aboutsummaryrefslogtreecommitdiffstats
path: root/modules/indexer/code
diff options
context:
space:
mode:
Diffstat (limited to 'modules/indexer/code')
-rw-r--r--modules/indexer/code/bleve/bleve.go16
-rw-r--r--modules/indexer/code/bleve/token/path/path.go12
-rw-r--r--modules/indexer/code/elasticsearch/elasticsearch.go8
-rw-r--r--modules/indexer/code/elasticsearch/elasticsearch_test.go4
-rw-r--r--modules/indexer/code/git.go4
-rw-r--r--modules/indexer/code/gitgrep/gitgrep.go5
-rw-r--r--modules/indexer/code/indexer_test.go2
-rw-r--r--modules/indexer/code/internal/indexer.go8
-rw-r--r--modules/indexer/code/search.go2
9 files changed, 27 insertions, 34 deletions
diff --git a/modules/indexer/code/bleve/bleve.go b/modules/indexer/code/bleve/bleve.go
index e1a381f992..70f0995a01 100644
--- a/modules/indexer/code/bleve/bleve.go
+++ b/modules/indexer/code/bleve/bleve.go
@@ -16,7 +16,6 @@ import (
"code.gitea.io/gitea/modules/analyze"
"code.gitea.io/gitea/modules/charset"
"code.gitea.io/gitea/modules/git"
- "code.gitea.io/gitea/modules/gitrepo"
"code.gitea.io/gitea/modules/indexer"
path_filter "code.gitea.io/gitea/modules/indexer/code/bleve/token/path"
"code.gitea.io/gitea/modules/indexer/code/internal"
@@ -30,7 +29,6 @@ import (
"github.com/blevesearch/bleve/v2"
analyzer_custom "github.com/blevesearch/bleve/v2/analysis/analyzer/custom"
analyzer_keyword "github.com/blevesearch/bleve/v2/analysis/analyzer/keyword"
- "github.com/blevesearch/bleve/v2/analysis/token/camelcase"
"github.com/blevesearch/bleve/v2/analysis/token/lowercase"
"github.com/blevesearch/bleve/v2/analysis/token/unicodenorm"
"github.com/blevesearch/bleve/v2/analysis/tokenizer/letter"
@@ -72,7 +70,7 @@ const (
filenameIndexerAnalyzer = "filenameIndexerAnalyzer"
filenameIndexerTokenizer = "filenameIndexerTokenizer"
repoIndexerDocType = "repoIndexerDocType"
- repoIndexerLatestVersion = 8
+ repoIndexerLatestVersion = 9
)
// generateBleveIndexMapping generates a bleve index mapping for the repo indexer
@@ -109,7 +107,7 @@ func generateBleveIndexMapping() (mapping.IndexMapping, error) {
"type": analyzer_custom.Name,
"char_filters": []string{},
"tokenizer": letter.Name,
- "token_filters": []string{unicodeNormalizeName, camelcase.Name, lowercase.Name},
+ "token_filters": []string{unicodeNormalizeName, lowercase.Name},
}); err != nil {
return nil, err
}
@@ -191,7 +189,8 @@ func (b *Indexer) addUpdate(ctx context.Context, batchWriter git.WriteCloserErro
return err
} else if !typesniffer.DetectContentType(fileContents).IsText() {
// FIXME: UTF-16 files will probably fail here
- return nil
+ // Even if the file is not recognized as a "text file", we could still put its name into the indexers to make the filename become searchable, while leave the content to empty.
