refactor: 重构文档处理流程和任务管理

This commit is contained in:
2026-04-09 09:11:43 +08:00
parent b6896f3fb4
commit 7f894745e9
34 changed files with 1216 additions and 1056 deletions

View File

@@ -3,15 +3,11 @@ package service
import (
"context"
"rag/common/eino"
"rag/consts/document"
"rag/consts/public"
"rag/common/task"
"rag/dao"
"rag/model/dto"
"rag/model/entity"
gmq "github.com/bjang03/gmq/core/gmq"
"github.com/bjang03/gmq/mq"
"github.com/bjang03/gmq/types"
"github.com/cloudwego/eino/components/indexer"
"github.com/cloudwego/eino/schema"
"github.com/gogf/gf/v2/frame/g"
@@ -22,10 +18,6 @@ var DocumentChunk = new(documentChunkService)
type documentChunkService struct{}
const (
DatasetIndexStatusReady = "ready"
)
// Update 更新文件块
func (s *documentChunkService) Update(ctx context.Context, req *dto.UpdateDocumentChunkReq) (err error) {
_, err = dao.DocumentChunk.Update(ctx, req)
@@ -60,32 +52,29 @@ func (s *documentChunkService) DocsChunkMsg(ctx context.Context, msg any) (err e
idx := eino.NewPGVectorIndexer(&eino.PGVectorIndexerOptions{
BatchSize: 10,
})
documentId := gconv.Int64(docs[0].MetaData[entity.DocumentChunkCol.DocumentId])
rows, err := idx.Store(ctx, docs, indexer.WithEmbedding(eino.EmbedderDashscope))
if err != nil || rows == 0 {
g.Log().Error(ctx, "DocsChunkMsg rows: , err:", rows, err)
// 写入任务进度失败 任务类型为sql存储
remark := " 向量存储数量: " + gconv.String(rows)
if err != nil {
remark = "向量存储失败: " + err.Error()
}
err = Task.WriteTaskProgress(ctx, &dto.WriteTaskProgressReq{
TaskId: documentId,
TaskType: task.TaskTypeGenerateVector,
Status: task.TaskStatusFailed,
Remark: remark,
})
return
}
tenantId := gconv.Uint64(docs[0].MetaData[entity.DocumentChunkCol.TenantId])
creator := gconv.String(docs[0].MetaData[entity.DocumentChunkCol.Creator])
documentId := gconv.Int64(docs[0].MetaData[entity.DocumentChunkCol.DocumentId])
err = s.publishKnowledgeDocumentMsg(ctx, tenantId, creator, documentId, document.VectorStatusCompleted.Code())
return
}
// publishKnowledgeDocumentMsg 发布消息
func (s *documentChunkService) publishKnowledgeDocumentMsg(ctx context.Context, tenantId uint64, creator string, documentId int64, vectorStatus document.VectorStatus) (err error) {
knowledgeDocumentMsg := dto.KnowledgeDocumentMsg{
TenantId: tenantId,
Creator: creator,
Id: documentId,
VectorStatus: vectorStatus,
}
err = gmq.GetGmq("primary").GmqPublish(ctx, &mq.RedisPubMessage{
PubMessage: types.PubMessage{
Topic: public.KnowledgeDocumentVectorStatusTopic,
Data: knowledgeDocumentMsg,
},
// 写入任务进度成功 任务类型为sql存储
err = Task.WriteTaskProgress(ctx, &dto.WriteTaskProgressReq{
TaskId: documentId,
TaskType: task.TaskTypeGenerateVector,
Status: task.TaskStatusCompleted,
Remark: "向量生成完成",
})
return
}