Files
backend/internal/service/chat_service.go

824 lines
27 KiB
Go
Raw Normal View History

package service
import (
"context"
"errors"
"fmt"
"slices"
"time"
"with_you/internal/cache"
"with_you/internal/dto"
"with_you/internal/model"
"with_you/internal/pkg/ws"
"with_you/internal/repository"
"go.uber.org/zap"
"gorm.io/gorm"
)
// 撤回消息的时间限制2分钟
const RecallMessageTimeout = 2 * time.Minute
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
// 消息幂等性 TTL24 小时内同一 client_msg_id 不会重复发送
const messageIdempotentTTL = 24 * time.Hour
// ChatService 聊天服务接口
type ChatService interface {
// 会话管理
GetOrCreateConversation(ctx context.Context, user1ID, user2ID string) (*model.Conversation, error)
GetConversationList(ctx context.Context, userID string, page, pageSize int) ([]*model.Conversation, int64, error)
GetConversationByID(ctx context.Context, conversationID string, userID string) (*model.Conversation, error)
DeleteConversationForSelf(ctx context.Context, conversationID string, userID string) error
SetConversationPinned(ctx context.Context, conversationID string, userID string, isPinned bool) error
SetConversationNotificationMuted(ctx context.Context, conversationID string, userID string, notificationMuted bool) error
// 消息操作
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
SendMessage(ctx context.Context, senderID string, conversationID string, segments model.MessageSegments, replyToID *string, clientMsgID string) (*model.Message, error)
GetMessages(ctx context.Context, conversationID string, userID string, page, pageSize int) ([]*model.Message, int64, error)
GetMessagesAfterSeq(ctx context.Context, conversationID string, userID string, afterSeq int64, limit int) ([]*model.Message, error)
GetMessagesBeforeSeq(ctx context.Context, conversationID string, userID string, beforeSeq int64, limit int) ([]*model.Message, error)
// 已读管理
MarkAsRead(ctx context.Context, conversationID string, userID string, seq int64) error
GetUnreadCount(ctx context.Context, conversationID string, userID string) (int64, error)
GetAllUnreadCount(ctx context.Context, userID string) (int64, error)
// 消息扩展功能
RecallMessage(ctx context.Context, messageID string, userID string) error
DeleteMessage(ctx context.Context, messageID string, userID string) error
// 实时事件相关
SendTyping(ctx context.Context, senderID string, conversationID string)
// 在线状态
IsUserOnline(userID string) bool
// 仅保存消息到数据库,不发送实时推送(供群聊等自行推送的场景使用)
SaveMessage(ctx context.Context, senderID string, conversationID string, segments model.MessageSegments, replyToID *string) (*model.Message, error)
}
// chatServiceImpl 聊天服务实现
type chatServiceImpl struct {
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
repo repository.MessageRepository
userRepo repository.UserRepository
sensitive SensitiveService
wsHub *ws.Hub
pushSvc PushService
cache cache.Cache
conversationCache *cache.ConversationCache
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
uploadService *UploadService
}
// NewChatService 创建聊天服务
func NewChatService(
repo repository.MessageRepository,
userRepo repository.UserRepository,
sensitive SensitiveService,
wsHub *ws.Hub,
cacheBackend cache.Cache,
uploadService *UploadService,
pushSvc PushService,
) ChatService {
// 创建适配器
convRepoAdapter := cache.NewConversationRepositoryAdapter(repo)
msgRepoAdapter := cache.NewMessageRepositoryAdapter(repo)
// 创建会话缓存
conversationCache := cache.NewConversationCache(
cacheBackend,
convRepoAdapter,
msgRepoAdapter,
cache.DefaultConversationCacheSettings(),
)
return &chatServiceImpl{
repo: repo,
userRepo: userRepo,
sensitive: sensitive,
wsHub: wsHub,
pushSvc: pushSvc,
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
cache: cacheBackend,
conversationCache: conversationCache,
uploadService: uploadService,
}
}
func (s *chatServiceImpl) publishToUsers(userIDs []string, event string, payload any) {
if s.wsHub == nil || len(userIDs) == 0 {
