Version: 0.9.14.dev.260410

后端：
  1. LLM 客户端从 newAgent/llm 提升为 infra/llm 基础设施层
     - 删除 backend/newAgent/llm/（ark.go / ark_adapter.go / client.go / json.go）
     - 等价迁移至 backend/infra/llm/，所有 newAgent node 与 service 统一改引用 infrallm
     - 消除 newAgent 对模型客户端的私有依赖，为 memory / websearch 等多模块复用铺路
  2. RAG 基础设施完成可运行态接入（factory / runtime / observer / service 四层成型）
     - 新建 backend/infra/rag/factory.go / runtime.go / observe.go / observer.go /
  service.go：工厂创建、运行时生命周期、轻量观测接口、检索服务门面
     - 更新 infra/rag/config/config.go：补齐 Milvus / Embed / Reranker 全部配置项与默认值
     - 更新 infra/rag/embed/eino_embedder.go：增强 Eino embedding 适配，支持 BaseURL / APIKey 环境变量 / 超时 /
  维度等参数
     - 更新 infra/rag/store/milvus_store.go：完整实现 Milvus 向量存储（建集合 / 建 Index / Upsert / Search /
  Delete），支持 COSINE / L2 / IP 度量
     - 更新 infra/rag/core/pipeline.go：适配 Runtime 接口，Pipeline 由 factory 注入而非手动拼装
     - 更新 infra/rag/corpus/memory_corpus.go / vector_store.go：对接 Memory 模块数据源与 Store 接口扩展
  3. Memory 模块从 Day1 骨架升级为 Day2 完整可运行态
     - 新建 memory/module.go：统一门面 Module，对外封装 EnqueueExtract / ReadService / ManageService / WithTx /
  StartWorker，启动层只依赖这一个入口
     - 新建 memory/orchestrator/llm_write_orchestrator.go：LLM 驱动的记忆抽取编排器，替代原 mock 抽取
     - 新建 memory/service/read_service.go：按用户开关过滤 + 轻量重排 + 访问时间刷新的读取链路
     - 新建 memory/service/manage_service.go：记忆管理面能力（列出 / 软删除 / 开关读写），删除同步写审计日志
     - 新建 memory/service/common.go：服务层公共工具
     - 新建 memory/worker/loop.go：后台轮询循环 RunPollingLoop，定时抢占 pending 任务并推进
     - 新建 memory/utils/audit.go / settings.go：审计日志构造、用户设置过滤等纯函数
     - 更新 memory/model/item.go / job.go / settings.go / config.go / status.go：补齐 DTO 字段与状态常量
     - 更新 memory/repo/item_repo.go / job_repo.go / audit_repo.go / settings_repo.go：补齐 CRUD 与查询能力
     - 更新 memory/worker/runner.go：Runner 对接 Module 与 LLM 抽取器，任务状态机完整化
     - 更新 memory/README.md：同步模块现状说明
  4. newAgent 接入 Memory 读取注入与工具注册依赖预埋
     - 新建 service/agentsvc/agent_memory.go：定义 MemoryReader 接口 + injectMemoryContext，在 graph
  执行前统一补充记忆上下文
     - 更新 service/agentsvc/agent.go：新增 memoryReader 字段与 SetMemoryReader 方法
     - 更新 service/agentsvc/agent_newagent.go：调用 injectMemoryContext 注入 pinned block，检索失败仅降级不阻断主链路
     - 更新 newAgent/tools/registry.go：新增 DefaultRegistryDeps（含 RAGRuntime），工具注册表支持依赖注入
  5. 启动流程与事件处理器接线更新
     - 更新 cmd/start.go：初始化 RAG Runtime → Memory Module → 注册事件处理器 → 启动 Worker 后台轮询
     - 更新 service/events/memory_extract_requested.go：改用 memory.Module.WithTx(tx) 统一门面，事件处理器不再直接依赖
  repo/service 内部包
  6. 缓存插件与配置同步
     - 更新 middleware/cache_deleter.go：静默忽略 MemoryJob / MemoryItem / MemoryAuditLog / MemoryUserSetting
  等新模型，避免日志刷屏；清理冗余注释
     - 更新 config.example.yaml：补齐 rag / memory / websearch 配置段及默认值
     - 更新 go.mod / go.sum：新增 eino-ext/openai / json-patch / go-openai 依赖
  前端：无 仓库：无

backend/newAgent/node/chat.go

@@ -9,7 +9,7 @@ import (
 
 	"github.com/cloudwego/eino/schema"
 