+ fileContents = nil
}
if _, err = batchReader.Discard(1); err != nil {
@@ -217,12 +216,7 @@ func (b *Indexer) addDelete(filename string, repo *repo_model.Repository, batch
func (b *Indexer) Index(ctx context.Context, repo *repo_model.Repository, sha string, changes *internal.RepoChanges) error {
batch := inner_bleve.NewFlushingBatch(b.inner.Indexer, maxBatchSize)
if len(changes.Updates) > 0 {
- r, err := gitrepo.OpenRepository(ctx, repo)
- if err != nil {
- return err
- }
- defer r.Close()
- gitBatch, err := r.NewBatch(ctx)
+ gitBatch, err := git.NewBatch(ctx, repo.RepoPath())
if err != nil {
return err
}
diff --git a/modules/indexer/code/bleve/token/path/path.go b/modules/indexer/code/bleve/token/path/path.go
index 107e0da109..6dfc12f146 100644
--- a/modules/indexer/code/bleve/token/path/path.go
+++ b/modules/indexer/code/bleve/token/path/path.go
@@ -51,13 +51,13 @@ func generatePathTokens(input analysis.TokenStream, reversed bool) analysis.Toke
slices.Reverse(input)
}
- for i := 0; i < len(input); i++ {
+ for i := range input {
var sb strings.Builder
- sb.WriteString(string(input[0].Term))
+ sb.Write(input[0].Term)
for j := 1; j < i; j++ {
sb.WriteString("/")
- sb.WriteString(string(input[j].Term))
+ sb.Write(input[j].Term)
}
term := sb.String()
@@ -97,5 +97,9 @@ func generatePathTokens(input analysis.TokenStream, reversed bool) analysis.Toke
}
func init() {
- registry.RegisterTokenFilter(Name, TokenFilterConstructor)
+ // FIXME: move it to the bleve's init function, but do not call it in global init
+ err := registry.RegisterTokenFilter(Name, TokenFilterConstructor)
+ if err != nil {
+ panic(err)
+ }
}
diff --git a/modules/indexer/code/elasticsearch/elasticsearch.go b/modules/indexer/code/elasticsearch/elasticsearch.go
index be8efad5fd..f925ce396a 100644
--- a/modules/indexer/code/elasticsearch/elasticsearch.go
+++ b/modules/indexer/code/elasticsearch/elasticsearch.go
@@ -15,7 +15,6 @@ import (
"code.gitea.io/gitea/modules/analyze"
"code.gitea.io/gitea/modules/charset"
"code.gitea.io/gitea/modules/git"
- "code.gitea.io/gitea/modules/gitrepo"
"code.gitea.io/gitea/modules/indexer"
"code.gitea.io/gitea/modules/indexer/code/internal"
indexer_internal "code.gitea.io/gitea/modules/indexer/internal"
@@ -209,12 +208,7 @@ func (b *Indexer) addDelete(filename string, repo *repo_model.Repository) elasti
func (b *Indexer) Index(ctx context.Context, repo *repo_model.Repository, sha string, changes *internal.RepoChanges) error {
reqs := make([]elastic.BulkableRequest, 0)
if len(changes.Updates) > 0 {
- r, err := gitrepo.OpenRepository(ctx, repo)
- if err != nil {
- return err
- }
- defer r.Close()
- batch, err := r.NewBatch(ctx)
+ batch, err := git.NewBatch(ctx, repo.RepoPath())
if err != nil {
return err
}
diff --git a/modules/indexer/code/elasticsearch/elasticsearch_test.go b/modules/indexer/code/elasticsearch/elasticsearch_test.go
index a6d2af92b2..e8f1f202ce 100644
--- a/modules/indexer/code/elasticsearch/elasticsearch_test.go
+++ b/modules/indexer/code/elasticsearch/elasticsearch_test.go
@@ -11,6 +11,6 @@ import (
func TestIndexPos(t *testing.T) {
startIdx, endIdx := contentMatchIndexPos("test index start and end", "start", "end")
- assert.EqualValues(t, 11, startIdx)
- assert.EqualValues(t, 15, endIdx)
+ assert.Equal(t, 11, startIdx)
+ assert.Equal(t, 15, endIdx)
}
diff --git a/modules/indexer/code/git.go b/modules/indexer/code/git.go
index 0089dd259f..41bc74e6ec 100644
--- a/modules/indexer/code/git.go
+++ b/modules/indexer/code/git.go
@@ -129,8 +129,8 @@ func nonGenesisChanges(ctx context.Context, repo *repo_model.Repository, revisio
changes.Updates = append(changes.Updates, updates...)
