diff --git a/cmd/gateway.go b/cmd/gateway.go index 129cc0d0..d094b591 100644 --- a/cmd/gateway.go +++ b/cmd/gateway.go @@ -616,6 +616,9 @@ func runGateway() { server.SetBuiltinToolsHandler(builtinToolsH) } if pendingMessagesH != nil { + if pc := cfg.Channels.PendingCompaction; pc != nil { + pendingMessagesH.SetKeepRecent(pc.KeepRecent) + } server.SetPendingMessagesHandler(pendingMessagesH) } @@ -678,6 +681,8 @@ func runGateway() { var instanceLoader *channels.InstanceLoader if pgStores.ChannelInstances != nil { instanceLoader = channels.NewInstanceLoader(pgStores.ChannelInstances, pgStores.Agents, channelMgr, msgBus, pgStores.Pairing) + instanceLoader.SetProviderRegistry(providerRegistry) + instanceLoader.SetPendingCompactionConfig(cfg.Channels.PendingCompaction) instanceLoader.RegisterFactory(channels.TypeTelegram, telegram.FactoryWithStores(pgStores.Agents, pgStores.Teams, pgStores.PendingMessages)) instanceLoader.RegisterFactory(channels.TypeDiscord, discord.FactoryWithPendingStore(pgStores.PendingMessages)) instanceLoader.RegisterFactory(channels.TypeFeishu, feishu.FactoryWithPendingStore(pgStores.PendingMessages)) diff --git a/cmd/gateway_http_handlers.go b/cmd/gateway_http_handlers.go index 0fa50e8d..b9e4d1e1 100644 --- a/cmd/gateway_http_handlers.go +++ b/cmd/gateway_http_handlers.go @@ -71,7 +71,7 @@ func wireHTTP(stores *store.Stores, token string, msgBus *bus.MessageBus, toolsR } if stores != nil && stores.PendingMessages != nil { - pendingMessagesH = httpapi.NewPendingMessagesHandler(stores.PendingMessages, token, providerReg) + pendingMessagesH = httpapi.NewPendingMessagesHandler(stores.PendingMessages, stores.Agents, token, providerReg) } return agentsH, skillsH, tracesH, mcpH, customToolsH, channelInstancesH, providersH, delegationsH, builtinToolsH, pendingMessagesH diff --git a/internal/channels/channel.go b/internal/channels/channel.go index f80f70b2..a7df05b7 100644 --- a/internal/channels/channel.go +++ b/internal/channels/channel.go @@ -308,6 +308,13 @@ func (c *BaseChannel) HandleMessage(senderID, chatID, content string, media []st c.bus.PublishInbound(msg) } +// PendingCompactable is optionally implemented by channels that have a PendingHistory +// supporting LLM-based compaction. InstanceLoader uses this to wire compaction config +// after channel creation. +type PendingCompactable interface { + SetPendingCompaction(cfg *CompactionConfig) +} + // Truncate shortens a string to maxLen, appending "..." if truncated. func Truncate(s string, maxLen int) string { if len(s) <= maxLen { diff --git a/internal/channels/discord/discord.go b/internal/channels/discord/discord.go index 6897e4b9..ea4305a3 100644 --- a/internal/channels/discord/discord.go +++ b/internal/channels/discord/discord.go @@ -98,6 +98,11 @@ func (c *Channel) Start(_ context.Context) error { // BlockReplyEnabled returns the per-channel block_reply override (nil = inherit gateway default). func (c *Channel) BlockReplyEnabled() *bool { return c.config.BlockReply } +// SetPendingCompaction configures LLM-based auto-compaction for pending messages. +func (c *Channel) SetPendingCompaction(cfg *channels.CompactionConfig) { + c.groupHistory.SetCompactionConfig(cfg) +} + // Stop closes the Discord gateway connection. func (c *Channel) Stop(_ context.Context) error { c.groupHistory.StopFlusher() diff --git a/internal/channels/feishu/feishu.go b/internal/channels/feishu/feishu.go index d47b06ba..94c70ac4 100644 --- a/internal/channels/feishu/feishu.go +++ b/internal/channels/feishu/feishu.go @@ -121,6 +121,11 @@ func (c *Channel) Start(ctx context.Context) error { // BlockReplyEnabled returns the per-channel block_reply override (nil = inherit gateway default). func (c *Channel) BlockReplyEnabled() *bool { return c.cfg.BlockReply } +// SetPendingCompaction configures LLM-based auto-compaction for pending messages. +func (c *Channel) SetPendingCompaction(cfg *channels.CompactionConfig) { + c.groupHistory.SetCompactionConfig(cfg) +} + // Stop shuts down the Feishu