-	newagentllm "github.com/LoveLosita/smartflow/backend/newAgent/llm"
+	infrallm "github.com/LoveLosita/smartflow/backend/infra/llm"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 	newagentprompt "github.com/LoveLosita/smartflow/backend/newAgent/prompt"
 	newagentstream "github.com/LoveLosita/smartflow/backend/newAgent/stream"
@@ -46,7 +46,7 @@ type ChatNodeInput struct {
 	ConversationContext *newagentmodel.ConversationContext
 	UserInput           string
 	ConfirmAction       string
-	Client              *newagentllm.Client
+	Client              *infrallm.Client
 	ChunkEmitter        *newagentstream.ChunkEmitter
 }
 
@@ -91,14 +91,14 @@ func RunChatNode(ctx context.Context, input ChatNodeInput) error {
 	}
 	messages := newagentprompt.BuildChatRoutingMessages(conversationContext, input.UserInput, flowState)
 
-	decision, rawResult, err := newagentllm.GenerateJSON[newagentmodel.ChatRoutingDecision](
+	decision, rawResult, err := infrallm.GenerateJSON[newagentmodel.ChatRoutingDecision](
 		ctx,
 		input.Client,
 		messages,
-		newagentllm.GenerateOptions{
+		infrallm.GenerateOptions{
 			Temperature: 0.1,
 			MaxTokens:   500,
-			Thinking:    newagentllm.ThinkingModeDisabled,
+			Thinking:    infrallm.ThinkingModeDisabled,
 			Metadata: map[string]any{
 				"stage": chatStageName,
 				"phase": "routing",
@@ -412,10 +412,10 @@ func handleDeepAnswer(
 
 	// 2. 第二次 LLM 调用：开 thinking，深度回答。
 	deepMessages := newagentprompt.BuildDeepAnswerMessages(conversationContext, input.UserInput)
-	deepResult, err := input.Client.GenerateText(ctx, deepMessages, newagentllm.GenerateOptions{
+	deepResult, err := input.Client.GenerateText(ctx, deepMessages, infrallm.GenerateOptions{
 		Temperature: 0.5,
 		MaxTokens:   2000,
-		Thinking:    newagentllm.ThinkingModeEnabled,
+		Thinking:    infrallm.ThinkingModeEnabled,
 		Metadata: map[string]any{
 			"stage": chatStageName,
 			"phase": "deep_answer",

backend/newAgent/node/deliver.go

@@ -8,7 +8,7 @@ import (
 
 	"github.com/cloudwego/eino/schema"
 
-	newagentllm "github.com/LoveLosita/smartflow/backend/newAgent/llm"
+	infrallm "github.com/LoveLosita/smartflow/backend/infra/llm"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 	newagentprompt "github.com/LoveLosita/smartflow/backend/newAgent/prompt"
 	newagentstream "github.com/LoveLosita/smartflow/backend/newAgent/stream"
@@ -30,7 +30,7 @@ const (
 type DeliverNodeInput struct {
 	RuntimeState        *newagentmodel.AgentRuntimeState
 	ConversationContext *newagentmodel.ConversationContext
-	Client              *newagentllm.Client
+	Client              *infrallm.Client
 	ChunkEmitter        *newagentstream.ChunkEmitter
 }
 
@@ -95,7 +95,7 @@ func RunDeliverNode(ctx context.Context, input DeliverNodeInput) error {
 // generateDeliverSummary 尝试调用 LLM 生成交付总结，失败时降级到机械格式化。
 func generateDeliverSummary(
 	ctx context.Context,
-	client *newagentllm.Client,
+	client *infrallm.Client,
 	flowState *newagentmodel.CommonState,
 	conversationContext *newagentmodel.ConversationContext,
 ) string {
@@ -116,10 +116,10 @@ func generateDeliverSummary(
 	result, err := client.GenerateText(
 		ctx,
 		messages,
-		newagentllm.GenerateOptions{
+		infrallm.GenerateOptions{
 			Temperature: 0.5,
 			MaxTokens:   800,
-			Thinking:    newagentllm.ThinkingModeDisabled,
+			Thinking:    infrallm.ThinkingModeDisabled,
 			Metadata: map[string]any{
 				"stage": deliverStageName,
 			},

backend/newAgent/node/execute.go

@@ -10,7 +10,7 @@ import (
 	"strings"
 	"time"
 
