Adding upstream version 2.5.1.
Signed-off-by: Daniel Baumann <daniel@debian.org>
This commit is contained in:
parent
c71cb8b61d
commit
982828099e
GPG key ID:
FBB4F0E80A80222F
783 changed files with 150650 additions and 0 deletions
60
analysis/lang/cjk/analyzer_cjk.go
Normal file
60
analysis/lang/cjk/analyzer_cjk.go
Normal file
|
|
@ -0,0 +1,60 @@
|
|||
// 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 cjk
|
||||
|
||||
import (
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
"github.com/blevesearch/bleve/v2/registry"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis/token/lowercase"
|
||||
"github.com/blevesearch/bleve/v2/analysis/tokenizer/unicode"
|
||||
)
|
||||
|
||||
const AnalyzerName = "cjk"
|
||||
|
||||
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.Analyzer, error) {
|
||||
tokenizer, err := cache.TokenizerNamed(unicode.Name)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
widthFilter, err := cache.TokenFilterNamed(WidthName)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
toLowerFilter, err := cache.TokenFilterNamed(lowercase.Name)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
bigramFilter, err := cache.TokenFilterNamed(BigramName)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
rv := analysis.DefaultAnalyzer{
|
||||
Tokenizer: tokenizer,
|
||||
TokenFilters: []analysis.TokenFilter{
|
||||
widthFilter,
|
||||
toLowerFilter,
|
||||
bigramFilter,
|
||||
},
|
||||
}
|
||||
return &rv, nil
|
||||
}
|
||||
|
||||
func init() {
|
||||
err := registry.RegisterAnalyzer(AnalyzerName, AnalyzerConstructor)
|
||||
if err != nil {
|
||||
panic(err)
|
||||
}
|
||||
}
|
||||
642
analysis/lang/cjk/analyzer_cjk_test.go
Normal file
642
analysis/lang/cjk/analyzer_cjk_test.go
Normal file
|
|
@ -0,0 +1,642 @@
|
|||
// 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 cjk
|
||||
|
||||
import (
|
||||
"reflect"
|
||||
"testing"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
"github.com/blevesearch/bleve/v2/registry"
|
||||
)
|
||||
|
||||
func TestCJKAnalyzer(t *testing.T) {
|
||||
tests := []struct {
|
||||
input []byte
|
||||
output analysis.TokenStream
|
||||
}{
|
||||
{
|
||||
input: []byte("こんにちは世界"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こん"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("んに"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("にち"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ちは"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は世"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世界"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 21,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一二三四五六七八九十"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一二"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("二三"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("三四"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("四五"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("五六"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("六七"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("七八"),
|
||||
Type: analysis.Double,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 24,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("八九"),
|
||||
Type: analysis.Double,
|
||||
Position: 8,
|
||||
Start: 21,
|
||||
End: 27,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("九十"),
|
||||
Type: analysis.Double,
|
||||
Position: 9,
|
||||
Start: 24,
|
||||
End: 30,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一 二三四 五六七八九 十"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("二三"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 10,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("三四"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 7,
|
||||
End: 13,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("五六"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 14,
|
||||
End: 20,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("六七"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 17,
|
||||
End: 23,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("七八"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 20,
|
||||
End: 26,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("八九"),
|
||||
Type: analysis.Double,
|
||||