channel. func (c *Channel) Stop(_ context.Context) error { c.groupHistory.StopFlusher() diff --git a/internal/channels/history_compaction.go b/internal/channels/history_compaction.go index a0054e07..735632ca 100644 --- a/internal/channels/history_compaction.go +++ b/internal/channels/history_compaction.go @@ -123,7 +123,7 @@ func (ph *PendingHistory) runCompaction(historyKey string, cfg *CompactionConfig // Force-flush buffer to ensure DB is consistent ph.flushNow() - ctx, cancel := context.WithTimeout(context.Background(), 45*time.Second) + ctx, cancel := context.WithTimeout(context.Background(), 180*time.Second) defer cancel() // Check threshold from DB (may have been cleared since trigger) diff --git a/internal/channels/instance_loader.go b/internal/channels/instance_loader.go index 3e2cfc72..7aded053 100644 --- a/internal/channels/instance_loader.go +++ b/internal/channels/instance_loader.go @@ -10,6 +10,8 @@ import ( "time" "github.com/nextlevelbuilder/goclaw/internal/bus" + "github.com/nextlevelbuilder/goclaw/internal/config" + "github.com/nextlevelbuilder/goclaw/internal/providers" "github.com/nextlevelbuilder/goclaw/internal/store" ) @@ -23,14 +25,16 @@ type ChannelFactory func(name string, creds json.RawMessage, cfg json.RawMessage // InstanceLoader loads channel instances from the database and registers them with the Manager. // Follows the DynamicToolLoader pattern: LoadAll at startup, Reload on cache invalidation. type InstanceLoader struct { - store store.ChannelInstanceStore - agentStore store.AgentStore - factories map[string]ChannelFactory - manager *Manager - msgBus *bus.MessageBus - pairingSvc store.PairingStore - mu sync.Mutex - loaded map[string]struct{} // channel names managed by this loader + store store.ChannelInstanceStore + agentStore store.AgentStore + providerReg *providers.Registry + pendingCompactCfg *config.PendingCompactionConfig + factories map[string]ChannelFactory + manager *Manager + msgBus *bus.MessageBus + pairingSvc store.PairingStore + mu sync.Mutex + loaded map[string]struct{} // channel names managed by this loader } // NewInstanceLoader creates a new InstanceLoader. @@ -52,6 +56,18 @@ func NewInstanceLoader( } } +// SetProviderRegistry sets the provider registry for pending message compaction. +// Must be called before LoadAll/Reload. +func (l *InstanceLoader) SetProviderRegistry(reg *providers.Registry) { + l.providerReg = reg +} + +// SetPendingCompactionConfig sets the global pending message compaction thresholds. +// Must be called before LoadAll/Reload. +func (l *InstanceLoader) SetPendingCompactionConfig(cfg *config.PendingCompactionConfig) { + l.pendingCompactCfg = cfg +} + // RegisterFactory registers a factory for a channel type (e.g., "telegram", "discord"). func (l *InstanceLoader) RegisterFactory(channelType string, factory ChannelFactory) { l.factories[channelType] = factory @@ -205,8 +221,10 @@ func (l *InstanceLoader) loadInstance(ctx context.Context, inst store.ChannelIns } // Resolve agent_key from UUID — the routing system (Router, session keys) uses agent_key, not UUID. + var ag *store.AgentData if base, ok := ch.(interface{ SetAgentID(string) }); ok { - ag, err := l.agentStore.GetByID(ctx, inst.AgentID) + var err error + ag, err = l.agentStore.GetByID(ctx, inst.AgentID) if err != nil { return fmt.Errorf("agent %s not found for channel %s: %w", inst.AgentID, inst.Name, err) } @@ -216,6 +234,30 @@ func (l *InstanceLoader) loadInstance(ctx context.Context, inst store.ChannelIns if base, ok := ch.(interface{ SetType(string) }); ok { base.SetType(inst.ChannelType) } + + // Wire pending message auto-compaction using the agent's configured provider+model. + if pc, ok := ch.