return
}
s.wsHub.PublishToUsers(userIDs, event, payload)
}
// GetOrCreateConversation 获取或创建私聊会话
func (s *chatServiceImpl) GetOrCreateConversation(ctx context.Context, user1ID, user2ID string) (*model.Conversation, error) {
conv, err := s.repo.GetOrCreatePrivateConversation(user1ID, user2ID)
if err != nil {
return nil, err
}
// 失效会话列表缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateConversationList(user1ID)
s.conversationCache.InvalidateConversationList(user2ID)
}
return conv, nil
}
// GetConversationList 获取用户的会话列表(带缓存)
func (s *chatServiceImpl) GetConversationList(ctx context.Context, userID string, page, pageSize int) ([]*model.Conversation, int64, error) {
// 优先使用缓存
if s.conversationCache != nil {
return s.conversationCache.GetConversationList(ctx, userID, page, pageSize)
}
return s.repo.GetConversations(userID, page, pageSize)
}
// GetConversationByID 获取会话详情(带缓存)
func (s *chatServiceImpl) GetConversationByID(ctx context.Context, conversationID string, userID string) (*model.Conversation, error) {
// 验证用户是否是会话参与者
participant, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("conversation not found or no permission")
}
return nil, fmt.Errorf("failed to get participant: %w", err)
}
// 获取会话信息(优先使用缓存)
var conv *model.Conversation
if s.conversationCache != nil {
conv, err = s.conversationCache.GetConversation(ctx, conversationID)
} else {
conv, err = s.repo.GetConversation(conversationID)
}
if err != nil {
return nil, fmt.Errorf("failed to get conversation: %w", err)
}
_ = participant // 可以用于返回已读位置等信息
return conv, nil
}
// getParticipant 获取参与者信息(优先使用缓存)
func (s *chatServiceImpl) getParticipant(ctx context.Context, conversationID, userID string) (*model.ConversationParticipant, error) {
if s.conversationCache != nil {
return s.conversationCache.GetParticipant(ctx, conversationID, userID)
}
return s.repo.GetParticipant(conversationID, userID)
}
// DeleteConversationForSelf 仅自己删除会话
func (s *chatServiceImpl) DeleteConversationForSelf(ctx context.Context, conversationID string, userID string) error {
participant, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return errors.New("conversation not found or no permission")
}
return fmt.Errorf("failed to get participant: %w", err)
}
if participant.ConversationID == "" {
return errors.New("conversation not found or no permission")
}
if err := s.repo.HideConversationForUser(conversationID, userID); err != nil {
return fmt.Errorf("failed to hide conversation: %w", err)
}
// 失效会话列表缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateConversationList(userID)
}
return nil
}
// SetConversationPinned 设置会话置顶(用户维度)
func (s *chatServiceImpl) SetConversationPinned(ctx context.Context, conversationID string, userID string, isPinned bool) error {
participant, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return errors.New("conversation not found or no permission")
}
return fmt.Errorf("failed to get participant: %w", err)
}
if participant.ConversationID == "" {
return errors.New("conversation not found or no permission")
}
if err := s.repo.UpdatePinned(conversationID, userID, isPinned); err != nil {
return fmt.Errorf("failed to update pinned status: %w", err)
}
// 失效缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateParticipant(conversationID, userID)
s.conversationCache.InvalidateConversationList(userID)
}
return nil
}
// SetConversationNotificationMuted 设置会话免打扰状态(用户维度)
func (s *chatServiceImpl) SetConversationNotificationMuted(ctx context.Context, conversationID string, userID string, notificationMuted bool) error {
participant, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return errors.New("conversation not found or no permission")
}
return fmt.Errorf("failed to get participant: %w", err)
}
if participant.ConversationID == "" {
return errors.New("conversation not found or no permission")
}