Position: 7,
|
||||
Start: 23,
|
||||
End: 29,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("十"),
|
||||
Type: analysis.Single,
|
||||
Position: 8,
|
||||
Start: 30,
|
||||
End: 33,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("abc defgh ijklmn opqrstu vwxy z"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("abc"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("defgh"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ijklmn"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 3,
|
||||
Start: 10,
|
||||
End: 16,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("opqrstu"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 4,
|
||||
Start: 17,
|
||||
End: 24,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("vwxy"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 5,
|
||||
Start: 25,
|
||||
End: 29,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("z"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 6,
|
||||
Start: 30,
|
||||
End: 31,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("あい"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("あい "),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("test"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("test"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 4,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("test "),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("test"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 4,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("あいtest"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("test"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 2,
|
||||
Start: 6,
|
||||
End: 10,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("testあい "),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("test"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 4,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 10,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("あいうえおabcかきくけこ"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("いう"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("うえ"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("えお"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("abc"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 5,
|
||||
Start: 15,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("かき"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 18,
|
||||
End: 24,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("きく"),
|
||||
Type: analysis.Double,
|
||||
Position: 7,
|
||||
Start: 21,
|
||||
End: 27,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("くけ"),
|
||||
Type: analysis.Double,
|
||||
Position: 8,
|
||||
Start: 24,
|
||||
End: 30,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("けこ"),
|
||||
Type: analysis.Double,
|
||||
Position: 9,
|
||||
Start: 27,
|
||||
End: 33,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("あいうえおabんcかきくけ こ"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("あい"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("いう"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("うえ"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("えお"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ab"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 5,
|
||||
Start: 15,
|
||||
End: 17,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Single,
|
||||
Position: 6,
|
||||
Start: 17,
|
||||
End: 20,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("c"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 7,
|
||||
Start: 20,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("かき"),
|
||||
Type: analysis.Double,
|
||||
Position: 8,
|
||||
Start: 21,
|
||||
End: 27,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("きく"),
|
||||
Type: analysis.Double,
|
||||
Position: 9,
|
||||
Start: 24,
|
||||
End: 30,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("くけ"),
|
||||
Type: analysis.Double,
|
||||
Position: 10,
|
||||
Start: 27,
|
||||
End: 33,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Single,
|
||||
Position: 11,
|
||||
Start: 34,
|
||||
