From 9117c2a4ec320fb8eb817117ceadd462d0194768 Mon Sep 17 00:00:00 2001 From: erio Date: Thu, 5 Mar 2026 19:37:38 +0800 Subject: [PATCH 01/53] fix: include upstream error details in usage API error response When FetchUsageWithOptions receives a non-200 response from the Anthropic API (e.g. 429 Rate Limited, 401 Unauthorized), the error was wrapped with fmt.Errorf which infraerrors.FromError cannot recognize, causing a generic "internal error" message with no details. Replace fmt.Errorf with infraerrors.New(500, "UPSTREAM_ERROR", msg) so the upstream error details (status code + body) are included in the 500 response message. The HTTP status remains 500 to avoid interfering with frontend auth routing (e.g. 401 would trigger JWT expiry redirect). --- backend/internal/repository/claude_usage_service.go | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/backend/internal/repository/claude_usage_service.go b/backend/internal/repository/claude_usage_service.go index f6054828c3..1264f6bbaf 100644 --- a/backend/internal/repository/claude_usage_service.go +++ b/backend/internal/repository/claude_usage_service.go @@ -8,6 +8,7 @@ import ( "net/http" "time" + infraerrors "github.com/Wei-Shaw/sub2api/internal/pkg/errors" "github.com/Wei-Shaw/sub2api/internal/pkg/httpclient" "github.com/Wei-Shaw/sub2api/internal/service" ) @@ -95,7 +96,8 @@ func (s *claudeUsageService) FetchUsageWithOptions(ctx context.Context, opts *se if resp.StatusCode != http.StatusOK { body, _ := io.ReadAll(resp.Body) - return nil, fmt.Errorf("API returned status %d: %s", resp.StatusCode, string(body)) + msg := fmt.Sprintf("API returned status %d: %s", resp.StatusCode, string(body)) + return nil, infraerrors.New(http.StatusInternalServerError, "UPSTREAM_ERROR", msg) } var usageResp service.ClaudeUsageResponse From 05527b13db3a562a87e4cf378f0680071f8668ba Mon Sep 17 00:00:00 2001 From: erio Date: Thu, 5 Mar 2026 20:54:37 +0800 Subject: [PATCH 02/53] feat: add quota limit for API key accounts - Add configurable spending limit (quota_limit) for apikey-type accounts - Atomic quota accumulation via PostgreSQL JSONB operations on TotalCost - Scheduler filters out over-quota accounts with outbox-triggered snapshot refresh - Display quota usage ($used / $limit) in account capacity column - Add "Reset Quota" action in account menu to reset usage to zero - Editing account settings preserves quota_used (no accidental reset) - Covers all 3 billing paths: Anthropic, Gemini, OpenAI RecordUsage chore: bump version to 0.1.90.4 --- .../internal/handler/admin/account_handler.go | 23 +++++++ .../handler/admin/admin_service_stub_test.go | 4 ++ backend/internal/handler/dto/mappers.go | 11 ++++ backend/internal/handler/dto/types.go | 4 ++ .../handler/sora_client_handler_test.go | 8 +++ .../handler/sora_gateway_handler_test.go | 8 +++ backend/internal/repository/account_repo.go | 57 ++++++++++++++++++ backend/internal/server/routes/admin.go | 1 + backend/internal/service/account.go | 32 ++++++++++ backend/internal/service/account_service.go | 4 ++ .../service/account_service_delete_test.go | 8 +++ backend/internal/service/admin_service.go | 9 +++ .../service/gateway_multiplatform_test.go | 8 +++ backend/internal/service/gateway_service.go | 53 ++++++++++++++-- .../service/gemini_multiplatform_test.go | 8 +++ .../service/openai_gateway_service.go | 7 +++ frontend/src/api/admin/accounts.ts | 13 ++++ .../account/AccountCapacityCell.vue | 60 +++++++++++++++++++ .../components/account/EditAccountModal.vue | 42 +++++++++++++ .../admin/account/AccountActionMenu.vue | 12 +++- frontend/src/i18n/locales/en.ts | 8 +++ frontend/src/i18n/locales/zh.ts | 8 +++ frontend/src/types/index.ts | 4 ++ frontend/src/views/admin/AccountsView.vue | 12 +++- 24 files changed, 398 insertions(+), 6 deletions(-) diff --git a/backend/internal/handler/admin/account_handler.go b/backend/internal/handler/admin/account_handler.go index 14f9e05d6a..c469a9979b 100644 --- a/backend/internal/handler/admin/account_handler.go +++ b/backend/internal/handler/admin/account_handler.go @@ -1328,6 +1328,29 @@ func (h *AccountHandler) ClearRateLimit(c *gin.Context) { response.Success(c, h.buildAccountResponseWithRuntime(c.Request.Context(), account)) } +// ResetQuota handles resetting account quota usage +// POST /api/v1/admin/accounts/:id/reset-quota +func (h *AccountHandler) ResetQuota(c *gin.Context) { + accountID, err := strconv.ParseInt(c.Param("id"), 10, 64) + if err != nil { + response.BadRequest(c, "Invalid account ID") + return + } + + if err := h.adminService.ResetAccountQuota(c.Request.Context(), accountID); err != nil { + response.InternalError(c, "Failed to reset account quota: "+err.Error()) + return + } + + account, err := h.adminService.GetAccount(c.Request.Context(), accountID) + if err != nil { + response.ErrorFrom(c, err) + return + } + + response.Success(c, h.buildAccountResponseWithRuntime(c.Request.Context(), account)) +} + // GetTempUnschedulable handles getting temporary unschedulable status // GET /api/v1/admin/accounts/:id/temp-unschedulable func (h *AccountHandler) GetTempUnschedulable(c *gin.Context) { diff --git a/backend/internal/handler/admin/admin_service_stub_test.go b/backend/internal/handler/admin/admin_service_stub_test.go index f3b99ddbe0..84a9f10252 100644 --- a/backend/internal/handler/admin/admin_service_stub_test.go +++ b/backend/internal/handler/admin/admin_service_stub_test.go @@ -425,5 +425,9 @@ func (s *stubAdminService) AdminUpdateAPIKeyGroupID(ctx context.Context, keyID i return nil, service.ErrAPIKeyNotFound } +func (s *stubAdminService) ResetAccountQuota(ctx context.Context, id int64) error { + return nil +} + // Ensure stub implements interface. var _ service.AdminService = (*stubAdminService)(nil) diff --git a/backend/internal/handler/dto/mappers.go b/backend/internal/handler/dto/mappers.go index fe2a1d773b..9e6fa4bc14 100644 --- a/backend/internal/handler/dto/mappers.go +++ b/backend/internal/handler/dto/mappers.go @@ -248,6 +248,17 @@ func AccountFromServiceShallow(a *service.Account) *Account { } } + // 提取 API Key 账号配额限制（仅 apikey 类型有效） + if a.Type == service.AccountTypeAPIKey { + if limit := a.GetQuotaLimit(); limit > 0 { + out.QuotaLimit = &limit + } + used := a.GetQuotaUsed() + if out.QuotaLimit != nil { + out.QuotaUsed = &used + } + } + return out } diff --git a/backend/internal/handler/dto/types.go b/backend/internal/handler/dto/types.go index 920615f708..a87db3c4f4 100644 --- a/backend/internal/handler/dto/types.go +++ b/backend/internal/handler/dto/types.go @@ -185,6 +185,10 @@ type Account struct { CacheTTLOverrideEnabled *bool `json:"cache_ttl_override_enabled,omitempty"` CacheTTLOverrideTarget *string `json:"cache_ttl_override_target,omitempty"` + // API Key 账号配额限制 + QuotaLimit *float64 `json:"quota_limit,omitempty"` + QuotaUsed *float64 `json:"quota_used,omitempty"` + Proxy *Proxy `json:"proxy,omitempty"` AccountGroups []AccountGroup `json:"account_groups,omitempty"` diff --git a/backend/internal/handler/sora_client_handler_test.go b/backend/internal/handler/sora_client_handler_test.go index d2d9790d26..d2a849b1cb 100644 --- a/backend/internal/handler/sora_client_handler_test.go +++ b/backend/internal/handler/sora_client_handler_test.go @@ -2132,6 +2132,14 @@ func (r *stubAccountRepoForHandler) BulkUpdate(context.Context, []int64, service return 0, nil } +func (r *stubAccountRepoForHandler) IncrementQuotaUsed(context.Context, int64, float64) error { + return nil +} + +func (r *stubAccountRepoForHandler) ResetQuotaUsed(context.Context, int64) error { + return nil +} + // ==================== Stub: SoraClient (用于 SoraGatewayService) ==================== var _ service.SoraClient = (*stubSoraClientForHandler)(nil) diff --git a/backend/internal/handler/sora_gateway_handler_test.go b/backend/internal/handler/sora_gateway_handler_test.go index b76ab67dac..637462ad50 100644 --- a/backend/internal/handler/sora_gateway_handler_test.go +++ b/backend/internal/handler/sora_gateway_handler_test.go @@ -216,6 +216,14 @@ func (r *stubAccountRepo) BulkUpdate(ctx context.Context, ids []int64, updates s return 0, nil } +func (r *stubAccountRepo) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return nil +} + +func (r *stubAccountRepo) ResetQuotaUsed(ctx context.Context, id int64) error { + return nil +} + func (r *stubAccountRepo) listSchedulable() []service.Account { var result []service.Account for _, acc := range r.accounts { diff --git a/backend/internal/repository/account_repo.go b/backend/internal/repository/account_repo.go index 6f0c5424aa..8fd819c731 100644 --- a/backend/internal/repository/account_repo.go +++ b/backend/internal/repository/account_repo.go @@ -1657,3 +1657,60 @@ func (r *accountRepository) FindByExtraField(ctx context.Context, key string, va return r.accountsToService(ctx, accounts) } + +// IncrementQuotaUsed 原子递增账号的 extra.quota_used 字段 +func (r *accountRepository) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + rows, err := r.sql.QueryContext(ctx, + `UPDATE accounts SET extra = jsonb_set( + COALESCE(extra, '{}'::jsonb), + '{quota_used}', + to_jsonb(COALESCE((extra->>'quota_used')::numeric, 0) + $1) + ), updated_at = NOW() + WHERE id = $2 AND deleted_at IS NULL + RETURNING + COALESCE((extra->>'quota_used')::numeric, 0), + COALESCE((extra->>'quota_limit')::numeric, 0)`, + amount, id) + if err != nil { + return err + } + defer func() { _ = rows.Close() }() + + var newUsed, limit float64 + if rows.Next() { + if err := rows.Scan(&newUsed, &limit); err != nil { + return err + } + } + if err := rows.Err(); err != nil { + return err + } + + // 配额刚超限时触发调度快照刷新，使账号及时从调度候选中移除 + if limit > 0 && newUsed >= limit && (newUsed-amount) < limit { + if err := enqueueSchedulerOutbox(ctx, r.sql, service.SchedulerOutboxEventAccountChanged, &id, nil, nil); err != nil { + logger.LegacyPrintf("repository.account", "[SchedulerOutbox] enqueue quota exceeded failed: account=%d err=%v", id, err) + } + } + return nil +} + +// ResetQuotaUsed 重置账号的 extra.quota_used 为 0 +func (r *accountRepository) ResetQuotaUsed(ctx context.Context, id int64) error { + _, err := r.sql.ExecContext(ctx, + `UPDATE accounts SET extra = jsonb_set( + COALESCE(extra, '{}'::jsonb), + '{quota_used}', + '0'::jsonb + ), updated_at = NOW() + WHERE id = $1 AND deleted_at IS NULL`, + id) + if err != nil { + return err + } + // 重置配额后触发调度快照刷新，使账号重新参与调度 + if err := enqueueSchedulerOutbox(ctx, r.sql, service.SchedulerOutboxEventAccountChanged, &id, nil, nil); err != nil { + logger.LegacyPrintf("repository.account", "[SchedulerOutbox] enqueue quota reset failed: account=%d err=%v", id, err) + } + return nil +} diff --git a/backend/internal/server/routes/admin.go b/backend/internal/server/routes/admin.go index e9f9bf62af..2e53feb36f 100644 --- a/backend/internal/server/routes/admin.go +++ b/backend/internal/server/routes/admin.go @@ -252,6 +252,7 @@ func registerAccountRoutes(admin *gin.RouterGroup, h *handler.Handlers) { accounts.GET("/:id/today-stats", h.Admin.Account.GetTodayStats) accounts.POST("/today-stats/batch", h.Admin.Account.GetBatchTodayStats) accounts.POST("/:id/clear-rate-limit", h.Admin.Account.ClearRateLimit) + accounts.POST("/:id/reset-quota", h.Admin.Account.ResetQuota) accounts.GET("/:id/temp-unschedulable", h.Admin.Account.GetTempUnschedulable) accounts.DELETE("/:id/temp-unschedulable", h.Admin.Account.ClearTempUnschedulable) accounts.POST("/:id/schedulable", h.Admin.Account.SetSchedulable) diff --git a/backend/internal/service/account.go b/backend/internal/service/account.go index 7d56b7541a..0f85c9bd8f 100644 --- a/backend/internal/service/account.go +++ b/backend/internal/service/account.go @@ -1117,6 +1117,38 @@ func (a *Account) GetCacheTTLOverrideTarget() string { return "5m" } +// GetQuotaLimit 获取 API Key 账号的配额限制（美元） +// 返回 0 表示未启用 +func (a *Account) GetQuotaLimit() float64 { + if a.Extra == nil { + return 0 + } + if v, ok := a.Extra["quota_limit"]; ok { + return parseExtraFloat64(v) + } + return 0 +} + +// GetQuotaUsed 获取 API Key 账号的已用配额（美元） +func (a *Account) GetQuotaUsed() float64 { + if a.Extra == nil { + return 0 + } + if v, ok := a.Extra["quota_used"]; ok { + return parseExtraFloat64(v) + } + return 0 +} + +// IsQuotaExceeded 检查 API Key 账号配额是否已超限 +func (a *Account) IsQuotaExceeded() bool { + limit := a.GetQuotaLimit() + if limit <= 0 { + return false + } + return a.GetQuotaUsed() >= limit +} + // GetWindowCostLimit 获取 5h 窗口费用阈值（美元） // 返回 0 表示未启用 func (a *Account) GetWindowCostLimit() float64 { diff --git a/backend/internal/service/account_service.go b/backend/internal/service/account_service.go index 18a70c5cc2..daa422123b 100644 --- a/backend/internal/service/account_service.go +++ b/backend/internal/service/account_service.go @@ -68,6 +68,10 @@ type AccountRepository interface { UpdateSessionWindow(ctx context.Context, id int64, start, end *time.Time, status string) error UpdateExtra(ctx context.Context, id int64, updates map[string]any) error BulkUpdate(ctx context.Context, ids []int64, updates AccountBulkUpdate) (int64, error) + // IncrementQuotaUsed 原子递增 API Key 账号的配额用量 + IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error + // ResetQuotaUsed 重置 API Key 账号的配额用量为 0 + ResetQuotaUsed(ctx context.Context, id int64) error } // AccountBulkUpdate describes the fields that can be updated in a bulk operation. diff --git a/backend/internal/service/account_service_delete_test.go b/backend/internal/service/account_service_delete_test.go index 768cf7b78e..c96b436f02 100644 --- a/backend/internal/service/account_service_delete_test.go +++ b/backend/internal/service/account_service_delete_test.go @@ -199,6 +199,14 @@ func (s *accountRepoStub) BulkUpdate(ctx context.Context, ids []int64, updates A panic("unexpected BulkUpdate call") } +func (s *accountRepoStub) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return nil +} + +func (s *accountRepoStub) ResetQuotaUsed(ctx context.Context, id int64) error { + return nil +} + // TestAccountService_Delete_NotFound 测试删除不存在的账号时返回正确的错误。 // 预期行为： // - ExistsByID 返回 false（账号不存在） diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index 67e7c783c1..2a2341b776 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -84,6 +84,7 @@ type AdminService interface { DeleteRedeemCode(ctx context.Context, id int64) error BatchDeleteRedeemCodes(ctx context.Context, ids []int64) (int64, error) ExpireRedeemCode(ctx context.Context, id int64) (*RedeemCode, error) + ResetAccountQuota(ctx context.Context, id int64) error } // CreateUserInput represents input for creating a new user via admin operations. @@ -1458,6 +1459,10 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U account.Credentials = input.Credentials } if len(input.Extra) > 0 { + // 保留 quota_used，防止编辑账号时意外重置配额用量 + if oldQuotaUsed, ok := account.Extra["quota_used"]; ok { + input.Extra["quota_used"] = oldQuotaUsed + } account.Extra = input.Extra } if input.ProxyID != nil { @@ -2439,3 +2444,7 @@ func (e *MixedChannelError) Error() string { return fmt.Sprintf("mixed_channel_warning: Group '%s' contains both %s and %s accounts. Using mixed channels in the same context may cause thinking block signature validation issues, which will fallback to non-thinking mode for historical messages.", e.GroupName, e.CurrentPlatform, e.OtherPlatform) } + +func (s *adminServiceImpl) ResetAccountQuota(ctx context.Context, id int64) error { + return s.accountRepo.ResetQuotaUsed(ctx, id) +} diff --git a/backend/internal/service/gateway_multiplatform_test.go b/backend/internal/service/gateway_multiplatform_test.go index 1cb3c61edb..320ceaa708 100644 --- a/backend/internal/service/gateway_multiplatform_test.go +++ b/backend/internal/service/gateway_multiplatform_test.go @@ -187,6 +187,14 @@ func (m *mockAccountRepoForPlatform) BulkUpdate(ctx context.Context, ids []int64 return 0, nil } +func (m *mockAccountRepoForPlatform) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return nil +} + +func (m *mockAccountRepoForPlatform) ResetQuotaUsed(ctx context.Context, id int64) error { + return nil +} + // Verify interface implementation var _ AccountRepository = (*mockAccountRepoForPlatform)(nil) diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index 132361f447..006d4bc38e 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -1228,6 +1228,10 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro modelScopeSkippedIDs = append(modelScopeSkippedIDs, account.ID) continue } + // 配额检查 + if !s.isAccountSchedulableForQuota(account) { + continue + } // 窗口费用检查（非粘性会话路径） if !s.isAccountSchedulableForWindowCost(ctx, account, false) { filteredWindowCost++ @@ -1260,6 +1264,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro s.isAccountAllowedForPlatform(stickyAccount, platform, useMixed) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, stickyAccount, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, stickyAccount, requestedModel) && + s.isAccountSchedulableForQuota(stickyAccount) && s.isAccountSchedulableForWindowCost(ctx, stickyAccount, true) && s.isAccountSchedulableForRPM(ctx, stickyAccount, true) { // 粘性会话窗口费用+RPM 检查 @@ -1416,6 +1421,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro s.isAccountAllowedForPlatform(account, platform, useMixed) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && + s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { // 粘性会话窗口费用+RPM 检查 @@ -1480,6 +1486,10 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) { continue } + // 配额检查 + if !s.isAccountSchedulableForQuota(acc) { + continue + } // 窗口费用检查（非粘性会话路径） if !s.isAccountSchedulableForWindowCost(ctx, acc, false) { continue @@ -2113,6 +2123,15 @@ func (s *GatewayService) withWindowCostPrefetch(ctx context.Context, accounts [] return context.WithValue(ctx, windowCostPrefetchContextKey, costs) } +// isAccountSchedulableForQuota 检查 API Key 账号是否在配额限制内 +// 仅适用于配置了 quota_limit 的 apikey 类型账号 +func (s *GatewayService) isAccountSchedulableForQuota(account *Account) bool { + if account.Type != AccountTypeAPIKey { + return true + } + return !account.IsQuotaExceeded() +} + // isAccountSchedulableForWindowCost 检查账号是否可根据窗口费用进行调度 // 仅适用于 Anthropic OAuth/SetupToken 账号 // 返回 true 表示可调度，false 表示不可调度 @@ -2590,7 +2609,7 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context, if clearSticky { _ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash) } - if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { + if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { if s.debugModelRoutingEnabled() { logger.LegacyPrintf("service.gateway", "[ModelRoutingDebug] legacy routed sticky hit: group_id=%v model=%s session=%s account=%d", derefGroupID(groupID), requestedModel, shortSessionHash(sessionHash), accountID) } @@ -2644,6 +2663,9 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context, if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) { continue } + if !s.isAccountSchedulableForQuota(acc) { + continue + } if !s.isAccountSchedulableForWindowCost(ctx, acc, false) { continue } @@ -2700,7 +2722,7 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context, if clearSticky { _ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash) } - if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { + if !clearSticky && s.isAccountInGroup(account, groupID) && account.Platform == platform && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { return account, nil } } @@ -2743,6 +2765,9 @@ func (s *GatewayService) selectAccountForModelWithPlatform(ctx context.Context, if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) { continue } + if !s.isAccountSchedulableForQuota(acc) { + continue + } if !s.isAccountSchedulableForWindowCost(ctx, acc, false) { continue } @@ -2818,7 +2843,7 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g if clearSticky { _ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash) } - if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { + if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { if account.Platform == nativePlatform || (account.Platform == PlatformAntigravity && account.IsMixedSchedulingEnabled()) { if s.debugModelRoutingEnabled() { logger.LegacyPrintf("service.gateway", "[ModelRoutingDebug] legacy mixed routed sticky hit: group_id=%v model=%s session=%s account=%d", derefGroupID(groupID), requestedModel, shortSessionHash(sessionHash), accountID) @@ -2874,6 +2899,9 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) { continue } + if !s.isAccountSchedulableForQuota(acc) { + continue + } if !s.isAccountSchedulableForWindowCost(ctx, acc, false) { continue } @@ -2930,7 +2958,7 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g if clearSticky { _ = s.cache.DeleteSessionAccountID(ctx, derefGroupID(groupID), sessionHash) } - if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { + if !clearSticky && s.isAccountInGroup(account, groupID) && (requestedModel == "" || s.isModelSupportedByAccountWithContext(ctx, account, requestedModel)) && s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) && s.isAccountSchedulableForQuota(account) && s.isAccountSchedulableForWindowCost(ctx, account, true) && s.isAccountSchedulableForRPM(ctx, account, true) { if account.Platform == nativePlatform || (account.Platform == PlatformAntigravity && account.IsMixedSchedulingEnabled()) { return account, nil } @@ -2975,6 +3003,9 @@ func (s *GatewayService) selectAccountWithMixedScheduling(ctx context.Context, g if !s.isAccountSchedulableForModelSelection(ctx, acc, requestedModel) { continue } + if !s.isAccountSchedulableForQuota(acc) { + continue + } if !s.isAccountSchedulableForWindowCost(ctx, acc, false) { continue } @@ -6578,6 +6609,13 @@ func (s *GatewayService) RecordUsage(ctx context.Context, input *RecordUsageInpu s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) } + // 更新 API Key 账号配额用量 + if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { + if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { + slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err) + } + } + // Schedule batch update for account last_used_at s.deferredService.ScheduleLastUsedUpdate(account.ID) @@ -6775,6 +6813,13 @@ func (s *GatewayService) RecordUsageWithLongContext(ctx context.Context, input * s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) } + // 更新 API Key 账号配额用量 + if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { + if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { + slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err) + } + } + // Schedule batch update for account last_used_at s.deferredService.ScheduleLastUsedUpdate(account.ID) diff --git a/backend/internal/service/gemini_multiplatform_test.go b/backend/internal/service/gemini_multiplatform_test.go index 9476e98496..b0b804eb00 100644 --- a/backend/internal/service/gemini_multiplatform_test.go +++ b/backend/internal/service/gemini_multiplatform_test.go @@ -176,6 +176,14 @@ func (m *mockAccountRepoForGemini) BulkUpdate(ctx context.Context, ids []int64, return 0, nil } +func (m *mockAccountRepoForGemini) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return nil +} + +func (m *mockAccountRepoForGemini) ResetQuotaUsed(ctx context.Context, id int64) error { + return nil +} + // Verify interface implementation var _ AccountRepository = (*mockAccountRepoForGemini)(nil) diff --git a/backend/internal/service/openai_gateway_service.go b/backend/internal/service/openai_gateway_service.go index d92b2ecf77..6752d18b72 100644 --- a/backend/internal/service/openai_gateway_service.go +++ b/backend/internal/service/openai_gateway_service.go @@ -3502,6 +3502,13 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) } + // 更新 API Key 账号配额用量 + if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { + if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { + logger.LegacyPrintf("service.openai_gateway", "increment account quota used failed: account_id=%d cost=%f error=%v", account.ID, cost.TotalCost, err) + } + } + // Schedule batch update for account last_used_at s.deferredService.ScheduleLastUsedUpdate(account.ID) diff --git a/frontend/src/api/admin/accounts.ts b/frontend/src/api/admin/accounts.ts index 25bb7b7b71..5524e0cbe1 100644 --- a/frontend/src/api/admin/accounts.ts +++ b/frontend/src/api/admin/accounts.ts @@ -240,6 +240,18 @@ export async function clearRateLimit(id: number): Promise { return data } +/** + * Reset account quota usage + * @param id - Account ID + * @returns Updated account + */ +export async function resetAccountQuota(id: number): Promise { + const { data } = await apiClient.post( + `/admin/accounts/${id}/reset-quota` + ) + return data +} + /** * Get temporary unschedulable status * @param id - Account ID @@ -576,6 +588,7 @@ export const accountsAPI = { getTodayStats, getBatchTodayStats, clearRateLimit, + resetAccountQuota, getTempUnschedulableStatus, resetTempUnschedulable, setSchedulable, diff --git a/frontend/src/components/account/AccountCapacityCell.vue b/frontend/src/components/account/AccountCapacityCell.vue index 2a4babf200..2001b185e3 100644 --- a/frontend/src/components/account/AccountCapacityCell.vue +++ b/frontend/src/components/account/AccountCapacityCell.vue @@ -71,6 +71,24 @@ {{ rpmStrategyTag }} + + +
+ + + + + ${{ formatCost(currentQuotaUsed) }} + / + ${{ formatCost(account.quota_limit) }} + +
@@ -286,6 +304,48 @@ const rpmTooltip = computed(() => { } }) +// 是否显示配额限制（仅 apikey 类型且设置了 quota_limit） +const showQuotaLimit = computed(() => { + return ( + props.account.type === 'apikey' && + props.account.quota_limit !== undefined && + props.account.quota_limit !== null && + props.account.quota_limit > 0 + ) +}) + +// 当前已用配额 +const currentQuotaUsed = computed(() => props.account.quota_used ?? 0) + +// 配额状态样式 +const quotaClass = computed(() => { + if (!showQuotaLimit.value) return '' + + const used = currentQuotaUsed.value + const limit = props.account.quota_limit || 0 + + if (used >= limit) { + return 'bg-red-100 text-red-700 dark:bg-red-900/30 dark:text-red-400' + } + if (used >= limit * 0.8) { + return 'bg-yellow-100 text-yellow-700 dark:bg-yellow-900/30 dark:text-yellow-400' + } + return 'bg-emerald-100 text-emerald-700 dark:bg-emerald-900/30 dark:text-emerald-400' +}) + +// 配额提示文字 +const quotaTooltip = computed(() => { + if (!showQuotaLimit.value) return '' + + const used = currentQuotaUsed.value + const limit = props.account.quota_limit || 0 + + if (used >= limit) { + return t('admin.accounts.capacity.quota.exceeded') + } + return t('admin.accounts.capacity.quota.normal') +}) + // 格式化费用显示 const formatCost = (value: number | null | undefined) => { if (value === null || value === undefined) return '0' diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 09b39bc060..665e4e950b 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -759,6 +759,26 @@ + +
+ +
+ $ + +
+

