fix: avoid stale knowledge records on upload failure

This commit is contained in:
wolfkill
2026-05-29 11:26:25 +08:00
committed by lyingbug
parent b603b1dcfa
commit 2a54bb2d45
2 changed files with 244 additions and 13 deletions
@@ -179,9 +179,10 @@ func (s *knowledgeService) CreateKnowledgeFromFile(ctx context.Context,
return nil, werrors.NewValidationError("文件名包含非法字符")
}
// Create knowledge record
logger.Info(ctx, "Creating knowledge record")
// Prepare knowledge record
logger.Info(ctx, "Preparing knowledge record")
knowledge := &types.Knowledge{
ID: uuid.New().String(),
TenantID: tenantID,
KnowledgeBaseID: kbID,
TagID: tagID, // 设置分类ID,用于知识分类管理
@@ -199,25 +200,24 @@ func (s *knowledgeService) CreateKnowledgeFromFile(ctx context.Context,
EmbeddingModelID: kb.EmbeddingModelID,
Metadata: metadataJSON,
}
// Save knowledge record to database
logger.Info(ctx, "Saving knowledge record to database")
if err := s.repo.CreateKnowledge(ctx, knowledge); err != nil {
logger.Errorf(ctx, "Failed to create knowledge record, ID: %s, error: %v", knowledge.ID, err)
return nil, err
}
// Save the file to storage (use KB-level storage engine if configured)
logger.Infof(ctx, "Saving file, knowledge ID: %s", knowledge.ID)
filePath, err := s.resolveFileService(ctx, kb).SaveFile(ctx, file, knowledge.TenantID, knowledge.ID)
fileSvc := s.resolveFileService(ctx, kb)
filePath, err := fileSvc.SaveFile(ctx, file, knowledge.TenantID, knowledge.ID)
if err != nil {
logger.Errorf(ctx, "Failed to save file, knowledge ID: %s, error: %v", knowledge.ID, err)
return nil, err
}
knowledge.FilePath = filePath
// Update knowledge record with file path
logger.Info(ctx, "Updating knowledge record with file path")
if err := s.repo.UpdateKnowledge(ctx, knowledge); err != nil {
logger.Errorf(ctx, "Failed to update knowledge with file path, ID: %s, error: %v", knowledge.ID, err)
// Save knowledge record to database after the file is safely stored.
logger.Info(ctx, "Saving knowledge record to database")
if err := s.repo.CreateKnowledge(ctx, knowledge); err != nil {
logger.Errorf(ctx, "Failed to create knowledge record, ID: %s, error: %v", knowledge.ID, err)
if deleteErr := fileSvc.DeleteFile(ctx, filePath); deleteErr != nil {
logger.Errorf(ctx, "Failed to delete saved file after knowledge creation failed, path: %s, error: %v", filePath, deleteErr)
}
return nil, err
}
@@ -0,0 +1,231 @@
package service
import (
"bytes"
"context"
"errors"
"io"
"mime/multipart"
"net/http/httptest"
"testing"
"github.com/Tencent/WeKnora/internal/types"
"github.com/Tencent/WeKnora/internal/types/interfaces"
"github.com/hibiken/asynq"
"github.com/stretchr/testify/require"
)
type createKnowledgeFileRepoStub struct {
interfaces.KnowledgeRepository
createCalls int
createErr error
createdKnowledge *types.Knowledge
}
func (r *createKnowledgeFileRepoStub) CheckKnowledgeExists(
ctx context.Context,
tenantID uint64,
kbID string,
params *types.KnowledgeCheckParams,
) (bool, *types.Knowledge, error) {
return false, nil, nil
}
func (r *createKnowledgeFileRepoStub) CreateKnowledge(ctx context.Context, knowledge *types.Knowledge) error {
r.createCalls++
copied := *knowledge
r.createdKnowledge = &copied
return r.createErr
}
type createKnowledgeFileKBServiceStub struct {
interfaces.KnowledgeBaseService
kb *types.KnowledgeBase
}
func (s *createKnowledgeFileKBServiceStub) GetKnowledgeBaseByID(
ctx context.Context,
id string,
) (*types.KnowledgeBase, error) {
return s.kb, nil
}
type createKnowledgeFileServiceStub struct {
saveErr error
saveCalls int
savedWithKnowledgeID string
deleteCalls int
deletedPath string
}
func (s *createKnowledgeFileServiceStub) CheckConnectivity(ctx context.Context) error {
return nil
}
func (s *createKnowledgeFileServiceStub) SaveFile(
ctx context.Context,
file *multipart.FileHeader,
tenantID uint64,
knowledgeID string,
) (string, error) {
s.saveCalls++
s.savedWithKnowledgeID = knowledgeID
if s.saveErr != nil {
return "", s.saveErr
}
return "stored/" + knowledgeID, nil
}
func (s *createKnowledgeFileServiceStub) SaveBytes(