End: 37,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一 روبرت موير"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("روبرت"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 14,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("موير"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 3,
|
||||
Start: 15,
|
||||
End: 23,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一 رُوبرت موير"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("رُوبرت"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 16,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("موير"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 3,
|
||||
Start: 17,
|
||||
End: 25,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("𩬅艱鍟䇹愯瀛"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("𩬅艱"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 7,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("艱鍟"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 4,
|
||||
End: 10,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("鍟䇹"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 7,
|
||||
End: 13,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("䇹愯"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 10,
|
||||
End: 16,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("愯瀛"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 13,
|
||||
End: 19,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: []byte("一丁丂"),
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("一丁"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("丁丂"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
cache := registry.NewCache()
|
||||
for _, test := range tests {
|
||||
analyzer, err := cache.AnalyzerNamed(AnalyzerName)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
actual := analyzer.Analyze(test.input)
|
||||
if !reflect.DeepEqual(actual, test.output) {
|
||||
t.Errorf("expected %v, got %v", test.output, actual)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func BenchmarkCJKAnalyzer(b *testing.B) {
|
||||
cache := registry.NewCache()
|
||||
analyzer, err := cache.AnalyzerNamed(AnalyzerName)
|
||||
if err != nil {
|
||||
b.Fatal(err)
|
||||
}
|
||||
|
||||
for i := 0; i < b.N; i++ {
|
||||
analyzer.Analyze(bleveWikiArticleJapanese)
|
||||
}
|
||||
}
|
||||
|
||||
var bleveWikiArticleJapanese = []byte(`加圧容器に貯蔵されている液体物質は、その時の気液平衡状態にあるが、火災により容器が加熱されていると容器内の液体は、その物質の大気圧のもとでの沸点より十分に高い温度まで加熱され、圧力も高くなる。この状態で容器が破裂すると容器内部の圧力は瞬間的に大気圧にまで低下する。
|
||||
この時に容器内の平衡状態が破られ、液体は突沸し、気体になることで爆発現象を起こす。液化石油ガスなどでは、さらに拡散して空気と混ざったガスが自由空間蒸気雲爆発を起こす。液化石油ガスなどの常温常圧で気体になる物を高い圧力で液化して収納している容器、あるいは、そのような液体を輸送するためのパイプラインや配管などが火災などによって破壊されたときに起きる。
|
||||
ブリーブという現象が明らかになったのは、フランス・リヨンの郊外にあるフェザンという町のフェザン製油所（ウニオン・ド・ゼネラル・ド・ペトロール）で大規模な爆発火災事故が発生したときだと言われている。
|
||||
中身の液体が高温高圧の水である場合には「水蒸気爆発」と呼ばれる。`)
|
||||
210
analysis/lang/cjk/cjk_bigram.go
Normal file
210
analysis/lang/cjk/cjk_bigram.go
Normal file
|
|
@ -0,0 +1,210 @@
|
|||
// 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 cjk
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"container/ring"
|
||||
"unicode/utf8"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
"github.com/blevesearch/bleve/v2/registry"
|
||||
)
|
||||
|
||||
const BigramName = "cjk_bigram"
|
||||
|
||||
type CJKBigramFilter struct {
|
||||
outputUnigram bool
|
||||
}
|
||||
|
||||
func NewCJKBigramFilter(outputUnigram bool) *CJKBigramFilter {
|
||||
return &CJKBigramFilter{
|
||||
outputUnigram: outputUnigram,
|
||||
}
|
||||
}
|
||||
|
||||
func (s *CJKBigramFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
|
||||
r := ring.New(2)
|
||||
itemsInRing := 0
|
||||
pos := 1
|
||||
outputPos := 1
|
||||
|
||||
rv := make(analysis.TokenStream, 0, len(input))
|
||||
|
||||
for _, tokout := range input {
|
||||
if tokout.Type == analysis.Ideographic {
|
||||
runes := bytes.Runes(tokout.Term)
|
||||
sofar := 0
|
||||
for _, run := range runes {
|
||||
rlen := utf8.RuneLen(run)
|
||||
token := &analysis.Token{
|
||||
Term: tokout.Term[sofar : sofar+rlen],
|
||||
Start: tokout.Start + sofar,
|
||||
End: tokout.Start + sofar + rlen,
|
||||
Position: pos,
|
||||
Type: tokout.Type,
|
||||
KeyWord: tokout.KeyWord,
|
||||
}
|
||||
pos++
|
||||
sofar += rlen
|
||||
if itemsInRing > 0 {
|
||||
// if items already buffered
|
||||
// check to see if this is aligned
|
||||
curr := r.Value.(*analysis.Token)
|
||||
if token.Start-curr.End != 0 {