{{ t('admin.accounts.quotaLimitHint') }}

+
+
(OPENAI_WS_MODE_OFF const openaiAPIKeyResponsesWebSocketV2Mode = ref(OPENAI_WS_MODE_OFF) const codexCLIOnlyEnabled = ref(false) const anthropicPassthroughEnabled = ref(false) +const editQuotaLimit = ref(null) const openAIWSModeOptions = computed(() => [ { value: OPENAI_WS_MODE_OFF, label: t('admin.accounts.openai.wsModeOff') }, // TODO: ctx_pool 选项暂时隐藏，待测试完成后恢复 @@ -1541,6 +1562,14 @@ watch( anthropicPassthroughEnabled.value = extra?.anthropic_passthrough === true } + // Load quota limit for apikey accounts + if (newAccount.type === 'apikey') { + const quotaVal = extra?.quota_limit as number | undefined + editQuotaLimit.value = (quotaVal && quotaVal > 0) ? quotaVal : null + } else { + editQuotaLimit.value = null + } + // Load antigravity model mapping (Antigravity 只支持映射模式) if (newAccount.platform === 'antigravity') { const credentials = newAccount.credentials as Record | undefined @@ -2283,6 +2312,19 @@ const handleSubmit = async () => { updatePayload.extra = newExtra } + // For apikey accounts, handle quota_limit in extra + if (props.account.type === 'apikey') { + const currentExtra = (updatePayload.extra as Record) || + (props.account.extra as Record) || {} + const newExtra: Record = { ...currentExtra } + if (editQuotaLimit.value != null && editQuotaLimit.value > 0) { + newExtra.quota_limit = editQuotaLimit.value + } else { + delete newExtra.quota_limit + } + updatePayload.extra = newExtra + } + const canContinue = await ensureAntigravityMixedChannelConfirmed(async () => { await submitUpdateAccount(accountID, updatePayload) }) diff --git a/frontend/src/components/admin/account/AccountActionMenu.vue b/frontend/src/components/admin/account/AccountActionMenu.vue index fbff0bede5..02596b9f7d 100644 --- a/frontend/src/components/admin/account/AccountActionMenu.vue +++ b/frontend/src/components/admin/account/AccountActionMenu.vue @@ -41,6 +41,10 @@ {{ t('admin.accounts.clearRateLimit') }} +
@@ -55,7 +59,7 @@ import { Icon } from '@/components/icons' import type { Account } from '@/types' const props = defineProps<{ show: boolean; account: Account | null; position: { top: number; left: number } | null }>() -const emit = defineEmits(['close', 'test', 'stats', 'schedule', 'reauth', 'refresh-token', 'reset-status', 'clear-rate-limit']) +const emit = defineEmits(['close', 'test', 'stats', 'schedule', 'reauth', 'refresh-token', 'reset-status', 'clear-rate-limit', 'reset-quota']) const { t } = useI18n() const isRateLimited = computed(() => { if (props.account?.rate_limit_reset_at && new Date(props.account.rate_limit_reset_at) > new Date()) { @@ -71,6 +75,12 @@ const isRateLimited = computed(() => { return false }) const isOverloaded = computed(() => props.account?.overload_until && new Date(props.account.overload_until) > new Date()) +const hasQuotaLimit = computed(() => { + return props.account?.type === 'apikey' && + props.account?.quota_limit !== undefined && + props.account?.quota_limit !== null && + props.account?.quota_limit > 0 +}) const handleKeydown = (event: KeyboardEvent) => { if (event.key === 'Escape') emit('close') diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index 055998a7fe..4be07f8508 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -1734,6 +1734,10 @@ export default { stickyExemptWarning: 'RPM limit (Sticky Exempt) - Approaching limit', stickyExemptOver: 'RPM limit (Sticky Exempt) - Over limit, sticky only' }, + quota: { + exceeded: 'Quota exceeded, account paused', + normal: 'Quota normal' + }, }, tempUnschedulable: { title: 'Temp Unschedulable', @@ -1779,6 +1783,10 @@ export default { } }, clearRateLimit: 'Clear Rate Limit', + resetQuota: 'Reset Quota', + quotaLimit: 'Quota Limit', + quotaLimitPlaceholder: '0 means unlimited', + quotaLimitHint: 'Set max spending limit (USD). Account will be paused when reached. Changing limit won\'t reset usage.', testConnection: 'Test Connection', reAuthorize: 'Re-Authorize', refreshToken: 'Refresh Token', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index cc203adbd6..ce262a2ace 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -1784,8 +1784,16 @@ export default { stickyExemptWarning: 'RPM 限制 (粘性豁免) - 接近阈值', stickyExemptOver: 'RPM 限制 (粘性豁免) - 超限，仅粘性会话' }, + quota: { + exceeded: '配额已用完，账号暂停调度', + normal: '配额正常' + }, }, clearRateLimit: '清除速率限制', + resetQuota: '重置配额', + quotaLimit: '配额限制', + quotaLimitPlaceholder: '0 表示不限制', + quotaLimitHint: '设置最大使用额度（美元），达到后账号暂停调度。修改限额不会重置已用额度。', testConnection: '测试连接', reAuthorize: '重新授权', refreshToken: '刷新令牌', diff --git a/frontend/src/types/index.ts b/frontend/src/types/index.ts index 915822f079..5a10abffa0 100644 --- a/frontend/src/types/index.ts +++ b/frontend/src/types/index.ts @@ -705,6 +705,10 @@ export interface Account { cache_ttl_override_enabled?: boolean | null cache_ttl_override_target?: string | null + // API Key 账号配额限制 + quota_limit?: number | null + quota_used?: number | null + // 运行时状态（仅当启用对应限制时返回） current_window_cost?: number | null // 当前窗口费用 active_sessions?: number | null // 当前活跃会话数 diff --git a/frontend/src/views/admin/AccountsView.vue b/frontend/src/views/admin/AccountsView.vue index 146b264701..0173ea0a86 100644 --- a/frontend/src/views/admin/AccountsView.vue +++ b/frontend/src/views/admin/AccountsView.vue @@ -261,7 +261,7 @@ - + @@ -1125,6 +1125,16 @@ const handleClearRateLimit = async (a: Account) => { console.error('Failed to clear rate limit:', error) } } +const handleResetQuota = async (a: Account) => { + try { + const updated = await adminAPI.accounts.resetAccountQuota(a.id) + patchAccountInList(updated) + enterAutoRefreshSilentWindow() + appStore.showSuccess(t('common.success')) + } catch (error) { + console.error('Failed to reset quota:', error) + } +} const handleDelete = (a: Account) => { deletingAcc.value = a; showDeleteDialog.value = true } const confirmDelete = async () => { if(!deletingAcc.value) return; try { await adminAPI.accounts.delete(deletingAcc.value.id); showDeleteDialog.value = false; deletingAcc.value = null; reload() } catch (error) { console.error('Failed to delete account:', error) } } const handleToggleSchedulable = async (a: Account) => { From 1893b0eb306555bac82659d347248d67b5af8288 Mon Sep 17 00:00:00 2001 From: erio Date: Thu, 5 Mar 2026 22:01:40 +0800 Subject: [PATCH 03/53] feat: restyle API Key quota limit UI to card/toggle format - Redesign quota limit section with card layout and toggle switch - Add watch to clear quota value when toggle is disabled - Add i18n keys for toggle labels and hints (zh/en) - Bump version to 0.1.90.5 --- .../components/account/EditAccountModal.vue | 75 +++++++++++++++---- frontend/src/i18n/locales/en.ts | 4 + frontend/src/i18n/locales/zh.ts | 4 + 3 files changed, 69 insertions(+), 14 deletions(-) diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 665e4e950b..ebdfb3b4c8 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -762,21 +762,58 @@
- -
- $ - +
+

{{ t('admin.accounts.quotaLimit') }}

+

+ {{ t('admin.accounts.quotaLimitHint') }} +

+
+ +
+
+
+ +

+ {{ t('admin.accounts.quotaLimitToggleHint') }} +

+
+ +
+ +
+
+ +
+ $ + +
+

{{ t('admin.accounts.quotaLimitAmountHint') }}

+
+
-

{{ t('admin.accounts.quotaLimitHint') }}

@@ -1407,6 +1444,7 @@ const openaiAPIKeyResponsesWebSocketV2Mode = ref(OPENAI_WS_MODE_OF const codexCLIOnlyEnabled = ref(false) const anthropicPassthroughEnabled = ref(false) const editQuotaLimit = ref(null) +const quotaLimitEnabled = ref(false) const openAIWSModeOptions = computed(() => [ { value: OPENAI_WS_MODE_OFF, label: t('admin.accounts.openai.wsModeOff') }, // TODO: ctx_pool 选项暂时隐藏，待测试完成后恢复 @@ -1435,6 +1473,13 @@ const isOpenAIModelRestrictionDisabled = computed(() => props.account?.platform === 'openai' && openaiPassthroughEnabled.value ) +// When quota limit toggle is turned off, clear the value +watch(quotaLimitEnabled, (enabled) => { + if (!enabled) { + editQuotaLimit.value = null + } +}) + // Computed: current preset mappings based on platform const presetMappings = computed(() => getPresetMappingsByPlatform(props.account?.platform || 'anthropic')) const tempUnschedPresets = computed(() => [ @@ -1565,8 +1610,10 @@ watch( // Load quota limit for apikey accounts if (newAccount.type === 'apikey') { const quotaVal = extra?.quota_limit as number | undefined + quotaLimitEnabled.value = !!(quotaVal && quotaVal > 0) editQuotaLimit.value = (quotaVal && quotaVal > 0) ? quotaVal : null } else { + quotaLimitEnabled.value = false editQuotaLimit.value = null } @@ -2317,7 +2364,7 @@ const handleSubmit = async () => { const currentExtra = (updatePayload.extra as Record) || (props.account.extra as Record) || {} const newExtra: Record = { ...currentExtra } - if (editQuotaLimit.value != null && editQuotaLimit.value > 0) { + if (quotaLimitEnabled.value && editQuotaLimit.value != null && editQuotaLimit.value > 0) { newExtra.quota_limit = editQuotaLimit.value } else { delete newExtra.quota_limit diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index 4be07f8508..84e83850d6 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -1787,6 +1787,10 @@ export default { quotaLimit: 'Quota Limit', quotaLimitPlaceholder: '0 means unlimited', quotaLimitHint: 'Set max spending limit (USD). Account will be paused when reached. Changing limit won\'t reset usage.', + quotaLimitToggle: 'Enable Quota Limit', + quotaLimitToggleHint: 'When enabled, account will be paused when usage reaches the set limit', + quotaLimitAmount: 'Limit Amount', + quotaLimitAmountHint: 'Maximum spending limit (USD). Account will be auto-paused when reached. Changing limit won\'t reset usage.', testConnection: 'Test Connection', reAuthorize: 'Re-Authorize', refreshToken: 'Refresh Token', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index ce262a2ace..c37b497a4c 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -1794,6 +1794,10 @@ export default { quotaLimit: '配额限制', quotaLimitPlaceholder: '0 表示不限制', quotaLimitHint: '设置最大使用额度（美元），达到后账号暂停调度。修改限额不会重置已用额度。', + quotaLimitToggle: '启用配额限制', + quotaLimitToggleHint: '开启后，当账号用量达到设定额度时自动暂停调度', + quotaLimitAmount: '限额金额', + quotaLimitAmountHint: '账号最大可用额度（美元），达到后自动暂停。修改限额不会重置已用额度。', testConnection: '测试连接', reAuthorize: '重新授权', refreshToken: '刷新令牌', From c826ac28ef61c96421f0807dbe72919743764f75 Mon Sep 17 00:00:00 2001 From: erio Date: Thu, 5 Mar 2026 22:13:56 +0800 Subject: [PATCH 04/53] refactor: extract QuotaLimitCard component for reuse in create and edit modals - Extract quota limit card/toggle UI into QuotaLimitCard.vue component - Use v-model pattern for clean parent-child data flow - Integrate into both EditAccountModal and CreateAccountModal - All apikey accounts (all platforms) now support quota limit on creation - Bump version to 0.1.90.6 --- .../components/account/CreateAccountModal.vue | 13 ++- .../components/account/EditAccountModal.vue | 69 +------------- .../src/components/account/QuotaLimitCard.vue | 92 +++++++++++++++++++ 3 files changed, 107 insertions(+), 67 deletions(-) create mode 100644 frontend/src/components/account/QuotaLimitCard.vue diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index 225b91f5f9..d58addb976 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -1227,6 +1227,9 @@
+ + +
@@ -2337,6 +2340,7 @@ import Icon from '@/components/icons/Icon.vue' import ProxySelector from '@/components/common/ProxySelector.vue' import GroupSelector from '@/components/common/GroupSelector.vue' import ModelWhitelistSelector from '@/components/account/ModelWhitelistSelector.vue' +import QuotaLimitCard from '@/components/account/QuotaLimitCard.vue' import { applyInterceptWarmup } from '@/components/account/credentialsBuilder' import { formatDateTimeLocalInput, parseDateTimeLocalInput } from '@/utils/format' import { createStableObjectKeyResolver } from '@/utils/stableObjectKey' @@ -2460,6 +2464,7 @@ const accountCategory = ref<'oauth-based' | 'apikey'>('oauth-based') // UI selec const addMethod = ref('oauth') // For oauth-based: 'oauth' or 'setup-token' const apiKeyBaseUrl = ref('https://api.anthropic.com') const apiKeyValue = ref('') +const editQuotaLimit = ref(null) const modelMappings = ref([]) const modelRestrictionMode = ref<'whitelist' | 'mapping'>('whitelist') const allowedModels = ref([]) @@ -3120,6 +3125,7 @@ const resetForm = () => { addMethod.value = 'oauth' apiKeyBaseUrl.value = 'https://api.anthropic.com' apiKeyValue.value = '' + editQuotaLimit.value = null modelMappings.value = [] modelRestrictionMode.value = 'whitelist' allowedModels.value = [...claudeModels] // Default fill related models @@ -3533,13 +3539,18 @@ const createAccountAndFinish = async ( if (!applyTempUnschedConfig(credentials)) { return } + // Inject quota_limit for apikey accounts + let finalExtra = extra + if (type === 'apikey' && editQuotaLimit.value != null && editQuotaLimit.value > 0) { + finalExtra = { ...(extra || {}), quota_limit: editQuotaLimit.value } + } await doCreateAccount({ name: form.name, notes: form.notes, platform, type, credentials, - extra, + extra: finalExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, priority: form.priority, diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index ebdfb3b4c8..6617bc3374 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -760,61 +760,7 @@
-
-
-

{{ t('admin.accounts.quotaLimit') }}

-

- {{ t('admin.accounts.quotaLimitHint') }} -

-
- -
-
-
- -

- {{ t('admin.accounts.quotaLimitToggleHint') }} -

-
- -
- -
-
- -
- $ - -
-

{{ t('admin.accounts.quotaLimitAmountHint') }}

-
-
-
-
+
(OPENAI_WS_MODE_OF const codexCLIOnlyEnabled = ref(false) const anthropicPassthroughEnabled = ref(false) const editQuotaLimit = ref(null) -const quotaLimitEnabled = ref(false) const openAIWSModeOptions = computed(() => [ { value: OPENAI_WS_MODE_OFF, label: t('admin.accounts.openai.wsModeOff') }, // TODO: ctx_pool 选项暂时隐藏，待测试完成后恢复 @@ -1473,13 +1419,6 @@ const isOpenAIModelRestrictionDisabled = computed(() => props.account?.platform === 'openai' && openaiPassthroughEnabled.value ) -// When quota limit toggle is turned off, clear the value -watch(quotaLimitEnabled, (enabled) => { - if (!enabled) { - editQuotaLimit.value = null - } -}) - // Computed: current preset mappings based on platform const presetMappings = computed(() => getPresetMappingsByPlatform(props.account?.platform || 'anthropic')) const tempUnschedPresets = computed(() => [ @@ -1610,10 +1549,8 @@ watch( // Load quota limit for apikey accounts if (newAccount.type === 'apikey') { const quotaVal = extra?.quota_limit as number | undefined - quotaLimitEnabled.value = !!(quotaVal && quotaVal > 0) editQuotaLimit.value = (quotaVal && quotaVal > 0) ? quotaVal : null } else { - quotaLimitEnabled.value = false editQuotaLimit.value = null } @@ -2364,7 +2301,7 @@ const handleSubmit = async () => { const currentExtra = (updatePayload.extra as Record) || (props.account.extra as Record) || {} const newExtra: Record = { ...currentExtra } - if (quotaLimitEnabled.value && editQuotaLimit.value != null && editQuotaLimit.value > 0) { + if (editQuotaLimit.value != null && editQuotaLimit.value > 0) { newExtra.quota_limit = editQuotaLimit.value } else { delete newExtra.quota_limit diff --git a/frontend/src/components/account/QuotaLimitCard.vue b/frontend/src/components/account/QuotaLimitCard.vue new file mode 100644 index 0000000000..1be73a257b --- /dev/null +++ b/frontend/src/components/account/QuotaLimitCard.vue @@ -0,0 +1,92 @@ + + + From c26f93c4a04a853d6507f1582c403b1027f58bc2 Mon Sep 17 00:00:00 2001 From: erio Date: Thu, 5 Mar 2026 22:10:00 +0800 Subject: [PATCH 05/53] fix: route antigravity apikey account test to native protocol Antigravity APIKey accounts were incorrectly routed to testAntigravityAccountConnection which calls AntigravityTokenProvider, but the token provider only handles OAuth and Upstream types, causing "not an antigravity oauth account" error. Extract routeAntigravityTest to route APIKey accounts to native Claude/Gemini test paths based on model prefix, matching the gateway_handler routing logic for normal requests. --- backend/internal/service/account_test_service.go | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/backend/internal/service/account_test_service.go b/backend/internal/service/account_test_service.go index 99046e30a1..9557e17564 100644 --- a/backend/internal/service/account_test_service.go +++ b/backend/internal/service/account_test_service.go @@ -180,7 +180,7 @@ func (s *AccountTestService) TestAccountConnection(c *gin.Context, accountID int } if account.Platform == PlatformAntigravity { - return s.testAntigravityAccountConnection(c, account, modelID) + return s.routeAntigravityTest(c, account, modelID) } if account.Platform == PlatformSora { @@ -1177,6 +1177,18 @@ func truncateSoraErrorBody(body []byte, max int) string { return soraerror.TruncateBody(body, max) } +// routeAntigravityTest 路由 Antigravity 账号的测试请求。 +// APIKey 类型走原生协议（与 gateway_handler 路由一致），OAuth/Upstream 走 CRS 中转。 +func (s *AccountTestService) routeAntigravityTest(c *gin.Context, account *Account, modelID string) error { + if account.Type == AccountTypeAPIKey { + if strings.HasPrefix(modelID, "gemini-") { + return s.testGeminiAccountConnection(c, account, modelID) + } + return s.testClaudeAccountConnection(c, account, modelID) + } + return s.testAntigravityAccountConnection(c, account, modelID) +} + // testAntigravityAccountConnection tests an Antigravity account's connection // 支持 Claude 和 Gemini 两种协议，使用非流式请求 func (s *AccountTestService) testAntigravityAccountConnection(c *gin.Context, account *Account, modelID string) error { From 02dea7b09b2126f2a91f14398279e88c7031bcf2 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 00:37:37 +0800 Subject: [PATCH 06/53] refactor: unify post-usage billing logic and fix account quota calculation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Extract postUsageBilling() to consolidate billing logic across GatewayService.RecordUsage, RecordUsageWithLongContext, and OpenAIGatewayService.RecordUsage, eliminating ~120 lines of duplicated code - Fix account quota to use TotalCost × accountRateMultiplier (was using raw TotalCost, inconsistent with account cost stats) - Fix RecordUsageWithLongContext API Key quota only updating in balance mode (now updates regardless of billing type) - Fix WebSocket client disconnect detection on Windows by adding "an established connection was aborted" to known disconnect errors --- backend/internal/service/gateway_service.go | 192 ++++++++++-------- .../service/openai_gateway_service.go | 57 ++---- .../internal/service/openai_ws_forwarder.go | 3 +- 3 files changed, 131 insertions(+), 121 deletions(-) diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index 006d4bc38e..177c46314c 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -6410,6 +6410,89 @@ type APIKeyQuotaUpdater interface { UpdateRateLimitUsage(ctx context.Context, apiKeyID int64, cost float64) error } +// postUsageBillingParams 统一扣费所需的参数 +type postUsageBillingParams struct { + Cost *CostBreakdown + User *User + APIKey *APIKey + Account *Account + Subscription *UserSubscription + IsSubscriptionBill bool + AccountRateMultiplier float64 + APIKeyService APIKeyQuotaUpdater +} + +// postUsageBilling 统一处理使用量记录后的扣费逻辑： +// - 订阅/余额扣费 +// - API Key 配额更新 +// - API Key 限速用量更新 +// - 账号配额用量更新（账号口径：TotalCost × 账号计费倍率） +func postUsageBilling(ctx context.Context, p *postUsageBillingParams, deps *billingDeps) { + cost := p.Cost + + // 1. 订阅 / 余额扣费 + if p.IsSubscriptionBill { + if cost.TotalCost > 0 { + if err := deps.userSubRepo.IncrementUsage(ctx, p.Subscription.ID, cost.TotalCost); err != nil { + slog.Error("increment subscription usage failed", "subscription_id", p.Subscription.ID, "error", err) + } + deps.billingCacheService.QueueUpdateSubscriptionUsage(p.User.ID, *p.APIKey.GroupID, cost.TotalCost) + } + } else { + if cost.ActualCost > 0 { + if err := deps.userRepo.DeductBalance(ctx, p.User.ID, cost.ActualCost); err != nil { + slog.Error("deduct balance failed", "user_id", p.User.ID, "error", err) + } + deps.billingCacheService.QueueDeductBalance(p.User.ID, cost.ActualCost) + } + } + + // 2. API Key 配额 + if cost.ActualCost > 0 && p.APIKey.Quota > 0 && p.APIKeyService != nil { + if err := p.APIKeyService.UpdateQuotaUsed(ctx, p.APIKey.ID, cost.ActualCost); err != nil { + slog.Error("update api key quota failed", "api_key_id", p.APIKey.ID, "error", err) + } + } + + // 3. API Key 限速用量 + if cost.ActualCost > 0 && p.APIKey.HasRateLimits() && p.APIKeyService != nil { + if err := p.APIKeyService.UpdateRateLimitUsage(ctx, p.APIKey.ID, cost.ActualCost); err != nil { + slog.Error("update api key rate limit usage failed", "api_key_id", p.APIKey.ID, "error", err) + } + deps.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(p.APIKey.ID, cost.ActualCost) + } + + // 4. 账号配额用量（账号口径：TotalCost × 账号计费倍率） + if cost.TotalCost > 0 && p.Account.Type == AccountTypeAPIKey && p.Account.GetQuotaLimit() > 0 { + accountCost := cost.TotalCost * p.AccountRateMultiplier + if err := deps.accountRepo.IncrementQuotaUsed(ctx, p.Account.ID, accountCost); err != nil { + slog.Error("increment account quota used failed", "account_id", p.Account.ID, "cost", accountCost, "error", err) + } + } + + // 5. 更新账号最近使用时间 + deps.deferredService.ScheduleLastUsedUpdate(p.Account.ID) +} + +// billingDeps 扣费逻辑依赖的服务（由各 gateway service 提供） +type billingDeps struct { + accountRepo AccountRepository + userRepo UserRepository + userSubRepo UserSubscriptionRepository + billingCacheService *BillingCacheService + deferredService *DeferredService +} + +func (s *GatewayService) billingDeps() *billingDeps { + return &billingDeps{ + accountRepo: s.accountRepo, + userRepo: s.userRepo, + userSubRepo: s.userSubRepo, + billingCacheService: s.billingCacheService, + deferredService: s.deferredService, + } +} + // RecordUsage 记录使用量并扣费（或更新订阅用量） func (s *GatewayService) RecordUsage(ctx context.Context, input *RecordUsageInput) error { result := input.Result @@ -6573,52 +6656,21 @@ func (s *GatewayService) RecordUsage(ctx context.Context, input *RecordUsageInpu shouldBill := inserted || err != nil - // 根据计费类型执行扣费 - if isSubscriptionBilling { - // 订阅模式：更新订阅用量（使用 TotalCost 原始费用，不考虑倍率） - if shouldBill && cost.TotalCost > 0 { - if err := s.userSubRepo.IncrementUsage(ctx, subscription.ID, cost.TotalCost); err != nil { - logger.LegacyPrintf("service.gateway", "Increment subscription usage failed: %v", err) - } - // 异步更新订阅缓存 - s.billingCacheService.QueueUpdateSubscriptionUsage(user.ID, *apiKey.GroupID, cost.TotalCost) - } + if shouldBill { + postUsageBilling(ctx, &postUsageBillingParams{ + Cost: cost, + User: user, + APIKey: apiKey, + Account: account, + Subscription: subscription, + IsSubscriptionBill: isSubscriptionBilling, + AccountRateMultiplier: accountRateMultiplier, + APIKeyService: input.APIKeyService, + }, s.billingDeps()) } else { - // 余额模式：扣除用户余额（使用 ActualCost 考虑倍率后的费用） - if shouldBill && cost.ActualCost > 0 { - if err := s.userRepo.DeductBalance(ctx, user.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Deduct balance failed: %v", err) - } - // 异步更新余额缓存 - s.billingCacheService.QueueDeductBalance(user.ID, cost.ActualCost) - } + s.deferredService.ScheduleLastUsedUpdate(account.ID) } - // 更新 API Key 配额（如果设置了配额限制） - if shouldBill && cost.ActualCost > 0 && apiKey.Quota > 0 && input.APIKeyService != nil { - if err := input.APIKeyService.UpdateQuotaUsed(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Update API key quota failed: %v", err) - } - } - - // Update API Key rate limit usage - if shouldBill && cost.ActualCost > 0 && apiKey.HasRateLimits() && input.APIKeyService != nil { - if err := input.APIKeyService.UpdateRateLimitUsage(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Update API key rate limit usage failed: %v", err) - } - s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) - } - - // 更新 API Key 账号配额用量 - if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { - if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { - slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err) - } - } - - // Schedule batch update for account last_used_at - s.deferredService.ScheduleLastUsedUpdate(account.ID) - return nil } @@ -6778,51 +6830,21 @@ func (s *GatewayService) RecordUsageWithLongContext(ctx context.Context, input * shouldBill := inserted || err != nil - // 根据计费类型执行扣费 - if isSubscriptionBilling { - // 订阅模式：更新订阅用量（使用 TotalCost 原始费用，不考虑倍率） - if shouldBill && cost.TotalCost > 0 { - if err := s.userSubRepo.IncrementUsage(ctx, subscription.ID, cost.TotalCost); err != nil { - logger.LegacyPrintf("service.gateway", "Increment subscription usage failed: %v", err) - } - // 异步更新订阅缓存 - s.billingCacheService.QueueUpdateSubscriptionUsage(user.ID, *apiKey.GroupID, cost.TotalCost) - } + if shouldBill { + postUsageBilling(ctx, &postUsageBillingParams{ + Cost: cost, + User: user, + APIKey: apiKey, + Account: account, + Subscription: subscription, + IsSubscriptionBill: isSubscriptionBilling, + AccountRateMultiplier: accountRateMultiplier, + APIKeyService: input.APIKeyService, + }, s.billingDeps()) } else { - // 余额模式：扣除用户余额（使用 ActualCost 考虑倍率后的费用） - if shouldBill && cost.ActualCost > 0 { - if err := s.userRepo.DeductBalance(ctx, user.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Deduct balance failed: %v", err) - } - // 异步更新余额缓存 - s.billingCacheService.QueueDeductBalance(user.ID, cost.ActualCost) - // API Key 独立配额扣费 - if input.APIKeyService != nil && apiKey.Quota > 0 { - if err := input.APIKeyService.UpdateQuotaUsed(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Add API key quota used failed: %v", err) - } - } - } + s.deferredService.ScheduleLastUsedUpdate(account.ID) } - // Update API Key rate limit usage - if shouldBill && cost.ActualCost > 0 && apiKey.HasRateLimits() && input.APIKeyService != nil { - if err := input.APIKeyService.UpdateRateLimitUsage(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.gateway", "Update API key rate limit usage failed: %v", err) - } - s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) - } - - // 更新 API Key 账号配额用量 - if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { - if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { - slog.Error("increment account quota used failed", "account_id", account.ID, "cost", cost.TotalCost, "error", err) - } - } - - // Schedule batch update for account last_used_at - s.deferredService.ScheduleLastUsedUpdate(account.ID) - return nil } diff --git a/backend/internal/service/openai_gateway_service.go b/backend/internal/service/openai_gateway_service.go index 6752d18b72..84fe351c58 100644 --- a/backend/internal/service/openai_gateway_service.go +++ b/backend/internal/service/openai_gateway_service.go @@ -319,6 +319,16 @@ func NewOpenAIGatewayService( return svc } +func (s *OpenAIGatewayService) billingDeps() *billingDeps { + return &billingDeps{ + accountRepo: s.accountRepo, + userRepo: s.userRepo, + userSubRepo: s.userSubRepo, + billingCacheService: s.billingCacheService, + deferredService: s.deferredService, + } +} + // CloseOpenAIWSPool 关闭 OpenAI WebSocket 连接池的后台 worker 和空闲连接。 // 应在应用优雅关闭时调用。 func (s *OpenAIGatewayService) CloseOpenAIWSPool() { @@ -3474,44 +3484,21 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec shouldBill := inserted || err != nil - // Deduct based on billing type - if isSubscriptionBilling { - if shouldBill && cost.TotalCost > 0 { - _ = s.userSubRepo.IncrementUsage(ctx, subscription.ID, cost.TotalCost) - s.billingCacheService.QueueUpdateSubscriptionUsage(user.ID, *apiKey.GroupID, cost.TotalCost) - } + if shouldBill { + postUsageBilling(ctx, &postUsageBillingParams{ + Cost: cost, + User: user, + APIKey: apiKey, + Account: account, + Subscription: subscription, + IsSubscriptionBill: isSubscriptionBilling, + AccountRateMultiplier: accountRateMultiplier, + APIKeyService: input.APIKeyService, + }, s.billingDeps()) } else { - if shouldBill && cost.ActualCost > 0 { - _ = s.userRepo.DeductBalance(ctx, user.ID, cost.ActualCost) - s.billingCacheService.QueueDeductBalance(user.ID, cost.ActualCost) - } + s.deferredService.ScheduleLastUsedUpdate(account.ID) } - // Update API key quota if applicable (only for balance mode with quota set) - if shouldBill && cost.ActualCost > 0 && apiKey.Quota > 0 && input.APIKeyService != nil { - if err := input.APIKeyService.UpdateQuotaUsed(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.openai_gateway", "Update API key quota failed: %v", err) - } - } - - // Update API Key rate limit usage - if shouldBill && cost.ActualCost > 0 && apiKey.HasRateLimits() && input.APIKeyService != nil { - if err := input.APIKeyService.UpdateRateLimitUsage(ctx, apiKey.ID, cost.ActualCost); err != nil { - logger.LegacyPrintf("service.openai_gateway", "Update API key rate limit usage failed: %v", err) - } - s.billingCacheService.QueueUpdateAPIKeyRateLimitUsage(apiKey.ID, cost.ActualCost) - } - - // 更新 API Key 账号配额用量 - if shouldBill && cost.TotalCost > 0 && account.Type == AccountTypeAPIKey && account.GetQuotaLimit() > 0 { - if err := s.accountRepo.IncrementQuotaUsed(ctx, account.ID, cost.TotalCost); err != nil { - logger.LegacyPrintf("service.openai_gateway", "increment account quota used failed: account_id=%d cost=%f error=%v", account.ID, cost.TotalCost, err) - } - } - - // Schedule batch update for account last_used_at - s.deferredService.ScheduleLastUsedUpdate(account.ID) - return nil } diff --git a/backend/internal/service/openai_ws_forwarder.go b/backend/internal/service/openai_ws_forwarder.go index a5c2fd7a27..7b6591fad4 100644 --- a/backend/internal/service/openai_ws_forwarder.go +++ b/backend/internal/service/openai_ws_forwarder.go @@ -864,7 +864,8 @@ func isOpenAIWSClientDisconnectError(err error) bool { strings.Contains(message, "unexpected eof") || strings.Contains(message, "use of closed network connection") || strings.Contains(message, "connection reset by peer") || - strings.Contains(message, "broken pipe") + strings.Contains(message, "broken pipe") || + strings.Contains(message, "an established connection was aborted") } func classifyOpenAIWSReadFallbackReason(err error) string { From 77701143bf9dab2ea5527d9f626bfbaa6b53b3b4 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 01:07:28 +0800 Subject: [PATCH 07/53] fix: use range assertion for time-sensitive ExpiresInDays test The test could flake depending on exact execution time near midnight boundaries. Use a range check (29 or 30) instead of exact equality. --- .../internal/service/subscription_calculate_progress_test.go | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backend/internal/service/subscription_calculate_progress_test.go b/backend/internal/service/subscription_calculate_progress_test.go index 22018bcd63..53e5c56806 100644 --- a/backend/internal/service/subscription_calculate_progress_test.go +++ b/backend/internal/service/subscription_calculate_progress_test.go @@ -34,7 +34,7 @@ func TestCalculateProgress_BasicFields(t *testing.T) { assert.Equal(t, int64(100), progress.ID) assert.Equal(t, "Premium", progress.GroupName) assert.Equal(t, sub.ExpiresAt, progress.ExpiresAt) - assert.Equal(t, 29, progress.ExpiresInDays) // 约 30 天 + assert.True(t, progress.ExpiresInDays == 29 || progress.ExpiresInDays == 30, "ExpiresInDays should be 29 or 30, got %d", progress.ExpiresInDays) assert.Nil(t, progress.Daily, "无日限额时 Daily 应为 nil") assert.Nil(t, progress.Weekly, "无周限额时 Weekly 应为 nil") assert.Nil(t, progress.Monthly, "无月限额时 Monthly 应为 nil") From 3d93a29b3511bfda29fe0026aa9f6a2a6e1bf6e4 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:16:47 +0800 Subject: [PATCH 08/53] fix: validate account status before update and update load factor hint MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Normalize non-standard status (e.g. "error") to "active" on edit load - Add pre-submit validation for status field to prevent 400 errors - Update load factor hint: "提高负载因子可以提高对账号的调度频率" --- backend/cmd/server/VERSION | 2 +- frontend/src/components/account/EditAccountModal.vue | 9 ++++++++- frontend/src/i18n/locales/en.ts | 3 ++- frontend/src/i18n/locales/zh.ts | 3 ++- 4 files changed, 13 insertions(+), 4 deletions(-) diff --git a/backend/cmd/server/VERSION b/backend/cmd/server/VERSION index 2619c8dd8f..42c952870a 100644 --- a/backend/cmd/server/VERSION +++ b/backend/cmd/server/VERSION @@ -1 +1 @@ -0.1.91.1 +0.1.91.2 diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 1b02e327b4..03f8367633 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -1513,7 +1513,9 @@ watch( form.load_factor = newAccount.load_factor ?? null form.priority = newAccount.priority form.rate_multiplier = newAccount.rate_multiplier ?? 1 - form.status = newAccount.status as 'active' | 'inactive' + form.status = (newAccount.status === 'active' || newAccount.status === 'inactive') + ? newAccount.status + : 'active' form.group_ids = newAccount.group_ids || [] form.expires_at = newAccount.expires_at ?? null @@ -2061,6 +2063,11 @@ const handleSubmit = async () => { if (!props.account) return const accountID = props.account.id + if (form.status !== 'active' && form.status !== 'inactive') { + appStore.showError(t('admin.accounts.pleaseSelectStatus')) + return + } + const updatePayload: Record = { ...form } try { // 后端期望 proxy_id: 0 表示清除代理，而不是 null diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index d055a4c836..4d51899c11 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -2012,7 +2012,7 @@ export default { noProxy: 'No Proxy', concurrency: 'Concurrency', loadFactor: 'Load Factor', - loadFactorHint: 'Defaults to concurrency', + loadFactorHint: 'Higher load factor increases scheduling frequency', priority: 'Priority', priorityHint: 'Lower value accounts are used first', billingRateMultiplier: 'Billing Rate Multiplier', @@ -2032,6 +2032,7 @@ export default { accountUpdated: 'Account updated successfully', failedToCreate: 'Failed to create account', failedToUpdate: 'Failed to update account', + pleaseSelectStatus: 'Please select a valid account status', mixedChannelWarningTitle: 'Mixed Channel Warning', mixedChannelWarning: 'Warning: Group "{groupName}" contains both {currentPlatform} and {otherPlatform} accounts. Mixing different channels may cause thinking block signature validation issues, which will fallback to non-thinking mode. Are you sure you want to continue?', pleaseEnterAccountName: 'Please enter account name', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index ecaff13deb..ec7836156a 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -2146,7 +2146,7 @@ export default { noProxy: '无代理', concurrency: '并发数', loadFactor: '负载因子', - loadFactorHint: '不填则等于并发数', + loadFactorHint: '提高负载因子可以提高对账号的调度频率', priority: '优先级', priorityHint: '优先级越小的账号优先使用', billingRateMultiplier: '账号计费倍率', @@ -2166,6 +2166,7 @@ export default { accountUpdated: '账号更新成功', failedToCreate: '创建账号失败', failedToUpdate: '更新账号失败', + pleaseSelectStatus: '请选择有效的账号状态', mixedChannelWarningTitle: '混合渠道警告', mixedChannelWarning: '警告：分组 "{groupName}" 中同时包含 {currentPlatform} 和 {otherPlatform} 账号。混合使用不同渠道可能导致 thinking block 签名验证问题，会自动回退到非 thinking 模式。确定要继续吗？', pleaseEnterAccountName: '请输入账号名称', From ee7d6bc19f0fef3d5c8ec018f96f40b68ce5cad9 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:27:41 +0800 Subject: [PATCH 09/53] fix: address load_factor code review findings - Fix bulk edit: send 0 instead of null/NaN to clear load_factor - Fix edit modal: explicit NaN check instead of implicit falsy - Fix create modal: use ?? instead of || for load_factor - Add load_factor upper limit validation (max 10000) - Add //go:build unit tag and self-contained intPtrHelper in test - Add design intent comments on WaitPlan.MaxConcurrency --- backend/cmd/server/VERSION | 2 +- .../service/account_load_factor_test.go | 12 ++++++++---- backend/internal/service/admin_service.go | 5 +++++ .../service/openai_account_scheduler.go | 2 ++ .../account/BulkEditAccountModal.vue | 4 +++- .../components/account/CreateAccountModal.vue | 18 +++++++++--------- .../components/account/EditAccountModal.vue | 5 +++-- 7 files changed, 31 insertions(+), 17 deletions(-) diff --git a/backend/cmd/server/VERSION b/backend/cmd/server/VERSION index 42c952870a..231369cfef 100644 --- a/backend/cmd/server/VERSION +++ b/backend/cmd/server/VERSION @@ -1 +1 @@ -0.1.91.2 +0.1.91.3 diff --git a/backend/internal/service/account_load_factor_test.go b/backend/internal/service/account_load_factor_test.go index 1cd1b17c9c..a4d78a4bba 100644 --- a/backend/internal/service/account_load_factor_test.go +++ b/backend/internal/service/account_load_factor_test.go @@ -1,3 +1,5 @@ +//go:build unit + package service import ( @@ -6,6 +8,8 @@ import ( "github.com/stretchr/testify/require" ) +func intPtrHelper(v int) *int { return &v } + func TestEffectiveLoadFactor_NilAccount(t *testing.T) { var a *Account require.Equal(t, 1, a.EffectiveLoadFactor()) @@ -22,21 +26,21 @@ func TestEffectiveLoadFactor_NilLoadFactor_ZeroConcurrency(t *testing.T) { } func TestEffectiveLoadFactor_PositiveLoadFactor(t *testing.T) { - a := &Account{Concurrency: 5, LoadFactor: intPtr(20)} + a := &Account{Concurrency: 5, LoadFactor: intPtrHelper(20)} require.Equal(t, 20, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_ZeroLoadFactor_FallbackToConcurrency(t *testing.T) { - a := &Account{Concurrency: 5, LoadFactor: intPtr(0)} + a := &Account{Concurrency: 5, LoadFactor: intPtrHelper(0)} require.Equal(t, 5, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_NegativeLoadFactor_FallbackToConcurrency(t *testing.T) { - a := &Account{Concurrency: 3, LoadFactor: intPtr(-1)} + a := &Account{Concurrency: 3, LoadFactor: intPtrHelper(-1)} require.Equal(t, 3, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_ZeroLoadFactor_ZeroConcurrency(t *testing.T) { - a := &Account{Concurrency: 0, LoadFactor: intPtr(0)} + a := &Account{Concurrency: 0, LoadFactor: intPtrHelper(0)} require.Equal(t, 1, a.EffectiveLoadFactor()) } diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index 11f850ee5f..8992e2505d 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -1437,6 +1437,9 @@ func (s *adminServiceImpl) CreateAccount(ctx context.Context, input *CreateAccou account.RateMultiplier = input.RateMultiplier } if input.LoadFactor != nil && *input.LoadFactor > 0 { + if *input.LoadFactor > 10000 { + return nil, errors.New("load_factor must be <= 10000") + } account.LoadFactor = input.LoadFactor } if err := s.accountRepo.Create(ctx, account); err != nil { @@ -1516,6 +1519,8 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U if input.LoadFactor != nil { if *input.LoadFactor <= 0 { account.LoadFactor = nil // 0 或负数表示清除 + } else if *input.LoadFactor > 10000 { + return nil, errors.New("load_factor must be <= 10000") } else { account.LoadFactor = input.LoadFactor } diff --git a/backend/internal/service/openai_account_scheduler.go b/backend/internal/service/openai_account_scheduler.go index 9cc6aae163..cf4bc26eb0 100644 --- a/backend/internal/service/openai_account_scheduler.go +++ b/backend/internal/service/openai_account_scheduler.go @@ -342,6 +342,7 @@ func (s *defaultOpenAIAccountScheduler) selectBySessionHash( } cfg := s.service.schedulingConfig() + // WaitPlan.MaxConcurrency 使用 Concurrency（非 EffectiveLoadFactor），因为 WaitPlan 控制的是 Redis 实际并发槽位等待。 if s.service.concurrencyService != nil { return &AccountSelectionResult{ Account: account, @@ -703,6 +704,7 @@ func (s *defaultOpenAIAccountScheduler) selectByLoadBalance( } cfg := s.service.schedulingConfig() + // WaitPlan.MaxConcurrency 使用 Concurrency（非 EffectiveLoadFactor），因为 WaitPlan 控制的是 Redis 实际并发槽位等待。 candidate := selectionOrder[0] return &AccountSelectionResult{ Account: candidate.account, diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 5e7d0ae4fe..9763d24ae6 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -1227,7 +1227,9 @@ const buildUpdatePayload = (): Record | null => { } if (enableLoadFactor.value) { - updates.load_factor = loadFactor.value + // 空值/NaN/0 时发送 0（后端约定 <= 0 表示清除） + const lf = loadFactor.value + updates.load_factor = (lf != null && !Number.isNaN(lf) && lf > 0) ? lf : 0 } if (enablePriority.value) { diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index 8013bbaf46..ad2357b739 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -3497,7 +3497,7 @@ const handleImportAccessToken = async (accessTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3562,7 +3562,7 @@ const createAccountAndFinish = async ( extra: finalExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3618,7 +3618,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3648,7 +3648,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3726,7 +3726,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3754,7 +3754,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3843,7 +3843,7 @@ const handleSoraValidateST = async (sessionTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3932,7 +3932,7 @@ const handleAntigravityValidateRT = async (refreshTokenInput: string) => { extra: {}, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -4256,7 +4256,7 @@ const handleCookieAuth = async (sessionKey: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 03f8367633..557c884902 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -2077,8 +2077,9 @@ const handleSubmit = async () => { if (form.expires_at === null) { updatePayload.expires_at = 0 } - // load_factor: 空值/0/NaN 时发送 0（后端约定 0 = 清除） - if (!form.load_factor || form.load_factor <= 0) { + // load_factor: 空值/NaN/0/负数 时发送 0（后端约定 <= 0 = 清除） + const lf = form.load_factor + if (lf == null || Number.isNaN(lf) || lf <= 0) { updatePayload.load_factor = 0 } updatePayload.auto_pause_on_expired = autoPauseOnExpired.value From dbe94847de8aa5d7ec76b9ee147be7659c349ee9 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:30:30 +0800 Subject: [PATCH 10/53] fix: prevent negative values for concurrency and load_factor inputs --- frontend/src/components/account/BulkEditAccountModal.vue | 2 ++ frontend/src/components/account/CreateAccountModal.vue | 6 ++++-- frontend/src/components/account/EditAccountModal.vue | 6 ++++-- 3 files changed, 10 insertions(+), 4 deletions(-) diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 9763d24ae6..8d27d100e0 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -496,6 +496,7 @@ class="input" :class="!enableConcurrency && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-concurrency-label" + @input="if (concurrency < 1) concurrency = 1" />
@@ -524,6 +525,7 @@ class="input" :class="!enableLoadFactor && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-load-factor-label" + @input="if (loadFactor != null && loadFactor < 1) loadFactor = null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index ad2357b739..b2ac56d724 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -1755,12 +1755,14 @@
- +
+ class="input" :placeholder="String(form.concurrency || 1)" + @input="if (form.load_factor != null && form.load_factor < 1) form.load_factor = null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 557c884902..bf36a41519 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -653,12 +653,14 @@
- +
+ class="input" :placeholder="String(form.concurrency || 1)" + @input="if (form.load_factor != null && form.load_factor < 1) form.load_factor = null" />