if err := s.repo.UpdateNotificationMuted(conversationID, userID, notificationMuted); err != nil {
return fmt.Errorf("failed to update notification_muted status: %w", err)
}
// 失效缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateParticipant(conversationID, userID)
s.conversationCache.InvalidateConversationList(userID)
}
return nil
}
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
func (s *chatServiceImpl) SendMessage(ctx context.Context, senderID string, conversationID string, segments model.MessageSegments, replyToID *string, clientMsgID string) (*model.Message, error) {
if clientMsgID != "" && s.cache != nil {
idemKey := cache.MessageIdempotentKey(senderID, clientMsgID)
if cachedMsgID, ok := cache.GetTyped[string](s.cache, idemKey); ok && cachedMsgID != "" {
if existing, err := s.repo.GetMessageByID(cachedMsgID); err == nil && existing != nil {
return existing, nil
}
}
}
conv, err := s.getConversation(ctx, conversationID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("会话不存在,请重新创建会话")
}
return nil, fmt.Errorf("failed to get conversation: %w", err)
}
// 拉黑限制:仅拦截"被拉黑方 -> 拉黑人"方向
if conv.Type == model.ConversationTypePrivate && s.userRepo != nil {
participants, pErr := s.getParticipants(ctx, conversationID)
if pErr != nil {
return nil, fmt.Errorf("failed to get participants: %w", pErr)
}
var sentCount *int64
for _, p := range participants {
if p.UserID == senderID {
continue
}
blocked, bErr := s.userRepo.IsBlocked(p.UserID, senderID)
if bErr != nil {
return nil, fmt.Errorf("failed to check block status: %w", bErr)
}
if blocked {
return nil, ErrUserBlocked
}
// 陌生人限制:对方未回关前,只允许发送一条文本消息,且禁止发送图片
isFollowedBack, fErr := s.userRepo.IsFollowing(p.UserID, senderID)
if fErr != nil {
return nil, fmt.Errorf("failed to check follow status: %w", fErr)
}
if !isFollowedBack {
if containsImageSegment(segments) {
return nil, errors.New("对方未关注你,暂不支持发送图片")
}
if sentCount == nil {
c, cErr := s.repo.CountMessagesBySenderInConversation(conversationID, senderID)
if cErr != nil {
return nil, fmt.Errorf("failed to count sender messages: %w", cErr)
}
sentCount = &c
}
if *sentCount >= 1 {
return nil, errors.New("对方未关注你前,仅允许发送一条消息")
}
}
}
}
// 验证用户是否是会话参与者
_, err = s.getParticipant(ctx, conversationID, senderID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("您不是该会话的参与者")
}
return nil, fmt.Errorf("failed to get participant: %w", err)
}
if s.uploadService != nil {
if err := s.uploadService.ValidateChatMessageImageSegments(segments); err != nil {
return nil, err
}
}
// 创建消息
message := &model.Message{
ConversationID: conversationID,
SenderID: senderID, // 直接使用string类型的UUID
Segments: segments,
ReplyToID: replyToID,
Status: model.MessageStatusNormal,
}
// 使用事务创建消息并更新seq
if err := s.repo.CreateMessageWithSeq(message); err != nil {
return nil, fmt.Errorf("failed to save message: %w", err)
}
// 新消息会改变分页结果,先失效分页缓存,避免读到旧列表
if s.conversationCache != nil {
s.conversationCache.InvalidateMessagePages(conversationID)
}
go func() {
if err := s.cacheMessage(context.Background(), conversationID, message); err != nil {
zap.L().Warn("async cache message failed",
zap.String("component", "ChatService"),
zap.String("convID", conversationID),
zap.String("msgID", message.ID),
zap.Error(err),
)
}
}()
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
if clientMsgID != "" && s.cache != nil {
s.cache.Set(cache.MessageIdempotentKey(senderID, clientMsgID), message.ID, messageIdempotentTTL)
}
// 获取会话中的参与者并发送消息
participants, err := s.getParticipants(ctx, conversationID)
if err == nil {
targetIDs := make([]string, 0, len(participants))
for _, p := range participants {
if conv.Type == model.ConversationTypePrivate && p.UserID == senderID {
continue
}
targetIDs = append(targetIDs, p.UserID)
}
detailType := "private"
if conv.Type == model.ConversationTypeGroup {
detailType = "group"
}
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