return nil
}
- lines := strings.Split(stdout, "\n")
- for _, line := range lines {
+ lines := strings.SplitSeq(stdout, "\n")
+ for line := range lines {
line = strings.TrimSpace(line)
if len(line) == 0 {
continue
diff --git a/modules/indexer/code/gitgrep/gitgrep.go b/modules/indexer/code/gitgrep/gitgrep.go
index 093c189ba3..6f6e0b47b9 100644
--- a/modules/indexer/code/gitgrep/gitgrep.go
+++ b/modules/indexer/code/gitgrep/gitgrep.go
@@ -26,9 +26,10 @@ func indexSettingToGitGrepPathspecList() (list []string) {
func PerformSearch(ctx context.Context, page int, repoID int64, gitRepo *git.Repository, ref git.RefName, keyword string, searchMode indexer.SearchModeType) (searchResults []*code_indexer.Result, total int, err error) {
grepMode := git.GrepModeWords
- if searchMode == indexer.SearchModeExact {
+ switch searchMode {
+ case indexer.SearchModeExact:
grepMode = git.GrepModeExact
- } else if searchMode == indexer.SearchModeRegexp {
+ case indexer.SearchModeRegexp:
grepMode = git.GrepModeRegexp
}
res, err := git.GrepSearch(ctx, gitRepo, keyword, git.GrepOptions{
diff --git a/modules/indexer/code/indexer_test.go b/modules/indexer/code/indexer_test.go
index 1d9bf99d40..78fea22f10 100644
--- a/modules/indexer/code/indexer_test.go
+++ b/modules/indexer/code/indexer_test.go
@@ -310,7 +310,7 @@ func TestESIndexAndSearch(t *testing.T) {
if indexer != nil {
indexer.Close()
}
- assert.FailNow(t, "Unable to init ES indexer Error: %v", err)
+ require.NoError(t, err, "Unable to init ES indexer")
}
defer indexer.Close()
diff --git a/modules/indexer/code/internal/indexer.go b/modules/indexer/code/internal/indexer.go
index 6c9a8af635..d58b028124 100644
--- a/modules/indexer/code/internal/indexer.go
+++ b/modules/indexer/code/internal/indexer.go
@@ -5,7 +5,7 @@ package internal
import (
"context"
- "fmt"
+ "errors"
"code.gitea.io/gitea/models/db"
repo_model "code.gitea.io/gitea/models/repo"
@@ -48,13 +48,13 @@ func (d *dummyIndexer) SupportedSearchModes() []indexer.SearchMode {
}
func (d *dummyIndexer) Index(ctx context.Context, repo *repo_model.Repository, sha string, changes *RepoChanges) error {
- return fmt.Errorf("indexer is not ready")
+ return errors.New("indexer is not ready")
}
func (d *dummyIndexer) Delete(ctx context.Context, repoID int64) error {
- return fmt.Errorf("indexer is not ready")
+ return errors.New("indexer is not ready")
}
func (d *dummyIndexer) Search(ctx context.Context, opts *SearchOptions) (int64, []*SearchResult, []*SearchResultLanguages, error) {
- return 0, nil, nil, fmt.Errorf("indexer is not ready")
+ return 0, nil, nil, errors.New("indexer is not ready")
}
diff --git a/modules/indexer/code/search.go b/modules/indexer/code/search.go
index e37aff8e59..a7a5d7d2e3 100644
--- a/modules/indexer/code/search.go
+++ b/modules/indexer/code/search.go
@@ -77,7 +77,7 @@ func HighlightSearchResultCode(filename, language string, lineNums []int, code s
// The lineNums outputted by highlight.Code might not match the original lineNums, because "highlight" removes the last `\n`
lines := make([]*ResultLine, min(len(highlightedLines), len(lineNums)))
- for i := 0; i < len(lines); i++ {
+ for i := range lines {
lines[i] = &ResultLine{
Num: lineNums[i],
FormattedContent: template.HTML(highlightedLines[i]),
generated by cgit v1.2.3 (git 2.39.1) at 2025-07-04 09:38:41 +0000