tdsds.com/gitea
1
1
Fork 0
mirror of https://github.com/go-gitea/gitea synced 2025-07-03 09:07:19 +00:00
Code Releases Activity

Fix bleve fuzziness search (#33078)

Browse Source 
Close #31565
This commit is contained in:
wxiaoguang
2025-01-03 00:32:02 +08:00
committed by GitHub
parent 9882917bce
commit 45973a100b
11 changed files with 83 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
modules/indexer/code/indexer.go
View File

@ -123,13 +123,12 @@ func Init() {
for _, indexerData := range items {
log.Trace("IndexerData Process Repo: %d", indexerData.RepoID)
if err := index(ctx, indexer, indexerData.RepoID); err != nil {
unhandled = append(unhandled, indexerData)
if !setting.IsInTesting {
log.Error("Codes indexer handler: index error for repo %v: %v", indexerData.RepoID, err)
}
}
}
return unhandled
return nil // do not re-queue the failed items, otherwise some broken repo will block the queue
}
indexerQueue = queue.CreateUniqueQueue(ctx, "code_indexer", handler)

4
modules/indexer/code/indexer_test.go
View File

@ -15,6 +15,8 @@ import (
"code.gitea.io/gitea/modules/indexer/code/bleve"
"code.gitea.io/gitea/modules/indexer/code/elasticsearch"
"code.gitea.io/gitea/modules/indexer/code/internal"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/test"
_ "code.gitea.io/gitea/models"
_ "code.gitea.io/gitea/models/actions"
@ -279,7 +281,7 @@ func testIndexer(name string, t *testing.T, indexer internal.Indexer) {
func TestBleveIndexAndSearch(t *testing.T) {
unittest.PrepareTestEnv(t)
defer test.MockVariableValue(&setting.Indexer.TypeBleveMaxFuzzniess, 2)()
dir := t.TempDir()
idx := bleve.NewIndexer(dir)

9
modules/indexer/internal/bleve/util.go
View File

@ -9,6 +9,7 @@ import (
"unicode"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/util"
"github.com/blevesearch/bleve/v2"
@ -54,9 +55,9 @@ func openIndexer(path string, latestVersion int) (bleve.Index, int, error) {
return index, 0, nil
}
// This method test the GuessFuzzinessByKeyword method. The fuzziness is based on the levenshtein distance and determines how many chars
// may be different on two string and they still be considered equivalent.
// Given a phrasse, its shortest word determines its fuzziness. If a phrase uses CJK (eg: `갃갃갃` `啊啊啊`), the fuzziness is zero.
// GuessFuzzinessByKeyword guesses fuzziness based on the levenshtein distance and determines how many chars
// may be different on two string, and they still be considered equivalent.
// Given a phrase, its shortest word determines its fuzziness. If a phrase uses CJK (eg: `갃갃갃` `啊啊啊`), the fuzziness is zero.
func GuessFuzzinessByKeyword(s string) int {
tokenizer := unicode_tokenizer.NewUnicodeTokenizer()
tokens := tokenizer.Tokenize([]byte(s))
@ -85,5 +86,5 @@ func guessFuzzinessByKeyword(s string) int {
return 0
}
}
return min(maxFuzziness, len(s)/4)
return min(min(setting.Indexer.TypeBleveMaxFuzzniess, maxFuzziness), len(s)/4)
}

7
modules/indexer/internal/bleve/util_test.go
View File

@ -7,10 +7,15 @@ import (
"fmt"
"testing"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/test"
"github.com/stretchr/testify/assert"
)
func TestBleveGuessFuzzinessByKeyword(t *testing.T) {
defer test.MockVariableValue(&setting.Indexer.TypeBleveMaxFuzzniess, 2)()
scenarios := []struct {
Input string
Fuzziness int // See util.go for the definition of fuzziness in this particular context
@ -46,7 +51,7 @@ func TestBleveGuessFuzzinessByKeyword(t *testing.T) {
}
for _, scenario := range scenarios {
t.Run(fmt.Sprintf("ensure fuzziness of '%s' is '%d'", scenario.Input, scenario.Fuzziness), func(t *testing.T) {
t.Run(fmt.Sprintf("Fuziniess:%s=%d", scenario.Input, scenario.Fuzziness), func(t *testing.T) {
assert.Equal(t, scenario.Fuzziness, GuessFuzzinessByKeyword(scenario.Input))
})
}

3
modules/setting/indexer.go
View File

@ -31,6 +31,8 @@ var Indexer = struct {
IncludePatterns []*GlobMatcher
ExcludePatterns []*GlobMatcher
ExcludeVendored bool
TypeBleveMaxFuzzniess int
}{
IssueType: "bleve",
IssuePath: "indexers/issues.bleve",
@ -88,6 +90,7 @@ func loadIndexerFrom(rootCfg ConfigProvider) {
Indexer.ExcludeVendored = sec.Key("REPO_INDEXER_EXCLUDE_VENDORED").MustBool(true)
Indexer.MaxIndexerFileSize = sec.Key("MAX_FILE_SIZE").MustInt64(1024 * 1024)
Indexer.StartupTimeout = sec.Key("STARTUP_TIMEOUT").MustDuration(30 * time.Second)
Indexer.TypeBleveMaxFuzzniess = sec.Key("TYPE_BLEVE_MAX_FUZZINESS").MustInt(0)
}
// IndexerGlobFromString parses a comma separated list of patterns and returns a glob.Glob slice suited for repo indexing