if len(targetIDs) > 20 && conv.Type == model.ConversationTypeGroup {
s.wsHub.PublishToUsersConcurrent(targetIDs, "chat_message", map[string]any{
"detail_type": detailType,
"message": dto.ConvertMessageToResponse(message),
})
} else {
s.publishToUsers(targetIDs, "chat_message", map[string]any{
"detail_type": detailType,
"message": dto.ConvertMessageToResponse(message),
})
}
for _, p := range participants {
if p.UserID == senderID {
continue
}
if totalUnread, uErr := s.repo.GetAllUnreadCount(p.UserID); uErr == nil {
s.publishToUsers([]string{p.UserID}, "conversation_unread", map[string]any{
"conversation_id": conversationID,
"total_unread": totalUnread,
})
}
}
}
if s.conversationCache != nil {
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
keys := make([]string, 0, len(participants)*2)
for _, p := range participants {
keys = append(keys,
cache.UnreadDetailKey(p.UserID, conversationID),
cache.UnreadConversationKey(p.UserID),
)
}
if len(keys) > 0 {
s.cache.DeleteBatch(keys)
}
for _, p := range participants {
s.conversationCache.InvalidateConversationList(p.UserID)
}
}
if s.pushSvc != nil && len(participants) > 0 {
sender := ChatMessageSender{ID: senderID}
if senderUser, sErr := s.userRepo.GetByID(senderID); sErr == nil {
sender.Name = senderUser.Nickname
sender.Avatar = senderUser.Avatar
}
convType := conv.Type
convName := ""
if conv.Type == model.ConversationTypeGroup && conv.Group != nil {
convName = conv.Group.Name
}
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
allTargetIDs := make([]string, 0, len(participants))
for _, p := range participants {
if p.UserID == senderID {
continue
}
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
allTargetIDs = append(allTargetIDs, p.UserID)
}
var offlineTargetIDs []string
if s.wsHub != nil && len(allTargetIDs) > 0 {
_, offlineTargetIDs = s.wsHub.FilterOnline(allTargetIDs)
} else {
offlineTargetIDs = allTargetIDs
}
if len(offlineTargetIDs) > 0 {
go func(ids []string, sender ChatMessageSender) {
for _, uid := range ids {
if pushErr := s.pushSvc.PushChatMessage(context.Background(), uid, conversationID, &sender, convType, convName, message); pushErr != nil {
zap.L().Debug("push chat message failed",
zap.String("userID", uid),
zap.String("conversationID", conversationID),
zap.Error(pushErr),
)
}
}
refactor: improve system stability, performance, and code structure This commit introduces several architectural improvements and optimizations across the codebase: - **Performance & Reliability**: - Implemented Redis pipelining in `ConversationCache.CacheMessage` to reduce network round-trips. - Added a circuit breaker to the JPush client to prevent cascading failures. - Introduced batch deletion and batch member addition capabilities in repositories. - Added message idempotency support using `client_msg_id` and a Redis-based cache. - Optimized WebSocket handling with connection limits (total and per-user) and improved error logging. - **Code Refactoring**: - Refactored `Router` to use a `RouterDeps` struct, simplifying the constructor and improving maintainability. - Unified model ID generation logic using new `id_helper.go` (supporting UUID and Snowflake). - Standardized JSON serialization/deserialization in models using `json_helper.go`. - Refactored DTO conversion logic, specifically for `UserResponse` (using functional options) and `Report` responses. - Removed redundant/deprecated DTOs like `PostDetailResponse` and `TradeItemDetailResponse`. - **Cache Improvements**: - Enhanced `LayeredCache` with `SetRaw` to avoid double-encoding when promoting values from Redis to local cache. - Added `DeleteBatch` support to the cache interface. - **Other Changes**: - Cleaned up `config.go` by removing redundant default values and explicit environment variable overrides. - Improved WebSocket registration flow to handle connection limits gracefully.