|
||||
// not aligned flush
|
||||
flushToken := s.flush(r, &itemsInRing, outputPos)
|
||||
if flushToken != nil {
|
||||
outputPos++
|
||||
rv = append(rv, flushToken)
|
||||
}
|
||||
}
|
||||
}
|
||||
// now we can add this token to the buffer
|
||||
r = r.Next()
|
||||
r.Value = token
|
||||
if itemsInRing < 2 {
|
||||
itemsInRing++
|
||||
}
|
||||
builtUnigram := false
|
||||
if itemsInRing > 1 && s.outputUnigram {
|
||||
unigram := s.buildUnigram(r, &itemsInRing, outputPos)
|
||||
if unigram != nil {
|
||||
builtUnigram = true
|
||||
rv = append(rv, unigram)
|
||||
}
|
||||
}
|
||||
bigramToken := s.outputBigram(r, &itemsInRing, outputPos)
|
||||
if bigramToken != nil {
|
||||
rv = append(rv, bigramToken)
|
||||
outputPos++
|
||||
}
|
||||
|
||||
// prev token should be removed if unigram was built
|
||||
if builtUnigram {
|
||||
itemsInRing--
|
||||
}
|
||||
}
|
||||
|
||||
} else {
|
||||
// flush anything already buffered
|
||||
flushToken := s.flush(r, &itemsInRing, outputPos)
|
||||
if flushToken != nil {
|
||||
rv = append(rv, flushToken)
|
||||
outputPos++
|
||||
}
|
||||
// output this token as is
|
||||
tokout.Position = outputPos
|
||||
rv = append(rv, tokout)
|
||||
outputPos++
|
||||
}
|
||||
}
|
||||
|
||||
// deal with possible trailing unigram
|
||||
if itemsInRing == 1 || s.outputUnigram {
|
||||
if itemsInRing == 2 {
|
||||
r = r.Next()
|
||||
}
|
||||
unigram := s.buildUnigram(r, &itemsInRing, outputPos)
|
||||
if unigram != nil {
|
||||
rv = append(rv, unigram)
|
||||
}
|
||||
}
|
||||
return rv
|
||||
}
|
||||
|
||||
func (s *CJKBigramFilter) flush(r *ring.Ring, itemsInRing *int, pos int) *analysis.Token {
|
||||
var rv *analysis.Token
|
||||
if *itemsInRing == 1 {
|
||||
rv = s.buildUnigram(r, itemsInRing, pos)
|
||||
}
|
||||
r.Value = nil
|
||||
*itemsInRing = 0
|
||||
|
||||
return rv
|
||||
}
|
||||
|
||||
func (s *CJKBigramFilter) outputBigram(r *ring.Ring, itemsInRing *int, pos int) *analysis.Token {
|
||||
if *itemsInRing == 2 {
|
||||
thisShingleRing := r.Move(-1)
|
||||
shingledBytes := make([]byte, 0)
|
||||
|
||||
// do first token
|
||||
prev := thisShingleRing.Value.(*analysis.Token)
|
||||
shingledBytes = append(shingledBytes, prev.Term...)
|
||||
|
||||
// do second token
|
||||
thisShingleRing = thisShingleRing.Next()
|
||||
curr := thisShingleRing.Value.(*analysis.Token)
|
||||
shingledBytes = append(shingledBytes, curr.Term...)
|
||||
|
||||
token := analysis.Token{
|
||||
Type: analysis.Double,
|
||||
Term: shingledBytes,
|
||||
Position: pos,
|
||||
Start: prev.Start,
|
||||
End: curr.End,
|
||||
}
|
||||
return &token
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *CJKBigramFilter) buildUnigram(r *ring.Ring, itemsInRing *int, pos int) *analysis.Token {
|
||||
switch *itemsInRing {
|
||||
case 2:
|
||||
thisShingleRing := r.Move(-1)
|
||||
// do first token
|
||||
prev := thisShingleRing.Value.(*analysis.Token)
|
||||
token := analysis.Token{
|
||||
Type: analysis.Single,
|
||||
Term: prev.Term,
|
||||
Position: pos,
|
||||
Start: prev.Start,
|
||||
End: prev.End,
|
||||
}
|
||||
return &token
|
||||
case 1:
|
||||
// do first token
|
||||
prev := r.Value.(*analysis.Token)
|
||||
token := analysis.Token{
|
||||
Type: analysis.Single,
|
||||
Term: prev.Term,
|
||||
Position: pos,
|
||||
Start: prev.Start,
|
||||
End: prev.End,
|
||||
}
|
||||
return &token
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func CJKBigramFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
|
||||
outputUnigram := false
|
||||
outVal, ok := config["output_unigram"].(bool)
|
||||
if ok {
|
||||
outputUnigram = outVal
|
||||
}
|
||||
return NewCJKBigramFilter(outputUnigram), nil
|
||||
}
|
||||
|
||||
func init() {
|
||||
err := registry.RegisterTokenFilter(BigramName, CJKBigramFilterConstructor)
|
||||
if err != nil {
|
||||
panic(err)
|
||||
}
|
||||
}
|
||||
848
analysis/lang/cjk/cjk_bigram_test.go
Normal file
848
analysis/lang/cjk/cjk_bigram_test.go
Normal file
|
|
@ -0,0 +1,848 @@
|
|||
// 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 cjk
|
||||
|
||||
import (
|
||||
"container/ring"
|
||||
"reflect"
|
||||
"testing"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
)
|
||||
|
||||