{{ t('admin.accounts.loadFactorHint') }}

From 02b78950c9ec903b9f568a941dba9c8757e7f852 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:33:55 +0800 Subject: [PATCH 11/53] fix: use HTML-safe expressions for @input handlers in Vue templates Replace `<` comparisons with Math.max/ternary+>= to avoid Vue template parser treating `<` as HTML tag start in attribute values. --- frontend/src/components/account/BulkEditAccountModal.vue | 4 ++-- frontend/src/components/account/CreateAccountModal.vue | 4 ++-- frontend/src/components/account/EditAccountModal.vue | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 8d27d100e0..e00924495e 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -496,7 +496,7 @@ class="input" :class="!enableConcurrency && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-concurrency-label" - @input="if (concurrency < 1) concurrency = 1" + @input="concurrency = Math.max(1, concurrency || 1)" />
@@ -525,7 +525,7 @@ class="input" :class="!enableLoadFactor && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-load-factor-label" - @input="if (loadFactor != null && loadFactor < 1) loadFactor = null" + @input="loadFactor = (loadFactor && loadFactor >= 1) ? loadFactor : null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index b2ac56d724..92725c0788 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -1756,13 +1756,13 @@
+ @input="form.concurrency = Math.max(1, form.concurrency || 1)" />
+ @input="form.load_factor = (form.load_factor && form.load_factor >= 1) ? form.load_factor : null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index bf36a41519..be7d2d45ce 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -654,13 +654,13 @@
+ @input="form.concurrency = Math.max(1, form.concurrency || 1)" />
+ @input="form.load_factor = (form.load_factor && form.load_factor >= 1) ? form.load_factor : null" />

