Support elastic search for code search (#10273)

* Support elastic search for code search

* Finished elastic search implementation and add some tests

* Enable test on drone and added docs

* Add new fields to elastic search

* Fix bug

* remove unused changes

* Use indexer alias to keep the gitea indexer version

* Improve codes

* Some code improvements

* The real indexer name changed to xxx.v1

Co-authored-by: zeripath <[email protected]>

Author: lunny

.drone.yml

@@ -209,6 +209,7 @@ steps:
       TAGS: bindata
       TEST_LDAP: 1
       USE_REPO_TEST_DIR: 1
+      TEST_INDEXER_CODE_ES_URL: "http://elastic:changeme@elasticsearch:9200"
     depends_on:
       - build
 

custom/conf/app.example.ini

@@ -428,7 +428,15 @@ STARTUP_TIMEOUT=30s
 
 ; repo indexer by default disabled, since it uses a lot of disk space
 REPO_INDEXER_ENABLED = false
+; Code search engine type, could be `bleve` or `elasticsearch`.
+REPO_INDEXER_TYPE = bleve
+; Index file used for code search.
 REPO_INDEXER_PATH = indexers/repos.bleve
+; Code indexer connection string, available when `REPO_INDEXER_TYPE` is elasticsearch. i.e. http://elastic:changeme@localhost:9200
+REPO_INDEXER_CONN_STR = 
+; Code indexer name, available when `REPO_INDEXER_TYPE` is elasticsearch
+REPO_INDEXER_NAME = gitea_codes
+
 UPDATE_BUFFER_LEN = 20
 MAX_FILE_SIZE = 1048576
 ; A comma separated list of glob patterns (see https://github.com/gobwas/glob) to include

docs/content/doc/advanced/config-cheat-sheet.en-us.md

@@ -270,7 +270,11 @@ relation to port exhaustion.
 - `ISSUE_INDEXER_QUEUE_BATCH_NUMBER`: **20**: Batch queue number.
 
 - `REPO_INDEXER_ENABLED`: **false**: Enables code search (uses a lot of disk space, about 6 times more than the repository size).
+- `REPO_INDEXER_TYPE`: **bleve**: Code search engine type, could be `bleve` or `elasticsearch`.
 - `REPO_INDEXER_PATH`: **indexers/repos.bleve**: Index file used for code search.
+- `REPO_INDEXER_CONN_STR`: ****: Code indexer connection string, available when `REPO_INDEXER_TYPE` is elasticsearch. i.e. http://elastic:changeme@localhost:9200
+- `REPO_INDEXER_NAME`: **gitea_codes**: Code indexer name, available when `REPO_INDEXER_TYPE` is elasticsearch
+
 - `REPO_INDEXER_INCLUDE`: **empty**: A comma separated list of glob patterns (see https://github.com/gobwas/glob) to **include** in the index. Use `**.txt` to match any files with .txt extension. An empty list means include all files.
 - `REPO_INDEXER_EXCLUDE`: **empty**: A comma separated list of glob patterns (see https://github.com/gobwas/glob) to **exclude** from the index. Files that match this list will not be indexed, even if they match in `REPO_INDEXER_INCLUDE`.
 - `REPO_INDEXER_EXCLUDE_VENDORED`: **true**: Exclude vendored files from index.

docs/content/doc/advanced/config-cheat-sheet.zh-cn.md

@@ -98,8 +98,12 @@ menu:
 - `ISSUE_INDEXER_QUEUE_CONN_STR`: **addrs=127.0.0.1:6379 db=0**: 当 `ISSUE_INDEXER_QUEUE_TYPE` 为 `redis` 时，保存Redis队列的连接字符串。
 - `ISSUE_INDEXER_QUEUE_BATCH_NUMBER`: **20**: 队列处理中批量提交数量。
 
-- `REPO_INDEXER_ENABLED`: **false**: 是否启用代码搜索（启用后会占用比较大的磁盘空间）。
+- `REPO_INDEXER_ENABLED`: **false**: 是否启用代码搜索（启用后会占用比较大的磁盘空间，如果是bleve可能需要占用约6倍存储空间）。
+- `REPO_INDEXER_TYPE`: **bleve**: 代码搜索引擎类型，可以为 `bleve` 或者 `elasticsearch`。
 - `REPO_INDEXER_PATH`: **indexers/repos.bleve**: 用于代码搜索的索引文件路径。
+- `REPO_INDEXER_CONN_STR`: ****: 代码搜索引擎连接字符串，当 `REPO_INDEXER_TYPE` 为 `elasticsearch` 时有效。例如： http://elastic:changeme@localhost:9200
+- `REPO_INDEXER_NAME`: **gitea_codes**: 代码搜索引擎的名字，当 `REPO_INDEXER_TYPE` 为 `elasticsearch` 时有效。
+
 - `UPDATE_BUFFER_LEN`: **20**: 代码索引请求的缓冲区长度。
 - `MAX_FILE_SIZE`: **1048576**: 进行解析的源代码文件的最大长度，小于该值时才会索引。
 

modules/indexer/code/bleve.go

@@ -58,10 +58,10 @@ func addUnicodeNormalizeTokenFilter(m *mapping.IndexMappingImpl) error {
 	})
 }
 