// Helper function to create a token
|
||||
func makeToken(term string, start, end, pos int) *analysis.Token {
|
||||
return &analysis.Token{
|
||||
Term: []byte(term),
|
||||
Start: start,
|
||||
End: end,
|
||||
Position: pos, // Note: buildUnigram uses the 'pos' argument, not the token's original pos
|
||||
Type: analysis.Ideographic,
|
||||
}
|
||||
}
|
||||
|
||||
func TestCJKBigramFilter_buildUnigram(t *testing.T) {
|
||||
filter := NewCJKBigramFilter(false)
|
||||
|
||||
tests := []struct {
|
||||
name string
|
||||
ringSetup func() (*ring.Ring, int) // Function to set up the ring and itemsInRing
|
||||
inputPos int // Position to pass to buildUnigram
|
||||
expectToken *analysis.Token
|
||||
}{
|
||||
{
|
||||
name: "itemsInRing == 2",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("一", 0, 3, 1) // Original pos 1
|
||||
token2 := makeToken("二", 3, 6, 2) // Original pos 2
|
||||
r.Value = token1
|
||||
r = r.Next()
|
||||
r.Value = token2
|
||||
// r currently points to token2, r.Move(-1) points to token1
|
||||
return r, 2
|
||||
},
|
||||
inputPos: 10, // Expected output position
|
||||
expectToken: &analysis.Token{
|
||||
Type: analysis.Single,
|
||||
Term: []byte("一"),
|
||||
Position: 10, // Should use inputPos
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 1 (ring points to the single item)",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("三", 6, 9, 3)
|
||||
r.Value = token1
|
||||
// r points to token1
|
||||
return r, 1
|
||||
},
|
||||
inputPos: 11,
|
||||
expectToken: &analysis.Token{
|
||||
Type: analysis.Single,
|
||||
Term: []byte("三"),
|
||||
Position: 11, // Should use inputPos
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 1 (ring points to nil, next is the single item)",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("四", 9, 12, 4)
|
||||
r = r.Next() // r points to nil initially
|
||||
r.Value = token1
|
||||
// r points to token1
|
||||
return r, 1
|
||||
},
|
||||
inputPos: 12,
|
||||
expectToken: &analysis.Token{
|
||||
Type: analysis.Single,
|
||||
Term: []byte("四"),
|
||||
Position: 12, // Should use inputPos
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 0",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
// Ring is empty
|
||||
return r, 0
|
||||
},
|
||||
inputPos: 13,
|
||||
expectToken: nil, // Expect nil when itemsInRing is not 1 or 2
|
||||
},
|
||||
{
|
||||
name: "itemsInRing > 2 (should behave like 0)",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("五", 12, 15, 5)
|
||||
token2 := makeToken("六", 15, 18, 6)
|
||||
r.Value = token1
|
||||
r = r.Next()
|
||||
r.Value = token2
|
||||
// Simulate incorrect itemsInRing count
|
||||
return r, 3
|
||||
},
|
||||
inputPos: 14,
|
||||
expectToken: nil, // Expect nil when itemsInRing is not 1 or 2
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
ringPtr, itemsInRing := tt.ringSetup()
|
||||
itemsInRingCopy := itemsInRing // Pass a pointer to a copy
|
||||
|
||||
gotToken := filter.buildUnigram(ringPtr, &itemsInRingCopy, tt.inputPos)
|
||||
|
||||
if !reflect.DeepEqual(gotToken, tt.expectToken) {
|
||||
t.Errorf("buildUnigram() got = %v, want %v", gotToken, tt.expectToken)
|
||||
}
|
||||
|
||||
// Check if itemsInRing was modified (it shouldn't be by buildUnigram)
|
||||
if itemsInRingCopy != itemsInRing {
|
||||
t.Errorf("buildUnigram() modified itemsInRing, got = %d, want %d", itemsInRingCopy, itemsInRing)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestCJKBigramFilter_outputBigram(t *testing.T) {
|
||||
// Create a filter instance (outputUnigram value doesn't matter for outputBigram)
|
||||
filter := NewCJKBigramFilter(false)
|
||||
|
||||
tests := []struct {
|
||||
name string
|
||||
ringSetup func() (*ring.Ring, int) // Function to set up the ring and itemsInRing
|
||||
inputPos int // Position to pass to outputBigram
|
||||
expectToken *analysis.Token
|
||||
}{
|
||||
{
|
||||
name: "itemsInRing == 2",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("一", 0, 3, 1) // Original pos 1
|
||||
token2 := makeToken("二", 3, 6, 2) // Original pos 2
|
||||
r.Value = token1
|
||||
r = r.Next()
|
||||
r.Value = token2
|
||||
// r currently points to token2, r.Move(-1) points to token1
|
||||
return r, 2
|
||||
},
|
||||
