Nulo/forgejo
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Move Bleve and Elastic code indexers to use a common cat-file --batch (#14781)

Browse source 
* Extract out the common cat-file batch calls

Signed-off-by: Andrew Thornton <art27@cantab.net>

* Move bleve and elastic indexers to use a common cat-file --batch when indexing

Signed-off-by: Andrew Thornton <art27@cantab.net>

* move catfilebatch to batch_reader and rename to batch_reader.go

Signed-off-by: Andrew Thornton <art27@cantab.net>

Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: Lauris BH <lauris@nix.lv>
This commit is contained in:
zeripath 2021-03-04 02:57:01 +00:00 committed by GitHub
parent 0044e80491
commit 523efa433b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 91 additions and 87 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

35
modules/git/batch_reader_nogogit.go → modules/git/batch_reader.go
View file

@ -2,17 +2,48 @@
// Use of this source code is governed by a MIT-style
// license that can be found in the LICENSE file.
// +build !gogit
package git
import (
"bufio"
"bytes"
"io"
"math"
"strconv"
"strings"
)
// CatFileBatch opens git cat-file --batch in the provided repo and returns a stdin pipe, a stdout reader and cancel function
func CatFileBatch(repoPath string) (*io.PipeWriter, *bufio.Reader, func()) {
// Next feed the commits in order into cat-file --batch, followed by their trees and sub trees as necessary.
// so let's create a batch stdin and stdout
batchStdinReader, batchStdinWriter := io.Pipe()
batchStdoutReader, batchStdoutWriter := io.Pipe()
cancel := func() {
_ = batchStdinReader.Close()
_ = batchStdinWriter.Close()
_ = batchStdoutReader.Close()
_ = batchStdoutWriter.Close()
}
go func() {
stderr := strings.Builder{}
err := NewCommand("cat-file", "--batch").RunInDirFullPipeline(repoPath, batchStdoutWriter, &stderr, batchStdinReader)
if err != nil {
_ = batchStdoutWriter.CloseWithError(ConcatenateError(err, (&stderr).String()))
_ = batchStdinReader.CloseWithError(ConcatenateError(err, (&stderr).String()))
} else {
_ = batchStdoutWriter.Close()
_ = batchStdinReader.Close()
}
}()
// For simplicities sake we'll us a buffered reader to read from the cat-file --batch
batchReader := bufio.NewReader(batchStdoutReader)
return batchStdinWriter, batchReader, cancel
}
// ReadBatchLine reads the header line from cat-file --batch
// We expect:
// <sha> SP <type> SP <size> LF

25
modules/git/commit_info_nogogit.go
View file

@ -141,29 +141,8 @@ func GetLastCommitForPaths(commit *Commit, treePath string, paths []string) ([]*
}
}()
// We feed the commits in order into cat-file --batch, followed by their trees and sub trees as necessary.
// so let's create a batch stdin and stdout
batchStdinReader, batchStdinWriter := io.Pipe()
batchStdoutReader, batchStdoutWriter := io.Pipe()
defer func() {
_ = batchStdinReader.Close()
_ = batchStdinWriter.Close()
_ = batchStdoutReader.Close()
_ = batchStdoutWriter.Close()
}()
go func() {
stderr := strings.Builder{}
err := NewCommand("cat-file", "--batch").RunInDirFullPipeline(commit.repo.Path, batchStdoutWriter, &stderr, batchStdinReader)
if err != nil {
_ = revListWriter.CloseWithError(ConcatenateError(err, (&stderr).String()))
} else {
_ = revListWriter.Close()
}
}()
// For simplicities sake we'll us a buffered reader
batchReader := bufio.NewReader(batchStdoutReader)
batchStdinWriter, batchReader, cancel := CatFileBatch(commit.repo.Path)
defer cancel()
mapsize := 4096
if len(paths) > mapsize {

23
modules/git/pipeline/lfs_nogogit.go
View file

@ -64,27 +64,8 @@ func FindLFSFile(repo *git.Repository, hash git.SHA1) ([]*LFSResult, error) {
// Next feed the commits in order into cat-file --batch, followed by their trees and sub trees as necessary.
// so let's create a batch stdin and stdout
batchStdinReader, batchStdinWriter := io.Pipe()
batchStdoutReader, batchStdoutWriter := io.Pipe()
defer func() {
_ = batchStdinReader.Close()
_ = batchStdinWriter.Close()
_ = batchStdoutReader.Close()
_ = batchStdoutWriter.Close()
}()
go func() {
stderr := strings.Builder{}
err := git.NewCommand("cat-file", "--batch").RunInDirFullPipeline(repo.Path, batchStdoutWriter, &stderr, batchStdinReader)
if err != nil {
_ = revListWriter.CloseWithError(git.ConcatenateError(err, (&stderr).String()))
} else {
_ = revListWriter.Close()
}
}()
// For simplicities sake we'll us a buffered reader to read from the cat-file --batch
batchReader := bufio.NewReader(batchStdoutReader)
batchStdinWriter, batchReader, cancel := git.CatFileBatch(repo.Path)
defer cancel()
// We'll use a scanner for the revList because it's simpler than a bufio.Reader
scan := bufio.NewScanner(revListReader)