-	newagentllm "github.com/LoveLosita/smartflow/backend/newAgent/llm"
+	infrallm "github.com/LoveLosita/smartflow/backend/infra/llm"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 	newagentprompt "github.com/LoveLosita/smartflow/backend/newAgent/prompt"
 	newagentstream "github.com/LoveLosita/smartflow/backend/newAgent/stream"
@@ -48,7 +48,7 @@ type ExecuteNodeInput struct {
 	RuntimeState          *newagentmodel.AgentRuntimeState
 	ConversationContext   *newagentmodel.ConversationContext
 	UserInput             string
-	Client                *newagentllm.Client
+	Client                *infrallm.Client
 	ChunkEmitter          *newagentstream.ChunkEmitter
 	ResumeNode            string
 	ToolRegistry          *newagenttools.ToolRegistry
@@ -188,14 +188,14 @@ func RunExecuteNode(ctx context.Context, input ExecuteNodeInput) error {
 		flowState.ConversationID,
 		flowState.RoundUsed,
 	)
-	decision, rawResult, err := newagentllm.GenerateJSON[newagentmodel.ExecuteDecision](
+	decision, rawResult, err := infrallm.GenerateJSON[newagentmodel.ExecuteDecision](
 		ctx,
 		input.Client,
 		messages,
-		newagentllm.GenerateOptions{
+		infrallm.GenerateOptions{
 			Temperature: 1.0,   // thinking 模式强制要求 temperature=1
 			MaxTokens:   16000, // 需为 thinking chain 留出足够预算
-			Thinking:    newagentllm.ThinkingModeEnabled,
+			Thinking:    infrallm.ThinkingModeEnabled,
 			Metadata: map[string]any{
 				"stage":      executeStageName,
 				"step_index": flowState.CurrentStep,

backend/newAgent/node/plan.go

@@ -8,7 +8,7 @@ import (
 
 	"github.com/google/uuid"
 
-	newagentllm "github.com/LoveLosita/smartflow/backend/newAgent/llm"
+	infrallm "github.com/LoveLosita/smartflow/backend/infra/llm"
 	newagentmodel "github.com/LoveLosita/smartflow/backend/newAgent/model"
 	newagentprompt "github.com/LoveLosita/smartflow/backend/newAgent/prompt"
 	newagentstream "github.com/LoveLosita/smartflow/backend/newAgent/stream"
@@ -31,7 +31,7 @@ type PlanNodeInput struct {
 	RuntimeState        *newagentmodel.AgentRuntimeState
 	ConversationContext *newagentmodel.ConversationContext
 	UserInput           string
-	Client              *newagentllm.Client
+	Client              *infrallm.Client
 	ChunkEmitter        *newagentstream.ChunkEmitter
 	ResumeNode          string
 	AlwaysExecute       bool // true 时计划生成后自动确认，不进入 confirm 节点
@@ -70,14 +70,14 @@ func RunPlanNode(ctx context.Context, input PlanNodeInput) error {
 	messages := newagentprompt.BuildPlanMessages(flowState, conversationContext, input.UserInput)
 
 	// 3. Phase 1：快速评估（开 thinking），让 LLM 同时产出复杂度评估和规划结果。
-	decision, rawResult, err := newagentllm.GenerateJSON[newagentmodel.PlanDecision](
+	decision, rawResult, err := infrallm.GenerateJSON[newagentmodel.PlanDecision](
 		ctx,
 		input.Client,
 		messages,
-		newagentllm.GenerateOptions{
+		infrallm.GenerateOptions{
 			Temperature: 0.2,
 			MaxTokens:   1600,
-			Thinking:    newagentllm.ThinkingModeEnabled,
+			Thinking:    infrallm.ThinkingModeEnabled,
 			Metadata: map[string]any{
 				"stage": planStageName,
 				"phase": "assessment",
@@ -108,14 +108,14 @@ func RunPlanNode(ctx context.Context, input PlanNodeInput) error {
 			return fmt.Errorf("深度规划状态推送失败: %w", err)
 		}
 
-		deepDecision, _, deepErr := newagentllm.GenerateJSON[newagentmodel.PlanDecision](
+		deepDecision, _, deepErr := infrallm.GenerateJSON[newagentmodel.PlanDecision](
 			ctx,
 			input.Client,
 			messages,
-			newagentllm.GenerateOptions{
+			infrallm.GenerateOptions{
 				Temperature: 0.2,
 				MaxTokens:   3200,
-				Thinking:    newagentllm.ThinkingModeEnabled,
+				Thinking:    infrallm.ThinkingModeEnabled,
 				Metadata: map[string]any{
 					"stage": planStageName,
 					"phase": "deep_planning",