inputPos: 10, // Expected output position
|
||||
expectToken: &analysis.Token{
|
||||
Type: analysis.Double,
|
||||
Term: []byte("一二"), // Combined term
|
||||
Position: 10, // Should use inputPos
|
||||
Start: 0, // Start of first token
|
||||
End: 6, // End of second token
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 2 with different terms",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("你好", 0, 6, 1)
|
||||
token2 := makeToken("世界", 6, 12, 2)
|
||||
r.Value = token1
|
||||
r = r.Next()
|
||||
r.Value = token2
|
||||
return r, 2
|
||||
},
|
||||
inputPos: 5,
|
||||
expectToken: &analysis.Token{
|
||||
Type: analysis.Double,
|
||||
Term: []byte("你好世界"),
|
||||
Position: 5,
|
||||
Start: 0,
|
||||
End: 12,
|
||||
},
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 1",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("三", 6, 9, 3)
|
||||
r.Value = token1
|
||||
return r, 1
|
||||
},
|
||||
inputPos: 11,
|
||||
expectToken: nil, // Expect nil when itemsInRing is not 2
|
||||
},
|
||||
{
|
||||
name: "itemsInRing == 0",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
// Ring is empty
|
||||
return r, 0
|
||||
},
|
||||
inputPos: 13,
|
||||
expectToken: nil, // Expect nil when itemsInRing is not 2
|
||||
},
|
||||
{
|
||||
name: "itemsInRing > 2 (should behave like 0)",
|
||||
ringSetup: func() (*ring.Ring, int) {
|
||||
r := ring.New(2)
|
||||
token1 := makeToken("五", 12, 15, 5)
|
||||
token2 := makeToken("六", 15, 18, 6)
|
||||
r.Value = token1
|
||||
r = r.Next()
|
||||
r.Value = token2
|
||||
// Simulate incorrect itemsInRing count
|
||||
return r, 3
|
||||
},
|
||||
inputPos: 14,
|
||||
expectToken: nil, // Expect nil when itemsInRing is not 2
|
||||
},
|
||||
}
|
||||
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
ringPtr, itemsInRing := tt.ringSetup()
|
||||
itemsInRingCopy := itemsInRing // Pass a pointer to a copy
|
||||
|
||||
gotToken := filter.outputBigram(ringPtr, &itemsInRingCopy, tt.inputPos)
|
||||
|
||||
if !reflect.DeepEqual(gotToken, tt.expectToken) {
|
||||
t.Errorf("outputBigram() got = %v, want %v", gotToken, tt.expectToken)
|
||||
}
|
||||
|
||||
// Check if itemsInRing was modified (it shouldn't be by outputBigram)
|
||||
if itemsInRingCopy != itemsInRing {
|
||||
t.Errorf("outputBigram() modified itemsInRing, got = %d, want %d", itemsInRingCopy, itemsInRing)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
func TestCJKBigramFilter(t *testing.T) {
|
||||
tests := []struct {
|
||||
outputUnigram bool
|
||||
input analysis.TokenStream
|
||||
output analysis.TokenStream
|
||||
}{
|
||||
// first test that non-adjacent terms are not combined
|
||||
{
|
||||
outputUnigram: false,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 2,
|
||||
Start: 5,
|
||||
End: 8,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Single,
|
||||
Position: 2,
|
||||
Start: 5,
|
||||
End: 8,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
outputUnigram: false,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こん"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("んに"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("にち"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ちは"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は世"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世界"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 21,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
outputUnigram: true,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("こん"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Single,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("んに"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Single,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("にち"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Single,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ちは"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Single,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は世"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Single,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 18,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世界"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 15,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Single,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