2026-05-04 13:07:03 +08:00
}(offlineTargetIDs, sender)
}
}
return message, nil
}
func (s *chatServiceImpl) getConversation(ctx context.Context, conversationID string) (*model.Conversation, error) {
if s.conversationCache != nil {
return s.conversationCache.GetConversation(ctx, conversationID)
}
return s.repo.GetConversation(conversationID)
}
// getParticipants 获取会话参与者(优先使用缓存)
func (s *chatServiceImpl) getParticipants(ctx context.Context, conversationID string) ([]*model.ConversationParticipant, error) {
if s.conversationCache != nil {
return s.conversationCache.GetParticipants(ctx, conversationID)
}
return s.repo.GetConversationParticipants(conversationID)
}
// cacheMessage 缓存消息(内部方法)
func (s *chatServiceImpl) cacheMessage(ctx context.Context, convID string, msg *model.Message) error {
if s.conversationCache == nil {
return nil
}
asyncCtx, cancel := context.WithTimeout(ctx, 5*time.Second)
defer cancel()
return s.conversationCache.CacheMessage(asyncCtx, convID, msg)
}
func containsImageSegment(segments model.MessageSegments) bool {
return slices.ContainsFunc(segments, func(seg model.MessageSegment) bool {
return seg.Type == string(model.ContentTypeImage) || seg.Type == "image"
})
}
// GetMessages 获取消息历史(分页,带缓存)
func (s *chatServiceImpl) GetMessages(ctx context.Context, conversationID string, userID string, page, pageSize int) ([]*model.Message, int64, error) {
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, 0, errors.New("conversation not found or no permission")
}
return nil, 0, fmt.Errorf("failed to get participant: %w", err)
}
// 优先使用缓存
if s.conversationCache != nil {
return s.conversationCache.GetMessages(ctx, conversationID, page, pageSize)
}
return s.repo.GetMessages(conversationID, page, pageSize)
}
// GetMessagesAfterSeq 获取指定seq之后的消息用于增量同步
func (s *chatServiceImpl) GetMessagesAfterSeq(ctx context.Context, conversationID string, userID string, afterSeq int64, limit int) ([]*model.Message, error) {
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("conversation not found or no permission")
}
return nil, fmt.Errorf("failed to get participant: %w", err)
}
if limit <= 0 {
limit = 100
}
return s.repo.GetMessagesAfterSeq(conversationID, afterSeq, limit)
}
// GetMessagesBeforeSeq 获取指定seq之前的历史消息用于下拉加载更多
func (s *chatServiceImpl) GetMessagesBeforeSeq(ctx context.Context, conversationID string, userID string, beforeSeq int64, limit int) ([]*model.Message, error) {
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("conversation not found or no permission")
}
return nil, fmt.Errorf("failed to get participant: %w", err)
}
if limit <= 0 {
limit = 20
}
return s.repo.GetMessagesBeforeSeq(conversationID, beforeSeq, limit)
}
// MarkAsRead 标记已读
func (s *chatServiceImpl) MarkAsRead(ctx context.Context, conversationID string, userID string, seq int64) error {
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return errors.New("conversation not found or no permission")
}
return fmt.Errorf("failed to get participant: %w", err)
}
// 1. 先写入DB保证数据一致性DB是唯一数据源
err = s.repo.UpdateLastReadSeq(conversationID, userID, seq)
if err != nil {
return fmt.Errorf("failed to update last read seq: %w", err)
}
// 2. DB 写入成功后失效缓存Cache-Aside 模式)
if s.conversationCache != nil {
// 失效参与者缓存,下次读取时会从 DB 加载最新数据
