0
0

fix issues with italian analyzer

switch it to not require icu/libstemmer
fixes #69
This commit is contained in:
Marty Schoch 2015-03-11 11:48:13 -04:00
parent eeaf514848
commit 7970f42c29
5 changed files with 196 additions and 29 deletions

View File

@ -7,9 +7,6 @@
// either express or implied. See the License for the specific language governing permissions
// and limitations under the License.
// +build libstemmer full
// +build icu full
package it
import (
@ -17,13 +14,13 @@ import (
"github.com/blevesearch/bleve/registry"
"github.com/blevesearch/bleve/analysis/token_filters/lower_case_filter"
"github.com/blevesearch/bleve/analysis/tokenizers/icu"
"github.com/blevesearch/bleve/analysis/tokenizers/unicode"
)
const AnalyzerName = "it"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
icuTokenizer, err := cache.TokenizerNamed(icu.Name)
tokenizer, err := cache.TokenizerNamed(unicode.Name)
if err != nil {
return nil, err
}
@ -39,12 +36,12 @@ func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (
if err != nil {
return nil, err
}
stemmerItFilter, err := cache.TokenFilterNamed(StemmerName)
stemmerItFilter, err := cache.TokenFilterNamed(LightStemmerName)
if err != nil {
return nil, err
}
rv := analysis.Analyzer{
Tokenizer: icuTokenizer,
Tokenizer: tokenizer,
TokenFilters: []analysis.TokenFilter{
elisionFilter,
toLowerFilter,

View File

@ -7,9 +7,6 @@
// either express or implied. See the License for the specific language governing permissions
// and limitations under the License.
// +build libstemmer full
// +build icu full
package it
import (
@ -26,29 +23,44 @@ func TestItalianAnalyzer(t *testing.T) {
output analysis.TokenStream
}{
// stemming
// fails, stemming discrepencies
// abbandon intead of abbandonat
// {
// input: []byte("abbandonata"),
// output: analysis.TokenStream{
// &analysis.Token{
// Term: []byte("abbandonat"),
// },
// },
// },
// {
// input: []byte("abbandonati"),
// output: analysis.TokenStream{
// &analysis.Token{
// Term: []byte("abbandonat"),
// },
// },
// },
{
input: []byte("abbandonata"),
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("abbandonat"),
},
},
},
{
input: []byte("abbandonati"),
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("abbandonat"),
},
},
},
// stop word
{
input: []byte("dallo"),
output: analysis.TokenStream{},
},
// contractions
{
input: []byte("dell'Italia"),
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("ital"),
},
},
},
{
input: []byte("l'Italiano"),
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("italian"),
},
},
},
}
cache := registry.NewCache()

View File

@ -17,7 +17,7 @@ import (
"github.com/blevesearch/bleve/registry"
)
func TestFrenchElision(t *testing.T) {
func TestItalianElision(t *testing.T) {
tests := []struct {
input analysis.TokenStream
output analysis.TokenStream

View File

@ -0,0 +1,96 @@
// Copyright (c) 2015 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 it
import (
"bytes"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const LightStemmerName = "stemmer_it_light"
type ItalianLightStemmerFilter struct {
}
func NewItalianLightStemmerFilterFilter() *ItalianLightStemmerFilter {
return &ItalianLightStemmerFilter{}
}
func (s *ItalianLightStemmerFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
for _, token := range input {
runes := bytes.Runes(token.Term)
runes = stem(runes)
token.Term = analysis.BuildTermFromRunes(runes)
}
return input
}
func stem(input []rune) []rune {
inputLen := len(input)
if inputLen < 6 {
return input
}
for i := 0; i < inputLen; i++ {
switch input[i] {
case 'à', 'á', 'â', 'ä':
input[i] = 'a'
case 'ò', 'ó', 'ô', 'ö':
input[i] = 'o'
case 'è', 'é', 'ê', 'ë':
input[i] = 'e'
case 'ù', 'ú', 'û', 'ü':
input[i] = 'u'
case 'ì', 'í', 'î', 'ï':
input[i] = 'i'
}
}
switch input[inputLen-1] {
case 'e':
if input[inputLen-2] == 'i' || input[inputLen-2] == 'h' {
return input[0 : inputLen-2]
} else {
return input[0 : inputLen-1]
}
case 'i':
if input[inputLen-2] == 'h' || input[inputLen-2] == 'i' {
return input[0 : inputLen-2]
} else {
return input[0 : inputLen-1]
}
case 'a':
if input[inputLen-2] == 'i' {
return input[0 : inputLen-2]
} else {
return input[0 : inputLen-1]
}
case 'o':
if input[inputLen-2] == 'i' {
return input[0 : inputLen-2]
} else {
return input[0 : inputLen-1]
}
}
return input
}
func ItalianLightStemmerFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
return NewItalianLightStemmerFilterFilter(), nil
}
func init() {
registry.RegisterTokenFilter(LightStemmerName, ItalianLightStemmerFilterConstructor)
}

View File

@ -0,0 +1,62 @@
// Copyright (c) 2015 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 it
import (
"reflect"
"testing"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
func TestItalianLightStemmer(t *testing.T) {
tests := []struct {
input analysis.TokenStream
output analysis.TokenStream
}{
{
input: analysis.TokenStream{
&analysis.Token{
Term: []byte("ragazzo"),
},
},
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("ragazz"),
},
},
},
{
input: analysis.TokenStream{
&analysis.Token{
Term: []byte("ragazzi"),
},
},
output: analysis.TokenStream{
&analysis.Token{
Term: []byte("ragazz"),
},
},
},
}
cache := registry.NewCache()
filter, err := cache.TokenFilterNamed(LightStemmerName)
if err != nil {
t.Fatal(err)
}
for _, test := range tests {
actual := filter.Filter(test.input)
if !reflect.DeepEqual(actual, test.output) {
t.Errorf("expected %s, got %s", test.output[0].Term, actual[0].Term)
}
}
}