-// openIndexer open the index at the specified path, checking for metadata
+// openBleveIndexer open the index at the specified path, checking for metadata
 // updates and bleve version updates.  If index needs to be created (or
 // re-created), returns (nil, nil)
-func openIndexer(path string, latestVersion int) (bleve.Index, error) {
+func openBleveIndexer(path string, latestVersion int) (bleve.Index, error) {
 	_, err := os.Stat(path)
 	if err != nil && os.IsNotExist(err) {
 		return nil, nil
@@ -104,54 +104,14 @@ func (d *RepoIndexerData) Type() string {
 	return repoIndexerDocType
 }
 
-func addUpdate(commitSha string, update fileUpdate, repo *models.Repository, batch rupture.FlushingBatch) error {
-	// Ignore vendored files in code search
-	if setting.Indexer.ExcludeVendored && enry.IsVendor(update.Filename) {
-		return nil
-	}
-	stdout, err := git.NewCommand("cat-file", "-s", update.BlobSha).
-		RunInDir(repo.RepoPath())
-	if err != nil {
-		return err
-	}
-	if size, err := strconv.Atoi(strings.TrimSpace(stdout)); err != nil {
-		return fmt.Errorf("Misformatted git cat-file output: %v", err)
-	} else if int64(size) > setting.Indexer.MaxIndexerFileSize {
-		return addDelete(update.Filename, repo, batch)
-	}
-
-	fileContents, err := git.NewCommand("cat-file", "blob", update.BlobSha).
-		RunInDirBytes(repo.RepoPath())
-	if err != nil {
-		return err
-	} else if !base.IsTextFile(fileContents) {
-		// FIXME: UTF-16 files will probably fail here
-		return nil
-	}
-
-	id := filenameIndexerID(repo.ID, update.Filename)
-	return batch.Index(id, &RepoIndexerData{
-		RepoID:    repo.ID,
-		CommitID:  commitSha,
-		Content:   string(charset.ToUTF8DropErrors(fileContents)),
-		Language:  analyze.GetCodeLanguage(update.Filename, fileContents),
-		UpdatedAt: time.Now().UTC(),
-	})
-}
-
-func addDelete(filename string, repo *models.Repository, batch rupture.FlushingBatch) error {
-	id := filenameIndexerID(repo.ID, filename)
-	return batch.Delete(id)
-}
-
 const (
 	repoIndexerAnalyzer      = "repoIndexerAnalyzer"
 	repoIndexerDocType       = "repoIndexerDocType"
 	repoIndexerLatestVersion = 5
 )
 
-// createRepoIndexer create a repo indexer if one does not already exist
-func createRepoIndexer(path string, latestVersion int) (bleve.Index, error) {
+// createBleveIndexer create a bleve repo indexer if one does not already exist
+func createBleveIndexer(path string, latestVersion int) (bleve.Index, error) {
 	docMapping := bleve.NewDocumentMapping()
 	numericFieldMapping := bleve.NewNumericFieldMapping()
 	numericFieldMapping.IncludeInAll = false
@@ -199,18 +159,6 @@ func createRepoIndexer(path string, latestVersion int) (bleve.Index, error) {
 	return indexer, nil
 }
 
-func filenameIndexerID(repoID int64, filename string) string {
-	return indexerID(repoID) + "_" + filename
-}
-
-func filenameOfIndexerID(indexerID string) string {
-	index := strings.IndexByte(indexerID, '_')
-	if index == -1 {
-		log.Error("Unexpected ID in repo indexer: %s", indexerID)
-	}
-	return indexerID[index+1:]
-}
-
 var (
 	_ Indexer = &BleveIndexer{}
 )
@@ -230,18 +178,59 @@ func NewBleveIndexer(indexDir string) (*BleveIndexer, bool, error) {
 	return indexer, created, err
 }
 
+func (b *BleveIndexer) addUpdate(commitSha string, update fileUpdate, repo *models.Repository, batch rupture.FlushingBatch) error {
+	// Ignore vendored files in code search
+	if setting.Indexer.ExcludeVendored && enry.IsVendor(update.Filename) {
+		return nil
+	}
+
+	stdout, err := git.NewCommand("cat-file", "-s", update.BlobSha).
+		RunInDir(repo.RepoPath())
+	if err != nil {
+		return err
+	}
+	if size, err := strconv.Atoi(strings.TrimSpace(stdout)); err != nil {
+		return fmt.Errorf("Misformatted git cat-file output: %v", err)
+	} else if int64(size) > setting.Indexer.MaxIndexerFileSize {
+		return b.addDelete(update.Filename, repo, batch)
+	}
+
+	fileContents, err := git.NewCommand("cat-file", "blob", update.BlobSha).
+		RunInDirBytes(repo.RepoPath())
+	if err != nil {
+		return err
+	} else if !base.IsTextFile(fileContents) {
+		// FIXME: UTF-16 files will probably fail here
+		return nil
+	}
+
+	id := filenameIndexerID(repo.ID, update.Filename)
+	return batch.Index(id, &RepoIndexerData{
+		RepoID:    repo.ID,
+		CommitID:  commitSha,
+		Content:   string(charset.ToUTF8DropErrors(fileContents)),
+		Language:  analyze.GetCodeLanguage(update.Filename, fileContents),
+		UpdatedAt: time.Now().UTC(),
+	})
+}
+
+func (b *BleveIndexer) addDelete(filename string, repo *models.Repository, batch rupture.FlushingBatch) error {
+	id := filenameIndexerID(repo.ID, filename)
+	return batch.Delete(id)
+}
+
 // init init the indexer
 func (b *BleveIndexer) init() (bool, error) {
 	var err error
-	b.indexer, err = openIndexer(b.indexDir, repoIndexerLatestVersion)
+	b.indexer, err = openBleveIndexer(b.indexDir, repoIndexerLatestVersion)
 	if err != nil {
 		return false, err
 	}
 	if b.indexer != nil {
 		return false, nil
 	}
 