27
modules/git/repo_language_stats_nogogit.go
View file

@ -11,7 +11,6 @@ import (
"bytes"
"io"
"math"
"strings"
"code.gitea.io/gitea/modules/analyze"
@ -22,30 +21,8 @@ import (
func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, error) {
// We will feed the commit IDs in order into cat-file --batch, followed by blobs as necessary.
// so let's create a batch stdin and stdout
batchStdinReader, batchStdinWriter := io.Pipe()
batchStdoutReader, batchStdoutWriter := io.Pipe()
defer func() {
_ = batchStdinReader.Close()
_ = batchStdinWriter.Close()
_ = batchStdoutReader.Close()
_ = batchStdoutWriter.Close()
}()
go func() {
stderr := strings.Builder{}
err := NewCommand("cat-file", "--batch").RunInDirFullPipeline(repo.Path, batchStdoutWriter, &stderr, batchStdinReader)
if err != nil {
_ = batchStdoutWriter.CloseWithError(ConcatenateError(err, (&stderr).String()))
_ = batchStdinReader.CloseWithError(ConcatenateError(err, (&stderr).String()))
} else {
_ = batchStdoutWriter.Close()
_ = batchStdinReader.Close()
}
}()
// For simplicities sake we'll us a buffered reader
batchReader := bufio.NewReader(batchStdoutReader)
batchStdinWriter, batchReader, cancel := CatFileBatch(repo.Path)
defer cancel()
writeID := func(id string) error {
_, err := batchStdinWriter.Write([]byte(id))

26
modules/indexer/code/bleve.go
View file

@ -5,7 +5,10 @@
package code
import (
"bufio"
"fmt"
"io"
"io/ioutil"
"os"
"strconv"
"strings"
@ -173,7 +176,7 @@ 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 {
func (b *BleveIndexer) addUpdate(batchWriter *io.PipeWriter, batchReader *bufio.Reader, 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
@ -196,8 +199,16 @@ func (b *BleveIndexer) addUpdate(commitSha string, update fileUpdate, repo *mode
return b.addDelete(update.Filename, repo, batch)
}
fileContents, err := git.NewCommand("cat-file", "blob", update.BlobSha).
RunInDirBytes(repo.RepoPath())
if _, err := batchWriter.Write([]byte(update.BlobSha + "\n")); err != nil {
return err
}
_, _, size, err := git.ReadBatchLine(batchReader)
if err != nil {
return err
}
fileContents, err := ioutil.ReadAll(io.LimitReader(batchReader, size))
if err != nil {
return err
} else if !base.IsTextFile(fileContents) {
@ -254,11 +265,18 @@ func (b *BleveIndexer) Close() {
// Index indexes the data
func (b *BleveIndexer) Index(repo *models.Repository, sha string, changes *repoChanges) error {
batch := rupture.NewFlushingBatch(b.indexer, maxBatchSize)
if len(changes.Updates) > 0 {
batchWriter, batchReader, cancel := git.CatFileBatch(repo.RepoPath())
defer cancel()
for _, update := range changes.Updates {
if err := b.addUpdate(sha, update, repo, batch); err != nil {
if err := b.addUpdate(batchWriter, batchReader, sha, update, repo, batch); err != nil {
return err
}
}
cancel()
}
for _, filename := range changes.RemovedFilenames {
if err := b.addDelete(filename, repo, batch); err != nil {
return err

26
modules/indexer/code/elastic_search.go
View file

@ -5,8 +5,11 @@
package code
import (
"bufio"
"context"
"fmt"
"io"
"io/ioutil"
"strconv"
"strings"
"time"
@ -172,7 +175,7 @@ func (b *ElasticSearchIndexer) init() (bool, error) {
return exists, nil
}
func (b *ElasticSearchIndexer) addUpdate(sha string, update fileUpdate, repo *models.Repository) ([]elastic.BulkableRequest, error) {
func (b *ElasticSearchIndexer) addUpdate(batchWriter *io.PipeWriter, batchReader *bufio.Reader, sha string, update fileUpdate, repo *models.Repository) ([]elastic.BulkableRequest, error) {
// Ignore vendored files in code search
if setting.Indexer.ExcludeVendored && enry.IsVendor(update.Filename) {
return nil, nil
@ -195,8 +198,16 @@ func (b *ElasticSearchIndexer) addUpdate(sha string, update fileUpdate, repo *mo
return []elastic.BulkableRequest{b.addDelete(update.Filename, repo)}, nil
}
fileContents, err := git.NewCommand("cat-file", "blob", update.BlobSha).
RunInDirBytes(repo.RepoPath())
if _, err := batchWriter.Write([]byte(update.BlobSha + "\n")); err != nil {
return nil, err
}
_, _, size, err := git.ReadBatchLine(batchReader)
if err != nil {
return nil, err
}
fileContents, err := ioutil.ReadAll(io.LimitReader(batchReader, size))
if err != nil {
return nil, err
} else if !base.IsTextFile(fileContents) {
@ -230,8 +241,13 @@ func (b *ElasticSearchIndexer) addDelete(filename string, repo *models.Repositor
// Index will save the index data
func (b *ElasticSearchIndexer) Index(repo *models.Repository, sha string, changes *repoChanges) error {
reqs := make([]elastic.BulkableRequest, 0)
if len(changes.Updates) > 0 {
batchWriter, batchReader, cancel := git.CatFileBatch(repo.RepoPath())
defer cancel()
for _, update := range changes.Updates {
updateReqs, err := b.addUpdate(sha, update, repo)
updateReqs, err := b.addUpdate(batchWriter, batchReader, sha, update, repo)
if err != nil {
return err
}
@ -239,6 +255,8 @@ func (b *ElasticSearchIndexer) Index(repo *models.Repository, sha string, change
reqs = append(reqs, updateReqs...)
}
}
cancel()
}
for _, filename := range changes.RemovedFilenames {
reqs = append(reqs, b.addDelete(filename, repo))