s.conversationCache.InvalidateParticipant(conversationID, userID)
// 失效未读数缓存
s.conversationCache.InvalidateUnreadCount(userID, conversationID)
// 失效会话列表缓存
s.conversationCache.InvalidateConversationList(userID)
}
participants, pErr := s.getParticipants(ctx, conversationID)
if pErr == nil {
detailType := "private"
groupID := ""
if conv, convErr := s.getConversation(ctx, conversationID); convErr == nil && conv.Type == model.ConversationTypeGroup {
detailType = "group"
if conv.GroupID != nil {
groupID = *conv.GroupID
}
}
targetIDs := make([]string, 0, len(participants))
for _, p := range participants {
targetIDs = append(targetIDs, p.UserID)
}
s.publishToUsers(targetIDs, "message_read", map[string]any{
"detail_type": detailType,
"conversation_id": conversationID,
"group_id": groupID,
"user_id": userID,
"seq": seq,
})
}
if totalUnread, uErr := s.repo.GetAllUnreadCount(userID); uErr == nil {
s.publishToUsers([]string{userID}, "conversation_unread", map[string]any{
"conversation_id": conversationID,
"total_unread": totalUnread,
})
}
return nil
}
// GetUnreadCount 获取指定会话的未读消息数(带缓存)
func (s *chatServiceImpl) GetUnreadCount(ctx context.Context, conversationID string, userID string) (int64, error) {
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, userID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return 0, errors.New("conversation not found or no permission")
}
return 0, fmt.Errorf("failed to get participant: %w", err)
}
// 优先使用缓存
if s.conversationCache != nil {
return s.conversationCache.GetUnreadCount(ctx, userID, conversationID)
}
return s.repo.GetUnreadCount(conversationID, userID)
}
// GetAllUnreadCount 获取所有会话的未读消息总数
func (s *chatServiceImpl) GetAllUnreadCount(ctx context.Context, userID string) (int64, error) {
return s.repo.GetAllUnreadCount(userID)
}
// RecallMessage 撤回消息2分钟内
func (s *chatServiceImpl) RecallMessage(ctx context.Context, messageID string, userID string) error {
// 获取消息
message, err := s.repo.GetMessageByID(messageID)
if err != nil {
return errors.New("message not found")
}
// 验证是否是消息发送者
if message.SenderIDStr() != userID {
return errors.New("can only recall your own messages")
}
// 验证消息是否已被撤回
if message.Status == model.MessageStatusRecalled {
return errors.New("message already recalled")
}
// 验证是否在2分钟内
if time.Since(message.CreatedAt) > RecallMessageTimeout {
return errors.New("message recall timeout (2 minutes)")
}
// 更新消息状态为已撤回,并清空原始消息内容,仅保留撤回占位
err = s.repo.RecallMessage(messageID, userID)
if err != nil {
return fmt.Errorf("failed to recall message: %w", err)
}
// 失效消息缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateConversation(message.ConversationID)
}
if participants, pErr := s.getParticipants(ctx, message.ConversationID); pErr == nil {
detailType := "private"
groupID := ""
if conv, convErr := s.getConversation(ctx, message.ConversationID); convErr == nil && conv.Type == model.ConversationTypeGroup {
detailType = "group"
if conv.GroupID != nil {
groupID = *conv.GroupID
}
}
targetIDs := make([]string, 0, len(participants))
for _, p := range participants {
targetIDs = append(targetIDs, p.UserID)
}
s.publishToUsers(targetIDs, "message_recall", map[string]any{
"detail_type": detailType,
"conversation_id": message.ConversationID,
"group_id": groupID,
"message_id": messageID,
"sender_id": userID,
})
}
return nil