{{ t('admin.accounts.loadFactorHint') }}

From 95e366b6c6672bed55ca7def05702cbfaa1f7fed Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:37:56 +0800 Subject: [PATCH 12/53] fix: add missing IncrementQuotaUsed and ResetQuotaUsed to stubAccountRepo in api_contract_test --- backend/internal/server/api_contract_test.go | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/backend/internal/server/api_contract_test.go b/backend/internal/server/api_contract_test.go index 40b2d592db..aafbbe21cb 100644 --- a/backend/internal/server/api_contract_test.go +++ b/backend/internal/server/api_contract_test.go @@ -1096,6 +1096,14 @@ func (s *stubAccountRepo) UpdateExtra(ctx context.Context, id int64, updates map return errors.New("not implemented") } +func (s *stubAccountRepo) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return errors.New("not implemented") +} + +func (s *stubAccountRepo) ResetQuotaUsed(ctx context.Context, id int64) error { + return errors.New("not implemented") +} + func (s *stubAccountRepo) BulkUpdate(ctx context.Context, ids []int64, updates service.AccountBulkUpdate) (int64, error) { s.bulkUpdateIDs = append([]int64{}, ids...) return int64(len(ids)), nil From 0d6c1c779053beb171b8158b1340f792eb5f9bd3 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 05:07:10 +0800 Subject: [PATCH 13/53] feat: add independent load_factor field for scheduling load calculation --- backend/ent/account.go | 16 ++- backend/ent/account/account.go | 8 ++ backend/ent/account/where.go | 55 +++++++++ backend/ent/account_create.go | 98 ++++++++++++++++ backend/ent/account_update.go | 72 ++++++++++++ backend/ent/migrate/schema.go | 23 ++-- backend/ent/mutation.go | 111 +++++++++++++++++- backend/ent/runtime/runtime.go | 12 +- backend/ent/schema/account.go | 2 + backend/go.sum | 15 +-- .../internal/handler/admin/account_handler.go | 7 ++ backend/internal/handler/dto/mappers.go | 1 + backend/internal/handler/dto/types.go | 1 + backend/internal/repository/account_repo.go | 18 +++ backend/internal/server/api_contract_test.go | 8 ++ backend/internal/service/account.go | 14 +++ .../service/account_load_factor_test.go | 42 +++++++ backend/internal/service/account_service.go | 1 + backend/internal/service/admin_service.go | 16 +++ backend/internal/service/gateway_service.go | 4 +- .../service/openai_account_scheduler.go | 2 +- .../service/openai_gateway_service.go | 2 +- backend/internal/service/ops_concurrency.go | 5 +- .../internal/service/ops_metrics_collector.go | 6 +- .../067_add_account_load_factor.sql | 1 + .../account/BulkEditAccountModal.vue | 40 ++++++- .../components/account/CreateAccountModal.vue | 33 +++++- .../components/account/EditAccountModal.vue | 21 +++- frontend/src/i18n/locales/en.ts | 4 +- frontend/src/i18n/locales/zh.ts | 4 +- frontend/src/types/index.ts | 3 + 31 files changed, 596 insertions(+), 49 deletions(-) create mode 100644 backend/internal/service/account_load_factor_test.go create mode 100644 backend/migrations/067_add_account_load_factor.sql diff --git a/backend/ent/account.go b/backend/ent/account.go index c77002b32e..2dbfc3a278 100644 --- a/backend/ent/account.go +++ b/backend/ent/account.go @@ -41,6 +41,8 @@ type Account struct { ProxyID *int64 `json:"proxy_id,omitempty"` // Concurrency holds the value of the "concurrency" field. Concurrency int `json:"concurrency,omitempty"` + // LoadFactor holds the value of the "load_factor" field. + LoadFactor *int `json:"load_factor,omitempty"` // Priority holds the value of the "priority" field. Priority int `json:"priority,omitempty"` // RateMultiplier holds the value of the "rate_multiplier" field. @@ -143,7 +145,7 @@ func (*Account) scanValues(columns []string) ([]any, error) { values[i] = new(sql.NullBool) case account.FieldRateMultiplier: values[i] = new(sql.NullFloat64) - case account.FieldID, account.FieldProxyID, account.FieldConcurrency, account.FieldPriority: + case account.FieldID, account.FieldProxyID, account.FieldConcurrency, account.FieldLoadFactor, account.FieldPriority: values[i] = new(sql.NullInt64) case account.FieldName, account.FieldNotes, account.FieldPlatform, account.FieldType, account.FieldStatus, account.FieldErrorMessage, account.FieldTempUnschedulableReason, account.FieldSessionWindowStatus: values[i] = new(sql.NullString) @@ -243,6 +245,13 @@ func (_m *Account) assignValues(columns []string, values []any) error { } else if value.Valid { _m.Concurrency = int(value.Int64) } + case account.FieldLoadFactor: + if value, ok := values[i].(*sql.NullInt64); !ok { + return fmt.Errorf("unexpected type %T for field load_factor", values[i]) + } else if value.Valid { + _m.LoadFactor = new(int) + *_m.LoadFactor = int(value.Int64) + } case account.FieldPriority: if value, ok := values[i].(*sql.NullInt64); !ok { return fmt.Errorf("unexpected type %T for field priority", values[i]) @@ -445,6 +454,11 @@ func (_m *Account) String() string { builder.WriteString("concurrency=") builder.WriteString(fmt.Sprintf("%v", _m.Concurrency)) builder.WriteString(", ") + if v := _m.LoadFactor; v != nil { + builder.WriteString("load_factor=") + builder.WriteString(fmt.Sprintf("%v", *v)) + } + builder.WriteString(", ") builder.WriteString("priority=") builder.WriteString(fmt.Sprintf("%v", _m.Priority)) builder.WriteString(", ") diff --git a/backend/ent/account/account.go b/backend/ent/account/account.go index 1fc34620de..4c1346490a 100644 --- a/backend/ent/account/account.go +++ b/backend/ent/account/account.go @@ -37,6 +37,8 @@ const ( FieldProxyID = "proxy_id" // FieldConcurrency holds the string denoting the concurrency field in the database. FieldConcurrency = "concurrency" + // FieldLoadFactor holds the string denoting the load_factor field in the database. + FieldLoadFactor = "load_factor" // FieldPriority holds the string denoting the priority field in the database. FieldPriority = "priority" // FieldRateMultiplier holds the string denoting the rate_multiplier field in the database. @@ -121,6 +123,7 @@ var Columns = []string{ FieldExtra, FieldProxyID, FieldConcurrency, + FieldLoadFactor, FieldPriority, FieldRateMultiplier, FieldStatus, @@ -250,6 +253,11 @@ func ByConcurrency(opts ...sql.OrderTermOption) OrderOption { return sql.OrderByField(FieldConcurrency, opts...).ToFunc() } +// ByLoadFactor orders the results by the load_factor field. +func ByLoadFactor(opts ...sql.OrderTermOption) OrderOption { + return sql.OrderByField(FieldLoadFactor, opts...).ToFunc() +} + // ByPriority orders the results by the priority field. func ByPriority(opts ...sql.OrderTermOption) OrderOption { return sql.OrderByField(FieldPriority, opts...).ToFunc() diff --git a/backend/ent/account/where.go b/backend/ent/account/where.go index 54db1dcb14..3749b45c55 100644 --- a/backend/ent/account/where.go +++ b/backend/ent/account/where.go @@ -100,6 +100,11 @@ func Concurrency(v int) predicate.Account { return predicate.Account(sql.FieldEQ(FieldConcurrency, v)) } +// LoadFactor applies equality check predicate on the "load_factor" field. It's identical to LoadFactorEQ. +func LoadFactor(v int) predicate.Account { + return predicate.Account(sql.FieldEQ(FieldLoadFactor, v)) +} + // Priority applies equality check predicate on the "priority" field. It's identical to PriorityEQ. func Priority(v int) predicate.Account { return predicate.Account(sql.FieldEQ(FieldPriority, v)) @@ -650,6 +655,56 @@ func ConcurrencyLTE(v int) predicate.Account { return predicate.Account(sql.FieldLTE(FieldConcurrency, v)) } +// LoadFactorEQ applies the EQ predicate on the "load_factor" field. +func LoadFactorEQ(v int) predicate.Account { + return predicate.Account(sql.FieldEQ(FieldLoadFactor, v)) +} + +// LoadFactorNEQ applies the NEQ predicate on the "load_factor" field. +func LoadFactorNEQ(v int) predicate.Account { + return predicate.Account(sql.FieldNEQ(FieldLoadFactor, v)) +} + +// LoadFactorIn applies the In predicate on the "load_factor" field. +func LoadFactorIn(vs ...int) predicate.Account { + return predicate.Account(sql.FieldIn(FieldLoadFactor, vs...)) +} + +// LoadFactorNotIn applies the NotIn predicate on the "load_factor" field. +func LoadFactorNotIn(vs ...int) predicate.Account { + return predicate.Account(sql.FieldNotIn(FieldLoadFactor, vs...)) +} + +// LoadFactorGT applies the GT predicate on the "load_factor" field. +func LoadFactorGT(v int) predicate.Account { + return predicate.Account(sql.FieldGT(FieldLoadFactor, v)) +} + +// LoadFactorGTE applies the GTE predicate on the "load_factor" field. +func LoadFactorGTE(v int) predicate.Account { + return predicate.Account(sql.FieldGTE(FieldLoadFactor, v)) +} + +// LoadFactorLT applies the LT predicate on the "load_factor" field. +func LoadFactorLT(v int) predicate.Account { + return predicate.Account(sql.FieldLT(FieldLoadFactor, v)) +} + +// LoadFactorLTE applies the LTE predicate on the "load_factor" field. +func LoadFactorLTE(v int) predicate.Account { + return predicate.Account(sql.FieldLTE(FieldLoadFactor, v)) +} + +// LoadFactorIsNil applies the IsNil predicate on the "load_factor" field. +func LoadFactorIsNil() predicate.Account { + return predicate.Account(sql.FieldIsNull(FieldLoadFactor)) +} + +// LoadFactorNotNil applies the NotNil predicate on the "load_factor" field. +func LoadFactorNotNil() predicate.Account { + return predicate.Account(sql.FieldNotNull(FieldLoadFactor)) +} + // PriorityEQ applies the EQ predicate on the "priority" field. func PriorityEQ(v int) predicate.Account { return predicate.Account(sql.FieldEQ(FieldPriority, v)) diff --git a/backend/ent/account_create.go b/backend/ent/account_create.go index 963ffee885..d6046c7977 100644 --- a/backend/ent/account_create.go +++ b/backend/ent/account_create.go @@ -139,6 +139,20 @@ func (_c *AccountCreate) SetNillableConcurrency(v *int) *AccountCreate { return _c } +// SetLoadFactor sets the "load_factor" field. +func (_c *AccountCreate) SetLoadFactor(v int) *AccountCreate { + _c.mutation.SetLoadFactor(v) + return _c +} + +// SetNillableLoadFactor sets the "load_factor" field if the given value is not nil. +func (_c *AccountCreate) SetNillableLoadFactor(v *int) *AccountCreate { + if v != nil { + _c.SetLoadFactor(*v) + } + return _c +} + // SetPriority sets the "priority" field. func (_c *AccountCreate) SetPriority(v int) *AccountCreate { _c.mutation.SetPriority(v) @@ -623,6 +637,10 @@ func (_c *AccountCreate) createSpec() (*Account, *sqlgraph.CreateSpec) { _spec.SetField(account.FieldConcurrency, field.TypeInt, value) _node.Concurrency = value } + if value, ok := _c.mutation.LoadFactor(); ok { + _spec.SetField(account.FieldLoadFactor, field.TypeInt, value) + _node.LoadFactor = &value + } if value, ok := _c.mutation.Priority(); ok { _spec.SetField(account.FieldPriority, field.TypeInt, value) _node.Priority = value @@ -936,6 +954,30 @@ func (u *AccountUpsert) AddConcurrency(v int) *AccountUpsert { return u } +// SetLoadFactor sets the "load_factor" field. +func (u *AccountUpsert) SetLoadFactor(v int) *AccountUpsert { + u.Set(account.FieldLoadFactor, v) + return u +} + +// UpdateLoadFactor sets the "load_factor" field to the value that was provided on create. +func (u *AccountUpsert) UpdateLoadFactor() *AccountUpsert { + u.SetExcluded(account.FieldLoadFactor) + return u +} + +// AddLoadFactor adds v to the "load_factor" field. +func (u *AccountUpsert) AddLoadFactor(v int) *AccountUpsert { + u.Add(account.FieldLoadFactor, v) + return u +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (u *AccountUpsert) ClearLoadFactor() *AccountUpsert { + u.SetNull(account.FieldLoadFactor) + return u +} + // SetPriority sets the "priority" field. func (u *AccountUpsert) SetPriority(v int) *AccountUpsert { u.Set(account.FieldPriority, v) @@ -1419,6 +1461,34 @@ func (u *AccountUpsertOne) UpdateConcurrency() *AccountUpsertOne { }) } +// SetLoadFactor sets the "load_factor" field. +func (u *AccountUpsertOne) SetLoadFactor(v int) *AccountUpsertOne { + return u.Update(func(s *AccountUpsert) { + s.SetLoadFactor(v) + }) +} + +// AddLoadFactor adds v to the "load_factor" field. +func (u *AccountUpsertOne) AddLoadFactor(v int) *AccountUpsertOne { + return u.Update(func(s *AccountUpsert) { + s.AddLoadFactor(v) + }) +} + +// UpdateLoadFactor sets the "load_factor" field to the value that was provided on create. +func (u *AccountUpsertOne) UpdateLoadFactor() *AccountUpsertOne { + return u.Update(func(s *AccountUpsert) { + s.UpdateLoadFactor() + }) +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (u *AccountUpsertOne) ClearLoadFactor() *AccountUpsertOne { + return u.Update(func(s *AccountUpsert) { + s.ClearLoadFactor() + }) +} + // SetPriority sets the "priority" field. func (u *AccountUpsertOne) SetPriority(v int) *AccountUpsertOne { return u.Update(func(s *AccountUpsert) { @@ -2113,6 +2183,34 @@ func (u *AccountUpsertBulk) UpdateConcurrency() *AccountUpsertBulk { }) } +// SetLoadFactor sets the "load_factor" field. +func (u *AccountUpsertBulk) SetLoadFactor(v int) *AccountUpsertBulk { + return u.Update(func(s *AccountUpsert) { + s.SetLoadFactor(v) + }) +} + +// AddLoadFactor adds v to the "load_factor" field. +func (u *AccountUpsertBulk) AddLoadFactor(v int) *AccountUpsertBulk { + return u.Update(func(s *AccountUpsert) { + s.AddLoadFactor(v) + }) +} + +// UpdateLoadFactor sets the "load_factor" field to the value that was provided on create. +func (u *AccountUpsertBulk) UpdateLoadFactor() *AccountUpsertBulk { + return u.Update(func(s *AccountUpsert) { + s.UpdateLoadFactor() + }) +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (u *AccountUpsertBulk) ClearLoadFactor() *AccountUpsertBulk { + return u.Update(func(s *AccountUpsert) { + s.ClearLoadFactor() + }) +} + // SetPriority sets the "priority" field. func (u *AccountUpsertBulk) SetPriority(v int) *AccountUpsertBulk { return u.Update(func(s *AccountUpsert) { diff --git a/backend/ent/account_update.go b/backend/ent/account_update.go index 875888e049..6f443c65e0 100644 --- a/backend/ent/account_update.go +++ b/backend/ent/account_update.go @@ -172,6 +172,33 @@ func (_u *AccountUpdate) AddConcurrency(v int) *AccountUpdate { return _u } +// SetLoadFactor sets the "load_factor" field. +func (_u *AccountUpdate) SetLoadFactor(v int) *AccountUpdate { + _u.mutation.ResetLoadFactor() + _u.mutation.SetLoadFactor(v) + return _u +} + +// SetNillableLoadFactor sets the "load_factor" field if the given value is not nil. +func (_u *AccountUpdate) SetNillableLoadFactor(v *int) *AccountUpdate { + if v != nil { + _u.SetLoadFactor(*v) + } + return _u +} + +// AddLoadFactor adds value to the "load_factor" field. +func (_u *AccountUpdate) AddLoadFactor(v int) *AccountUpdate { + _u.mutation.AddLoadFactor(v) + return _u +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (_u *AccountUpdate) ClearLoadFactor() *AccountUpdate { + _u.mutation.ClearLoadFactor() + return _u +} + // SetPriority sets the "priority" field. func (_u *AccountUpdate) SetPriority(v int) *AccountUpdate { _u.mutation.ResetPriority() @@ -684,6 +711,15 @@ func (_u *AccountUpdate) sqlSave(ctx context.Context) (_node int, err error) { if value, ok := _u.mutation.AddedConcurrency(); ok { _spec.AddField(account.FieldConcurrency, field.TypeInt, value) } + if value, ok := _u.mutation.LoadFactor(); ok { + _spec.SetField(account.FieldLoadFactor, field.TypeInt, value) + } + if value, ok := _u.mutation.AddedLoadFactor(); ok { + _spec.AddField(account.FieldLoadFactor, field.TypeInt, value) + } + if _u.mutation.LoadFactorCleared() { + _spec.ClearField(account.FieldLoadFactor, field.TypeInt) + } if value, ok := _u.mutation.Priority(); ok { _spec.SetField(account.FieldPriority, field.TypeInt, value) } @@ -1063,6 +1099,33 @@ func (_u *AccountUpdateOne) AddConcurrency(v int) *AccountUpdateOne { return _u } +// SetLoadFactor sets the "load_factor" field. +func (_u *AccountUpdateOne) SetLoadFactor(v int) *AccountUpdateOne { + _u.mutation.ResetLoadFactor() + _u.mutation.SetLoadFactor(v) + return _u +} + +// SetNillableLoadFactor sets the "load_factor" field if the given value is not nil. +func (_u *AccountUpdateOne) SetNillableLoadFactor(v *int) *AccountUpdateOne { + if v != nil { + _u.SetLoadFactor(*v) + } + return _u +} + +// AddLoadFactor adds value to the "load_factor" field. +func (_u *AccountUpdateOne) AddLoadFactor(v int) *AccountUpdateOne { + _u.mutation.AddLoadFactor(v) + return _u +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (_u *AccountUpdateOne) ClearLoadFactor() *AccountUpdateOne { + _u.mutation.ClearLoadFactor() + return _u +} + // SetPriority sets the "priority" field. func (_u *AccountUpdateOne) SetPriority(v int) *AccountUpdateOne { _u.mutation.ResetPriority() @@ -1605,6 +1668,15 @@ func (_u *AccountUpdateOne) sqlSave(ctx context.Context) (_node *Account, err er if value, ok := _u.mutation.AddedConcurrency(); ok { _spec.AddField(account.FieldConcurrency, field.TypeInt, value) } + if value, ok := _u.mutation.LoadFactor(); ok { + _spec.SetField(account.FieldLoadFactor, field.TypeInt, value) + } + if value, ok := _u.mutation.AddedLoadFactor(); ok { + _spec.AddField(account.FieldLoadFactor, field.TypeInt, value) + } + if _u.mutation.LoadFactorCleared() { + _spec.ClearField(account.FieldLoadFactor, field.TypeInt) + } if value, ok := _u.mutation.Priority(); ok { _spec.SetField(account.FieldPriority, field.TypeInt, value) } diff --git a/backend/ent/migrate/schema.go b/backend/ent/migrate/schema.go index 85e9407237..8e54f31c7b 100644 --- a/backend/ent/migrate/schema.go +++ b/backend/ent/migrate/schema.go @@ -106,6 +106,7 @@ var ( {Name: "credentials", Type: field.TypeJSON, SchemaType: map[string]string{"postgres": "jsonb"}}, {Name: "extra", Type: field.TypeJSON, SchemaType: map[string]string{"postgres": "jsonb"}}, {Name: "concurrency", Type: field.TypeInt, Default: 3}, + {Name: "load_factor", Type: field.TypeInt, Nullable: true}, {Name: "priority", Type: field.TypeInt, Default: 50}, {Name: "rate_multiplier", Type: field.TypeFloat64, Default: 1, SchemaType: map[string]string{"postgres": "decimal(10,4)"}}, {Name: "status", Type: field.TypeString, Size: 20, Default: "active"}, @@ -132,7 +133,7 @@ var ( ForeignKeys: []*schema.ForeignKey{ { Symbol: "accounts_proxies_proxy", - Columns: []*schema.Column{AccountsColumns[27]}, + Columns: []*schema.Column{AccountsColumns[28]}, RefColumns: []*schema.Column{ProxiesColumns[0]}, OnDelete: schema.SetNull, }, @@ -151,52 +152,52 @@ var ( { Name: "account_status", Unique: false, - Columns: []*schema.Column{AccountsColumns[13]}, + Columns: []*schema.Column{AccountsColumns[14]}, }, { Name: "account_proxy_id", Unique: false, - Columns: []*schema.Column{AccountsColumns[27]}, + Columns: []*schema.Column{AccountsColumns[28]}, }, { Name: "account_priority", Unique: false, - Columns: []*schema.Column{AccountsColumns[11]}, + Columns: []*schema.Column{AccountsColumns[12]}, }, { Name: "account_last_used_at", Unique: false, - Columns: []*schema.Column{AccountsColumns[15]}, + Columns: []*schema.Column{AccountsColumns[16]}, }, { Name: "account_schedulable", Unique: false, - Columns: []*schema.Column{AccountsColumns[18]}, + Columns: []*schema.Column{AccountsColumns[19]}, }, { Name: "account_rate_limited_at", Unique: false, - Columns: []*schema.Column{AccountsColumns[19]}, + Columns: []*schema.Column{AccountsColumns[20]}, }, { Name: "account_rate_limit_reset_at", Unique: false, - Columns: []*schema.Column{AccountsColumns[20]}, + Columns: []*schema.Column{AccountsColumns[21]}, }, { Name: "account_overload_until", Unique: false, - Columns: []*schema.Column{AccountsColumns[21]}, + Columns: []*schema.Column{AccountsColumns[22]}, }, { Name: "account_platform_priority", Unique: false, - Columns: []*schema.Column{AccountsColumns[6], AccountsColumns[11]}, + Columns: []*schema.Column{AccountsColumns[6], AccountsColumns[12]}, }, { Name: "account_priority_status", Unique: false, - Columns: []*schema.Column{AccountsColumns[11], AccountsColumns[13]}, + Columns: []*schema.Column{AccountsColumns[12], AccountsColumns[14]}, }, { Name: "account_deleted_at", diff --git a/backend/ent/mutation.go b/backend/ent/mutation.go index 85e2ea716e..6c6194a63c 100644 --- a/backend/ent/mutation.go +++ b/backend/ent/mutation.go @@ -2260,6 +2260,8 @@ type AccountMutation struct { extra *map[string]interface{} concurrency *int addconcurrency *int + load_factor *int + addload_factor *int priority *int addpriority *int rate_multiplier *float64 @@ -2845,6 +2847,76 @@ func (m *AccountMutation) ResetConcurrency() { m.addconcurrency = nil } +// SetLoadFactor sets the "load_factor" field. +func (m *AccountMutation) SetLoadFactor(i int) { + m.load_factor = &i + m.addload_factor = nil +} + +// LoadFactor returns the value of the "load_factor" field in the mutation. +func (m *AccountMutation) LoadFactor() (r int, exists bool) { + v := m.load_factor + if v == nil { + return + } + return *v, true +} + +// OldLoadFactor returns the old "load_factor" field's value of the Account entity. +// If the Account object wasn't provided to the builder, the object is fetched from the database. +// An error is returned if the mutation operation is not UpdateOne, or the database query fails. +func (m *AccountMutation) OldLoadFactor(ctx context.Context) (v *int, err error) { + if !m.op.Is(OpUpdateOne) { + return v, errors.New("OldLoadFactor is only allowed on UpdateOne operations") + } + if m.id == nil || m.oldValue == nil { + return v, errors.New("OldLoadFactor requires an ID field in the mutation") + } + oldValue, err := m.oldValue(ctx) + if err != nil { + return v, fmt.Errorf("querying old value for OldLoadFactor: %w", err) + } + return oldValue.LoadFactor, nil +} + +// AddLoadFactor adds i to the "load_factor" field. +func (m *AccountMutation) AddLoadFactor(i int) { + if m.addload_factor != nil { + *m.addload_factor += i + } else { + m.addload_factor = &i + } +} + +// AddedLoadFactor returns the value that was added to the "load_factor" field in this mutation. +func (m *AccountMutation) AddedLoadFactor() (r int, exists bool) { + v := m.addload_factor + if v == nil { + return + } + return *v, true +} + +// ClearLoadFactor clears the value of the "load_factor" field. +func (m *AccountMutation) ClearLoadFactor() { + m.load_factor = nil + m.addload_factor = nil + m.clearedFields[account.FieldLoadFactor] = struct{}{} +} + +// LoadFactorCleared returns if the "load_factor" field was cleared in this mutation. +func (m *AccountMutation) LoadFactorCleared() bool { + _, ok := m.clearedFields[account.FieldLoadFactor] + return ok +} + +// ResetLoadFactor resets all changes to the "load_factor" field. +func (m *AccountMutation) ResetLoadFactor() { + m.load_factor = nil + m.addload_factor = nil + delete(m.clearedFields, account.FieldLoadFactor) +} + // SetPriority sets the "priority" field. func (m *AccountMutation) SetPriority(i int) { m.priority = &i @@ -3773,7 +3845,7 @@ func (m *AccountMutation) Type() string { // order to get all numeric fields that were incremented/decremented, call // AddedFields(). func (m *AccountMutation) Fields() []string { - fields := make([]string, 0, 27) + fields := make([]string, 0, 28) if m.created_at != nil { fields = append(fields, account.FieldCreatedAt) } @@ -3807,6 +3879,9 @@ func (m *AccountMutation) Fields() []string { if m.concurrency != nil { fields = append(fields, account.FieldConcurrency) } + if m.load_factor != nil { + fields = append(fields, account.FieldLoadFactor) + } if m.priority != nil { fields = append(fields, account.FieldPriority) } @@ -3885,6 +3960,8 @@ func (m *AccountMutation) Field(name string) (ent.Value, bool) { return m.ProxyID() case account.FieldConcurrency: return m.Concurrency() + case account.FieldLoadFactor: + return m.LoadFactor() case account.FieldPriority: return m.Priority() case account.FieldRateMultiplier: @@ -3948,6 +4025,8 @@ func (m *AccountMutation) OldField(ctx context.Context, name string) (ent.Value, return m.OldProxyID(ctx) case account.FieldConcurrency: return m.OldConcurrency(ctx) + case account.FieldLoadFactor: + return m.OldLoadFactor(ctx) case account.FieldPriority: return m.OldPriority(ctx) case account.FieldRateMultiplier: @@ -4066,6 +4145,13 @@ func (m *AccountMutation) SetField(name string, value ent.Value) error { } m.SetConcurrency(v) return nil + case account.FieldLoadFactor: + v, ok := value.(int) + if !ok { + return fmt.Errorf("unexpected type %T for field %s", value, name) + } + m.SetLoadFactor(v) + return nil case account.FieldPriority: v, ok := value.(int) if !ok { @@ -4189,6 +4275,9 @@ func (m *AccountMutation) AddedFields() []string { if m.addconcurrency != nil { fields = append(fields, account.FieldConcurrency) } + if m.addload_factor != nil { + fields = append(fields, account.FieldLoadFactor) + } if m.addpriority != nil { fields = append(fields, account.FieldPriority) } @@ -4205,6 +4294,8 @@ func (m *AccountMutation) AddedField(name string) (ent.Value, bool) { switch name { case account.FieldConcurrency: return m.AddedConcurrency() + case account.FieldLoadFactor: + return m.AddedLoadFactor() case account.FieldPriority: return m.AddedPriority() case account.FieldRateMultiplier: @@ -4225,6 +4316,13 @@ func (m *AccountMutation) AddField(name string, value ent.Value) error { } m.AddConcurrency(v) return nil + case account.FieldLoadFactor: + v, ok := value.(int) + if !ok { + return fmt.Errorf("unexpected type %T for field %s", value, name) + } + m.AddLoadFactor(v) + return nil case account.FieldPriority: v, ok := value.(int) if !ok { @@ -4256,6 +4354,9 @@ func (m *AccountMutation) ClearedFields() []string { if m.FieldCleared(account.FieldProxyID) { fields = append(fields, account.FieldProxyID) } + if m.FieldCleared(account.FieldLoadFactor) { + fields = append(fields, account.FieldLoadFactor) + } if m.FieldCleared(account.FieldErrorMessage) { fields = append(fields, account.FieldErrorMessage) } @@ -4312,6 +4413,9 @@ func (m *AccountMutation) ClearField(name string) error { case account.FieldProxyID: m.ClearProxyID() return nil + case account.FieldLoadFactor: + m.ClearLoadFactor() + return nil case account.FieldErrorMessage: m.ClearErrorMessage() return nil @@ -4386,6 +4490,9 @@ func (m *AccountMutation) ResetField(name string) error { case account.FieldConcurrency: m.ResetConcurrency() return nil + case account.FieldLoadFactor: + m.ResetLoadFactor() + return nil case account.FieldPriority: m.ResetPriority() return nil @@ -10191,7 +10298,7 @@ func (m *GroupMutation) Type() string { // order to get all numeric fields that were incremented/decremented, call // AddedFields(). func (m *GroupMutation) Fields() []string { - fields := make([]string, 0, 30) + fields := make([]string, 0, 31) if m.created_at != nil { fields = append(fields, group.FieldCreatedAt) } diff --git a/backend/ent/runtime/runtime.go b/backend/ent/runtime/runtime.go index 2c7467f67c..7ae4d25314 100644 --- a/backend/ent/runtime/runtime.go +++ b/backend/ent/runtime/runtime.go @@ -212,29 +212,29 @@ func init() { // account.DefaultConcurrency holds the default value on creation for the concurrency field. account.DefaultConcurrency = accountDescConcurrency.Default.(int) // accountDescPriority is the schema descriptor for priority field. - accountDescPriority := accountFields[8].Descriptor() + accountDescPriority := accountFields[9].Descriptor() // account.DefaultPriority holds the default value on creation for the priority field. account.DefaultPriority = accountDescPriority.Default.(int) // accountDescRateMultiplier is the schema descriptor for rate_multiplier field. - accountDescRateMultiplier := accountFields[9].Descriptor() + accountDescRateMultiplier := accountFields[10].Descriptor() // account.DefaultRateMultiplier holds the default value on creation for the rate_multiplier field. account.DefaultRateMultiplier = accountDescRateMultiplier.Default.(float64) // accountDescStatus is the schema descriptor for status field. - accountDescStatus := accountFields[10].Descriptor() + accountDescStatus := accountFields[11].Descriptor() // account.DefaultStatus holds the default value on creation for the status field. account.DefaultStatus = accountDescStatus.Default.(string) // account.StatusValidator is a validator for the "status" field. It is called by the builders before save. account.StatusValidator = accountDescStatus.Validators[0].(func(string) error) // accountDescAutoPauseOnExpired is the schema descriptor for auto_pause_on_expired field. - accountDescAutoPauseOnExpired := accountFields[14].Descriptor() + accountDescAutoPauseOnExpired := accountFields[15].Descriptor() // account.DefaultAutoPauseOnExpired holds the default value on creation for the auto_pause_on_expired field. account.DefaultAutoPauseOnExpired = accountDescAutoPauseOnExpired.Default.(bool) // accountDescSchedulable is the schema descriptor for schedulable field. - accountDescSchedulable := accountFields[15].Descriptor() + accountDescSchedulable := accountFields[16].Descriptor() // account.DefaultSchedulable holds the default value on creation for the schedulable field. account.DefaultSchedulable = accountDescSchedulable.Default.(bool) // accountDescSessionWindowStatus is the schema descriptor for session_window_status field. - accountDescSessionWindowStatus := accountFields[23].Descriptor() + accountDescSessionWindowStatus := accountFields[24].Descriptor() // account.SessionWindowStatusValidator is a validator for the "session_window_status" field. It is called by the builders before save. account.SessionWindowStatusValidator = accountDescSessionWindowStatus.Validators[0].(func(string) error) accountgroupFields := schema.AccountGroup{}.Fields() diff --git a/backend/ent/schema/account.go b/backend/ent/schema/account.go index 443f9e09b7..5616d39915 100644 --- a/backend/ent/schema/account.go +++ b/backend/ent/schema/account.go @@ -97,6 +97,8 @@ func (Account) Fields() []ent.Field { field.Int("concurrency"). Default(3), + field.Int("load_factor").Optional().Nillable(), + // priority: 账户优先级，数值越小优先级越高 // 调度器会优先使用高优先级的账户 field.Int("priority"). diff --git a/backend/go.sum b/backend/go.sum index 32e389a76b..101613876f 100644 --- a/backend/go.sum +++ b/backend/go.sum @@ -124,6 +124,8 @@ github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/r github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc= github.com/distribution/reference v0.6.0 h1:0IXCQ5g4/QMHHkarYzh5l+u8T3t73zM5QvfrDyIgxBk= github.com/distribution/reference v0.6.0/go.mod h1:BbU0aIcezP1/5jX/8MP0YiH4SdvB5Y4f/wlDRiLyi3E= +github.com/dlclark/regexp2 v1.10.0 h1:+/GIL799phkJqYW+3YbOd8LCcbHzT0Pbo8zl70MHsq0= +github.com/dlclark/regexp2 v1.10.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8= github.com/docker/docker v28.5.1+incompatible h1:Bm8DchhSD2J6PsFzxC35TZo4TLGR2PdW/E69rU45NhM= github.com/docker/docker v28.5.1+incompatible/go.mod h1:eEKB0N0r5NX/I1kEveEz05bcu8tLC/8azJZsviup8Sk= github.com/docker/go-connections v0.6.0 h1:LlMG9azAe1TqfR7sO+NJttz1gy6KO7VJBh+pMmjSD94= @@ -171,8 +173,6 @@ github.com/goccy/go-json v0.10.2 h1:CrxCmQqYDkv1z7lO7Wbh2HN93uovUHgrECaO5ZrCXAU= github.com/goccy/go-json v0.10.2/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MGFi0w8I= github.com/golang-jwt/jwt/v5 v5.2.2 h1:Rl4B7itRWVtYIHFrSNd7vhTiz9UpLdi6gZhZ3wEeDy8= github.com/golang-jwt/jwt/v5 v5.2.2/go.mod h1:pqrtFR0X4osieyHYxtmOUWsAWrfe1Q5UVIyoH402zdk= -github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek= -github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps= github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE= github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE= github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8= @@ -182,7 +182,6 @@ github.com/google/go-querystring v1.1.0/go.mod h1:Kcdr2DB4koayq7X8pmAG4sNG59So17 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg= github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e h1:ijClszYn+mADRFY17kjQEVQ1XRhq2/JR1M3sGqeJoxs= github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e/go.mod h1:boTsfXsheKC2y+lKOCMpSfarhxDeIzfZG1jqGcPl3cA= -github.com/google/subcommands v1.2.0 h1:vWQspBTo2nEqTUFita5/KeEWlUL8kQObDFbub/EN9oE= github.com/google/subcommands v1.2.0/go.mod h1:ZjhPrFU+Olkh9WazFPsl27BQ4UPiG37m3yTrtFlrHVk= github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0= github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo= @@ -203,6 +202,8 @@ github.com/icholy/digest v1.1.0 h1:HfGg9Irj7i+IX1o1QAmPfIBNu/Q5A5Tu3n/MED9k9H4= github.com/icholy/digest v1.1.0/go.mod h1:QNrsSGQ5v7v9cReDI0+eyjsXGUoRSUZQHeQ5C4XLa0Y= github.com/imroc/req/v3 v3.57.0 h1:LMTUjNRUybUkTPn8oJDq8Kg3JRBOBTcnDhKu7mzupKI= github.com/imroc/req/v3 v3.57.0/go.mod h1:JL62ey1nvSLq81HORNcosvlf7SxZStONNqOprg0Pz00= +github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8= +github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw= github.com/jackc/pgpassfile v1.0.0 h1:/6Hmqy13Ss2zCq62VdNG8tM1wchn8zjSGOBJ6icpsIM= github.com/jackc/pgpassfile v1.0.0/go.mod h1:CEx0iS5ambNFdcRtxPj5JhEz+xB6uRky5eyVu/W2HEg= github.com/jackc/pgservicefile v0.0.0-20240606120523-5a60cdf6a761 h1:iCEnooe7UlwOQYpKFhBabPMi4aNAfoODPEFNiAnClxo= @@ -285,6 +286,10 @@ github.com/pelletier/go-toml/v2 v2.2.2 h1:aYUidT7k73Pcl9nb2gScu7NSrKCSHIDE89b3+6 github.com/pelletier/go-toml/v2 v2.2.2/go.mod h1:1t835xjRzz80PqgE6HHgN2JOsmgYu/h4qDAS4n929Rs= github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4= github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0= +github.com/pkoukk/tiktoken-go v0.1.8 h1:85ENo+3FpWgAACBaEUVp+lctuTcYUO7BtmfhlN/QTRo= +github.com/pkoukk/tiktoken-go v0.1.8/go.mod h1:9NiV+i9mJKGj1rYOT+njbv+ZwA/zJxYdewGl6qVatpg= +github.com/pkoukk/tiktoken-go-loader v0.0.2 h1:LUKws63GV3pVHwH1srkBplBv+7URgmOmhSkRxsIvsK4= +github.com/pkoukk/tiktoken-go-loader v0.0.2/go.mod h1:4mIkYyZooFlnenDlormIo6cd5wrlUKNr97wp9nGgEKo= github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4= github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U= github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4= @@ -398,8 +403,6 @@ go.opentelemetry.io/otel/metric v1.37.0 h1:mvwbQS5m0tbmqML4NqK+e3aDiO02vsf/Wgbsd go.opentelemetry.io/otel/metric v1.37.0/go.mod h1:04wGrZurHYKOc+RKeye86GwKiTb9FKm1WHtO+4EVr2E= go.opentelemetry.io/otel/sdk v1.37.0 h1:ItB0QUqnjesGRvNcmAcU0LyvkVyGJ2xftD29bWdDvKI= go.opentelemetry.io/otel/sdk v1.37.0/go.mod h1:VredYzxUvuo2q3WRcDnKDjbdvmO0sCzOvVAiY+yUkAg= -go.opentelemetry.io/otel/sdk/metric v1.37.0 h1:90lI228XrB9jCMuSdA0673aubgRobVZFhbjxHHspCPc= -go.opentelemetry.io/otel/sdk/metric v1.37.0/go.mod h1:cNen4ZWfiD37l5NhS+Keb5RXVWZWpRE+9WyVCpbo5ps= go.opentelemetry.io/otel/trace v1.37.0 h1:HLdcFNbRQBE2imdSEgm/kwqmQj1Or1l/7bW6mxVK7z4= go.opentelemetry.io/otel/trace v1.37.0/go.mod h1:TlgrlQ+PtQO5XFerSPUYG0JSgGyryXewPGyayAWSBS0= go.opentelemetry.io/proto/otlp v1.3.1 h1:TrMUixzpM0yuc/znrFTP9MMRh8trP93mkCiDVeXrui0= @@ -455,8 +458,6 @@ golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGm golang.org/x/tools v0.41.0 h1:a9b8iMweWG+S0OBnlU36rzLp20z1Rp10w+IY2czHTQc= golang.org/x/tools v0.41.0/go.mod h1:XSY6eDqxVNiYgezAVqqCeihT4j1U2CCsqvH3WhQpnlg= golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0= -gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk= -gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E= google.golang.org/genproto v0.0.0-20231106174013-bbf56f31fb17 h1:wpZ8pe2x1Q3f2KyT5f8oP/fa9rHAKgFPr/HZdNuS+PQ= google.golang.org/genproto/googleapis/api v0.0.0-20250929231259-57b25ae835d4 h1:8XJ4pajGwOlasW+L13MnEGA8W4115jJySQtVfS2/IBU= google.golang.org/genproto/googleapis/api v0.0.0-20250929231259-57b25ae835d4/go.mod h1:NnuHhy+bxcg30o7FnVAZbXsPHUDQ9qKWAQKCD7VxFtk= diff --git a/backend/internal/handler/admin/account_handler.go b/backend/internal/handler/admin/account_handler.go index 14f9e05d6a..6c1be84a82 100644 --- a/backend/internal/handler/admin/account_handler.go +++ b/backend/internal/handler/admin/account_handler.go @@ -102,6 +102,7 @@ type CreateAccountRequest struct { Concurrency int `json:"concurrency"` Priority int `json:"priority"` RateMultiplier *float64 `json:"rate_multiplier"` + LoadFactor *int `json:"load_factor"` GroupIDs []int64 `json:"group_ids"` ExpiresAt *int64 `json:"expires_at"` AutoPauseOnExpired *bool `json:"auto_pause_on_expired"` @@ -120,6 +121,7 @@ type UpdateAccountRequest struct { Concurrency *int `json:"concurrency"` Priority *int `json:"priority"` RateMultiplier *float64 `json:"rate_multiplier"` + LoadFactor *int `json:"load_factor"` Status string `json:"status" binding:"omitempty,oneof=active inactive"` GroupIDs *[]int64 `json:"group_ids"` ExpiresAt *int64 `json:"expires_at"` @@ -135,6 +137,7 @@ type BulkUpdateAccountsRequest struct { Concurrency *int `json:"concurrency"` Priority *int `json:"priority"` RateMultiplier *float64 `json:"rate_multiplier"` + LoadFactor *int `json:"load_factor"` Status string `json:"status" binding:"omitempty,oneof=active inactive error"` Schedulable *bool `json:"schedulable"` GroupIDs *[]int64 `json:"group_ids"` @@ -506,6 +509,7 @@ func (h *AccountHandler) Create(c *gin.Context) { Concurrency: req.Concurrency, Priority: req.Priority, RateMultiplier: req.RateMultiplier, + LoadFactor: req.LoadFactor, GroupIDs: req.GroupIDs, ExpiresAt: req.ExpiresAt, AutoPauseOnExpired: req.AutoPauseOnExpired, @@ -575,6 +579,7 @@ func (h *AccountHandler) Update(c *gin.Context) { Concurrency: req.Concurrency, // 指针类型，nil 表示未提供 Priority: req.Priority, // 指针类型，nil 表示未提供 RateMultiplier: req.RateMultiplier, + LoadFactor: req.LoadFactor, Status: req.Status, GroupIDs: req.GroupIDs, ExpiresAt: req.ExpiresAt, @@ -1101,6 +1106,7 @@ func (h *AccountHandler) BulkUpdate(c *gin.Context) { req.Concurrency != nil || req.Priority != nil || req.RateMultiplier != nil || + req.LoadFactor != nil || req.Status != "" || req.Schedulable != nil || req.GroupIDs != nil || @@ -1119,6 +1125,7 @@ func (h *AccountHandler) BulkUpdate(c *gin.Context) { Concurrency: req.Concurrency, Priority: req.Priority, RateMultiplier: req.RateMultiplier, + LoadFactor: req.LoadFactor, Status: req.Status, Schedulable: req.Schedulable, GroupIDs: req.GroupIDs, diff --git a/backend/internal/handler/dto/mappers.go b/backend/internal/handler/dto/mappers.go index fe2a1d773b..c0a5105651 100644 --- a/backend/internal/handler/dto/mappers.go +++ b/backend/internal/handler/dto/mappers.go @@ -183,6 +183,7 @@ func AccountFromServiceShallow(a *service.Account) *Account { Extra: a.Extra, ProxyID: a.ProxyID, Concurrency: a.Concurrency, + LoadFactor: a.LoadFactor, Priority: a.Priority, RateMultiplier: a.BillingRateMultiplier(), Status: a.Status, diff --git a/backend/internal/handler/dto/types.go b/backend/internal/handler/dto/types.go index 920615f708..f4339bb651 100644 --- a/backend/internal/handler/dto/types.go +++ b/backend/internal/handler/dto/types.go @@ -131,6 +131,7 @@ type Account struct { Extra map[string]any `json:"extra"` ProxyID *int64 `json:"proxy_id"` Concurrency int `json:"concurrency"` + LoadFactor *int `json:"load_factor,omitempty"` Priority int `json:"priority"` RateMultiplier float64 `json:"rate_multiplier"` Status string `json:"status"` diff --git a/backend/internal/repository/account_repo.go b/backend/internal/repository/account_repo.go index 6f0c5424aa..30b1a864b9 100644 --- a/backend/internal/repository/account_repo.go +++ b/backend/internal/repository/account_repo.go @@ -84,6 +84,9 @@ func (r *accountRepository) Create(ctx context.Context, account *service.Account if account.RateMultiplier != nil { builder.SetRateMultiplier(*account.RateMultiplier) } + if account.LoadFactor != nil { + builder.SetLoadFactor(*account.LoadFactor) + } if account.ProxyID != nil { builder.SetProxyID(*account.ProxyID) @@ -318,6 +321,11 @@ func (r *accountRepository) Update(ctx context.Context, account *service.Account if account.RateMultiplier != nil { builder.SetRateMultiplier(*account.RateMultiplier) } + if account.LoadFactor != nil { + builder.SetLoadFactor(*account.LoadFactor) + } else { + builder.ClearLoadFactor() + } if account.ProxyID != nil { builder.SetProxyID(*account.ProxyID) @@ -1223,6 +1231,15 @@ func (r *accountRepository) BulkUpdate(ctx context.Context, ids []int64, updates args = append(args, *updates.RateMultiplier) idx++ } + if updates.LoadFactor != nil { + if *updates.LoadFactor <= 0 { + setClauses = append(setClauses, "load_factor = NULL") + } else { + setClauses = append(setClauses, "load_factor = $"+itoa(idx)) + args = append(args, *updates.LoadFactor) + idx++ + } + } if updates.Status != nil { setClauses = append(setClauses, "status = $"+itoa(idx)) args = append(args, *updates.Status) @@ -1545,6 +1562,7 @@ func accountEntityToService(m *dbent.Account) *service.Account { Concurrency: m.Concurrency, Priority: m.Priority, RateMultiplier: &rateMultiplier, + LoadFactor: m.LoadFactor, Status: m.Status, ErrorMessage: derefString(m.ErrorMessage), LastUsedAt: m.LastUsedAt, diff --git a/backend/internal/server/api_contract_test.go b/backend/internal/server/api_contract_test.go index 40b2d592db..aafbbe21cb 100644 --- a/backend/internal/server/api_contract_test.go +++ b/backend/internal/server/api_contract_test.go @@ -1096,6 +1096,14 @@ func (s *stubAccountRepo) UpdateExtra(ctx context.Context, id int64, updates map return errors.New("not implemented") } +func (s *stubAccountRepo) IncrementQuotaUsed(ctx context.Context, id int64, amount float64) error { + return errors.New("not implemented") +} + +func (s *stubAccountRepo) ResetQuotaUsed(ctx context.Context, id int64) error { + return errors.New("not implemented") +} + func (s *stubAccountRepo) BulkUpdate(ctx context.Context, ids []int64, updates service.AccountBulkUpdate) (int64, error) { s.bulkUpdateIDs = append([]int64{}, ids...) return int64(len(ids)), nil diff --git a/backend/internal/service/account.go b/backend/internal/service/account.go index 7d56b7541a..2bf9c0bf10 100644 --- a/backend/internal/service/account.go +++ b/backend/internal/service/account.go @@ -28,6 +28,7 @@ type Account struct { // RateMultiplier 账号计费倍率（>=0，允许 0 表示该账号计费为 0）。 // 使用指针用于兼容旧版本调度缓存（Redis）中缺字段的情况：nil 表示按 1.0 处理。 RateMultiplier *float64 + LoadFactor *int // 调度负载因子；nil 表示使用 Concurrency Status string ErrorMessage string LastUsedAt *time.Time @@ -88,6 +89,19 @@ func (a *Account) BillingRateMultiplier() float64 { return *a.RateMultiplier } +func (a *Account) EffectiveLoadFactor() int { + if a == nil { + return 1 + } + if a.LoadFactor != nil && *a.LoadFactor > 0 { + return *a.LoadFactor + } + if a.Concurrency > 0 { + return a.Concurrency + } + return 1 +} + func (a *Account) IsSchedulable() bool { if !a.IsActive() || !a.Schedulable { return false diff --git a/backend/internal/service/account_load_factor_test.go b/backend/internal/service/account_load_factor_test.go new file mode 100644 index 0000000000..1cd1b17c9c --- /dev/null +++ b/backend/internal/service/account_load_factor_test.go @@ -0,0 +1,42 @@ +package service + +import ( + "testing" + + "github.com/stretchr/testify/require" +) + +func TestEffectiveLoadFactor_NilAccount(t *testing.T) { + var a *Account + require.Equal(t, 1, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_NilLoadFactor_PositiveConcurrency(t *testing.T) { + a := &Account{Concurrency: 5} + require.Equal(t, 5, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_NilLoadFactor_ZeroConcurrency(t *testing.T) { + a := &Account{Concurrency: 0} + require.Equal(t, 1, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_PositiveLoadFactor(t *testing.T) { + a := &Account{Concurrency: 5, LoadFactor: intPtr(20)} + require.Equal(t, 20, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_ZeroLoadFactor_FallbackToConcurrency(t *testing.T) { + a := &Account{Concurrency: 5, LoadFactor: intPtr(0)} + require.Equal(t, 5, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_NegativeLoadFactor_FallbackToConcurrency(t *testing.T) { + a := &Account{Concurrency: 3, LoadFactor: intPtr(-1)} + require.Equal(t, 3, a.EffectiveLoadFactor()) +} + +func TestEffectiveLoadFactor_ZeroLoadFactor_ZeroConcurrency(t *testing.T) { + a := &Account{Concurrency: 0, LoadFactor: intPtr(0)} + require.Equal(t, 1, a.EffectiveLoadFactor()) +} diff --git a/backend/internal/service/account_service.go b/backend/internal/service/account_service.go index 18a70c5cc2..cfe20eb13e 100644 --- a/backend/internal/service/account_service.go +++ b/backend/internal/service/account_service.go @@ -78,6 +78,7 @@ type AccountBulkUpdate struct { Concurrency *int Priority *int RateMultiplier *float64 + LoadFactor *int Status *string Schedulable *bool Credentials map[string]any diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index 67e7c783c1..e2d4b4580f 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -195,6 +195,7 @@ type CreateAccountInput struct { Concurrency int Priority int RateMultiplier *float64 // 账号计费倍率（>=0，允许 0） + LoadFactor *int GroupIDs []int64 ExpiresAt *int64 AutoPauseOnExpired *bool @@ -215,6 +216,7 @@ type UpdateAccountInput struct { Concurrency *int // 使用指针区分"未提供"和"设置为0" Priority *int // 使用指针区分"未提供"和"设置为0" RateMultiplier *float64 // 账号计费倍率（>=0，允许 0） + LoadFactor *int Status string GroupIDs *[]int64 ExpiresAt *int64 @@ -230,6 +232,7 @@ type BulkUpdateAccountsInput struct { Concurrency *int Priority *int RateMultiplier *float64 // 账号计费倍率（>=0，允许 0） + LoadFactor *int Status string Schedulable *bool GroupIDs *[]int64 @@ -1413,6 +1416,9 @@ func (s *adminServiceImpl) CreateAccount(ctx context.Context, input *CreateAccou } account.RateMultiplier = input.RateMultiplier } + if input.LoadFactor != nil && *input.LoadFactor > 0 { + account.LoadFactor = input.LoadFactor + } if err := s.accountRepo.Create(ctx, account); err != nil { return nil, err } @@ -1483,6 +1489,13 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U } account.RateMultiplier = input.RateMultiplier } + if input.LoadFactor != nil { + if *input.LoadFactor <= 0 { + account.LoadFactor = nil // 0 或负数表示清除 + } else { + account.LoadFactor = input.LoadFactor + } + } if input.Status != "" { account.Status = input.Status } @@ -1616,6 +1629,9 @@ func (s *adminServiceImpl) BulkUpdateAccounts(ctx context.Context, input *BulkUp if input.RateMultiplier != nil { repoUpdates.RateMultiplier = input.RateMultiplier } + if input.LoadFactor != nil { + repoUpdates.LoadFactor = input.LoadFactor + } if input.Status != "" { repoUpdates.Status = &input.Status } diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index 132361f447..6d2922bc2d 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -1311,7 +1311,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro for _, acc := range routingCandidates { routingLoads = append(routingLoads, AccountWithConcurrency{ ID: acc.ID, - MaxConcurrency: acc.Concurrency, + MaxConcurrency: acc.EffectiveLoadFactor(), }) } routingLoadMap, _ := s.concurrencyService.GetAccountsLoadBatch(ctx, routingLoads) @@ -1499,7 +1499,7 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro for _, acc := range candidates { accountLoads = append(accountLoads, AccountWithConcurrency{ ID: acc.ID, - MaxConcurrency: acc.Concurrency, + MaxConcurrency: acc.EffectiveLoadFactor(), }) } diff --git a/backend/internal/service/openai_account_scheduler.go b/backend/internal/service/openai_account_scheduler.go index 99013ce555..9cc6aae163 100644 --- a/backend/internal/service/openai_account_scheduler.go +++ b/backend/internal/service/openai_account_scheduler.go @@ -590,7 +590,7 @@ func (s *defaultOpenAIAccountScheduler) selectByLoadBalance( filtered = append(filtered, account) loadReq = append(loadReq, AccountWithConcurrency{ ID: account.ID, - MaxConcurrency: account.Concurrency, + MaxConcurrency: account.EffectiveLoadFactor(), }) } if len(filtered) == 0 { diff --git a/backend/internal/service/openai_gateway_service.go b/backend/internal/service/openai_gateway_service.go index d92b2ecf77..ceca4510f6 100644 --- a/backend/internal/service/openai_gateway_service.go +++ b/backend/internal/service/openai_gateway_service.go @@ -1242,7 +1242,7 @@ func (s *OpenAIGatewayService) SelectAccountWithLoadAwareness(ctx context.Contex for _, acc := range candidates { accountLoads = append(accountLoads, AccountWithConcurrency{ ID: acc.ID, - MaxConcurrency: acc.Concurrency, + MaxConcurrency: acc.EffectiveLoadFactor(), }) } diff --git a/backend/internal/service/ops_concurrency.go b/backend/internal/service/ops_concurrency.go index 92b37e7313..c03108c4b9 100644 --- a/backend/internal/service/ops_concurrency.go +++ b/backend/internal/service/ops_concurrency.go @@ -64,8 +64,9 @@ func (s *OpsService) getAccountsLoadMapBestEffort(ctx context.Context, accounts if acc.ID <= 0 { continue } - if prev, ok := unique[acc.ID]; !ok || acc.Concurrency > prev { - unique[acc.ID] = acc.Concurrency + lf := acc.EffectiveLoadFactor() + if prev, ok := unique[acc.ID]; !ok || lf > prev { + unique[acc.ID] = lf } } diff --git a/backend/internal/service/ops_metrics_collector.go b/backend/internal/service/ops_metrics_collector.go index 30adaae066..6c33707138 100644 --- a/backend/internal/service/ops_metrics_collector.go +++ b/backend/internal/service/ops_metrics_collector.go @@ -389,13 +389,9 @@ func (c *OpsMetricsCollector) collectConcurrencyQueueDepth(parentCtx context.Con if acc.ID <= 0 { continue } - maxConc := acc.Concurrency - if maxConc < 0 { - maxConc = 0 - } batch = append(batch, AccountWithConcurrency{ ID: acc.ID, - MaxConcurrency: maxConc, + MaxConcurrency: acc.EffectiveLoadFactor(), }) } if len(batch) == 0 { diff --git a/backend/migrations/067_add_account_load_factor.sql b/backend/migrations/067_add_account_load_factor.sql new file mode 100644 index 0000000000..6805e8c286 --- /dev/null +++ b/backend/migrations/067_add_account_load_factor.sql @@ -0,0 +1 @@ +ALTER TABLE accounts ADD COLUMN IF NOT EXISTS load_factor INTEGER; diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 1c83e65831..d7853b235a 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -469,7 +469,7 @@
-
+
+
+
+ + +
+ +