// Assuming that `、` is removed by unicode tokenizer from `こんにちは、世界`
|
||||
outputUnigram: true,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 8,
|
||||
Start: 21,
|
||||
End: 24,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Single,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("こん"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Single,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("んに"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Single,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("にち"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Single,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ちは"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Single,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Single,
|
||||
Position: 6,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世界"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 18,
|
||||
End: 24,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Single,
|
||||
Position: 7,
|
||||
Start: 21,
|
||||
End: 24,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
outputUnigram: false,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こ"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 3,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ん"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("に"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ち"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("は"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("cat"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 6,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 7,
|
||||
Start: 18,
|
||||
End: 21,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("界"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 8,
|
||||
Start: 21,
|
||||
End: 24,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("こん"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("んに"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("にち"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ちは"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("cat"),
|
||||
Type: analysis.AlphaNumeric,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("世界"),
|
||||
Type: analysis.Double,
|
||||
Position: 6,
|
||||
Start: 18,
|
||||
End: 24,
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
outputUnigram: false,
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("パイプライン"),
|
||||
Type: analysis.Ideographic,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 18,
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("パイ"),
|
||||
Type: analysis.Double,
|
||||
Position: 1,
|
||||
Start: 0,
|
||||
End: 6,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("イプ"),
|
||||
Type: analysis.Double,
|
||||
Position: 2,
|
||||
Start: 3,
|
||||
End: 9,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("プラ"),
|
||||
Type: analysis.Double,
|
||||
Position: 3,
|
||||
Start: 6,
|
||||
End: 12,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("ライ"),
|
||||
Type: analysis.Double,
|
||||
Position: 4,
|
||||
Start: 9,
|
||||
End: 15,
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("イン"),
|
||||
Type: analysis.Double,
|
||||
Position: 5,
|
||||
Start: 12,
|
||||
End: 18,
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
for _, test := range tests {
|
||||
cjkBigramFilter := NewCJKBigramFilter(test.outputUnigram)
|
||||
actual := cjkBigramFilter.Filter(test.input)
|
||||
if !reflect.DeepEqual(actual, test.output) {
|
||||
t.Errorf("expected %s, got %s", test.output, actual)
|
||||
}
|
||||
}
|
||||
}
|
||||
104
analysis/lang/cjk/cjk_width.go
Normal file
104
analysis/lang/cjk/cjk_width.go
Normal file
|
|
@ -0,0 +1,104 @@
|
|||
// Copyright (c) 2016 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 cjk
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"unicode/utf8"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
"github.com/blevesearch/bleve/v2/registry"
|
||||
)
|
||||
|
||||
const WidthName = "cjk_width"
|
||||
|
||||
type CJKWidthFilter struct{}
|
||||
|
||||
func NewCJKWidthFilter() *CJKWidthFilter {
|
||||
return &CJKWidthFilter{}
|
||||
}
|
||||
|
||||
func (s *CJKWidthFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