(PendingCompactable); ok && ag != nil && l.providerReg != nil && ag.Provider != "" { + if p, err := l.providerReg.Get(ag.Provider); err == nil { + model := ag.Model + if model == "" { + model = p.DefaultModel() + } + if model != "" { + cc := &CompactionConfig{ + Provider: p, + Model: model, + } + // Apply global threshold/keepRecent from config if set. + if l.pendingCompactCfg != nil { + cc.Threshold = l.pendingCompactCfg.Threshold + cc.KeepRecent = l.pendingCompactCfg.KeepRecent + } + pc.SetPendingCompaction(cc) + slog.Debug("pending compaction configured", "channel", inst.Name, "provider", ag.Provider, "model", model, + "threshold", cc.Threshold, "keep_recent", cc.KeepRecent) + } + } + } l.manager.RegisterChannel(inst.Name, ch) l.loaded[inst.Name] = struct{}{} diff --git a/internal/channels/slack/channel.go b/internal/channels/slack/channel.go index 1a43dfdc..add9e1c2 100644 --- a/internal/channels/slack/channel.go +++ b/internal/channels/slack/channel.go @@ -280,6 +280,11 @@ func (c *Channel) handleEvent(evt socketmode.Event) { } } +// SetPendingCompaction configures LLM-based auto-compaction for pending messages. +func (c *Channel) SetPendingCompaction(cfg *channels.CompactionConfig) { + c.groupHistory.SetCompactionConfig(cfg) +} + // Stop gracefully shuts down the Slack channel. func (c *Channel) Stop(_ context.Context) error { c.groupHistory.StopFlusher() diff --git a/internal/channels/telegram/channel.go b/internal/channels/telegram/channel.go index 8c322d5d..6a3fca46 100644 --- a/internal/channels/telegram/channel.go +++ b/internal/channels/telegram/channel.go @@ -206,6 +206,11 @@ func (c *Channel) StreamEnabled(isGroup bool) bool { // BlockReplyEnabled returns the per-channel block_reply override (nil = inherit gateway default). func (c *Channel) BlockReplyEnabled() *bool { return c.config.BlockReply } +// SetPendingCompaction configures LLM-based auto-compaction for pending messages. +func (c *Channel) SetPendingCompaction(cfg *channels.CompactionConfig) { + c.groupHistory.SetCompactionConfig(cfg) +} + // Stop shuts down the Telegram bot by cancelling the long polling context // and waiting for the polling goroutine to exit. func (c *Channel) Stop(_ context.Context) error { diff --git a/internal/channels/zalo/personal/channel.go b/internal/channels/zalo/personal/channel.go index 2b87304d..156e5d7d 100644 --- a/internal/channels/zalo/personal/channel.go +++ b/internal/channels/zalo/personal/channel.go @@ -123,6 +123,11 @@ func (c *Channel) Start(ctx context.Context) error { return nil } +// SetPendingCompaction configures LLM-based auto-compaction for pending messages. +func (c *Channel) SetPendingCompaction(cfg *channels.CompactionConfig) { + c.groupHistory.SetCompactionConfig(cfg) +} + // Stop gracefully shuts down the Zalo Personal channel. func (c *Channel) Stop(_ context.Context) error { c.groupHistory.StopFlusher() diff --git a/internal/config/config_channels.go b/internal/config/config_channels.go index 95b32c67..e6634a4d 100644 --- a/internal/config/config_channels.go +++ b/internal/config/config_channels.go @@ -1,14 +1,23 @@ package config +// PendingCompactionConfig configures LLM-based compaction of pending group messages. +// When a group accumulates more than Threshold pending messages, older messages are +// summarized by an LLM and replaced with a compact summary, keeping KeepRecent raw messages. +type PendingCompactionConfig struct { + Threshold int `json:"threshold,omitempty"` // trigger compaction when entries exceed this (default 50) + KeepRecent int `json:"keep_recent,omitempty"` // keep this many recent raw messages after compaction (default 15) +} + // ChannelsConfig contains per-channel configuration. type ChannelsConfig struct { - Telegram TelegramConfig `json:"telegram"` - Discord DiscordConfig `json:"discord"` - Slack SlackConfig `json:"slack"` - WhatsApp WhatsAppConfig `json:"whatsapp"` - Zalo ZaloConfig `json:"zalo"` - ZaloPersonal ZaloPersonalConfig `json:"zalo_personal"` - Feishu FeishuConfig `json:"feishu"` + Telegram TelegramConfig `json:"telegram"` + Discord DiscordConfig `json:"discord"` + Slack SlackConfig `json:"slack"` + WhatsApp WhatsAppConfig `json:"whatsapp"` + Zalo ZaloConfig `json:"zalo"` + ZaloPersonal ZaloPersonalConfig `json:"zalo_personal"` + Feishu FeishuConfig `json:"feishu"` + PendingCompaction *PendingCompactionConfig `json:"pending_compaction,omitempty"` // global pending message compaction settings } type TelegramConfig struct { diff --git a/internal/http/pending_messages.go b/internal/http/pending_messages.go index af7d19ab..27766ddb 100644 --- a/internal/http/pending_messages.go +++ b/internal/http/pending_messages.go @@ -16,14 +16,19 @@ import ( // PendingMessagesHandler handles pending message HTTP endpoints. type PendingMessagesHandler struct { store store.PendingMessageStore + agentStore store.AgentStore token string providerReg *providers.Registry + keepRecent int // global keepRecent from config (0 = use default 15) } -func NewPendingMessagesHandler(s store.PendingMessageStore, token string, providerReg *providers.Registry) *PendingMessagesHandler { - return &PendingMessagesHandler{store: s, token: token, providerReg: providerReg} +func NewPendingMessagesHandler(s store.PendingMessageStore, agentStore store.AgentStore, token string, providerReg *providers.Registry) *PendingMessagesHandler { + return &PendingMessagesHandler{store: s, agentStore: agentStore, token: token, providerReg: providerReg} } +// SetKeepRecent sets the global keepRecent value from config. +func (h *PendingMessagesHandler) SetKeepRecent(n int) { h.keepRecent = n } + func (h *PendingMessagesHandler) RegisterRoutes(mux *http.ServeMux) { mux.HandleFunc("GET /v1/pending-messages", h.authMiddleware(h.handleListGroups)) mux.HandleFunc("GET /v1/pending-messages/messages", h.authMiddleware(h.handleListMessages)) @@ -121,8 +126,8 @@ func (h *PendingMessagesHandler) handleCompact(w http.ResponseWriter, r *http.Re return } - // Resolve an LLM provider for summarization - provider := h.resolveProvider() + // Resolve an LLM provider for summarization using the default agent's config + provider, model := h.resolveProviderAndModel() if provider == nil { // Fallback: hard delete if no provider available slog.Warn("compact.no_provider", "channel", req.ChannelName, "key", req.HistoryKey) @@ -134,10 +139,14 @@ func (h *PendingMessagesHandler) handleCompact(w http.ResponseWriter, r *http.Re return } - ctx, cancel := context.WithTimeout(r.Context(), 45*time.Second) + ctx, cancel := context.WithTimeout(r.Context(), 180*time.Second) defer cancel() - remaining, err := channels.CompactGroup(ctx, h.store, req.ChannelName, req.HistoryKey, provider, provider.DefaultModel(), 15) + keepRecent := h.keepRecent + if keepRecent <= 0 { + keepRecent = 15 + } + remaining, err := channels.CompactGroup(ctx, h.store, req.ChannelName, req.HistoryKey, provider, model, keepRecent) if err != nil { slog.Warn("compact.failed", "channel", req.ChannelName, "key", req.HistoryKey, "error", err) writeJSON(w, http.StatusInternalServerError, map[string]string{"error": err.Error()}) @@ -146,18 +155,38 @@ func (h *PendingMessagesHandler) handleCompact(w http.ResponseWriter, r *http.Re writeJSON(w, http.StatusOK, map[string]interface{}{"status": "ok", "method": "summarized", "remaining": remaining}) } -// resolveProvider returns the first available LLM provider, or nil. -func (h *PendingMessagesHandler) resolveProvider() providers.Provider { +// resolveProviderAndModel resolves the LLM provider+model for pending message compaction. +// Uses the default agent's configured provider and model so compaction uses the same +// LLM as the agent that processes these messages. +func (h *PendingMessagesHandler) resolveProviderAndModel() (providers.Provider, string) { if h.providerReg == nil { - return nil + return nil, "" } - names := h.providerReg.List() - if len(names) == 0 { - return nil + + // Use the default agent's provider+model + if h.agentStore != nil { + if ag, err := h.agentStore.GetDefault(context.Background()); err == nil && ag.Provider != "" { + if p, err := h.providerReg.Get(ag.Provider); err == nil { + model := ag.Model + if model == "" { + model = p.DefaultModel() + } + if model != "" { + return p, model + } + } + } } - p, err := h.providerReg.Get(names[0]) - if err != nil { - return nil + + // Fallback: first provider with a valid default model + for _, name := range h.providerReg.List() { + p, err := h.providerReg.Get(name) + if err != nil { + continue + } + if p.DefaultModel() != "" { + return p, p.DefaultModel() + } } - return p + return nil, "" } diff --git a/ui/web/src/i18n/locales/en/config.json b/ui/web/src/i18n/locales/en/config.json index 22e0b5b8..08a9c8e6 100644 --- a/ui/web/src/i18n/locales/en/config.json +++ b/ui/web/src/i18n/locales/en/config.json @@ -124,6 +124,13 @@ "behavior.blockReplyInfo": "Only the final response is sent. Intermediate tool status messages are suppressed.", "behavior.intentClassifyHint": "Classify user intent before routing to reduce unnecessary agent invocations.", "behavior.intentClassifyInfo": "Agent will only be invoked when the classifier detects actionable intent.", + "behavior.pendingCompactionTitle": "Pending Message Compaction", + "behavior.pendingCompactionDescription": "Automatically summarize old group messages using LLM when the buffer exceeds a threshold", + "behavior.pendingCompactionThreshold": "Threshold", + "behavior.pendingCompactionThresholdTip": "Trigger compaction when a group accumulates more than this many pending messages. Set to 0 to use the default (50).", + "behavior.pendingCompactionKeepRecent": "Keep Recent", + "behavior.pendingCompactionKeepRecentTip": "Number of recent raw messages to keep after compaction. Older messages are replaced with an LLM summary. Default is 15.", + "behavior.pendingCompactionInfo": "When a group chat exceeds the threshold, the LLM automatically summarizes older messages into a single compact entry, keeping the most recent ones intact for context.", "behavior.rateLimitTitle": "Rate Limiting", "behavior.rateLimitDescription": "Inbound message size limits and per-user request throttling", "behavior.sessionsTitle": "Session Scoping", diff --git a/ui/web/src/i18n/locales/en/pending-messages.json b/ui/web/src/i18n/locales/en/pending-messages.json index b370357d..107c5e7b 100644 --- a/ui/web/src/i18n/locales/en/pending-messages.json +++ b/ui/web/src/i18n/locales/en/pending-messages.json @@ -3,6 +3,14 @@ "description": "Buffered channel messages awaiting agent processing", "emptyTitle": "No pending messages", "emptyDescription": "No buffered message groups found. Messages appear here when channels buffer incoming messages before agent processing.", + "howItWorks": { + "title": "How it works", + "step1": "Messages from channels (Telegram, Discord, etc.) are buffered here while the agent is busy or offline.", + "step2": "When the buffer reaches the threshold (default: 50 messages), auto-compaction kicks in — an LLM summarizes older messages, keeping only the 15 most recent.", + "step3": "On the agent's next turn, the compacted summary + recent messages are injected as context, so the agent catches up without being overwhelmed.", + "compactAction": "Compact — Manually trigger LLM summarization now. Useful when you want the agent to process a clean summary instead of raw message flood.", + "clearAction": "Clear — Permanently delete all buffered messages for a group. Use when messages are no longer relevant." + }, "columns": { "channel": "Channel", "group": "Group", @@ -28,5 +36,13 @@ "title": "Messages — {{name}}", "noMessages": "No messages found.", "summary": "Summary" + }, + "toast": { + "compacted": "Compaction complete", + "compactedSummarized": "Summarized to {{count}} messages", + "compactedDeleted": "All messages deleted (no LLM provider available)", + "failedCompact": "Compaction failed", + "cleared": "Messages cleared", + "failedClear": "Failed to clear messages" } } diff --git a/ui/web/src/i18n/locales/vi/config.json b/ui/web/src/i18n/locales/vi/config.json index ba3333b4..692a868e 100644 --- a/ui/web/src/i18n/locales/vi/config.json +++ b/ui/web/src/i18n/locales/vi/config.json @@ -124,6 +124,13 @@ "behavior.blockReplyInfo": "Chỉ phản hồi cuối được gửi. Thông báo trạng thái công cụ trung gian bị