{{ t('admin.accounts.loadFactorHint') }}

+
-
+
+
+ + +

{{ t('admin.accounts.loadFactorHint') }}

+
, proxy_id: null as number | null, concurrency: 10, + load_factor: null as number | null, priority: 1, rate_multiplier: 1, group_ids: [] as number[], @@ -3112,6 +3119,7 @@ const resetForm = () => { form.credentials = {} form.proxy_id = null form.concurrency = 10 + form.load_factor = null form.priority = 1 form.rate_multiplier = 1 form.group_ids = [] @@ -3483,6 +3491,7 @@ const handleImportAccessToken = async (accessTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3542,6 +3551,7 @@ const createAccountAndFinish = async ( extra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3597,6 +3607,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3626,6 +3637,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3703,6 +3715,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3730,6 +3743,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3818,6 +3832,7 @@ const handleSoraValidateST = async (sessionTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3906,6 +3921,7 @@ const handleAntigravityValidateRT = async (refreshTokenInput: string) => { extra: {}, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -4064,8 +4080,11 @@ const handleAnthropicExchange = async (authCode: string) => { } // Add RPM limit settings - if (rpmLimitEnabled.value && baseRpm.value != null && baseRpm.value > 0) { - extra.base_rpm = baseRpm.value + if (rpmLimitEnabled.value) { + const DEFAULT_BASE_RPM = 15 + extra.base_rpm = (baseRpm.value != null && baseRpm.value > 0) + ? baseRpm.value + : DEFAULT_BASE_RPM extra.rpm_strategy = rpmStrategy.value if (rpmStickyBuffer.value != null && rpmStickyBuffer.value > 0) { extra.rpm_sticky_buffer = rpmStickyBuffer.value @@ -4176,8 +4195,11 @@ const handleCookieAuth = async (sessionKey: string) => { } // Add RPM limit settings - if (rpmLimitEnabled.value && baseRpm.value != null && baseRpm.value > 0) { - extra.base_rpm = baseRpm.value + if (rpmLimitEnabled.value) { + const DEFAULT_BASE_RPM = 15 + extra.base_rpm = (baseRpm.value != null && baseRpm.value > 0) + ? baseRpm.value + : DEFAULT_BASE_RPM extra.rpm_strategy = rpmStrategy.value if (rpmStickyBuffer.value != null && rpmStickyBuffer.value > 0) { extra.rpm_sticky_buffer = rpmStickyBuffer.value @@ -4223,6 +4245,7 @@ const handleCookieAuth = async (sessionKey: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, + load_factor: form.load_factor || undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 09b39bc060..764fd9bef0 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -650,11 +650,17 @@
-
+
+
+ + +

{{ t('admin.accounts.loadFactorHint') }}

+
{ if (form.expires_at === null) { updatePayload.expires_at = 0 } + // load_factor: 空值/0/NaN 时发送 0（后端约定 0 = 清除） + if (!form.load_factor || form.load_factor <= 0) { + updatePayload.load_factor = 0 + } updatePayload.auto_pause_on_expired = autoPauseOnExpired.value // For apikey type, handle credentials update @@ -2188,8 +2200,11 @@ const handleSubmit = async () => { } // RPM limit settings - if (rpmLimitEnabled.value && baseRpm.value != null && baseRpm.value > 0) { - newExtra.base_rpm = baseRpm.value + if (rpmLimitEnabled.value) { + const DEFAULT_BASE_RPM = 15 + newExtra.base_rpm = (baseRpm.value != null && baseRpm.value > 0) + ? baseRpm.value + : DEFAULT_BASE_RPM newExtra.rpm_strategy = rpmStrategy.value if (rpmStickyBuffer.value != null && rpmStickyBuffer.value > 0) { newExtra.rpm_sticky_buffer = rpmStickyBuffer.value diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index 055998a7fe..d08aed4d38 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -1991,10 +1991,12 @@ export default { proxy: 'Proxy', noProxy: 'No Proxy', concurrency: 'Concurrency', + loadFactor: 'Load Factor', + loadFactorHint: 'Defaults to concurrency', priority: 'Priority', priorityHint: 'Lower value accounts are used first', billingRateMultiplier: 'Billing Rate Multiplier', - billingRateMultiplierHint: '>=0, 0 means free. Affects account billing only', + billingRateMultiplierHint: '0 = free, affects account billing only', expiresAt: 'Expires At', expiresAtHint: 'Leave empty for no expiration', higherPriorityFirst: 'Lower value means higher priority', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index cc203adbd6..3dc70d1e1a 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -2133,10 +2133,12 @@ export default { proxy: '代理', noProxy: '无代理', concurrency: '并发数', + loadFactor: '负载因子', + loadFactorHint: '不填则等于并发数', priority: '优先级', priorityHint: '优先级越小的账号优先使用', billingRateMultiplier: '账号计费倍率', - billingRateMultiplierHint: '>=0，0 表示该账号计费为 0；仅影响账号计费口径', + billingRateMultiplierHint: '0 表示不计费，仅影响账号计费', expiresAt: '过期时间', expiresAtHint: '留空表示不过期', higherPriorityFirst: '数值越小优先级越高', diff --git a/frontend/src/types/index.ts b/frontend/src/types/index.ts index 915822f079..46725965d4 100644 --- a/frontend/src/types/index.ts +++ b/frontend/src/types/index.ts @@ -653,6 +653,7 @@ export interface Account { } & Record) proxy_id: number | null concurrency: number + load_factor?: number | null current_concurrency?: number // Real-time concurrency count from Redis priority: number rate_multiplier?: number // Account billing multiplier (>=0, 0 means free) @@ -783,6 +784,7 @@ export interface CreateAccountRequest { extra?: Record proxy_id?: number | null concurrency?: number + load_factor?: number | null priority?: number rate_multiplier?: number // Account billing multiplier (>=0, 0 means free) group_ids?: number[] @@ -799,6 +801,7 @@ export interface UpdateAccountRequest { extra?: Record proxy_id?: number | null concurrency?: number + load_factor?: number | null priority?: number rate_multiplier?: number // Account billing multiplier (>=0, 0 means free) schedulable?: boolean From 01b3a09d7dd0c6a5b4bc3adcc369912795769084 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:16:47 +0800 Subject: [PATCH 14/53] fix: validate account status before update and update load factor hint MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Normalize non-standard status (e.g. "error") to "active" on edit load - Add pre-submit validation for status field to prevent 400 errors - Update load factor hint: "提高负载因子可以提高对账号的调度频率" --- frontend/src/components/account/EditAccountModal.vue | 9 ++++++++- frontend/src/i18n/locales/en.ts | 3 ++- frontend/src/i18n/locales/zh.ts | 3 ++- 3 files changed, 12 insertions(+), 3 deletions(-) diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 764fd9bef0..196e346cb0 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -1508,7 +1508,9 @@ watch( form.load_factor = newAccount.load_factor ?? null form.priority = newAccount.priority form.rate_multiplier = newAccount.rate_multiplier ?? 1 - form.status = newAccount.status as 'active' | 'inactive' + form.status = (newAccount.status === 'active' || newAccount.status === 'inactive') + ? newAccount.status + : 'active' form.group_ids = newAccount.group_ids || [] form.expires_at = newAccount.expires_at ?? null @@ -2048,6 +2050,11 @@ const handleSubmit = async () => { if (!props.account) return const accountID = props.account.id + if (form.status !== 'active' && form.status !== 'inactive') { + appStore.showError(t('admin.accounts.pleaseSelectStatus')) + return + } + const updatePayload: Record = { ...form } try { // 后端期望 proxy_id: 0 表示清除代理，而不是 null diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index d08aed4d38..89a4f0a041 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -1992,7 +1992,7 @@ export default { noProxy: 'No Proxy', concurrency: 'Concurrency', loadFactor: 'Load Factor', - loadFactorHint: 'Defaults to concurrency', + loadFactorHint: 'Higher load factor increases scheduling frequency', priority: 'Priority', priorityHint: 'Lower value accounts are used first', billingRateMultiplier: 'Billing Rate Multiplier', @@ -2012,6 +2012,7 @@ export default { accountUpdated: 'Account updated successfully', failedToCreate: 'Failed to create account', failedToUpdate: 'Failed to update account', + pleaseSelectStatus: 'Please select a valid account status', mixedChannelWarningTitle: 'Mixed Channel Warning', mixedChannelWarning: 'Warning: Group "{groupName}" contains both {currentPlatform} and {otherPlatform} accounts. Mixing different channels may cause thinking block signature validation issues, which will fallback to non-thinking mode. Are you sure you want to continue?', pleaseEnterAccountName: 'Please enter account name', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index 3dc70d1e1a..67f50e83de 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -2134,7 +2134,7 @@ export default { noProxy: '无代理', concurrency: '并发数', loadFactor: '负载因子', - loadFactorHint: '不填则等于并发数', + loadFactorHint: '提高负载因子可以提高对账号的调度频率', priority: '优先级', priorityHint: '优先级越小的账号优先使用', billingRateMultiplier: '账号计费倍率', @@ -2154,6 +2154,7 @@ export default { accountUpdated: '账号更新成功', failedToCreate: '创建账号失败', failedToUpdate: '更新账号失败', + pleaseSelectStatus: '请选择有效的账号状态', mixedChannelWarningTitle: '混合渠道警告', mixedChannelWarning: '警告：分组 "{groupName}" 中同时包含 {currentPlatform} 和 {otherPlatform} 账号。混合使用不同渠道可能导致 thinking block 签名验证问题，会自动回退到非 thinking 模式。确定要继续吗？', pleaseEnterAccountName: '请输入账号名称', From 12b33f4ea41ac5ac959f3fb9d858f863254aa927 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:27:41 +0800 Subject: [PATCH 15/53] fix: address load_factor code review findings - Fix bulk edit: send 0 instead of null/NaN to clear load_factor - Fix edit modal: explicit NaN check instead of implicit falsy - Fix create modal: use ?? instead of || for load_factor - Add load_factor upper limit validation (max 10000) - Add //go:build unit tag and self-contained intPtrHelper in test - Add design intent comments on WaitPlan.MaxConcurrency --- .../service/account_load_factor_test.go | 12 ++++++++---- backend/internal/service/admin_service.go | 5 +++++ .../service/openai_account_scheduler.go | 2 ++ .../account/BulkEditAccountModal.vue | 4 +++- .../components/account/CreateAccountModal.vue | 18 +++++++++--------- .../components/account/EditAccountModal.vue | 5 +++-- 6 files changed, 30 insertions(+), 16 deletions(-) diff --git a/backend/internal/service/account_load_factor_test.go b/backend/internal/service/account_load_factor_test.go index 1cd1b17c9c..a4d78a4bba 100644 --- a/backend/internal/service/account_load_factor_test.go +++ b/backend/internal/service/account_load_factor_test.go @@ -1,3 +1,5 @@ +//go:build unit + package service import ( @@ -6,6 +8,8 @@ import ( "github.com/stretchr/testify/require" ) +func intPtrHelper(v int) *int { return &v } + func TestEffectiveLoadFactor_NilAccount(t *testing.T) { var a *Account require.Equal(t, 1, a.EffectiveLoadFactor()) @@ -22,21 +26,21 @@ func TestEffectiveLoadFactor_NilLoadFactor_ZeroConcurrency(t *testing.T) { } func TestEffectiveLoadFactor_PositiveLoadFactor(t *testing.T) { - a := &Account{Concurrency: 5, LoadFactor: intPtr(20)} + a := &Account{Concurrency: 5, LoadFactor: intPtrHelper(20)} require.Equal(t, 20, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_ZeroLoadFactor_FallbackToConcurrency(t *testing.T) { - a := &Account{Concurrency: 5, LoadFactor: intPtr(0)} + a := &Account{Concurrency: 5, LoadFactor: intPtrHelper(0)} require.Equal(t, 5, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_NegativeLoadFactor_FallbackToConcurrency(t *testing.T) { - a := &Account{Concurrency: 3, LoadFactor: intPtr(-1)} + a := &Account{Concurrency: 3, LoadFactor: intPtrHelper(-1)} require.Equal(t, 3, a.EffectiveLoadFactor()) } func TestEffectiveLoadFactor_ZeroLoadFactor_ZeroConcurrency(t *testing.T) { - a := &Account{Concurrency: 0, LoadFactor: intPtr(0)} + a := &Account{Concurrency: 0, LoadFactor: intPtrHelper(0)} require.Equal(t, 1, a.EffectiveLoadFactor()) } diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index e2d4b4580f..e5ceac078a 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -1417,6 +1417,9 @@ func (s *adminServiceImpl) CreateAccount(ctx context.Context, input *CreateAccou account.RateMultiplier = input.RateMultiplier } if input.LoadFactor != nil && *input.LoadFactor > 0 { + if *input.LoadFactor > 10000 { + return nil, errors.New("load_factor must be <= 10000") + } account.LoadFactor = input.LoadFactor } if err := s.accountRepo.Create(ctx, account); err != nil { @@ -1492,6 +1495,8 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U if input.LoadFactor != nil { if *input.LoadFactor <= 0 { account.LoadFactor = nil // 0 或负数表示清除 + } else if *input.LoadFactor > 10000 { + return nil, errors.New("load_factor must be <= 10000") } else { account.LoadFactor = input.LoadFactor } diff --git a/backend/internal/service/openai_account_scheduler.go b/backend/internal/service/openai_account_scheduler.go index 9cc6aae163..cf4bc26eb0 100644 --- a/backend/internal/service/openai_account_scheduler.go +++ b/backend/internal/service/openai_account_scheduler.go @@ -342,6 +342,7 @@ func (s *defaultOpenAIAccountScheduler) selectBySessionHash( } cfg := s.service.schedulingConfig() + // WaitPlan.MaxConcurrency 使用 Concurrency（非 EffectiveLoadFactor），因为 WaitPlan 控制的是 Redis 实际并发槽位等待。 if s.service.concurrencyService != nil { return &AccountSelectionResult{ Account: account, @@ -703,6 +704,7 @@ func (s *defaultOpenAIAccountScheduler) selectByLoadBalance( } cfg := s.service.schedulingConfig() + // WaitPlan.MaxConcurrency 使用 Concurrency（非 EffectiveLoadFactor），因为 WaitPlan 控制的是 Redis 实际并发槽位等待。 candidate := selectionOrder[0] return &AccountSelectionResult{ Account: candidate.account, diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index d7853b235a..12ca5a703c 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -1227,7 +1227,9 @@ const buildUpdatePayload = (): Record | null => { } if (enableLoadFactor.value) { - updates.load_factor = loadFactor.value + // 空值/NaN/0 时发送 0（后端约定 <= 0 表示清除） + const lf = loadFactor.value + updates.load_factor = (lf != null && !Number.isNaN(lf) && lf > 0) ? lf : 0 } if (enablePriority.value) { diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index a1113267fb..c5420ca498 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -3491,7 +3491,7 @@ const handleImportAccessToken = async (accessTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3551,7 +3551,7 @@ const createAccountAndFinish = async ( extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3607,7 +3607,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3637,7 +3637,7 @@ const handleOpenAIExchange = async (authCode: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3715,7 +3715,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3743,7 +3743,7 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3832,7 +3832,7 @@ const handleSoraValidateST = async (sessionTokenInput: string) => { extra: soraExtra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -3921,7 +3921,7 @@ const handleAntigravityValidateRT = async (refreshTokenInput: string) => { extra: {}, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, @@ -4245,7 +4245,7 @@ const handleCookieAuth = async (sessionKey: string) => { extra, proxy_id: form.proxy_id, concurrency: form.concurrency, - load_factor: form.load_factor || undefined, + load_factor: form.load_factor ?? undefined, priority: form.priority, rate_multiplier: form.rate_multiplier, group_ids: form.group_ids, diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 196e346cb0..b516c4b31d 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -2064,8 +2064,9 @@ const handleSubmit = async () => { if (form.expires_at === null) { updatePayload.expires_at = 0 } - // load_factor: 空值/0/NaN 时发送 0（后端约定 0 = 清除） - if (!form.load_factor || form.load_factor <= 0) { + // load_factor: 空值/NaN/0/负数 时发送 0（后端约定 <= 0 = 清除） + const lf = form.load_factor + if (lf == null || Number.isNaN(lf) || lf <= 0) { updatePayload.load_factor = 0 } updatePayload.auto_pause_on_expired = autoPauseOnExpired.value From 8a8cdeebb41ea0044b38dcbbf8e6dc13e580b322 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:30:30 +0800 Subject: [PATCH 16/53] fix: prevent negative values for concurrency and load_factor inputs --- frontend/src/components/account/BulkEditAccountModal.vue | 2 ++ frontend/src/components/account/CreateAccountModal.vue | 6 ++++-- frontend/src/components/account/EditAccountModal.vue | 6 ++++-- 3 files changed, 10 insertions(+), 4 deletions(-) diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 12ca5a703c..36bc94b1d7 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -496,6 +496,7 @@ class="input" :class="!enableConcurrency && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-concurrency-label" + @input="if (concurrency < 1) concurrency = 1" />
@@ -524,6 +525,7 @@ class="input" :class="!enableLoadFactor && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-load-factor-label" + @input="if (loadFactor != null && loadFactor < 1) loadFactor = null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index c5420ca498..e5e4450859 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -1752,12 +1752,14 @@
- +
+ class="input" :placeholder="String(form.concurrency || 1)" + @input="if (form.load_factor != null && form.load_factor < 1) form.load_factor = null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index b516c4b31d..b170169ad1 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -653,12 +653,14 @@
- +
+ class="input" :placeholder="String(form.concurrency || 1)" + @input="if (form.load_factor != null && form.load_factor < 1) form.load_factor = null" />

{{ t('admin.accounts.loadFactorHint') }}

From 18f2e21414c3704ec666a54d0441a5e1738d05b9 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 04:33:55 +0800 Subject: [PATCH 17/53] fix: use HTML-safe expressions for @input handlers in Vue templates Replace `<` comparisons with Math.max/ternary+>= to avoid Vue template parser treating `<` as HTML tag start in attribute values. --- frontend/src/components/account/BulkEditAccountModal.vue | 4 ++-- frontend/src/components/account/CreateAccountModal.vue | 4 ++-- frontend/src/components/account/EditAccountModal.vue | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 36bc94b1d7..5bc338f77f 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -496,7 +496,7 @@ class="input" :class="!enableConcurrency && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-concurrency-label" - @input="if (concurrency < 1) concurrency = 1" + @input="concurrency = Math.max(1, concurrency || 1)" />
@@ -525,7 +525,7 @@ class="input" :class="!enableLoadFactor && 'cursor-not-allowed opacity-50'" aria-labelledby="bulk-edit-load-factor-label" - @input="if (loadFactor != null && loadFactor < 1) loadFactor = null" + @input="loadFactor = (loadFactor && loadFactor >= 1) ? loadFactor : null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/CreateAccountModal.vue b/frontend/src/components/account/CreateAccountModal.vue index e5e4450859..efb3318361 100644 --- a/frontend/src/components/account/CreateAccountModal.vue +++ b/frontend/src/components/account/CreateAccountModal.vue @@ -1753,13 +1753,13 @@
+ @input="form.concurrency = Math.max(1, form.concurrency || 1)" />
+ @input="form.load_factor = (form.load_factor && form.load_factor >= 1) ? form.load_factor : null" />

