx = i
|
|
|
|
// advance the positions where [0 <= x < i], since we
|
|
// know they were equal to the former max entry
|
|
maxID = s.currs[s.maxIDIdx].IndexInternalID
|
|
for x := 0; x < i; x++ {
|
|
err = s.advanceChild(ctx, x, maxID)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
|
|
continue OUTER
|
|
}
|
|
|
|
// maxID > currs[i], so need to advance searchers[i]
|
|
err = s.advanceChild(ctx, i, maxID)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
// don't bump i, so that we'll examine the just-advanced
|
|
// currs[i] again
|
|
}
|
|
|
|
// if we get here, a doc matched all readers, so score and add it
|
|
rv = s.scorer.Score(ctx, s.currs)
|
|
|
|
// we know all the searchers are pointing at the same thing
|
|
// so they all need to be bumped
|
|
for i, termSearcher := range s.searchers {
|
|
if s.currs[i] != rv {
|
|
ctx.DocumentMatchPool.Put(s.currs[i])
|
|
}
|
|
s.currs[i], err = termSearcher.Next(ctx)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
|
|
// don't continue now, wait for the next call to Next()
|
|
break
|
|
}
|
|
return rv, nil
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Advance(ctx *search.SearchContext, ID index.IndexInternalID) (*search.DocumentMatch, error) {
|
|
if !s.initialized {
|
|
err := s.initSearchers(ctx)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
for i := range s.searchers {
|
|
err := s.advanceChild(ctx, i, ID)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
return s.Next(ctx)
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) advanceChild(ctx *search.SearchContext, i int, ID index.IndexInternalID) (err error) {
|
|
if s.currs[i] != nil {
|
|
ctx.DocumentMatchPool.Put(s.currs[i])
|
|
}
|
|
s.currs[i], err = s.searchers[i].Advance(ctx, ID)
|
|
return err
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Count() uint64 {
|
|
// for now return a worst case
|
|
var sum uint64
|
|
for _, searcher := range s.searchers {
|
|
sum += searcher.Count()
|
|
}
|
|
return sum
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Close() (rv error) {
|
|
for _, searcher := range s.searchers {
|
|
err := searcher.Close()
|
|
if err != nil && rv == nil {
|
|
rv = err
|
|
}
|
|
}
|
|
return rv
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) Min() int {
|
|
return 0
|
|
}
|
|
|
|
func (s *ConjunctionSearcher) DocumentMatchPoolSize() int {
|
|
rv := len(s.currs)
|
|
for _, s := range s.searchers {
|
|
rv += s.DocumentMatchPoolSize()
|
|
}
|
|
return rv
|
|
}
|