|
||||
for _, token := range input {
|
||||
runeCount := utf8.RuneCount(token.Term)
|
||||
runes := bytes.Runes(token.Term)
|
||||
for i := 0; i < runeCount; i++ {
|
||||
ch := runes[i]
|
||||
if ch >= 0xFF01 && ch <= 0xFF5E {
|
||||
// fullwidth ASCII variants
|
||||
runes[i] -= 0xFEE0
|
||||
} else if ch >= 0xFF65 && ch <= 0xFF9F {
|
||||
// halfwidth Katakana variants
|
||||
if (ch == 0xFF9E || ch == 0xFF9F) && i > 0 && combine(runes, i, ch) {
|
||||
runes = analysis.DeleteRune(runes, i)
|
||||
i--
|
||||
runeCount = len(runes)
|
||||
} else {
|
||||
runes[i] = kanaNorm[ch-0xFF65]
|
||||
}
|
||||
}
|
||||
}
|
||||
token.Term = analysis.BuildTermFromRunes(runes)
|
||||
}
|
||||
|
||||
return input
|
||||
}
|
||||
|
||||
var kanaNorm = []rune{
|
||||
0x30fb, 0x30f2, 0x30a1, 0x30a3, 0x30a5, 0x30a7, 0x30a9, 0x30e3, 0x30e5,
|
||||
0x30e7, 0x30c3, 0x30fc, 0x30a2, 0x30a4, 0x30a6, 0x30a8, 0x30aa, 0x30ab,
|
||||
0x30ad, 0x30af, 0x30b1, 0x30b3, 0x30b5, 0x30b7, 0x30b9, 0x30bb, 0x30bd,
|
||||
0x30bf, 0x30c1, 0x30c4, 0x30c6, 0x30c8, 0x30ca, 0x30cb, 0x30cc, 0x30cd,
|
||||
0x30ce, 0x30cf, 0x30d2, 0x30d5, 0x30d8, 0x30db, 0x30de, 0x30df, 0x30e0,
|
||||
0x30e1, 0x30e2, 0x30e4, 0x30e6, 0x30e8, 0x30e9, 0x30ea, 0x30eb, 0x30ec,
|
||||
0x30ed, 0x30ef, 0x30f3, 0x3099, 0x309A,
|
||||
}
|
||||
|
||||
var kanaCombineVoiced = []rune{
|
||||
78, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1,
|
||||
0, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1,
|
||||
0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
||||
0, 8, 8, 8, 8, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
|
||||
}
|
||||
var kanaCombineHalfVoiced = []rune{
|
||||
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
||||
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 2, 0, 0, 2,
|
||||
0, 0, 2, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
||||
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
||||
}
|
||||
|
||||
func combine(text []rune, pos int, r rune) bool {
|
||||
prev := text[pos-1]
|
||||
if prev >= 0x30A6 && prev <= 0x30FD {
|
||||
if r == 0xFF9F {
|
||||
text[pos-1] += kanaCombineHalfVoiced[prev-0x30A6]
|
||||
} else {
|
||||
text[pos-1] += kanaCombineVoiced[prev-0x30A6]
|
||||
}
|
||||
return text[pos-1] != prev
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
func CJKWidthFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
|
||||
return NewCJKWidthFilter(), nil
|
||||
}
|
||||
|
||||
func init() {
|
||||
err := registry.RegisterTokenFilter(WidthName, CJKWidthFilterConstructor)
|
||||
if err != nil {
|
||||
panic(err)
|
||||
}
|
||||
}
|
||||
93
analysis/lang/cjk/cjk_width_test.go
Normal file
93
analysis/lang/cjk/cjk_width_test.go
Normal file
|
|
@ -0,0 +1,93 @@
|
|||
// Copyright (c) 2016 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 cjk
|
||||
|
||||
import (
|
||||
"reflect"
|
||||
"testing"
|
||||
|
||||
"github.com/blevesearch/bleve/v2/analysis"
|
||||
)
|
||||
|
||||
func TestCJKWidthFilter(t *testing.T) {
|
||||
|
||||
tests := []struct {
|
||||
input analysis.TokenStream
|
||||
output analysis.TokenStream
|
||||
}{
|
||||
{
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("Ｔｅｓｔ"),
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("１２３４"),
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("Test"),
|
||||
},
|
||||
&analysis.Token{
|
||||
Term: []byte("1234"),
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("ｶﾀｶﾅ"),
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("カタカナ"),
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("ｳﾞｨｯﾂ"),
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("ヴィッツ"),
|
||||
},
|
||||
},
|
||||
},
|
||||
{
|
||||
input: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("ﾊﾟﾅｿﾆｯｸ"),
|
||||
},
|
||||
},
|
||||
output: analysis.TokenStream{
|
||||
&analysis.Token{
|
||||
Term: []byte("パナソニック"),
|
||||
},
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
for _, test := range tests {
|
||||
cjkWidthFilter := NewCJKWidthFilter()
|
||||
actual := cjkWidthFilter.Filter(test.input)
|
||||
if !reflect.DeepEqual(actual, test.output) {
|
||||
t.Errorf("expected %s, got %s", test.output, actual)
|
||||
}
|
||||
}
|
||||
}
|
||||
Loading…
Add table
Add a link
Reference in a new issue