1
0
Fork 0
forked from forgejo/forgejo

Improve issue search (#2387)

* Improve issue indexer

* Fix new issue sqlite bug

* Different test indexer paths for each db

* Add integration indexer paths to make clean
This commit is contained in:
Ethan Koenig 2017-09-16 13:16:21 -07:00 committed by Lauris BH
parent 52e11b24bf
commit b0f7457d9e
122 changed files with 15280 additions and 1458 deletions

View file

@ -0,0 +1,145 @@
// 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 custom
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
)
const Name = "custom"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
var err error
var charFilters []analysis.CharFilter
charFiltersValue, ok := config["char_filters"]
if ok {
switch charFiltersValue := charFiltersValue.(type) {
case []string:
charFilters, err = getCharFilters(charFiltersValue, cache)
if err != nil {
return nil, err
}
case []interface{}:
charFiltersNames, err := convertInterfaceSliceToStringSlice(charFiltersValue, "char filter")
if err != nil {
return nil, err
}
charFilters, err = getCharFilters(charFiltersNames, cache)
if err != nil {
return nil, err
}
default:
return nil, fmt.Errorf("unsupported type for char_filters, must be slice")
}
}
var tokenizerName string
tokenizerValue, ok := config["tokenizer"]
if ok {
tokenizerName, ok = tokenizerValue.(string)
if !ok {
return nil, fmt.Errorf("must specify tokenizer as string")
}
} else {
return nil, fmt.Errorf("must specify tokenizer")
}
tokenizer, err := cache.TokenizerNamed(tokenizerName)
if err != nil {
return nil, err
}
var tokenFilters []analysis.TokenFilter
tokenFiltersValue, ok := config["token_filters"]
if ok {
switch tokenFiltersValue := tokenFiltersValue.(type) {
case []string:
tokenFilters, err = getTokenFilters(tokenFiltersValue, cache)
if err != nil {
return nil, err
}
case []interface{}:
tokenFiltersNames, err := convertInterfaceSliceToStringSlice(tokenFiltersValue, "token filter")
if err != nil {
return nil, err
}
tokenFilters, err = getTokenFilters(tokenFiltersNames, cache)
if err != nil {
return nil, err
}
default:
return nil, fmt.Errorf("unsupported type for token_filters, must be slice")
}
}
rv := analysis.Analyzer{
Tokenizer: tokenizer,
}
if charFilters != nil {
rv.CharFilters = charFilters
}
if tokenFilters != nil {
rv.TokenFilters = tokenFilters
}
return &rv, nil
}
func init() {
registry.RegisterAnalyzer(Name, AnalyzerConstructor)
}
func getCharFilters(charFilterNames []string, cache *registry.Cache) ([]analysis.CharFilter, error) {
charFilters := make([]analysis.CharFilter, len(charFilterNames))
for i, charFilterName := range charFilterNames {
charFilter, err := cache.CharFilterNamed(charFilterName)
if err != nil {
return nil, err
}
charFilters[i] = charFilter
}
return charFilters, nil
}
func getTokenFilters(tokenFilterNames []string, cache *registry.Cache) ([]analysis.TokenFilter, error) {
tokenFilters := make([]analysis.TokenFilter, len(tokenFilterNames))
for i, tokenFilterName := range tokenFilterNames {
tokenFilter, err := cache.TokenFilterNamed(tokenFilterName)
if err != nil {
return nil, err
}
tokenFilters[i] = tokenFilter
}
return tokenFilters, nil
}
func convertInterfaceSliceToStringSlice(interfaceSlice []interface{}, objType string) ([]string, error) {
stringSlice := make([]string, len(interfaceSlice))
for i, interfaceObj := range interfaceSlice {
stringObj, ok := interfaceObj.(string)
if ok {
stringSlice[i] = stringObj
} else {
return nil, fmt.Errorf(objType + " name must be a string")
}
}
return stringSlice, nil
}