{{ t('admin.accounts.loadFactorHint') }}

diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index b170169ad1..f57f8abc9b 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -654,13 +654,13 @@
+ @input="form.concurrency = Math.max(1, form.concurrency || 1)" />
+ @input="form.load_factor = (form.load_factor && form.load_factor >= 1) ? form.load_factor : null" />

{{ t('admin.accounts.loadFactorHint') }}

From af3a5076d6169c1ab1b1a8e956dfc1420d5f8983 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 05:17:52 +0800 Subject: [PATCH 18/53] fix: add load_factor upper bound validation to BulkUpdateAccounts --- backend/internal/service/admin_service.go | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index e5ceac078a..5d96db45f2 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -1635,7 +1635,13 @@ func (s *adminServiceImpl) BulkUpdateAccounts(ctx context.Context, input *BulkUp repoUpdates.RateMultiplier = input.RateMultiplier } if input.LoadFactor != nil { - repoUpdates.LoadFactor = input.LoadFactor + if *input.LoadFactor <= 0 { + repoUpdates.LoadFactor = nil // 0 或负数表示清除 + } else if *input.LoadFactor > 10000 { + return nil, errors.New("load_factor must be <= 10000") + } else { + repoUpdates.LoadFactor = input.LoadFactor + } } if input.Status != "" { repoUpdates.Status = &input.Status From c87e6526c1ed563619bdde21c79e2b14015052da Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 05:18:45 +0800 Subject: [PATCH 19/53] fix: use real Concurrency instead of LoadFactor for metrics and monitoring LoadFactor should only affect scheduling weight, not load rate reporting. --- backend/internal/service/ops_concurrency.go | 9 ++++++--- backend/internal/service/ops_metrics_collector.go | 2 +- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/backend/internal/service/ops_concurrency.go b/backend/internal/service/ops_concurrency.go index c03108c4b9..a571dd4df4 100644 --- a/backend/internal/service/ops_concurrency.go +++ b/backend/internal/service/ops_concurrency.go @@ -64,9 +64,12 @@ func (s *OpsService) getAccountsLoadMapBestEffort(ctx context.Context, accounts if acc.ID <= 0 { continue } - lf := acc.EffectiveLoadFactor() - if prev, ok := unique[acc.ID]; !ok || lf > prev { - unique[acc.ID] = lf + c := acc.Concurrency + if c <= 0 { + c = 1 + } + if prev, ok := unique[acc.ID]; !ok || c > prev { + unique[acc.ID] = c } } diff --git a/backend/internal/service/ops_metrics_collector.go b/backend/internal/service/ops_metrics_collector.go index 6c33707138..f93481e7ff 100644 --- a/backend/internal/service/ops_metrics_collector.go +++ b/backend/internal/service/ops_metrics_collector.go @@ -391,7 +391,7 @@ func (c *OpsMetricsCollector) collectConcurrencyQueueDepth(parentCtx context.Con } batch = append(batch, AccountWithConcurrency{ ID: acc.ID, - MaxConcurrency: acc.EffectiveLoadFactor(), + MaxConcurrency: acc.Concurrency, }) } if len(batch) == 0 { From e74270f8ebd3cb317540c9fbdcf8df31060fd2bf Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 05:29:07 +0800 Subject: [PATCH 20/53] chore: bump version to 0.1.91.4 --- backend/cmd/server/VERSION | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backend/cmd/server/VERSION b/backend/cmd/server/VERSION index 231369cfef..adaf52a822 100644 --- a/backend/cmd/server/VERSION +++ b/backend/cmd/server/VERSION @@ -1 +1 @@ -0.1.91.3 +0.1.91.4 From 95fcb02fe27054372763c65da9719d85c1f63dff Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 05:34:03 +0800 Subject: [PATCH 21/53] fix: restore SimulateClaudeMaxEnabled and ResetAccountQuota lost during merge --- backend/internal/service/admin_service.go | 48 ++++++++++++++++++----- 1 file changed, 38 insertions(+), 10 deletions(-) diff --git a/backend/internal/service/admin_service.go b/backend/internal/service/admin_service.go index 5d96db45f2..4a60882afa 100644 --- a/backend/internal/service/admin_service.go +++ b/backend/internal/service/admin_service.go @@ -84,6 +84,7 @@ type AdminService interface { DeleteRedeemCode(ctx context.Context, id int64) error BatchDeleteRedeemCodes(ctx context.Context, ids []int64) (int64, error) ExpireRedeemCode(ctx context.Context, id int64) (*RedeemCode, error) + ResetAccountQuota(ctx context.Context, id int64) error } // CreateUserInput represents input for creating a new user via admin operations. @@ -137,9 +138,10 @@ type CreateGroupInput struct { // 无效请求兜底分组 ID（仅 anthropic 平台使用） FallbackGroupIDOnInvalidRequest *int64 // 模型路由配置（仅 anthropic 平台使用） - ModelRouting map[string][]int64 - ModelRoutingEnabled bool // 是否启用模型路由 - MCPXMLInject *bool + ModelRouting map[string][]int64 + ModelRoutingEnabled bool // 是否启用模型路由 + MCPXMLInject *bool + SimulateClaudeMaxEnabled *bool // 支持的模型系列（仅 antigravity 平台使用） SupportedModelScopes []string // Sora 存储配额 @@ -173,9 +175,10 @@ type UpdateGroupInput struct { // 无效请求兜底分组 ID（仅 anthropic 平台使用） FallbackGroupIDOnInvalidRequest *int64 // 模型路由配置（仅 anthropic 平台使用） - ModelRouting map[string][]int64 - ModelRoutingEnabled *bool // 是否启用模型路由 - MCPXMLInject *bool + ModelRouting map[string][]int64 + ModelRoutingEnabled *bool // 是否启用模型路由 + MCPXMLInject *bool + SimulateClaudeMaxEnabled *bool // 支持的模型系列（仅 antigravity 平台使用） SupportedModelScopes *[]string // Sora 存储配额 @@ -356,6 +359,10 @@ type ProxyExitInfoProber interface { ProbeProxy(ctx context.Context, proxyURL string) (*ProxyExitInfo, int64, error) } +type groupExistenceBatchReader interface { + ExistsByIDs(ctx context.Context, ids []int64) (map[int64]bool, error) +} + type proxyQualityTarget struct { Target string URL string @@ -431,10 +438,6 @@ type userGroupRateBatchReader interface { GetByUserIDs(ctx context.Context, userIDs []int64) (map[int64]map[int64]float64, error) } -type groupExistenceBatchReader interface { - ExistsByIDs(ctx context.Context, ids []int64) (map[int64]bool, error) -} - // NewAdminService creates a new AdminService func NewAdminService( userRepo UserRepository, @@ -850,6 +853,13 @@ func (s *adminServiceImpl) CreateGroup(ctx context.Context, input *CreateGroupIn if input.MCPXMLInject != nil { mcpXMLInject = *input.MCPXMLInject } + simulateClaudeMaxEnabled := false + if input.SimulateClaudeMaxEnabled != nil { + if platform != PlatformAnthropic && *input.SimulateClaudeMaxEnabled { + return nil, fmt.Errorf("simulate_claude_max_enabled only supported for anthropic groups") + } + simulateClaudeMaxEnabled = *input.SimulateClaudeMaxEnabled + } // 如果指定了复制账号的源分组，先获取账号 ID 列表 var accountIDsToCopy []int64 @@ -906,6 +916,7 @@ func (s *adminServiceImpl) CreateGroup(ctx context.Context, input *CreateGroupIn FallbackGroupIDOnInvalidRequest: fallbackOnInvalidRequest, ModelRouting: input.ModelRouting, MCPXMLInject: mcpXMLInject, + SimulateClaudeMaxEnabled: simulateClaudeMaxEnabled, SupportedModelScopes: input.SupportedModelScopes, SoraStorageQuotaBytes: input.SoraStorageQuotaBytes, } @@ -1115,6 +1126,15 @@ func (s *adminServiceImpl) UpdateGroup(ctx context.Context, id int64, input *Upd if input.MCPXMLInject != nil { group.MCPXMLInject = *input.MCPXMLInject } + if input.SimulateClaudeMaxEnabled != nil { + if group.Platform != PlatformAnthropic && *input.SimulateClaudeMaxEnabled { + return nil, fmt.Errorf("simulate_claude_max_enabled only supported for anthropic groups") + } + group.SimulateClaudeMaxEnabled = *input.SimulateClaudeMaxEnabled + } + if group.Platform != PlatformAnthropic { + group.SimulateClaudeMaxEnabled = false + } // 支持的模型系列（仅 antigravity 平台使用） if input.SupportedModelScopes != nil { @@ -1467,6 +1487,10 @@ func (s *adminServiceImpl) UpdateAccount(ctx context.Context, id int64, input *U account.Credentials = input.Credentials } if len(input.Extra) > 0 { + // 保留 quota_used，防止编辑账号时意外重置配额用量 + if oldQuotaUsed, ok := account.Extra["quota_used"]; ok { + input.Extra["quota_used"] = oldQuotaUsed + } account.Extra = input.Extra } if input.ProxyID != nil { @@ -2466,3 +2490,7 @@ func (e *MixedChannelError) Error() string { return fmt.Sprintf("mixed_channel_warning: Group '%s' contains both %s and %s accounts. Using mixed channels in the same context may cause thinking block signature validation issues, which will fallback to non-thinking mode for historical messages.", e.GroupName, e.CurrentPlatform, e.OtherPlatform) } + +func (s *adminServiceImpl) ResetAccountQuota(ctx context.Context, id int64) error { + return s.accountRepo.ResetQuotaUsed(ctx, id) +} From 1a0d4ed668dfe1dee317e924f52c7384aec10b42 Mon Sep 17 00:00:00 2001 From: yangjianbo Date: Fri, 6 Mar 2026 08:14:04 +0800 Subject: [PATCH 22/53] =?UTF-8?q?feat(openai):=20=E5=A2=9E=E5=8A=A0=20gpt-?= =?UTF-8?q?5.4=20=E6=A8=A1=E5=9E=8B=E6=94=AF=E6=8C=81=E4=B8=8E=E5=AE=9A?= =?UTF-8?q?=E4=BB=B7=E9=85=8D=E7=BD=AE?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 接入 gpt-5.4 模型识别与规范化，补充默认模型列表 - 增加 gpt-5.4 输入/缓存命中/输出价格与计费兜底逻辑 - 同步前端模型白名单与 OpenCode 上下文窗口（1050000/128000） Co-Authored-By: Claude Opus 4.6 (cherry picked from commit 924476dcac6181cd0f3ee731ec7b73672ff03793) --- backend/internal/pkg/openai/constants.go | 1 + backend/internal/service/billing_service.go | 48 ++++++++++++- .../internal/service/billing_service_test.go | 67 ++++++++++++++++++- .../service/openai_codex_transform.go | 10 +++ .../service/openai_codex_transform_test.go | 4 ++ backend/internal/service/pricing_service.go | 21 +++++- .../internal/service/pricing_service_test.go | 14 ++++ .../model_prices_and_context_window.json | 33 +++++++++ .../account/BulkEditAccountModal.vue | 7 ++ frontend/src/components/keys/UseKeyModal.vue | 16 +++++ frontend/src/composables/useModelWhitelist.ts | 3 + 11 files changed, 218 insertions(+), 6 deletions(-) diff --git a/backend/internal/pkg/openai/constants.go b/backend/internal/pkg/openai/constants.go index 4bbc68e773..b0a31a5f38 100644 --- a/backend/internal/pkg/openai/constants.go +++ b/backend/internal/pkg/openai/constants.go @@ -15,6 +15,7 @@ type Model struct { // DefaultModels OpenAI models list var DefaultModels = []Model{ + {ID: "gpt-5.4", Object: "model", Created: 1738368000, OwnedBy: "openai", Type: "model", DisplayName: "GPT-5.4"}, {ID: "gpt-5.3-codex", Object: "model", Created: 1735689600, OwnedBy: "openai", Type: "model", DisplayName: "GPT-5.3 Codex"}, {ID: "gpt-5.3-codex-spark", Object: "model", Created: 1735689600, OwnedBy: "openai", Type: "model", DisplayName: "GPT-5.3 Codex Spark"}, {ID: "gpt-5.2", Object: "model", Created: 1733875200, OwnedBy: "openai", Type: "model", DisplayName: "GPT-5.2"}, diff --git a/backend/internal/service/billing_service.go b/backend/internal/service/billing_service.go index 5d67c80862..fbc06fd4e3 100644 --- a/backend/internal/service/billing_service.go +++ b/backend/internal/service/billing_service.go @@ -161,6 +161,32 @@ func (s *BillingService) initFallbackPricing() { CacheReadPricePerToken: 0.2e-6, // $0.20 per MTok SupportsCacheBreakdown: false, } + + // OpenAI GPT-5.1（本地兜底，防止动态定价不可用时拒绝计费） + s.fallbackPrices["gpt-5.1"] = &ModelPricing{ + InputPricePerToken: 1.25e-6, // $1.25 per MTok + OutputPricePerToken: 10e-6, // $10 per MTok + CacheCreationPricePerToken: 1.25e-6, // $1.25 per MTok + CacheReadPricePerToken: 0.125e-6, + SupportsCacheBreakdown: false, + } + // OpenAI GPT-5.4（业务指定价格） + s.fallbackPrices["gpt-5.4"] = &ModelPricing{ + InputPricePerToken: 2.5e-6, // $2.5 per MTok + OutputPricePerToken: 15e-6, // $15 per MTok + CacheCreationPricePerToken: 2.5e-6, // $2.5 per MTok + CacheReadPricePerToken: 0.25e-6, // $0.25 per MTok + SupportsCacheBreakdown: false, + } + // Codex 族兜底统一按 GPT-5.1 Codex 价格计费 + s.fallbackPrices["gpt-5.1-codex"] = &ModelPricing{ + InputPricePerToken: 1.5e-6, // $1.5 per MTok + OutputPricePerToken: 12e-6, // $12 per MTok + CacheCreationPricePerToken: 1.5e-6, // $1.5 per MTok + CacheReadPricePerToken: 0.15e-6, + SupportsCacheBreakdown: false, + } + s.fallbackPrices["gpt-5.3-codex"] = s.fallbackPrices["gpt-5.1-codex"] } // getFallbackPricing 根据模型系列获取回退价格 @@ -189,12 +215,30 @@ func (s *BillingService) getFallbackPricing(model string) *ModelPricing { } return s.fallbackPrices["claude-3-haiku"] } + // Claude 未知型号统一回退到 Sonnet，避免计费中断。 + if strings.Contains(modelLower, "claude") { + return s.fallbackPrices["claude-sonnet-4"] + } if strings.Contains(modelLower, "gemini-3.1-pro") || strings.Contains(modelLower, "gemini-3-1-pro") { return s.fallbackPrices["gemini-3.1-pro"] } - // 默认使用Sonnet价格 - return s.fallbackPrices["claude-sonnet-4"] + // OpenAI 仅匹配已知 GPT-5/Codex 族，避免未知 OpenAI 型号误计价。 + if strings.Contains(modelLower, "gpt-5") || strings.Contains(modelLower, "codex") { + normalized := normalizeCodexModel(modelLower) + switch normalized { + case "gpt-5.4": + return s.fallbackPrices["gpt-5.4"] + case "gpt-5.3-codex": + return s.fallbackPrices["gpt-5.3-codex"] + case "gpt-5.1-codex", "gpt-5.1-codex-max", "gpt-5.1-codex-mini", "codex-mini-latest": + return s.fallbackPrices["gpt-5.1-codex"] + case "gpt-5.1": + return s.fallbackPrices["gpt-5.1"] + } + } + + return nil } // GetModelPricing 获取模型价格配置 diff --git a/backend/internal/service/billing_service_test.go b/backend/internal/service/billing_service_test.go index 5eb278f647..3ae6dc8562 100644 --- a/backend/internal/service/billing_service_test.go +++ b/backend/internal/service/billing_service_test.go @@ -133,7 +133,7 @@ func TestGetModelPricing_CaseInsensitive(t *testing.T) { require.Equal(t, p1.InputPricePerToken, p2.InputPricePerToken) } -func TestGetModelPricing_UnknownModelFallsBackToSonnet(t *testing.T) { +func TestGetModelPricing_UnknownClaudeModelFallsBackToSonnet(t *testing.T) { svc := newTestBillingService() // 不包含 opus/sonnet/haiku 关键词的 Claude 模型会走默认 Sonnet 价格 @@ -142,6 +142,71 @@ func TestGetModelPricing_UnknownModelFallsBackToSonnet(t *testing.T) { require.InDelta(t, 3e-6, pricing.InputPricePerToken, 1e-12) } +func TestGetModelPricing_UnknownOpenAIModelReturnsError(t *testing.T) { + svc := newTestBillingService() + + pricing, err := svc.GetModelPricing("gpt-unknown-model") + require.Error(t, err) + require.Nil(t, pricing) + require.Contains(t, err.Error(), "pricing not found") +} + +func TestGetModelPricing_OpenAIGPT51Fallback(t *testing.T) { + svc := newTestBillingService() + + pricing, err := svc.GetModelPricing("gpt-5.1") + require.NoError(t, err) + require.NotNil(t, pricing) + require.InDelta(t, 1.25e-6, pricing.InputPricePerToken, 1e-12) +} + +func TestGetModelPricing_OpenAIGPT54Fallback(t *testing.T) { + svc := newTestBillingService() + + pricing, err := svc.GetModelPricing("gpt-5.4") + require.NoError(t, err) + require.NotNil(t, pricing) + require.InDelta(t, 2.5e-6, pricing.InputPricePerToken, 1e-12) + require.InDelta(t, 15e-6, pricing.OutputPricePerToken, 1e-12) + require.InDelta(t, 0.25e-6, pricing.CacheReadPricePerToken, 1e-12) +} + +func TestGetFallbackPricing_FamilyMatching(t *testing.T) { + svc := newTestBillingService() + + tests := []struct { + name string + model string + expectedInput float64 + expectNilPricing bool + }{ + {name: "empty model", model: " ", expectNilPricing: true}, + {name: "claude opus 4.6", model: "claude-opus-4.6-20260201", expectedInput: 5e-6}, + {name: "claude opus 4.5 alt separator", model: "claude-opus-4-5-20260101", expectedInput: 5e-6}, + {name: "claude generic model fallback sonnet", model: "claude-foo-bar", expectedInput: 3e-6}, + {name: "gemini explicit fallback", model: "gemini-3-1-pro", expectedInput: 2e-6}, + {name: "gemini unknown no fallback", model: "gemini-2.0-pro", expectNilPricing: true}, + {name: "openai gpt5.1", model: "gpt-5.1", expectedInput: 1.25e-6}, + {name: "openai gpt5.4", model: "gpt-5.4", expectedInput: 2.5e-6}, + {name: "openai gpt5.3 codex", model: "gpt-5.3-codex", expectedInput: 1.5e-6}, + {name: "openai gpt5.1 codex max alias", model: "gpt-5.1-codex-max", expectedInput: 1.5e-6}, + {name: "openai codex mini latest alias", model: "codex-mini-latest", expectedInput: 1.5e-6}, + {name: "openai unknown no fallback", model: "gpt-unknown-model", expectNilPricing: true}, + {name: "non supported family", model: "qwen-max", expectNilPricing: true}, + } + + for _, tt := range tests { + t.Run(tt.name, func(t *testing.T) { + pricing := svc.getFallbackPricing(tt.model) + if tt.expectNilPricing { + require.Nil(t, pricing) + return + } + require.NotNil(t, pricing) + require.InDelta(t, tt.expectedInput, pricing.InputPricePerToken, 1e-12) + }) + } +} func TestCalculateCostWithLongContext_BelowThreshold(t *testing.T) { svc := newTestBillingService() diff --git a/backend/internal/service/openai_codex_transform.go b/backend/internal/service/openai_codex_transform.go index 16befb8247..9bc48cf6ba 100644 --- a/backend/internal/service/openai_codex_transform.go +++ b/backend/internal/service/openai_codex_transform.go @@ -9,6 +9,13 @@ import ( var codexCLIInstructions string var codexModelMap = map[string]string{ + "gpt-5.4": "gpt-5.4", + "gpt-5.4-none": "gpt-5.4", + "gpt-5.4-low": "gpt-5.4", + "gpt-5.4-medium": "gpt-5.4", + "gpt-5.4-high": "gpt-5.4", + "gpt-5.4-xhigh": "gpt-5.4", + "gpt-5.4-chat-latest": "gpt-5.4", "gpt-5.3": "gpt-5.3-codex", "gpt-5.3-none": "gpt-5.3-codex", "gpt-5.3-low": "gpt-5.3-codex", @@ -154,6 +161,9 @@ func normalizeCodexModel(model string) string { normalized := strings.ToLower(modelID) + if strings.Contains(normalized, "gpt-5.4") || strings.Contains(normalized, "gpt 5.4") { + return "gpt-5.4" + } if strings.Contains(normalized, "gpt-5.2-codex") || strings.Contains(normalized, "gpt 5.2 codex") { return "gpt-5.2-codex" } diff --git a/backend/internal/service/openai_codex_transform_test.go b/backend/internal/service/openai_codex_transform_test.go index 27093f6c37..7ee4bbc8d5 100644 --- a/backend/internal/service/openai_codex_transform_test.go +++ b/backend/internal/service/openai_codex_transform_test.go @@ -167,6 +167,10 @@ func TestApplyCodexOAuthTransform_EmptyInput(t *testing.T) { func TestNormalizeCodexModel_Gpt53(t *testing.T) { cases := map[string]string{ + "gpt-5.4": "gpt-5.4", + "gpt-5.4-high": "gpt-5.4", + "gpt-5.4-chat-latest": "gpt-5.4", + "gpt 5.4": "gpt-5.4", "gpt-5.3": "gpt-5.3-codex", "gpt-5.3-codex": "gpt-5.3-codex", "gpt-5.3-codex-xhigh": "gpt-5.3-codex", diff --git a/backend/internal/service/pricing_service.go b/backend/internal/service/pricing_service.go index 41e8b5eb0c..b81a160a75 100644 --- a/backend/internal/service/pricing_service.go +++ b/backend/internal/service/pricing_service.go @@ -21,8 +21,16 @@ import ( ) var ( - openAIModelDatePattern = regexp.MustCompile(`-\d{8}$`) - openAIModelBasePattern = regexp.MustCompile(`^(gpt-\d+(?:\.\d+)?)(?:-|$)`) + openAIModelDatePattern = regexp.MustCompile(`-\d{8}$`) + openAIModelBasePattern = regexp.MustCompile(`^(gpt-\d+(?:\.\d+)?)(?:-|$)`) + openAIGPT54FallbackPricing = &LiteLLMModelPricing{ + InputCostPerToken: 2.5e-06, // $2.5 per MTok + OutputCostPerToken: 1.5e-05, // $15 per MTok + CacheReadInputTokenCost: 2.5e-07, // $0.25 per MTok + LiteLLMProvider: "openai", + Mode: "chat", + SupportsPromptCaching: true, + } ) // LiteLLMModelPricing LiteLLM价格数据结构 @@ -660,7 +668,8 @@ func (s *PricingService) matchByModelFamily(model string) *LiteLLMModelPricing { // 2. gpt-5.2-codex -> gpt-5.2（去掉后缀如 -codex, -mini, -max 等） // 3. gpt-5.2-20251222 -> gpt-5.2（去掉日期版本号） // 4. gpt-5.3-codex -> gpt-5.2-codex -// 5. 最终回退到 DefaultTestModel (gpt-5.1-codex) +// 5. gpt-5.4* -> 业务静态兜底价 +// 6. 最终回退到 DefaultTestModel (gpt-5.1-codex) func (s *PricingService) matchOpenAIModel(model string) *LiteLLMModelPricing { if strings.HasPrefix(model, "gpt-5.3-codex-spark") { if pricing, ok := s.pricingData["gpt-5.1-codex"]; ok { @@ -690,6 +699,12 @@ func (s *PricingService) matchOpenAIModel(model string) *LiteLLMModelPricing { } } + if strings.HasPrefix(model, "gpt-5.4") { + logger.With(zap.String("component", "service.pricing")). + Info(fmt.Sprintf("[Pricing] OpenAI fallback matched %s -> %s", model, "gpt-5.4(static)")) + return openAIGPT54FallbackPricing + } + // 最终回退到 DefaultTestModel defaultModel := strings.ToLower(openai.DefaultTestModel) if pricing, ok := s.pricingData[defaultModel]; ok { diff --git a/backend/internal/service/pricing_service_test.go b/backend/internal/service/pricing_service_test.go index 127ff342c6..de29528372 100644 --- a/backend/internal/service/pricing_service_test.go +++ b/backend/internal/service/pricing_service_test.go @@ -51,3 +51,17 @@ func TestGetModelPricing_OpenAIFallbackMatchedLoggedAsInfo(t *testing.T) { require.True(t, logSink.ContainsMessageAtLevel("[Pricing] OpenAI fallback matched gpt-5.3-codex -> gpt-5.2-codex", "info")) require.False(t, logSink.ContainsMessageAtLevel("[Pricing] OpenAI fallback matched gpt-5.3-codex -> gpt-5.2-codex", "warn")) } + +func TestGetModelPricing_Gpt54UsesStaticFallbackWhenRemoteMissing(t *testing.T) { + svc := &PricingService{ + pricingData: map[string]*LiteLLMModelPricing{ + "gpt-5.1-codex": &LiteLLMModelPricing{InputCostPerToken: 1.25e-6}, + }, + } + + got := svc.GetModelPricing("gpt-5.4") + require.NotNil(t, got) + require.InDelta(t, 2.5e-6, got.InputCostPerToken, 1e-12) + require.InDelta(t, 1.5e-5, got.OutputCostPerToken, 1e-12) + require.InDelta(t, 2.5e-7, got.CacheReadInputTokenCost, 1e-12) +} diff --git a/backend/resources/model-pricing/model_prices_and_context_window.json b/backend/resources/model-pricing/model_prices_and_context_window.json index 650e128e9a..72860bf9e7 100644 --- a/backend/resources/model-pricing/model_prices_and_context_window.json +++ b/backend/resources/model-pricing/model_prices_and_context_window.json @@ -5140,6 +5140,39 @@ "supports_vision": true, "supports_web_search": true }, + "gpt-5.4": { + "cache_read_input_token_cost": 2.5e-07, + "input_cost_per_token": 2.5e-06, + "litellm_provider": "openai", + "max_input_tokens": 1050000, + "max_output_tokens": 128000, + "max_tokens": 128000, + "mode": "chat", + "output_cost_per_token": 1.5e-05, + "supported_endpoints": [ + "/v1/chat/completions", + "/v1/responses" + ], + "supported_modalities": [ + "text", + "image" + ], + "supported_output_modalities": [ + "text", + "image" + ], + "supports_function_calling": true, + "supports_native_streaming": true, + "supports_parallel_function_calling": true, + "supports_pdf_input": true, + "supports_prompt_caching": true, + "supports_reasoning": true, + "supports_response_schema": true, + "supports_service_tier": true, + "supports_system_messages": true, + "supports_tool_choice": true, + "supports_vision": true + }, "gpt-5.3-codex": { "cache_read_input_token_cost": 1.75e-07, "cache_read_input_token_cost_priority": 3.5e-07, diff --git a/frontend/src/components/account/BulkEditAccountModal.vue b/frontend/src/components/account/BulkEditAccountModal.vue index 5bc338f77f..1d6f32febd 100644 --- a/frontend/src/components/account/BulkEditAccountModal.vue +++ b/frontend/src/components/account/BulkEditAccountModal.vue @@ -951,6 +951,7 @@ const allModels = [ { value: 'claude-3-haiku-20240307', label: 'Claude 3 Haiku' }, { value: 'gpt-5.3-codex', label: 'GPT-5.3 Codex' }, { value: 'gpt-5.3-codex-spark', label: 'GPT-5.3 Codex Spark' }, + { value: 'gpt-5.4', label: 'GPT-5.4' }, { value: 'gpt-5.2-2025-12-11', label: 'GPT-5.2' }, { value: 'gpt-5.2-codex', label: 'GPT-5.2 Codex' }, { value: 'gpt-5.1-codex-max', label: 'GPT-5.1 Codex Max' }, @@ -1065,6 +1066,12 @@ const presetMappings = [ to: 'gpt-5.3-codex-spark', color: 'bg-emerald-100 text-emerald-700 hover:bg-emerald-200 dark:bg-emerald-900/30 dark:text-emerald-400' }, + { + label: 'GPT-5.4', + from: 'gpt-5.4', + to: 'gpt-5.4', + color: 'bg-rose-100 text-rose-700 hover:bg-rose-200 dark:bg-rose-900/30 dark:text-rose-400' + }, { label: '5.2→5.3', from: 'gpt-5.2-codex', diff --git a/frontend/src/components/keys/UseKeyModal.vue b/frontend/src/components/keys/UseKeyModal.vue index a0233742a1..a61ce6b459 100644 --- a/frontend/src/components/keys/UseKeyModal.vue +++ b/frontend/src/components/keys/UseKeyModal.vue @@ -670,6 +670,22 @@ function generateOpenCodeConfig(platform: string, baseUrl: string, apiKey: strin xhigh: {} } }, + 'gpt-5.4': { + name: 'GPT-5.4', + limit: { + context: 1050000, + output: 128000 + }, + options: { + store: false + }, + variants: { + low: {}, + medium: {}, + high: {}, + xhigh: {} + } + }, 'gpt-5.3-codex-spark': { name: 'GPT-5.3 Codex Spark', limit: { diff --git a/frontend/src/composables/useModelWhitelist.ts b/frontend/src/composables/useModelWhitelist.ts index 444e4b917e..8e525873dd 100644 --- a/frontend/src/composables/useModelWhitelist.ts +++ b/frontend/src/composables/useModelWhitelist.ts @@ -24,6 +24,8 @@ const openaiModels = [ // GPT-5.2 系列 'gpt-5.2', 'gpt-5.2-2025-12-11', 'gpt-5.2-chat-latest', 'gpt-5.2-codex', 'gpt-5.2-pro', 'gpt-5.2-pro-2025-12-11', + // GPT-5.4 系列 + 'gpt-5.4', // GPT-5.3 系列 'gpt-5.3-codex', 'gpt-5.3-codex-spark', 'chatgpt-4o-latest', @@ -277,6 +279,7 @@ const openaiPresetMappings = [ { label: 'GPT-5.3 Codex Spark', from: 'gpt-5.3-codex-spark', to: 'gpt-5.3-codex-spark', color: 'bg-teal-100 text-teal-700 hover:bg-teal-200 dark:bg-teal-900/30 dark:text-teal-400' }, { label: 'GPT-5.1', from: 'gpt-5.1', to: 'gpt-5.1', color: 'bg-orange-100 text-orange-700 hover:bg-orange-200 dark:bg-orange-900/30 dark:text-orange-400' }, { label: 'GPT-5.2', from: 'gpt-5.2', to: 'gpt-5.2', color: 'bg-red-100 text-red-700 hover:bg-red-200 dark:bg-red-900/30 dark:text-red-400' }, + { label: 'GPT-5.4', from: 'gpt-5.4', to: 'gpt-5.4', color: 'bg-rose-100 text-rose-700 hover:bg-rose-200 dark:bg-rose-900/30 dark:text-rose-400' }, { label: 'GPT-5.1 Codex', from: 'gpt-5.1-codex', to: 'gpt-5.1-codex', color: 'bg-cyan-100 text-cyan-700 hover:bg-cyan-200 dark:bg-cyan-900/30 dark:text-cyan-400' } ] From f3660264358dce5a639f66e205e132ab6c2b2b7c Mon Sep 17 00:00:00 2001 From: yangjianbo Date: Fri, 6 Mar 2026 09:04:58 +0800 Subject: [PATCH 23/53] =?UTF-8?q?fix(openai):=20=E4=BF=AE=E5=A4=8D=20gpt-5?= =?UTF-8?q?.4=20=E9=95=BF=E4=B8=8A=E4=B8=8B=E6=96=87=E8=AE=A1=E8=B4=B9?= =?UTF-8?q?=E4=B8=8E=E5=BF=AB=E7=85=A7=E7=99=BD=E5=90=8D=E5=8D=95?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 补齐 gpt-5.4 fallback 的长上下文计费元信息，\n确保超过 272000 输入 token 时对整次会话应用\n2x 输入与 1.5x 输出计费规则。\n\n同时将官方快照 gpt-5.4-2026-03-05 加入前端\n白名单候选与回归测试，避免 whitelist 模式误拦截。\n\nCo-Authored-By: Claude Opus 4.6 (cherry picked from commit d95497af87f608c6dadcbe7d6e851de9413ae147) --- backend/internal/service/billing_service.go | 108 ++++++++++++++---- .../internal/service/billing_service_test.go | 22 ++++ backend/internal/service/pricing_service.go | 18 ++- .../internal/service/pricing_service_test.go | 3 + .../__tests__/useModelWhitelist.spec.ts | 15 +++ frontend/src/composables/useModelWhitelist.ts | 2 +- 6 files changed, 137 insertions(+), 31 deletions(-) diff --git a/backend/internal/service/billing_service.go b/backend/internal/service/billing_service.go index fbc06fd4e3..d058c25af7 100644 --- a/backend/internal/service/billing_service.go +++ b/backend/internal/service/billing_service.go @@ -43,15 +43,24 @@ type BillingCache interface { // ModelPricing 模型价格配置（per-token价格，与LiteLLM格式一致） type ModelPricing struct { - InputPricePerToken float64 // 每token输入价格 (USD) - OutputPricePerToken float64 // 每token输出价格 (USD) - CacheCreationPricePerToken float64 // 缓存创建每token价格 (USD) - CacheReadPricePerToken float64 // 缓存读取每token价格 (USD) - CacheCreation5mPrice float64 // 5分钟缓存创建每token价格 (USD) - CacheCreation1hPrice float64 // 1小时缓存创建每token价格 (USD) - SupportsCacheBreakdown bool // 是否支持详细的缓存分类 + InputPricePerToken float64 // 每token输入价格 (USD) + OutputPricePerToken float64 // 每token输出价格 (USD) + CacheCreationPricePerToken float64 // 缓存创建每token价格 (USD) + CacheReadPricePerToken float64 // 缓存读取每token价格 (USD) + CacheCreation5mPrice float64 // 5分钟缓存创建每token价格 (USD) + CacheCreation1hPrice float64 // 1小时缓存创建每token价格 (USD) + SupportsCacheBreakdown bool // 是否支持详细的缓存分类 + LongContextInputThreshold int // 超过阈值后按整次会话提升输入价格 + LongContextInputMultiplier float64 // 长上下文整次会话输入倍率 + LongContextOutputMultiplier float64 // 长上下文整次会话输出倍率 } +const ( + openAIGPT54LongContextInputThreshold = 272000 + openAIGPT54LongContextInputMultiplier = 2.0 + openAIGPT54LongContextOutputMultiplier = 1.5 +) + // UsageTokens 使用的token数量 type UsageTokens struct { InputTokens int @@ -172,11 +181,14 @@ func (s *BillingService) initFallbackPricing() { } // OpenAI GPT-5.4（业务指定价格） s.fallbackPrices["gpt-5.4"] = &ModelPricing{ - InputPricePerToken: 2.5e-6, // $2.5 per MTok - OutputPricePerToken: 15e-6, // $15 per MTok - CacheCreationPricePerToken: 2.5e-6, // $2.5 per MTok - CacheReadPricePerToken: 0.25e-6, // $0.25 per MTok - SupportsCacheBreakdown: false, + InputPricePerToken: 2.5e-6, // $2.5 per MTok + OutputPricePerToken: 15e-6, // $15 per MTok + CacheCreationPricePerToken: 2.5e-6, // $2.5 per MTok + CacheReadPricePerToken: 0.25e-6, // $0.25 per MTok + SupportsCacheBreakdown: false, + LongContextInputThreshold: openAIGPT54LongContextInputThreshold, + LongContextInputMultiplier: openAIGPT54LongContextInputMultiplier, + LongContextOutputMultiplier: openAIGPT54LongContextOutputMultiplier, } // Codex 族兜底统一按 GPT-5.1 Codex 价格计费 s.fallbackPrices["gpt-5.1-codex"] = &ModelPricing{ @@ -256,15 +268,18 @@ func (s *BillingService) GetModelPricing(model string) (*ModelPricing, error) { price5m := litellmPricing.CacheCreationInputTokenCost price1h := litellmPricing.CacheCreationInputTokenCostAbove1hr enableBreakdown := price1h > 0 && price1h > price5m - return &ModelPricing{ - InputPricePerToken: litellmPricing.InputCostPerToken, - OutputPricePerToken: litellmPricing.OutputCostPerToken, - CacheCreationPricePerToken: litellmPricing.CacheCreationInputTokenCost, - CacheReadPricePerToken: litellmPricing.CacheReadInputTokenCost, - CacheCreation5mPrice: price5m, - CacheCreation1hPrice: price1h, - SupportsCacheBreakdown: enableBreakdown, - }, nil + return s.applyModelSpecificPricingPolicy(model, &ModelPricing{ + InputPricePerToken: litellmPricing.InputCostPerToken, + OutputPricePerToken: litellmPricing.OutputCostPerToken, + CacheCreationPricePerToken: litellmPricing.CacheCreationInputTokenCost, + CacheReadPricePerToken: litellmPricing.CacheReadInputTokenCost, + CacheCreation5mPrice: price5m, + CacheCreation1hPrice: price1h, + SupportsCacheBreakdown: enableBreakdown, + LongContextInputThreshold: litellmPricing.LongContextInputTokenThreshold, + LongContextInputMultiplier: litellmPricing.LongContextInputCostMultiplier, + LongContextOutputMultiplier: litellmPricing.LongContextOutputCostMultiplier, + }), nil } } @@ -272,7 +287,7 @@ func (s *BillingService) GetModelPricing(model string) (*ModelPricing, error) { fallback := s.getFallbackPricing(model) if fallback != nil { log.Printf("[Billing] Using fallback pricing for model: %s", model) - return fallback, nil + return s.applyModelSpecificPricingPolicy(model, fallback), nil } return nil, fmt.Errorf("pricing not found for model: %s", model) @@ -286,12 +301,18 @@ func (s *BillingService) CalculateCost(model string, tokens UsageTokens, rateMul } breakdown := &CostBreakdown{} + inputPricePerToken := pricing.InputPricePerToken + outputPricePerToken := pricing.OutputPricePerToken + if s.shouldApplySessionLongContextPricing(tokens, pricing) { + inputPricePerToken *= pricing.LongContextInputMultiplier + outputPricePerToken *= pricing.LongContextOutputMultiplier + } // 计算输入token费用（使用per-token价格） - breakdown.InputCost = float64(tokens.InputTokens) * pricing.InputPricePerToken + breakdown.InputCost = float64(tokens.InputTokens) * inputPricePerToken // 计算输出token费用 - breakdown.OutputCost = float64(tokens.OutputTokens) * pricing.OutputPricePerToken + breakdown.OutputCost = float64(tokens.OutputTokens) * outputPricePerToken // 计算缓存费用 if pricing.SupportsCacheBreakdown && (pricing.CacheCreation5mPrice > 0 || pricing.CacheCreation1hPrice > 0) { @@ -323,6 +344,45 @@ func (s *BillingService) CalculateCost(model string, tokens UsageTokens, rateMul return breakdown, nil } +func (s *BillingService) applyModelSpecificPricingPolicy(model string, pricing *ModelPricing) *ModelPricing { + if pricing == nil { + return nil + } + if !isOpenAIGPT54Model(model) { + return pricing + } + if pricing.LongContextInputThreshold > 0 && pricing.LongContextInputMultiplier > 0 && pricing.LongContextOutputMultiplier > 0 { + return pricing + } + cloned := *pricing + if cloned.LongContextInputThreshold <= 0 { + cloned.LongContextInputThreshold = openAIGPT54LongContextInputThreshold + } + if cloned.LongContextInputMultiplier <= 0 { + cloned.LongContextInputMultiplier = openAIGPT54LongContextInputMultiplier + } + if cloned.LongContextOutputMultiplier <= 0 { + cloned.LongContextOutputMultiplier = openAIGPT54LongContextOutputMultiplier + } + return &cloned +} + +func (s *BillingService) shouldApplySessionLongContextPricing(tokens UsageTokens, pricing *ModelPricing) bool { + if pricing == nil || pricing.LongContextInputThreshold <= 0 { + return false + } + if pricing.LongContextInputMultiplier <= 1 && pricing.LongContextOutputMultiplier <= 1 { + return false + } + totalInputTokens := tokens.InputTokens + tokens.CacheReadTokens + return totalInputTokens > pricing.LongContextInputThreshold +} + +func isOpenAIGPT54Model(model string) bool { + normalized := normalizeCodexModel(strings.TrimSpace(strings.ToLower(model))) + return normalized == "gpt-5.4" +} + // CalculateCostWithConfig 使用配置中的默认倍率计算费用 func (s *BillingService) CalculateCostWithConfig(model string, tokens UsageTokens) (*CostBreakdown, error) { multiplier := s.cfg.Default.RateMultiplier diff --git a/backend/internal/service/billing_service_test.go b/backend/internal/service/billing_service_test.go index 3ae6dc8562..0ba52e5640 100644 --- a/backend/internal/service/billing_service_test.go +++ b/backend/internal/service/billing_service_test.go @@ -169,6 +169,28 @@ func TestGetModelPricing_OpenAIGPT54Fallback(t *testing.T) { require.InDelta(t, 2.5e-6, pricing.InputPricePerToken, 1e-12) require.InDelta(t, 15e-6, pricing.OutputPricePerToken, 1e-12) require.InDelta(t, 0.25e-6, pricing.CacheReadPricePerToken, 1e-12) + require.Equal(t, 272000, pricing.LongContextInputThreshold) + require.InDelta(t, 2.0, pricing.LongContextInputMultiplier, 1e-12) + require.InDelta(t, 1.5, pricing.LongContextOutputMultiplier, 1e-12) +} + +func TestCalculateCost_OpenAIGPT54LongContextAppliesWholeSessionMultipliers(t *testing.T) { + svc := newTestBillingService() + + tokens := UsageTokens{ + InputTokens: 300000, + OutputTokens: 4000, + } + + cost, err := svc.CalculateCost("gpt-5.4-2026-03-05", tokens, 1.0) + require.NoError(t, err) + + expectedInput := float64(tokens.InputTokens) * 2.5e-6 * 2.0 + expectedOutput := float64(tokens.OutputTokens) * 15e-6 * 1.5 + require.InDelta(t, expectedInput, cost.InputCost, 1e-10) + require.InDelta(t, expectedOutput, cost.OutputCost, 1e-10) + require.InDelta(t, expectedInput+expectedOutput, cost.TotalCost, 1e-10) + require.InDelta(t, expectedInput+expectedOutput, cost.ActualCost, 1e-10) } func TestGetFallbackPricing_FamilyMatching(t *testing.T) { diff --git a/backend/internal/service/pricing_service.go b/backend/internal/service/pricing_service.go index b81a160a75..897623d633 100644 --- a/backend/internal/service/pricing_service.go +++ b/backend/internal/service/pricing_service.go @@ -24,12 +24,15 @@ var ( openAIModelDatePattern = regexp.MustCompile(`-\d{8}$`) openAIModelBasePattern = regexp.MustCompile(`^(gpt-\d+(?:\.