}
// DeleteMessage 删除消息(仅对自己可见)
func (s *chatServiceImpl) DeleteMessage(ctx context.Context, messageID string, userID string) error {
// 获取消息
message, err := s.repo.GetMessageByID(messageID)
if err != nil {
return errors.New("message not found")
}
// 验证用户是否是会话参与者
_, err = s.getParticipant(ctx, message.ConversationID, userID)
if err != nil {
return errors.New("no permission to delete this message")
}
// 对于删除消息,我们使用软删除,但需要确保只对当前用户隐藏
// 这里简化处理:只有发送者可以删除自己的消息
if message.SenderIDStr() != userID {
return errors.New("can only delete your own messages")
}
// 更新消息状态为已删除
err = s.repo.UpdateMessageStatus(message.ID, model.MessageStatusDeleted)
if err != nil {
return fmt.Errorf("failed to delete message: %w", err)
}
// 失效消息缓存
if s.conversationCache != nil {
s.conversationCache.InvalidateConversation(message.ConversationID)
}
return nil
}
// SendTyping 发送正在输入状态
func (s *chatServiceImpl) SendTyping(ctx context.Context, senderID string, conversationID string) {
if s.wsHub == nil {
return
}
// 验证用户是否是会话参与者
_, err := s.getParticipant(ctx, conversationID, senderID)
if err != nil {
return
}
// 获取会话中的其他参与者
participants, err := s.getParticipants(ctx, conversationID)
if err != nil {
return
}
detailType := "private"
if conv, convErr := s.getConversation(ctx, conversationID); convErr == nil && conv.Type == model.ConversationTypeGroup {
detailType = "group"
}
for _, p := range participants {
if p.UserID == senderID {
continue
}
if s.wsHub != nil {
s.wsHub.PublishToUser(p.UserID, "typing", map[string]any{
"detail_type": detailType,
"conversation_id": conversationID,
"user_id": senderID,
"is_typing": true,
})
}
}
}
// IsUserOnline 检查用户是否在线
func (s *chatServiceImpl) IsUserOnline(userID string) bool {
if s.wsHub != nil {
return s.wsHub.HasClients(userID)
}
return false
}
// SaveMessage 仅保存消息到数据库,不发送实时推送
// 适用于群聊等由调用方自行负责推送的场景
func (s *chatServiceImpl) SaveMessage(ctx context.Context, senderID string, conversationID string, segments model.MessageSegments, replyToID *string) (*model.Message, error) {
// 验证会话是否存在
_, err := s.getConversation(ctx, conversationID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("会话不存在,请重新创建会话")
}
return nil, fmt.Errorf("failed to get conversation: %w", err)
}
// 验证用户是否是会话参与者
_, err = s.getParticipant(ctx, conversationID, senderID)
if err != nil {
if errors.Is(err, gorm.ErrRecordNotFound) {
return nil, errors.New("您不是该会话的参与者")
}
return nil, fmt.Errorf("failed to get participant: %w", err)
}
if s.uploadService != nil {
if err := s.uploadService.ValidateChatMessageImageSegments(segments); err != nil {
return nil, err
}
}
message := &model.Message{
ConversationID: conversationID,
SenderID: senderID,
Segments: segments,
ReplyToID: replyToID,
Status: model.MessageStatusNormal,
}
if err := s.repo.CreateMessageWithSeq(message); err != nil {
return nil, fmt.Errorf("failed to save message: %w", err)
}
// 新消息会改变分页结果,先失效分页缓存,避免读到旧列表
if s.conversationCache != nil {
s.conversationCache.InvalidateMessagePages(conversationID)
}
// 异步写入缓存
go func() {
if err := s.cacheMessage(context.Background(), conversationID, message); err != nil {
zap.L().Warn("async cache message failed",
zap.String("component", "ChatService"),
zap.String("convID", conversationID),
zap.String("msgID", message.ID),
zap.Error(err),
)
}
}()
return message, nil
}