View file

@ -1,46 +0,0 @@
// 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 simple
import (
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/token/lowercase"
"github.com/blevesearch/bleve/analysis/tokenizer/letter"
"github.com/blevesearch/bleve/registry"
)
const Name = "simple"
func AnalyzerConstructor(config map[string]interface{}, cache *registry.Cache) (*analysis.Analyzer, error) {
tokenizer, err := cache.TokenizerNamed(letter.Name)
if err != nil {
return nil, err
}
toLowerFilter, err := cache.TokenFilterNamed(lowercase.Name)
if err != nil {
return nil, err
}
rv := analysis.Analyzer{
Tokenizer: tokenizer,
TokenFilters: []analysis.TokenFilter{
toLowerFilter,
},
}
return &rv, nil
}
func init() {
registry.RegisterAnalyzer(Name, AnalyzerConstructor)
}

View file

@ -0,0 +1,79 @@
// 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 unicodenorm
import (
"fmt"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/registry"
"golang.org/x/text/unicode/norm"
)
const Name = "normalize_unicode"
const NFC = "nfc"
const NFD = "nfd"
const NFKC = "nfkc"
const NFKD = "nfkd"
var forms = map[string]norm.Form{
NFC: norm.NFC,
NFD: norm.NFD,
NFKC: norm.NFKC,
NFKD: norm.NFKD,
}
type UnicodeNormalizeFilter struct {
form norm.Form
}
func NewUnicodeNormalizeFilter(formName string) (*UnicodeNormalizeFilter, error) {
form, ok := forms[formName]
if !ok {
return nil, fmt.Errorf("no form named %s", formName)
}
return &UnicodeNormalizeFilter{
form: form,
}, nil
}
func MustNewUnicodeNormalizeFilter(formName string) *UnicodeNormalizeFilter {
filter, err := NewUnicodeNormalizeFilter(formName)
if err != nil {
panic(err)
}
return filter
}
func (s *UnicodeNormalizeFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
for _, token := range input {
token.Term = s.form.Bytes(token.Term)
}
return input
}
func UnicodeNormalizeFilterConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.TokenFilter, error) {
formVal, ok := config["form"].(string)
if !ok {
return nil, fmt.Errorf("must specify form")
}
form := formVal
return NewUnicodeNormalizeFilter(form)
}
func init() {
registry.RegisterTokenFilter(Name, UnicodeNormalizeFilterConstructor)
}

View file

@ -1,76 +0,0 @@
// 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 character
import (
"unicode/utf8"
"github.com/blevesearch/bleve/analysis"
)
type IsTokenRune func(r rune) bool
type CharacterTokenizer struct {
isTokenRun IsTokenRune
}
func NewCharacterTokenizer(f IsTokenRune) *CharacterTokenizer {
return &CharacterTokenizer{
isTokenRun: f,
}
}
func (c *CharacterTokenizer) Tokenize(input []byte) analysis.TokenStream {
rv := make(analysis.TokenStream, 0, 1024)
offset := 0
start := 0
end := 0
count := 0
for currRune, size := utf8.DecodeRune(input[offset:]); currRune != utf8.RuneError; currRune, size = utf8.DecodeRune(input[offset:]) {
isToken := c.isTokenRun(currRune)
if isToken {
end = offset + size
} else {
if end-start > 0 {
// build token
rv = append(rv, &analysis.Token{
Term: input[start:end],
Start: start,
End: end,
Position: count + 1,
Type: analysis.AlphaNumeric,
})
count++
}
start = offset + size
end = start
}
offset += size
}
// if we ended in the middle of a token, finish it
if end-start > 0 {
// build token
rv = append(rv, &analysis.Token{
Term: input[start:end],
Start: start,
End: end,
Position: count + 1,
Type: analysis.AlphaNumeric,
})
}
return rv
}

View file

@ -1,33 +0,0 @@
// 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 letter
import (
"unicode"
"github.com/blevesearch/bleve/analysis"
"github.com/blevesearch/bleve/analysis/tokenizer/character"
"github.com/blevesearch/bleve/registry"
)
const Name = "letter"
func TokenizerConstructor(config map[string]interface{}, cache *registry.Cache) (analysis.Tokenizer, error) {
return character.NewCharacterTokenizer(unicode.IsLetter), nil
}
func init() {
registry.RegisterTokenizer(Name, TokenizerConstructor)
}