feat: 多功能升级 — 流式逐字渲染、对话缓存、会话组织优化、记忆管理修复、性能仪表盘

- 前端消息流式逐字渲染 (AI-Core ChatStream → SSE → Gateway → WebSocket stream_chunk → fadeInUp + cursorBlink)
- 后端对话缓存 (conversationCache sync.Map, GET /sessions/:id/messages)
- 前端侧边栏历史多轮对话显示
- DevTools 性能监控图标移至首页仪表盘
- DevTools 用户记忆查询/删减功能修复 (补全 DELETE 数据链路)
- 后端和 DevTools 按用户分类组织实时活动会话 (map[userID]map[sessionID]*Client)
- 新增 docs/api-reference/ 路由参考文档
- 新增 docs/message-flow-architecture.md 消息链路架构文档

backend/gateway/internal/ws/hub.go

@@ -1,6 +1,7 @@
 package ws
 
 import (
+	"fmt"
 	"log"
 	"sync"
 	"time"
@@ -24,6 +25,13 @@ type SessionMessage struct {
 	Timestamp int64  `json:"timestamp"`
 }
 
+// Message 完整对话消息（用于缓存）
+type Message struct {
+	Role      string `json:"role"`
+	Content   string `json:"content"`
+	Timestamp int64  `json:"timestamp"`
+}
+
 const maxRecentMessages = 20
 
 // Hub WebSocket连接池
@@ -39,6 +47,10 @@ type Hub struct {
 
 	// 会话状态追踪 (sessionID -> SessionState)
 	sessions map[string]*SessionState
+
+	// 对话缓存：key = "userID:sessionID", value = []Message
+	conversationCache sync.Map
+	convCacheMu       sync.Mutex
 }
 
 // NewHub 创建WebSocket Hub
@@ -203,6 +215,42 @@ func (h *Hub) GetActiveSessions() []*SessionState {
 	return result
 }
 
+// GetActiveSessionsByUser 返回按用户分组的活跃会话列表
+func (h *Hub) GetActiveSessionsByUser() map[string][]*SessionState {
+	h.mu.RLock()
+	defer h.mu.RUnlock()
+
+	result := make(map[string][]*SessionState)
+	for _, s := range h.sessions {
+		cp := *s
+		cp.RecentMessages = nil
+		result[s.UserID] = append(result[s.UserID], &cp)
+	}
+	return result
+}
+
+// GetUserSessions 获取某用户的所有活跃会话
+func (h *Hub) GetUserSessions(userID string) []*SessionState {
+	h.mu.RLock()
+	defer h.mu.RUnlock()
+
+	var result []*SessionState
+	if clients, ok := h.userClients[userID]; ok {
+		seen := make(map[string]bool)
+		for c := range clients {
+			if !seen[c.SessionID] {
+				if s, ok := h.sessions[c.SessionID]; ok {
+					cp := *s
+					cp.RecentMessages = nil
+					result = append(result, &cp)
+					seen[c.SessionID] = true
+				}
+			}
+		}
+	}
+	return result
+}
+
 // GetSession 返回指定会话的详细信息（含最近消息）
 func (h *Hub) GetSession(sessionID string) *SessionState {
 	h.mu.RLock()
@@ -263,3 +311,48 @@ func (h *Hub) RecordMessage(sessionID, role, content string) {
 		s.RecentMessages = s.RecentMessages[len(s.RecentMessages)-maxRecentMessages:]
 	}
 }
+
+// ========== 对话缓存方法 ==========
+
+// cacheKey 生成对话缓存 key
+func cacheKey(userID, sessionID string) string {
+	return fmt.Sprintf("%s:%s", userID, sessionID)
+}
+
+// CacheMessage 缓存单条消息到对话历史
+func (h *Hub) CacheMessage(userID, sessionID string, msg Message) {
+	key := cacheKey(userID, sessionID)
+
+	h.convCacheMu.Lock()
+	defer h.convCacheMu.Unlock()
+
+	existing, _ := h.conversationCache.Load(key)
+	var messages []Message
+	if existing != nil {
+		messages = existing.([]Message)
+	}
+	messages = append(messages, msg)
+	h.conversationCache.Store(key, messages)
+}
+
+// GetConversation 获取完整对话历史
+func (h *Hub) GetConversation(userID, sessionID string) []Message {
+	key := cacheKey(userID, sessionID)
+
+	val, ok := h.conversationCache.Load(key)
+	if !ok {
+		return []Message{}
+	}
+
+	messages, ok := val.([]Message)
+	if !ok {
+		return []Message{}
+	}
+	return messages
+}
+
+// DeleteConversation 删除对话缓存
+func (h *Hub) DeleteConversation(userID, sessionID string) {
+	key := cacheKey(userID, sessionID)
+	h.conversationCache.Delete(key)
+}