ẩn.", "behavior.intentClassifyHint": "Phân loại ý định người dùng trước khi định tuyến để giảm các lần gọi agent không cần thiết.", "behavior.intentClassifyInfo": "Agent chỉ được gọi khi bộ phân loại phát hiện ý định có thể thực hiện.", + "behavior.pendingCompactionTitle": "Nén tin nhắn chờ", + "behavior.pendingCompactionDescription": "Tự động tóm tắt tin nhắn nhóm cũ bằng LLM khi bộ đệm vượt ngưỡng", + "behavior.pendingCompactionThreshold": "Ngưỡng", + "behavior.pendingCompactionThresholdTip": "Kích hoạt nén khi nhóm tích lũy nhiều hơn số tin nhắn chờ này. Đặt 0 để dùng mặc định (50).", + "behavior.pendingCompactionKeepRecent": "Giữ gần đây", + "behavior.pendingCompactionKeepRecentTip": "Số tin nhắn gần đây giữ lại sau khi nén. Tin cũ hơn được thay bằng bản tóm tắt LLM. Mặc định là 15.", + "behavior.pendingCompactionInfo": "Khi nhóm chat vượt ngưỡng, LLM tự động tóm tắt các tin nhắn cũ thành một mục rút gọn, giữ nguyên các tin gần đây nhất làm ngữ cảnh.", "behavior.rateLimitTitle": "Giới hạn tốc độ", "behavior.rateLimitDescription": "Giới hạn kích thước tin nhắn đến và kiểm soát tốc độ yêu cầu mỗi người dùng", "behavior.sessionsTitle": "Phạm vi phiên", diff --git a/ui/web/src/i18n/locales/vi/pending-messages.json b/ui/web/src/i18n/locales/vi/pending-messages.json index f4cb2b4e..5aeefc11 100644 --- a/ui/web/src/i18n/locales/vi/pending-messages.json +++ b/ui/web/src/i18n/locales/vi/pending-messages.json @@ -3,6 +3,14 @@ "description": "Tin nhắn channel đã được lưu đệm đang chờ xử lý từ agent", "emptyTitle": "Không có tin nhắn đang chờ", "emptyDescription": "Không tìm thấy nhóm tin nhắn đệm. Tin nhắn xuất hiện ở đây khi channel lưu đệm tin nhắn đến trước khi agent xử lý.", + "howItWorks": { + "title": "Cơ chế hoạt động", + "step1": "Tin nhắn từ các channel (Telegram, Discord, v.v.) được lưu đệm tại đây khi agent đang bận hoặc offline.", + "step2": "Khi bộ đệm đạt ngưỡng (mặc định: 50 tin nhắn), hệ thống tự động tóm tắt — LLM tóm gọn các tin nhắn cũ, chỉ giữ lại 15 tin mới nhất.", + "step3": "Ở lượt tiếp theo của agent, bản tóm tắt + tin nhắn gần đây được inject làm ngữ cảnh, giúp agent nắm bắt nhanh mà không bị quá tải.", + "compactAction": "Tóm tắt — Kích hoạt LLM tóm tắt thủ công ngay. Hữu ích khi bạn muốn agent xử lý bản tóm tắt sạch thay vì hàng loạt tin nhắn thô.", + "clearAction": "Xóa — Xóa vĩnh viễn tất cả tin nhắn đệm của một nhóm. Dùng khi tin nhắn không còn liên quan." + }, "columns": { "channel": "Channel", "group": "Nhóm", @@ -28,5 +36,13 @@ "title": "Tin nhắn — {{name}}", "noMessages": "Không tìm thấy tin nhắn.", "summary": "Tóm tắt" + }, + "toast": { + "compacted": "Tóm tắt hoàn tất", + "compactedSummarized": "Đã tóm tắt còn {{count}} tin nhắn", + "compactedDeleted": "Đã xóa tất cả (không có LLM provider)", + "failedCompact": "Tóm tắt thất bại", + "cleared": "Đã xóa tin nhắn", + "failedClear": "Không thể xóa tin nhắn" } } diff --git a/ui/web/src/i18n/locales/zh/config.json b/ui/web/src/i18n/locales/zh/config.json index ca3181d7..e0dfe327 100644 --- a/ui/web/src/i18n/locales/zh/config.json +++ b/ui/web/src/i18n/locales/zh/config.json @@ -124,6 +124,13 @@ "behavior.blockReplyInfo": "仅发送最终响应，中间工具状态消息被抑制。", "behavior.intentClassifyHint": "在路由之前对用户意图进行分类，以减少不必要的 Agent 调用。", "behavior.intentClassifyInfo": "仅当分类器检测到可执行意图时才调用 Agent。", + "behavior.pendingCompactionTitle": "待处理消息压缩", + "behavior.pendingCompactionDescription": "当缓冲区超过阈值时，自动使用 LLM 总结旧群组消息", + "behavior.pendingCompactionThreshold": "阈值", + "behavior.pendingCompactionThresholdTip": "当群组累积超过此数量的待处理消息时触发压缩。设为 0 使用默认值（50）。", + "behavior.pendingCompactionKeepRecent": "保留最近", + "behavior.pendingCompactionKeepRecentTip": "压缩后保留的最近原始消息数量。较旧的消息将替换为 LLM 摘要。默认为 15。", + "behavior.pendingCompactionInfo": "当群聊超过阈值时，LLM 自动将旧消息总结为一条精简条目，保留最近的消息作为上下文。", "behavior.rateLimitTitle": "速率限制", "behavior.rateLimitDescription": "入站消息大小限制和每用户请求限流", "behavior.sessionsTitle": "会话范围", diff --git a/ui/web/src/i18n/locales/zh/pending-messages.json b/ui/web/src/i18n/locales/zh/pending-messages.json index 12cea04b..89dc514a 100644 --- a/ui/web/src/i18n/locales/zh/pending-messages.json +++ b/ui/web/src/i18n/locales/zh/pending-messages.json @@ -3,6 +3,14 @@ "description": "已缓冲的Channel消息，等待Agent处理", "emptyTitle": "暂无待处理消息", "emptyDescription": "未找到缓冲消息组。