\d+)?)(?:-|$)`) openAIGPT54FallbackPricing = &LiteLLMModelPricing{ - InputCostPerToken: 2.5e-06, // $2.5 per MTok - OutputCostPerToken: 1.5e-05, // $15 per MTok - CacheReadInputTokenCost: 2.5e-07, // $0.25 per MTok - LiteLLMProvider: "openai", - Mode: "chat", - SupportsPromptCaching: true, + InputCostPerToken: 2.5e-06, // $2.5 per MTok + OutputCostPerToken: 1.5e-05, // $15 per MTok + CacheReadInputTokenCost: 2.5e-07, // $0.25 per MTok + LongContextInputTokenThreshold: 272000, + LongContextInputCostMultiplier: 2.0, + LongContextOutputCostMultiplier: 1.5, + LiteLLMProvider: "openai", + Mode: "chat", + SupportsPromptCaching: true, } ) @@ -41,6 +44,9 @@ type LiteLLMModelPricing struct { CacheCreationInputTokenCost float64 `json:"cache_creation_input_token_cost"` CacheCreationInputTokenCostAbove1hr float64 `json:"cache_creation_input_token_cost_above_1hr"` CacheReadInputTokenCost float64 `json:"cache_read_input_token_cost"` + LongContextInputTokenThreshold int `json:"long_context_input_token_threshold,omitempty"` + LongContextInputCostMultiplier float64 `json:"long_context_input_cost_multiplier,omitempty"` + LongContextOutputCostMultiplier float64 `json:"long_context_output_cost_multiplier,omitempty"` LiteLLMProvider string `json:"litellm_provider"` Mode string `json:"mode"` SupportsPromptCaching bool `json:"supports_prompt_caching"` diff --git a/backend/internal/service/pricing_service_test.go b/backend/internal/service/pricing_service_test.go index de29528372..6b67c55a06 100644 --- a/backend/internal/service/pricing_service_test.go +++ b/backend/internal/service/pricing_service_test.go @@ -64,4 +64,7 @@ func TestGetModelPricing_Gpt54UsesStaticFallbackWhenRemoteMissing(t *testing.T) require.InDelta(t, 2.5e-6, got.InputCostPerToken, 1e-12) require.InDelta(t, 1.5e-5, got.OutputCostPerToken, 1e-12) require.InDelta(t, 2.5e-7, got.CacheReadInputTokenCost, 1e-12) + require.Equal(t, 272000, got.LongContextInputTokenThreshold) + require.InDelta(t, 2.0, got.LongContextInputCostMultiplier, 1e-12) + require.InDelta(t, 1.5, got.LongContextOutputCostMultiplier, 1e-12) } diff --git a/frontend/src/composables/__tests__/useModelWhitelist.spec.ts b/frontend/src/composables/__tests__/useModelWhitelist.spec.ts index 4088e5a4a3..79c88a2908 100644 --- a/frontend/src/composables/__tests__/useModelWhitelist.spec.ts +++ b/frontend/src/composables/__tests__/useModelWhitelist.spec.ts @@ -2,6 +2,13 @@ import { describe, expect, it } from 'vitest' import { buildModelMappingObject, getModelsByPlatform } from '../useModelWhitelist' describe('useModelWhitelist', () => { + it('openai 模型列表包含 GPT-5.4 官方快照', () => { + const models = getModelsByPlatform('openai') + + expect(models).toContain('gpt-5.4') + expect(models).toContain('gpt-5.4-2026-03-05') + }) + it('antigravity 模型列表包含图片模型兼容项', () => { const models = getModelsByPlatform('antigravity') @@ -15,4 +22,12 @@ describe('useModelWhitelist', () => { 'gemini-3.1-flash-image': 'gemini-3.1-flash-image' }) }) + + it('whitelist 模式会保留 GPT-5.4 官方快照的精确映射', () => { + const mapping = buildModelMappingObject('whitelist', ['gpt-5.4-2026-03-05'], []) + + expect(mapping).toEqual({ + 'gpt-5.4-2026-03-05': 'gpt-5.4-2026-03-05' + }) + }) }) diff --git a/frontend/src/composables/useModelWhitelist.ts b/frontend/src/composables/useModelWhitelist.ts index 8e525873dd..986a99f4b4 100644 --- a/frontend/src/composables/useModelWhitelist.ts +++ b/frontend/src/composables/useModelWhitelist.ts @@ -25,7 +25,7 @@ const openaiModels = [ 'gpt-5.2', 'gpt-5.2-2025-12-11', 'gpt-5.2-chat-latest', 'gpt-5.2-codex', 'gpt-5.2-pro', 'gpt-5.2-pro-2025-12-11', // GPT-5.4 系列 - 'gpt-5.4', + 'gpt-5.4', 'gpt-5.4-2026-03-05', // GPT-5.3 系列 'gpt-5.3-codex', 'gpt-5.3-codex-spark', 'chatgpt-4o-latest', From 491a744481b51bedc9f74cc93247edf56a4774c7 Mon Sep 17 00:00:00 2001 From: shaw Date: Fri, 6 Mar 2026 10:23:22 +0800 Subject: [PATCH 24/53] =?UTF-8?q?fix:=20=E4=BF=AE=E5=A4=8D=E8=B4=A6?= =?UTF-8?q?=E5=8F=B7=E5=88=97=E8=A1=A8=E9=A6=96=E6=AC=A1=E5=8A=A0=E8=BD=BD?= =?UTF-8?q?=E7=AA=97=E5=8F=A3=E8=B4=B9=E7=94=A8=E6=98=BE=E7=A4=BA=20$0.00?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit lite 模式下从快照缓存读取窗口费用，非 lite 模式查询后写入缓存 --- .../internal/handler/admin/account_handler.go | 64 ++++++++++++------- .../admin/account_window_cost_cache.go | 25 ++++++++ 2 files changed, 65 insertions(+), 24 deletions(-) create mode 100644 backend/internal/handler/admin/account_window_cost_cache.go diff --git a/backend/internal/handler/admin/account_handler.go b/backend/internal/handler/admin/account_handler.go index 23ebc8e62c..f42159a857 100644 --- a/backend/internal/handler/admin/account_handler.go +++ b/backend/internal/handler/admin/account_handler.go @@ -288,32 +288,48 @@ func (h *AccountHandler) List(c *gin.Context) { } } - // 仅非 lite 模式获取窗口费用（PostgreSQL 聚合查询，高开销） - if !lite && len(windowCostAccountIDs) > 0 { - windowCosts = make(map[int64]float64) - var mu sync.Mutex - g, gctx := errgroup.WithContext(c.Request.Context()) - g.SetLimit(10) // 限制并发数 - - for i := range accounts { - acc := &accounts[i] - if !acc.IsAnthropicOAuthOrSetupToken() || acc.GetWindowCostLimit() <= 0 { - continue - } - accCopy := acc // 闭包捕获 - g.Go(func() error { - // 使用统一的窗口开始时间计算逻辑（考虑窗口过期情况） - startTime := accCopy.GetCurrentWindowStartTime() - stats, err := h.accountUsageService.GetAccountWindowStats(gctx, accCopy.ID, startTime) - if err == nil && stats != nil { - mu.Lock() - windowCosts[accCopy.ID] = stats.StandardCost // 使用标准费用 - mu.Unlock() + // 窗口费用获取：lite 模式从快照缓存读取，非 lite 模式执行 PostgreSQL 查询后写入缓存 + if len(windowCostAccountIDs) > 0 { + if lite { + // lite 模式：尝试从快照缓存读取 + cacheKey := buildWindowCostCacheKey(windowCostAccountIDs) + if cached, ok := accountWindowCostCache.Get(cacheKey); ok { + if costs, ok := cached.Payload.(map[int64]float64); ok { + windowCosts = costs } - return nil // 不返回错误，允许部分失败 - }) + } + // 缓存未命中则 windowCosts 保持 nil（仅发生在服务刚启动时） + } else { + // 非 lite 模式：执行 PostgreSQL 聚合查询（高开销） + windowCosts = make(map[int64]float64) + var mu sync.Mutex + g, gctx := errgroup.WithContext(c.Request.Context()) + g.SetLimit(10) // 限制并发数 + + for i := range accounts { + acc := &accounts[i] + if !acc.IsAnthropicOAuthOrSetupToken() || acc.GetWindowCostLimit() <= 0 { + continue + } + accCopy := acc // 闭包捕获 + g.Go(func() error { + // 使用统一的窗口开始时间计算逻辑（考虑窗口过期情况） + startTime := accCopy.GetCurrentWindowStartTime() + stats, err := h.accountUsageService.GetAccountWindowStats(gctx, accCopy.ID, startTime) + if err == nil && stats != nil { + mu.Lock() + windowCosts[accCopy.ID] = stats.StandardCost // 使用标准费用 + mu.Unlock() + } + return nil // 不返回错误，允许部分失败 + }) + } + _ = g.Wait() + + // 查询完毕后写入快照缓存，供 lite 模式使用 + cacheKey := buildWindowCostCacheKey(windowCostAccountIDs) + accountWindowCostCache.Set(cacheKey, windowCosts) } - _ = g.Wait() } // Build response with concurrency info diff --git a/backend/internal/handler/admin/account_window_cost_cache.go b/backend/internal/handler/admin/account_window_cost_cache.go new file mode 100644 index 0000000000..3271b63036 --- /dev/null +++ b/backend/internal/handler/admin/account_window_cost_cache.go @@ -0,0 +1,25 @@ +package admin + +import ( + "strconv" + "strings" + "time" +) + +var accountWindowCostCache = newSnapshotCache(30 * time.Second) + +func buildWindowCostCacheKey(accountIDs []int64) string { + if len(accountIDs) == 0 { + return "accounts_window_cost_empty" + } + var b strings.Builder + b.Grow(len(accountIDs) * 6) + _, _ = b.WriteString("accounts_window_cost:") + for i, id := range accountIDs { + if i > 0 { + _ = b.WriteByte(',') + } + _, _ = b.WriteString(strconv.FormatInt(id, 10)) + } + return b.String() +} From 225fd035ae7462d8d7bf0b8d375c61166fee80a4 Mon Sep 17 00:00:00 2001 From: shaw Date: Fri, 6 Mar 2026 10:55:09 +0800 Subject: [PATCH 25/53] =?UTF-8?q?chore:=20=E6=9B=B4=E6=96=B0codex=E9=85=8D?= =?UTF-8?q?=E7=BD=AE=E9=83=A8=E5=88=86=E6=94=AF=E6=8C=81gpt-5.4=E7=9A=84?= =?UTF-8?q?=E9=95=BF=E4=B8=8A=E4=B8=8B=E6=96=87?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- frontend/src/components/keys/UseKeyModal.vue | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/frontend/src/components/keys/UseKeyModal.vue b/frontend/src/components/keys/UseKeyModal.vue index a61ce6b459..5ae2fa3f65 100644 --- a/frontend/src/components/keys/UseKeyModal.vue +++ b/frontend/src/components/keys/UseKeyModal.vue @@ -512,12 +512,14 @@ function generateOpenAIFiles(baseUrl: string, apiKey: string): FileConfig[] { // config.toml content const configContent = `model_provider = "OpenAI" -model = "gpt-5.3-codex" -review_model = "gpt-5.3-codex" -model_reasoning_effort = "xhigh" +model = "gpt-5.4" +review_model = "gpt-5.4" +model_reasoning_effort = "high" disable_response_storage = true network_access = "enabled" windows_wsl_setup_acknowledged = true +model_context_window = 1000000 +model_auto_compact_token_limit = 900000 [model_providers.OpenAI] name = "OpenAI" @@ -549,12 +551,14 @@ function generateOpenAIWsFiles(baseUrl: string, apiKey: string): FileConfig[] { // config.toml content with WebSocket v2 const configContent = `model_provider = "OpenAI" -model = "gpt-5.3-codex" -review_model = "gpt-5.3-codex" -model_reasoning_effort = "xhigh" +model = "gpt-5.4" +review_model = "gpt-5.4" +model_reasoning_effort = "high" disable_response_storage = true network_access = "enabled" windows_wsl_setup_acknowledged = true +model_context_window = 1000000 +model_auto_compact_token_limit = 900000 [model_providers.OpenAI] name = "OpenAI" From 0cce0a887792bf8b6a7af8ab53c588bc87e3c90e Mon Sep 17 00:00:00 2001 From: shaw Date: Fri, 6 Mar 2026 11:29:43 +0800 Subject: [PATCH 26/53] =?UTF-8?q?chore:=20gpt-5.4=E7=A4=BA=E4=BE=8B?= =?UTF-8?q?=E9=85=8D=E7=BD=AE=E4=BF=AE=E6=94=B9model=5Freasoning=5Feffort?= =?UTF-8?q?=E4=B8=BAxhigh?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- frontend/src/components/keys/UseKeyModal.vue | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/frontend/src/components/keys/UseKeyModal.vue b/frontend/src/components/keys/UseKeyModal.vue index 5ae2fa3f65..99d78f6974 100644 --- a/frontend/src/components/keys/UseKeyModal.vue +++ b/frontend/src/components/keys/UseKeyModal.vue @@ -514,7 +514,7 @@ function generateOpenAIFiles(baseUrl: string, apiKey: string): FileConfig[] { const configContent = `model_provider = "OpenAI" model = "gpt-5.4" review_model = "gpt-5.4" -model_reasoning_effort = "high" +model_reasoning_effort = "xhigh" disable_response_storage = true network_access = "enabled" windows_wsl_setup_acknowledged = true @@ -553,7 +553,7 @@ function generateOpenAIWsFiles(baseUrl: string, apiKey: string): FileConfig[] { const configContent = `model_provider = "OpenAI" model = "gpt-5.4" review_model = "gpt-5.4" -model_reasoning_effort = "high" +model_reasoning_effort = "xhigh" disable_response_storage = true network_access = "enabled" windows_wsl_setup_acknowledged = true From 896fb2b0b4392204b0e089e72cbd08adbc41e9ef Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 12:43:22 +0800 Subject: [PATCH 27/53] feat: add client affinity scheduling for Anthropic/Antigravity accounts MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 新增客户端亲和调度功能，新会话优先调度到客户端之前使用过的账号，减少账号频繁切换。 - 调度层 Layer 1.6：粘性会话未命中时，从 Redis Sorted Set 按最近使用时间降序 查找亲和账号，通过全部可调度检查后尝试获取槽位 - 账号级 opt-in：通过 Extra 字段 client_affinity_enabled 控制，未启用的账号 既不写入亲和关系也不被亲和选中 - Redis 数据结构：ZSET client_affinity:{groupID}:{clientID}，Lua 脚本原子清理 过期成员 + 查询/更新，per-member TTL 通过 score 实现 - 仅 Claude Code/Console 客户端生效（通过 metadata.user_id 格式识别） - DTO/前端：新增开关 UI，支持 Anthropic OAuth/SetupToken 和 Antigravity 平台 --- backend/internal/handler/dto/mappers.go | 6 ++ backend/internal/handler/dto/types.go | 4 + backend/internal/repository/gateway_cache.go | 73 ++++++++++++- backend/internal/service/account.go | 18 ++++ .../service/antigravity_smart_retry_test.go | 10 ++ .../gateway_hotpath_optimization_test.go | 9 ++ .../service/gateway_multiplatform_test.go | 10 ++ backend/internal/service/gateway_service.go | 100 +++++++++++++++++- .../service/gemini_multiplatform_test.go | 10 ++ .../service/openai_gateway_service_test.go | 10 ++ .../service/openai_ws_state_store_test.go | 10 ++ backend/internal/testutil/stubs.go | 9 ++ .../components/account/EditAccountModal.vue | 52 +++++++++ frontend/src/i18n/locales/en.ts | 4 + frontend/src/i18n/locales/zh.ts | 4 + frontend/src/types/index.ts | 4 + 16 files changed, 331 insertions(+), 2 deletions(-) diff --git a/backend/internal/handler/dto/mappers.go b/backend/internal/handler/dto/mappers.go index d3fc51ed66..816550ab04 100644 --- a/backend/internal/handler/dto/mappers.go +++ b/backend/internal/handler/dto/mappers.go @@ -250,6 +250,12 @@ func AccountFromServiceShallow(a *service.Account) *Account { } } + // 客户端亲和调度（适用于 Anthropic/Antigravity 平台） + if a.IsClientAffinityEnabled() { + enabled := true + out.ClientAffinityEnabled = &enabled + } + // 提取 API Key 账号配额限制（仅 apikey 类型有效） if a.Type == service.AccountTypeAPIKey { if limit := a.GetQuotaLimit(); limit > 0 { diff --git a/backend/internal/handler/dto/types.go b/backend/internal/handler/dto/types.go index f0d13d3f02..e5310a21f5 100644 --- a/backend/internal/handler/dto/types.go +++ b/backend/internal/handler/dto/types.go @@ -188,6 +188,10 @@ type Account struct { CacheTTLOverrideEnabled *bool `json:"cache_ttl_override_enabled,omitempty"` CacheTTLOverrideTarget *string `json:"cache_ttl_override_target,omitempty"` + // 客户端亲和调度（仅 Anthropic/Antigravity 平台有效） + // 启用后新会话会优先调度到客户端之前使用过的账号 + ClientAffinityEnabled *bool `json:"client_affinity_enabled,omitempty"` + // API Key 账号配额限制 QuotaLimit *float64 `json:"quota_limit,omitempty"` QuotaUsed *float64 `json:"quota_used,omitempty"` diff --git a/backend/internal/repository/gateway_cache.go b/backend/internal/repository/gateway_cache.go index 58291b6652..8fa6d0c401 100644 --- a/backend/internal/repository/gateway_cache.go +++ b/backend/internal/repository/gateway_cache.go @@ -3,13 +3,17 @@ package repository import ( "context" "fmt" + "strconv" "time" "github.com/Wei-Shaw/sub2api/internal/service" "github.com/redis/go-redis/v9" ) -const stickySessionPrefix = "sticky_session:" +const ( + stickySessionPrefix = "sticky_session:" + clientAffinityPrefix = "client_affinity:" +) type gatewayCache struct { rdb *redis.Client @@ -51,3 +55,70 @@ func (c *gatewayCache) DeleteSessionAccountID(ctx context.Context, groupID int64 key := buildSessionKey(groupID, sessionHash) return c.rdb.Del(ctx, key).Err() } + +// buildAffinityKey 构建客户端亲和 key +// 格式: client_affinity:{groupID}:{clientID} +func buildAffinityKey(groupID int64, clientID string) string { + return fmt.Sprintf("%s%d:%s", clientAffinityPrefix, groupID, clientID) +} + +// getAffinityScript: 清理过期成员后返回亲和账号列表（按最近使用降序） +// KEYS[1] = client_affinity:{groupID}:{clientID} +// ARGV[1] = 过期阈值时间戳 (now - ttl) +var getAffinityScript = redis.NewScript(` +redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[1]) +return redis.call('ZREVRANGE', KEYS[1], 0, -1) +`) + +// updateAffinityScript: 清理过期成员、添加/更新亲和关系、刷新 TTL +// KEYS[1] = client_affinity:{groupID}:{clientID} +// ARGV[1] = 当前时间戳 +// ARGV[2] = TTL 秒数 +// ARGV[3] = accountID +// ARGV[4] = 过期阈值时间戳 (now - ttl) +var updateAffinityScript = redis.NewScript(` +redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[4]) +redis.call('ZADD', KEYS[1], ARGV[1], ARGV[3]) +redis.call('EXPIRE', KEYS[1], ARGV[2]) +return 1 +`) + +func (c *gatewayCache) GetClientAffinityAccounts(ctx context.Context, groupID int64, clientID string, ttl time.Duration) ([]int64, error) { + key := buildAffinityKey(groupID, clientID) + now := time.Now().Unix() + expireThreshold := now - int64(ttl.Seconds()) + + result, err := getAffinityScript.Run(ctx, c.rdb, []string{key}, expireThreshold).StringSlice() + if err != nil { + if err == redis.Nil { + return nil, nil + } + return nil, err + } + + accountIDs := make([]int64, 0, len(result)) + for _, s := range result { + id, err := strconv.ParseInt(s, 10, 64) + if err != nil { + continue + } + accountIDs = append(accountIDs, id) + } + return accountIDs, nil +} + +func (c *gatewayCache) UpdateClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64, ttl time.Duration) error { + key := buildAffinityKey(groupID, clientID) + now := time.Now().Unix() + ttlSeconds := int64(ttl.Seconds()) + expireThreshold := now - ttlSeconds + + return updateAffinityScript.Run(ctx, c.rdb, []string{key}, + now, ttlSeconds, accountID, expireThreshold, + ).Err() +} + +func (c *gatewayCache) RemoveClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64) error { + key := buildAffinityKey(groupID, clientID) + return c.rdb.ZRem(ctx, key, accountID).Err() +} diff --git a/backend/internal/service/account.go b/backend/internal/service/account.go index 8eb3748c69..390790a325 100644 --- a/backend/internal/service/account.go +++ b/backend/internal/service/account.go @@ -1099,6 +1099,24 @@ func (a *Account) IsSessionIDMaskingEnabled() bool { return false } +// IsClientAffinityEnabled 检查是否启用客户端亲和调度 +// 仅适用于 Anthropic 平台账号（包括 anthropic 和 antigravity 平台） +// 启用后，新会话会优先调度到之前使用过的账号 +func (a *Account) IsClientAffinityEnabled() bool { + if a.Platform != PlatformAnthropic && a.Platform != PlatformAntigravity { + return false + } + if a.Extra == nil { + return false + } + if v, ok := a.Extra["client_affinity_enabled"]; ok { + if enabled, ok := v.(bool); ok { + return enabled + } + } + return false +} + // IsCacheTTLOverrideEnabled 检查是否启用缓存 TTL 强制替换 // 仅适用于 Anthropic OAuth/SetupToken 类型账号 // 启用后将所有 cache creation tokens 归入指定的 TTL 类型（5m 或 1h） diff --git a/backend/internal/service/antigravity_smart_retry_test.go b/backend/internal/service/antigravity_smart_retry_test.go index 432c80e59f..7f836bf47c 100644 --- a/backend/internal/service/antigravity_smart_retry_test.go +++ b/backend/internal/service/antigravity_smart_retry_test.go @@ -9,6 +9,7 @@ import ( "net/http" "strings" "testing" + "time" "github.com/stretchr/testify/require" ) @@ -29,6 +30,15 @@ func (c *stubSmartRetryCache) DeleteSessionAccountID(_ context.Context, groupID c.deleteCalls = append(c.deleteCalls, deleteSessionCall{groupID: groupID, sessionHash: sessionHash}) return nil } +func (c *stubSmartRetryCache) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (c *stubSmartRetryCache) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (c *stubSmartRetryCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} // mockSmartRetryUpstream 用于 handleSmartRetry 测试的 mock upstream type mockSmartRetryUpstream struct { diff --git a/backend/internal/service/gateway_hotpath_optimization_test.go b/backend/internal/service/gateway_hotpath_optimization_test.go index 161c4ba4b1..137490e8db 100644 --- a/backend/internal/service/gateway_hotpath_optimization_test.go +++ b/backend/internal/service/gateway_hotpath_optimization_test.go @@ -143,6 +143,15 @@ func (s *stickyGatewayCacheHotpathStub) RefreshSessionTTL(ctx context.Context, g func (s *stickyGatewayCacheHotpathStub) DeleteSessionAccountID(ctx context.Context, groupID int64, sessionHash string) error { return nil } +func (s *stickyGatewayCacheHotpathStub) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (s *stickyGatewayCacheHotpathStub) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (s *stickyGatewayCacheHotpathStub) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} func (s *modelsListAccountRepoStub) ListSchedulableByGroupID(ctx context.Context, groupID int64) ([]Account, error) { s.listByGroupCalls.Add(1) diff --git a/backend/internal/service/gateway_multiplatform_test.go b/backend/internal/service/gateway_multiplatform_test.go index 320ceaa708..466ff294bb 100644 --- a/backend/internal/service/gateway_multiplatform_test.go +++ b/backend/internal/service/gateway_multiplatform_test.go @@ -235,6 +235,16 @@ func (m *mockGatewayCacheForPlatform) DeleteSessionAccountID(ctx context.Context return nil } +func (m *mockGatewayCacheForPlatform) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (m *mockGatewayCacheForPlatform) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (m *mockGatewayCacheForPlatform) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} + type mockGroupRepoForGateway struct { groups map[int64]*Group getByIDCalls int diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index ccccdf4d10..0fe2545ec0 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -304,6 +304,10 @@ var ( sessionIDRegex = regexp.MustCompile(`session_([a-f0-9-]{36})`) claudeCliUserAgentRe = regexp.MustCompile(`^claude-cli/\d+\.\d+\.\d+`) + // clientIDFromMetadataRegex 从 metadata.user_id 中提取客户端 ID（64位 hex） + // 格式: user_{64位hex}_account_... + clientIDFromMetadataRegex = regexp.MustCompile(`^user_([a-f0-9]{64})_account_`) + // claudeCodePromptPrefixes 用于检测 Claude Code 系统提示词的前缀列表 // 支持多种变体：标准版、Agent SDK 版、Explore Agent 版、Compact 版等 // 注意：前缀之间不应存在包含关系，否则会导致冗余匹配 @@ -364,6 +368,13 @@ type GatewayCache interface { // DeleteSessionAccountID 删除粘性会话绑定，用于账号不可用时主动清理 // Delete sticky session binding, used to proactively clean up when account becomes unavailable DeleteSessionAccountID(ctx context.Context, groupID int64, sessionHash string) error + + // GetClientAffinityAccounts 获取客户端亲和账号列表（按最近使用降序），同时清理过期成员 + GetClientAffinityAccounts(ctx context.Context, groupID int64, clientID string, ttl time.Duration) ([]int64, error) + // UpdateClientAffinity 添加/更新客户端亲和关系（更新 score 为当前时间戳，刷新 key TTL） + UpdateClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64, ttl time.Duration) error + // RemoveClientAffinity 删除单个客户端亲和关系 + RemoveClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64) error } // derefGroupID safely dereferences *int64 to int64, returning 0 if nil @@ -434,6 +445,20 @@ func shouldClearStickySession(account *Account, requestedModel string) bool { return false } +// extractClientIDFromMetadata 从 metadata.user_id 中提取客户端 ID（64位 hex）。 +// 格式: user_{64位hex}_account_..._session_... +// 返回空字符串表示无法提取（非 Claude Code/Console 客户端）。 +func extractClientIDFromMetadata(metadataUserID string) string { + if metadataUserID == "" { + return "" + } + matches := clientIDFromMetadataRegex.FindStringSubmatch(metadataUserID) + if matches == nil { + return "" + } + return matches[1] +} + type AccountWaitPlan struct { AccountID int64 MaxConcurrency int @@ -1046,7 +1071,7 @@ func (s *GatewayService) SelectAccountForModelWithExclusions(ctx context.Context } // SelectAccountWithLoadAwareness selects account with load-awareness and wait plan. -// metadataUserID: 已废弃参数，会话限制现在统一使用 sessionHash +// metadataUserID: 用于客户端亲和调度，从中提取客户端 ID func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, groupID *int64, sessionHash string, requestedModel string, excludedIDs map[int64]struct{}, metadataUserID string) (*AccountSelectionResult, error) { // 调试日志：记录调度入口参数 excludedIDsList := make([]int64, 0, len(excludedIDs)) @@ -1077,6 +1102,9 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro } } + // 提取客户端 ID（用于客户端亲和调度） + affinityClientID := extractClientIDFromMetadata(metadataUserID) + if s.debugModelRoutingEnabled() && requestedModel != "" { groupPlatform := "" if group != nil { @@ -1374,6 +1402,9 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro if sessionHash != "" && s.cache != nil { _ = s.cache.SetSessionAccountID(ctx, derefGroupID(groupID), sessionHash, item.account.ID, stickySessionTTL) } + if affinityClientID != "" && s.cache != nil && item.account.IsClientAffinityEnabled() { + _ = s.cache.UpdateClientAffinity(ctx, derefGroupID(groupID), affinityClientID, item.account.ID, stickySessionTTL) + } if s.debugModelRoutingEnabled() { logger.LegacyPrintf("service.gateway", "[ModelRoutingDebug] routed select: group_id=%v model=%s session=%s account=%d", derefGroupID(groupID), requestedModel, shortSessionHash(sessionHash), item.account.ID) } @@ -1470,6 +1501,66 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro } } + // ============ Layer 1.6: 客户端亲和（仅在粘性会话未命中时生效） ============ + if affinityClientID != "" && s.cache != nil && stickyAccountID <= 0 { + affinityAccountIDs, err := s.cache.GetClientAffinityAccounts(ctx, derefGroupID(groupID), affinityClientID, stickySessionTTL) + if err == nil && len(affinityAccountIDs) > 0 { + for _, affinityAccID := range affinityAccountIDs { + if isExcluded(affinityAccID) { + continue + } + account, ok := accountByID[affinityAccID] + if !ok || !s.isAccountSchedulableForSelection(account) { + continue + } + if !account.IsClientAffinityEnabled() { + continue + } + if !s.isAccountAllowedForPlatform(account, platform, useMixed) { + continue + } + if requestedModel != "" && !s.isModelSupportedByAccountWithContext(ctx, account, requestedModel) { + continue + } + if !s.isAccountSchedulableForModelSelection(ctx, account, requestedModel) { + continue + } + if !s.isAccountSchedulableForQuota(account) { + continue + } + if !s.isAccountSchedulableForWindowCost(ctx, account, false) { + continue + } + if !s.isAccountSchedulableForRPM(ctx, account, false) { + continue + } + + result, err := s.tryAcquireAccountSlot(ctx, affinityAccID, account.Concurrency) + if err == nil && result.Acquired { + if !s.checkAndRegisterSession(ctx, account, sessionHash) { + result.ReleaseFunc() + continue + } + // 亲和命中：更新亲和 score + 绑定粘性会话 + _ = s.cache.UpdateClientAffinity(ctx, derefGroupID(groupID), affinityClientID, affinityAccID, stickySessionTTL) + if sessionHash != "" { + _ = s.cache.SetSessionAccountID(ctx, derefGroupID(groupID), sessionHash, affinityAccID, stickySessionTTL) + } + slog.Debug("client_affinity_hit", + "group_id", derefGroupID(groupID), + "client_id", affinityClientID[:8]+"...", + "account_id", affinityAccID) + return &AccountSelectionResult{ + Account: account, + Acquired: true, + ReleaseFunc: result.ReleaseFunc, + }, nil + } + } + // 所有亲和账号不可用，继续到 Layer 2 + } + } + // ============ Layer 2: 负载感知选择 ============ candidates := make([]*Account, 0, len(accounts)) for i := range accounts { @@ -1522,6 +1613,9 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro loadMap, err := s.concurrencyService.GetAccountsLoadBatch(ctx, accountLoads) if err != nil { if result, ok := s.tryAcquireByLegacyOrder(ctx, candidates, groupID, sessionHash, preferOAuth); ok { + if affinityClientID != "" && s.cache != nil && result.Account != nil && result.Account.IsClientAffinityEnabled() { + _ = s.cache.UpdateClientAffinity(ctx, derefGroupID(groupID), affinityClientID, result.Account.ID, stickySessionTTL) + } return result, nil } } else { @@ -1560,6 +1654,10 @@ func (s *GatewayService) SelectAccountWithLoadAwareness(ctx context.Context, gro if sessionHash != "" && s.cache != nil { _ = s.cache.SetSessionAccountID(ctx, derefGroupID(groupID), sessionHash, selected.account.ID, stickySessionTTL) } + // 更新客户端亲和关系 + if affinityClientID != "" && s.cache != nil && selected.account.IsClientAffinityEnabled() { + _ = s.cache.UpdateClientAffinity(ctx, derefGroupID(groupID), affinityClientID, selected.account.ID, stickySessionTTL) + } return &AccountSelectionResult{ Account: selected.account, Acquired: true, diff --git a/backend/internal/service/gemini_multiplatform_test.go b/backend/internal/service/gemini_multiplatform_test.go index b0b804eb00..56f94d5f1f 100644 --- a/backend/internal/service/gemini_multiplatform_test.go +++ b/backend/internal/service/gemini_multiplatform_test.go @@ -288,6 +288,16 @@ func (m *mockGatewayCacheForGemini) DeleteSessionAccountID(ctx context.Context, return nil } +func (m *mockGatewayCacheForGemini) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (m *mockGatewayCacheForGemini) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (m *mockGatewayCacheForGemini) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} + // TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_GeminiPlatform 测试 Gemini 单平台选择 func TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_GeminiPlatform(t *testing.T) { ctx := context.Background() diff --git a/backend/internal/service/openai_gateway_service_test.go b/backend/internal/service/openai_gateway_service_test.go index 4f5f7f3c12..7e53c8996e 100644 --- a/backend/internal/service/openai_gateway_service_test.go +++ b/backend/internal/service/openai_gateway_service_test.go @@ -265,6 +265,16 @@ func (c *stubGatewayCache) DeleteSessionAccountID(ctx context.Context, groupID i return nil } +func (c *stubGatewayCache) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (c *stubGatewayCache) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (c *stubGatewayCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} + func TestOpenAISelectAccountWithLoadAwareness_FiltersUnschedulable(t *testing.T) { now := time.Now() resetAt := now.Add(10 * time.Minute) diff --git a/backend/internal/service/openai_ws_state_store_test.go b/backend/internal/service/openai_ws_state_store_test.go index 235d42331d..6f77f64356 100644 --- a/backend/internal/service/openai_ws_state_store_test.go +++ b/backend/internal/service/openai_ws_state_store_test.go @@ -193,6 +193,16 @@ func (c *openAIWSStateStoreTimeoutProbeCache) DeleteSessionAccountID(ctx context return nil } +func (c *openAIWSStateStoreTimeoutProbeCache) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (c *openAIWSStateStoreTimeoutProbeCache) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (c *openAIWSStateStoreTimeoutProbeCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} + func TestOpenAIWSStateStore_RedisOpsUseShortTimeout(t *testing.T) { probe := &openAIWSStateStoreTimeoutProbeCache{} store := NewOpenAIWSStateStore(probe) diff --git a/backend/internal/testutil/stubs.go b/backend/internal/testutil/stubs.go index 217a5f5699..669372ef51 100644 --- a/backend/internal/testutil/stubs.go +++ b/backend/internal/testutil/stubs.go @@ -97,6 +97,15 @@ func (c StubGatewayCache) RefreshSessionTTL(_ context.Context, _ int64, _ string func (c StubGatewayCache) DeleteSessionAccountID(_ context.Context, _ int64, _ string) error { return nil } +func (c StubGatewayCache) GetClientAffinityAccounts(_ context.Context, _ int64, _ string, _ time.Duration) ([]int64, error) { + return nil, nil +} +func (c StubGatewayCache) UpdateClientAffinity(_ context.Context, _ int64, _ string, _ int64, _ time.Duration) error { + return nil +} +func (c StubGatewayCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { + return nil +} // ============================================================ // StubSessionLimitCache — service.SessionLimitCache 的空实现 diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index be7d2d45ce..6026eb9570 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -1165,6 +1165,38 @@
+ +
+
+
+
+ +