-	b.indexer, err = createRepoIndexer(b.indexDir, repoIndexerLatestVersion)
+	b.indexer, err = createBleveIndexer(b.indexDir, repoIndexerLatestVersion)
 	if err != nil {
 		return false, err
 	}
@@ -262,38 +251,19 @@ func (b *BleveIndexer) Close() {
 }
 
 // Index indexes the data
-func (b *BleveIndexer) Index(repoID int64) error {
-	repo, err := models.GetRepositoryByID(repoID)
-	if err != nil {
-		return err
-	}
-
-	sha, err := getDefaultBranchSha(repo)
-	if err != nil {
-		return err
-	}
-	changes, err := getRepoChanges(repo, sha)
-	if err != nil {
-		return err
-	} else if changes == nil {
-		return nil
-	}
-
+func (b *BleveIndexer) Index(repo *models.Repository, sha string, changes *repoChanges) error {
 	batch := rupture.NewFlushingBatch(b.indexer, maxBatchSize)
 	for _, update := range changes.Updates {
-		if err := addUpdate(sha, update, repo, batch); err != nil {
+		if err := b.addUpdate(sha, update, repo, batch); err != nil {
 			return err
 		}
 	}
 	for _, filename := range changes.RemovedFilenames {
-		if err := addDelete(filename, repo, batch); err != nil {
+		if err := b.addDelete(filename, repo, batch); err != nil {
 			return err
 		}
 	}
-	if err = batch.Flush(); err != nil {
-		return err
-	}
-	return repo.UpdateIndexerStatus(models.RepoIndexerTypeCode, sha)
+	return batch.Flush()
 }
 
 // Delete deletes indexes by ids

modules/indexer/code/bleve_test.go

@@ -6,21 +6,15 @@ package code
 
 import (
 	"io/ioutil"
-	"path/filepath"
 	"testing"
 
 	"code.gitea.io/gitea/models"
-	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/util"
 
 	"github.com/stretchr/testify/assert"
 )
 
-func TestMain(m *testing.M) {
-	models.MainTest(m, filepath.Join("..", "..", ".."))
-}
-
-func TestIndexAndSearch(t *testing.T) {
+func TestBleveIndexAndSearch(t *testing.T) {
 	models.PrepareTestEnv(t)
 
 	dir, err := ioutil.TempDir("", "bleve.index")
@@ -31,56 +25,15 @@ func TestIndexAndSearch(t *testing.T) {
 	}
 	defer util.RemoveAll(dir)
 
-	setting.Indexer.RepoIndexerEnabled = true
 	idx, _, err := NewBleveIndexer(dir)
 	if err != nil {
-		assert.Fail(t, "Unable to create indexer Error: %v", err)
+		assert.Fail(t, "Unable to create bleve indexer Error: %v", err)
 		if idx != nil {
 			idx.Close()
 		}
 		return
 	}
 	defer idx.Close()
 
-	err = idx.Index(1)
-	assert.NoError(t, err)
-
-	var (
-		keywords = []struct {
-			Keyword string
-			IDs     []int64
-			Langs   int
-		}{
-			{
-				Keyword: "Description",
-				IDs:     []int64{1},
-				Langs:   1,
-			},
-			{
-				Keyword: "repo1",
-				IDs:     []int64{1},
-				Langs:   1,
-			},
-			{
-				Keyword: "non-exist",
-				IDs:     []int64{},
-				Langs:   0,
-			},
-		}
-	)
-
-	for _, kw := range keywords {
-		total, res, langs, err := idx.Search(nil, "", kw.Keyword, 1, 10)
-		assert.NoError(t, err)
-		assert.EqualValues(t, len(kw.IDs), total)
-
-		assert.NotNil(t, langs)
-		assert.Len(t, langs, kw.Langs)
-
-		var ids = make([]int64, 0, len(res))
-		for _, hit := range res {
-			ids = append(ids, hit.RepoID)
-		}
-		assert.EqualValues(t, kw.IDs, ids)
-	}
+	testIndexer("beleve", t, idx)
 }