ctx context.Context,
data []byte,
tenantID uint64,
fileName string,
temp bool,
) (string, error) {
return "", errors.New("not implemented")
}
func (s *createKnowledgeFileServiceStub) GetFile(ctx context.Context, filePath string) (io.ReadCloser, error) {
return nil, errors.New("not implemented")
}
func (s *createKnowledgeFileServiceStub) GetFileURL(ctx context.Context, filePath string) (string, error) {
return "", errors.New("not implemented")
}
func (s *createKnowledgeFileServiceStub) DeleteFile(ctx context.Context, filePath string) error {
s.deleteCalls++
s.deletedPath = filePath
return nil
}
type createKnowledgeTaskEnqueuerStub struct {
calls int
}
func (s *createKnowledgeTaskEnqueuerStub) Enqueue(
task *asynq.Task,
opts ...asynq.Option,
) (*asynq.TaskInfo, error) {
s.calls++
return &asynq.TaskInfo{ID: "task-1", Queue: "default"}, nil
}
func TestCreateKnowledgeFromFileDoesNotPersistWhenStorageSaveFails(t *testing.T) {
t.Parallel()
repo := &createKnowledgeFileRepoStub{}
fileSvc := &createKnowledgeFileServiceStub{saveErr: errors.New("storage unavailable")}
svc := &knowledgeService{
repo: repo,
kbService: &createKnowledgeFileKBServiceStub{kb: &types.KnowledgeBase{ID: "kb-1"}},
fileSvc: fileSvc,
}
knowledge, err := svc.CreateKnowledgeFromFile(
newCreateKnowledgeFileContext(),
"kb-1",
newMultipartFileHeader(t, "doc.txt", "hello"),
nil,
nil,
"",
"",
"",
)
require.Error(t, err)
require.Nil(t, knowledge)
require.Equal(t, 1, fileSvc.saveCalls)
require.Zero(t, repo.createCalls)
}
func TestCreateKnowledgeFromFilePersistsStoredFilePathOnCreate(t *testing.T) {
t.Parallel()
repo := &createKnowledgeFileRepoStub{}
fileSvc := &createKnowledgeFileServiceStub{}
task := &createKnowledgeTaskEnqueuerStub{}
svc := &knowledgeService{
repo: repo,
kbService: &createKnowledgeFileKBServiceStub{kb: &types.KnowledgeBase{ID: "kb-1"}},
fileSvc: fileSvc,
task: task,
}
knowledge, err := svc.CreateKnowledgeFromFile(
newCreateKnowledgeFileContext(),
"kb-1",
newMultipartFileHeader(t, "doc.txt", "hello"),
nil,
nil,
"",
"",
"",
)
require.NoError(t, err)
require.NotNil(t, knowledge)
require.Equal(t, 1, fileSvc.saveCalls)
require.NotEmpty(t, fileSvc.savedWithKnowledgeID)
require.Equal(t, fileSvc.savedWithKnowledgeID, knowledge.ID)
require.Equal(t, 1, repo.createCalls)
require.NotNil(t, repo.createdKnowledge)
require.Equal(t, "stored/"+knowledge.ID, repo.createdKnowledge.FilePath)
require.Equal(t, 1, task.calls)
}
func TestCreateKnowledgeFromFileDeletesStoredFileWhenCreateFails(t *testing.T) {
t.Parallel()
repo := &createKnowledgeFileRepoStub{createErr: errors.New("database unavailable")}
fileSvc := &createKnowledgeFileServiceStub{}
svc := &knowledgeService{
repo: repo,
kbService: &createKnowledgeFileKBServiceStub{kb: &types.KnowledgeBase{ID: "kb-1"}},
fileSvc: fileSvc,
}
knowledge, err := svc.CreateKnowledgeFromFile(
newCreateKnowledgeFileContext(),
"kb-1",
newMultipartFileHeader(t, "doc.txt", "hello"),
nil,
nil,
"",
"",
"",
)
require.EqualError(t, err, "database unavailable")
require.Nil(t, knowledge)
require.Equal(t, 1, fileSvc.saveCalls)
require.Equal(t, 1, repo.createCalls)
require.Equal(t, 1, fileSvc.deleteCalls)
require.Equal(t, "stored/"+fileSvc.savedWithKnowledgeID, fileSvc.deletedPath)
}
func newCreateKnowledgeFileContext() context.Context {
ctx := context.WithValue(context.Background(), types.TenantIDContextKey, uint64(1))
ctx = context.WithValue(ctx, types.TenantInfoContextKey, &types.Tenant{})
return ctx
}
func newMultipartFileHeader(t *testing.T, filename string, content string) *multipart.FileHeader {
t.Helper()
var body bytes.Buffer
writer := multipart.NewWriter(&body)
part, err := writer.CreateFormFile("file", filename)
require.NoError(t, err)
_, err = part.Write([]byte(content))
require.NoError(t, err)
require.NoError(t, writer.Close())
req := httptest.NewRequest("POST", "/", &body)
req.Header.Set("Content-Type", writer.FormDataContentType())
require.NoError(t, req.ParseMultipartForm(1024))
return req.MultipartForm.File["file"][0]
}