当Channel在Agent处理前缓冲传入消息时，消息将显示在此处。", + "howItWorks": { + "title": "工作机制", + "step1": "来自各Channel（Telegram、Discord等）的消息在Agent忙碌或离线时缓冲在此。", + "step2": "当缓冲区达到阈值（默认：50条消息）时，系统自动压缩 — LLM将旧消息总结为摘要，仅保留最近15条。", + "step3": "在Agent的下一轮回复中，压缩摘要 + 最近消息作为上下文注入，让Agent快速了解情况而不会被淹没。", + "compactAction": "压缩 — 立即手动触发LLM摘要。当您希望Agent处理简洁摘要而非大量原始消息时使用。", + "clearAction": "清除 — 永久删除某组的所有缓冲消息。当消息不再相关时使用。" + }, "columns": { "channel": "Channel", "group": "组", @@ -28,5 +36,13 @@ "title": "消息 — {{name}}", "noMessages": "未找到消息。", "summary": "摘要" + }, + "toast": { + "compacted": "压缩完成", + "compactedSummarized": "已压缩为 {{count}} 条消息", + "compactedDeleted": "已删除全部（无可用LLM Provider）", + "failedCompact": "压缩失败", + "cleared": "消息已清除", + "failedClear": "清除消息失败" } } diff --git a/ui/web/src/pages/config/sections/behavior-pending-compaction-card.tsx b/ui/web/src/pages/config/sections/behavior-pending-compaction-card.tsx new file mode 100644 index 00000000..c1e013fd --- /dev/null +++ b/ui/web/src/pages/config/sections/behavior-pending-compaction-card.tsx @@ -0,0 +1,99 @@ +import { Archive, Clock, Info } from "lucide-react"; +import { useTranslation } from "react-i18next"; +import { Input } from "@/components/ui/input"; +import { Label } from "@/components/ui/label"; +import { + Card, + CardContent, + CardDescription, + CardHeader, + CardTitle, +} from "@/components/ui/card"; + +export interface PendingCompactionValues { + threshold?: number; + keep_recent?: number; +} + +interface Props { + value: PendingCompactionValues; + onChange: (v: PendingCompactionValues) => void; +} + +/** Global pending message compaction thresholds with visual emphasis. */ +export function BehaviorPendingCompactionCard({ value, onChange }: Props) { + const { t } = useTranslation("config"); + + const update = (patch: Partial) => + onChange({ ...value, ...patch }); + + return ( + + + + {t("behavior.pendingCompactionTitle")} + + + {t("behavior.pendingCompactionDescription")} + + + + {/* Threshold */} +
+
+
+ +
+ +

+ {t("behavior.pendingCompactionThresholdTip")} +

+
+
+ update({ threshold: Number(e.target.value) })} + placeholder="50" + min={0} + className="w-24 shrink-0" + /> +
+
+ + {/* Keep Recent */} +
+
+
+ +
+ +

+ {t("behavior.pendingCompactionKeepRecentTip")} +

+
+
+ update({ keep_recent: Number(e.target.value) })} + placeholder="15" + min={1} + className="w-24 shrink-0" + /> +
+
+ + {/* Info banner */} +
+ + {t("behavior.pendingCompactionInfo")} +
+ + + ); +} diff --git a/ui/web/src/pages/config/sections/behavior-section.tsx b/ui/web/src/pages/config/sections/behavior-section.tsx index 020b6b58..ea53322f 100644 --- a/ui/web/src/pages/config/sections/behavior-section.tsx +++ b/ui/web/src/pages/config/sections/behavior-section.tsx @@ -6,6 +6,7 @@ import { BehaviorUxCard } from "./behavior-ux-card"; import { BehaviorRateCard } from "./behavior-rate-card"; import { BehaviorSessionsCard } from "./behavior-sessions-card"; import { BehaviorSecurityCard } from "./behavior-security-card"; +import { BehaviorPendingCompactionCard, type PendingCompactionValues } from "./behavior-pending-compaction-card"; /* eslint-disable @typescript-eslint/no-explicit-any */ @@ -22,6 +23,7 @@ export function BehaviorSection({ config, onPatch, saving }: Props) { const ag = config.agents?.defaults ?? {}; const tl = config.tools ?? {}; const ss = config.sessions ?? {}; + const ch = config.channels ?? {}; // UX toggles (from gateway + agents.defaults) const [ux, setUx] = useState({ @@ -49,6 +51,11 @@ export function BehaviorSection({ config, onPatch, saving }: Props) { scrub_credentials: tl.scrub_credentials, }); + // Pending compaction (from channels.pending_compaction) + const [pendingCompaction, setPendingCompaction] = useState( + ch.pending_compaction ?? {}, + ); + const [dirty, setDirty] = useState(false); // Reset drafts when external config changes @@ -68,6 +75,7 @@ export function BehaviorSection({ config, onPatch, saving }: Props) { injection_action: gw.injection_action, scrub_credentials: tl.scrub_credentials, }); + setPendingCompaction(ch.pending_compaction ?? {}); setDirty(false); }, [config]); // eslint-disable-line react-hooks/exhaustive-deps @@ -91,6 +99,7 @@ export function BehaviorSection({ config, onPatch, saving }: Props) { }, tools: { ...tl, scrub_credentials: security.scrub_credentials }, sessions: { ...ss, ...sessions }, + channels: { ...ch, pending_compaction: pendingCompaction }, }); }; @@ -100,6 +109,7 @@ export function BehaviorSection({ config, onPatch, saving }: Props) { + {dirty && (
diff --git a/ui/web/src/pages/pending-messages/hooks/use-pending-messages.ts b/ui/web/src/pages/pending-messages/hooks/use-pending-messages.ts index 9fdc37b4..ae278423 100644 --- a/ui/web/src/pages/pending-messages/hooks/use-pending-messages.ts +++ b/ui/web/src/pages/pending-messages/hooks/use-pending-messages.ts @@ -1,5 +1,7 @@ import { useState, useCallback } from "react"; +import i18next from "i18next"; import { useHttp } from "@/hooks/use-ws"; +import { toast } from "@/stores/use-toast-store"; import type { PendingMessageGroup, PendingMessage } from "../types"; export function usePendingMessages() { @@ -42,12 +44,23 @@ export function usePendingMessages() { const compactGroup = useCallback( async (channel: string, key: string) => { try { - await http.post("/v1/pending-messages/compact", { - channel_name: channel, - history_key: key, - }); + const res = await http.post<{ status: string; method?: string; remaining?: number }>( + "/v1/pending-messages/compact", + { channel_name: channel, history_key: key }, + ); + const method = res?.method ?? "summarized"; + toast.success( + i18next.t("pending-messages:toast.compacted"), + method === "deleted" + ? i18next.t("pending-messages:toast.compactedDeleted") + : i18next.t("pending-messages:toast.compactedSummarized", { count: res?.remaining ?? 0 }), + ); return true; - } catch { + } catch (err) { + toast.error( + i18next.t("pending-messages:toast.failedCompact"), + err instanceof Error ? err.message : "", + ); return false; } }, @@ -58,8 +71,13 @@ export function usePendingMessages() { async (channel: string, key: string) => { try { await http.delete(`/v1/pending-messages?channel=${encodeURIComponent(channel)}&key=${encodeURIComponent(key)}`); + toast.success(i18next.t("pending-messages:toast.cleared")); return true; - } catch { + } catch (err) { + toast.error( + i18next.t("pending-messages:toast.failedClear"), + err instanceof Error ? err.message : "", + ); return false; } }, diff --git a/ui/web/src/pages/pending-messages/pending-messages-page.tsx b/ui/web/src/pages/pending-messages/pending-messages-page.tsx index 05a16f46..45ba5066 100644 --- a/ui/web/src/pages/pending-messages/pending-messages-page.tsx +++ b/ui/web/src/pages/pending-messages/pending-messages-page.tsx @@ -1,6 +1,6 @@ import { useState, useEffect } from "react"; import { useTranslation } from "react-i18next"; -import { Inbox, RefreshCw, Trash2, Archive, Loader2 } from "lucide-react"; +import { Inbox, RefreshCw, Trash2, Archive, Loader2, Info, ChevronDown, ChevronUp } from "lucide-react"; import { Button } from "@/components/ui/button"; import { Badge } from "@/components/ui/badge"; import { PageHeader } from "@/components/shared/page-header"; @@ -72,6 +72,8 @@ export function PendingMessagesPage() { } /> + +
{showSkeleton ? ( @@ -181,6 +183,50 @@ export function PendingMessagesPage() { ); } +function HowItWorksCard() { + const { t } = useTranslation("pending-messages"); + const [open, setOpen] = useState(true); + + return ( +
+ + {open && ( +
+
+ 1 +

{t("howItWorks.step1")}

+
+
+ 2 +

{t("howItWorks.step2")}

+
+
+ 3 +

{t("howItWorks.step3")}

+
+
+
+

{t("howItWorks.compactAction")}

+

{t("howItWorks.clearAction")}

+
+
+ )} +
+ ); +} + function ConfirmClearDialog({ group, onConfirm,