+ {{ t('admin.accounts.quotaControl.clientAffinity.hint') }} +

+
+ +
+
+
+
@@ -1391,6 +1423,7 @@ const tlsFingerprintEnabled = ref(false) const sessionIdMaskingEnabled = ref(false) const cacheTTLOverrideEnabled = ref(false) const cacheTTLOverrideTarget = ref('5m') +const clientAffinityEnabled = ref(false) // OpenAI 自动透传开关（OAuth/API Key） const openaiPassthroughEnabled = ref(false) @@ -1872,6 +1905,12 @@ function loadQuotaControlSettings(account: Account) { sessionIdMaskingEnabled.value = false cacheTTLOverrideEnabled.value = false cacheTTLOverrideTarget.value = '5m' + clientAffinityEnabled.value = false + + // Client affinity applies to both Anthropic and Antigravity platforms + if (account.client_affinity_enabled === true) { + clientAffinityEnabled.value = true + } // Only applies to Anthropic OAuth/SetupToken accounts if (account.platform !== 'anthropic' || (account.type !== 'oauth' && account.type !== 'setup-token')) { @@ -2196,6 +2235,12 @@ const handleSubmit = async () => { } else { delete newExtra.mixed_scheduling } + // Client affinity setting + if (clientAffinityEnabled.value) { + newExtra.client_affinity_enabled = true + } else { + delete newExtra.client_affinity_enabled + } updatePayload.extra = newExtra } @@ -2271,6 +2316,13 @@ const handleSubmit = async () => { delete newExtra.cache_ttl_override_target } + // Client affinity setting + if (clientAffinityEnabled.value) { + newExtra.client_affinity_enabled = true + } else { + delete newExtra.client_affinity_enabled + } + updatePayload.extra = newExtra } diff --git a/frontend/src/i18n/locales/en.ts b/frontend/src/i18n/locales/en.ts index 4d51899c11..16b0bcd594 100644 --- a/frontend/src/i18n/locales/en.ts +++ b/frontend/src/i18n/locales/en.ts @@ -2005,6 +2005,10 @@ export default { hint: 'Force all cache creation tokens to be billed as the selected TTL tier (5m or 1h)', target: 'Target TTL', targetHint: 'Select the TTL tier for billing' + }, + clientAffinity: { + label: 'Client Affinity Scheduling', + hint: 'When enabled, new sessions prefer accounts previously used by this client to reduce account switching' } }, expired: 'Expired', diff --git a/frontend/src/i18n/locales/zh.ts b/frontend/src/i18n/locales/zh.ts index ec7836156a..5ddef97620 100644 --- a/frontend/src/i18n/locales/zh.ts +++ b/frontend/src/i18n/locales/zh.ts @@ -2139,6 +2139,10 @@ export default { hint: '将所有缓存创建 token 强制按指定的 TTL 类型（5分钟或1小时）计费', target: '目标 TTL', targetHint: '选择计费使用的 TTL 类型' + }, + clientAffinity: { + label: '客户端亲和调度', + hint: '启用后，新会话会优先调度到该客户端之前使用过的账号，避免频繁切换账号' } }, expired: '已过期', diff --git a/frontend/src/types/index.ts b/frontend/src/types/index.ts index 7da1b0c866..ff61120d7c 100644 --- a/frontend/src/types/index.ts +++ b/frontend/src/types/index.ts @@ -710,6 +710,10 @@ export interface Account { cache_ttl_override_enabled?: boolean | null cache_ttl_override_target?: string | null + // 客户端亲和调度（仅 Anthropic/Antigravity 平台有效） + // 启用后新会话会优先调度到客户端之前使用过的账号 + client_affinity_enabled?: boolean | null + // API Key 账号配额限制 quota_limit?: number | null quota_used?: number | null From 180c515f9a7aceabb77057d1ffe19ed2cf2addd2 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 13:02:04 +0800 Subject: [PATCH 28/53] fix: race condition in TestOpsSystemLogSink_StartStopAndFlushSuccess MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit done 信号在 BatchInsertSystemLogsFn 内发出，但 writtenCount 在函数返回后 才原子递增，导致测试主 goroutine 收到信号后立即检查 WrittenCount 时可能为 0。 改用短轮询等待 WrittenCount > 0，消除竞态。 --- backend/internal/service/ops_system_log_sink_test.go | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/backend/internal/service/ops_system_log_sink_test.go b/backend/internal/service/ops_system_log_sink_test.go index 12a2ec0c7d..137ee33c72 100644 --- a/backend/internal/service/ops_system_log_sink_test.go +++ b/backend/internal/service/ops_system_log_sink_test.go @@ -183,6 +183,15 @@ func TestOpsSystemLogSink_StartStopAndFlushSuccess(t *testing.T) { if strings.TrimSpace(item.Message) == "" { t.Fatalf("message should not be empty") } + // writtenCount is incremented after BatchInsertSystemLogsFn returns, + // so poll briefly to avoid a race between the done signal and the atomic add. + deadline := time.Now().Add(time.Second) + for time.Now().Before(deadline) { + if sink.Health().WrittenCount > 0 { + break + } + time.Sleep(time.Millisecond) + } health := sink.Health() if health.WrittenCount == 0 { t.Fatalf("written_count should be >0") From 75f1faea1475f3bc6771792bc3fbf424404b7b24 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 13:08:32 +0800 Subject: [PATCH 29/53] chore: bump version to 0.1.92.1 --- backend/cmd/server/VERSION | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backend/cmd/server/VERSION b/backend/cmd/server/VERSION index adaf52a822..b28c5259d4 100644 --- a/backend/cmd/server/VERSION +++ b/backend/cmd/server/VERSION @@ -1 +1 @@ -0.1.91.4 +0.1.92.1 From 312f2ecb8d87579da8e1b4f3cbd223d54f39fb70 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 13:54:25 +0800 Subject: [PATCH 30/53] fix: OpenAI passthrough accounts bypass model mapping check MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 透传模式账号仅替换认证，应允许所有模型通过。之前调度阶段的 isModelSupportedByAccount 不感知透传模式，导致 model_mapping 中未配置的新模型（如 gpt-5.4）被拒绝返回 503。 --- backend/internal/service/gateway_service.go | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index 0fe2545ec0..f6c993a89c 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -3424,6 +3424,10 @@ func (s *GatewayService) isModelSupportedByAccount(account *Account, requestedMo if account.Platform == PlatformSora { return s.isSoraModelSupportedByAccount(account, requestedModel) } + // OpenAI 透传模式：仅替换认证，允许所有模型 + if account.Platform == PlatformOpenAI && account.IsOpenAIPassthroughEnabled() { + return true + } // OAuth/SetupToken 账号使用 Anthropic 标准映射（短ID → 长ID） if account.Platform == PlatformAnthropic && account.Type != AccountTypeAPIKey { requestedModel = claude.NormalizeModelID(requestedModel) From 489a6d483926b498e7ca8fb76ed16e1ebd7d2962 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 14:20:08 +0800 Subject: [PATCH 31/53] fix: OpenAI scheduler also bypass model check for passthrough accounts MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 之前只在 gateway_service.isModelSupportedByAccount 中添加了透传短路， 但 OpenAI 调度器 (openai_account_scheduler.go) 直接调用 account.IsModelSupported()，未经过该函数，导致 gpt-5.4 仍然 503。 --- backend/internal/service/openai_account_scheduler.go | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/backend/internal/service/openai_account_scheduler.go b/backend/internal/service/openai_account_scheduler.go index cf4bc26eb0..9e2b81035a 100644 --- a/backend/internal/service/openai_account_scheduler.go +++ b/backend/internal/service/openai_account_scheduler.go @@ -323,7 +323,7 @@ func (s *defaultOpenAIAccountScheduler) selectBySessionHash( _ = s.service.deleteStickySessionAccountID(ctx, req.GroupID, sessionHash) return nil, nil } - if req.RequestedModel != "" && !account.IsModelSupported(req.RequestedModel) { + if req.RequestedModel != "" && !account.IsOpenAIPassthroughEnabled() && !account.IsModelSupported(req.RequestedModel) { return nil, nil } if !s.isAccountTransportCompatible(account, req.RequiredTransport) { @@ -582,7 +582,7 @@ func (s *defaultOpenAIAccountScheduler) selectByLoadBalance( if !account.IsSchedulable() || !account.IsOpenAI() { continue } - if req.RequestedModel != "" && !account.IsModelSupported(req.RequestedModel) { + if req.RequestedModel != "" && !account.IsOpenAIPassthroughEnabled() && !account.IsModelSupported(req.RequestedModel) { continue } if !s.isAccountTransportCompatible(account, req.RequiredTransport) { From 69abdff7322d8cf5102f749c60dacc052174ab3a Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 14:24:58 +0800 Subject: [PATCH 32/53] chore: bump version to 0.1.92.2 --- backend/cmd/server/VERSION | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/backend/cmd/server/VERSION b/backend/cmd/server/VERSION index b28c5259d4..4e611a941e 100644 --- a/backend/cmd/server/VERSION +++ b/backend/cmd/server/VERSION @@ -1 +1 @@ -0.1.92.1 +0.1.92.2 From c1816f39f2a3e40c91a3eee29384f5c15f24d553 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 14:30:44 +0800 Subject: [PATCH 33/53] revert: remove passthrough check from isModelSupportedByAccount MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 该函数不被 OpenAI 调度路径调用，透传短路应只加在 openai_account_scheduler.go 中。 --- backend/internal/service/gateway_service.go | 4 ---- 1 file changed, 4 deletions(-) diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index f6c993a89c..0fe2545ec0 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -3424,10 +3424,6 @@ func (s *GatewayService) isModelSupportedByAccount(account *Account, requestedMo if account.Platform == PlatformSora { return s.isSoraModelSupportedByAccount(account, requestedModel) } - // OpenAI 透传模式：仅替换认证，允许所有模型 - if account.Platform == PlatformOpenAI && account.IsOpenAIPassthroughEnabled() { - return true - } // OAuth/SetupToken 账号使用 Anthropic 标准映射（短ID → 长ID） if account.Platform == PlatformAnthropic && account.Type != AccountTypeAPIKey { requestedModel = claude.NormalizeModelID(requestedModel) From 63dff626732383ff8e8d474978943e4f0b3a4bac Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 14:46:47 +0800 Subject: [PATCH 34/53] fix: restrict client affinity to Anthropic OAuth/SetupToken accounts only MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 客户端亲和仅对 Claude Code 账号（Anthropic OAuth/SetupToken）生效， 移除 Antigravity 平台支持。UI toggle 移动到拦截预热请求下方。 --- backend/internal/handler/dto/mappers.go | 2 +- backend/internal/handler/dto/types.go | 2 +- backend/internal/service/account.go | 4 +- .../components/account/EditAccountModal.vue | 78 +++++++++---------- 4 files changed, 39 insertions(+), 47 deletions(-) diff --git a/backend/internal/handler/dto/mappers.go b/backend/internal/handler/dto/mappers.go index 816550ab04..068771ac8c 100644 --- a/backend/internal/handler/dto/mappers.go +++ b/backend/internal/handler/dto/mappers.go @@ -250,7 +250,7 @@ func AccountFromServiceShallow(a *service.Account) *Account { } } - // 客户端亲和调度（适用于 Anthropic/Antigravity 平台） + // 客户端亲和调度（仅 Anthropic OAuth/SetupToken 账号） if a.IsClientAffinityEnabled() { enabled := true out.ClientAffinityEnabled = &enabled diff --git a/backend/internal/handler/dto/types.go b/backend/internal/handler/dto/types.go index e5310a21f5..a7e04cd2ea 100644 --- a/backend/internal/handler/dto/types.go +++ b/backend/internal/handler/dto/types.go @@ -188,7 +188,7 @@ type Account struct { CacheTTLOverrideEnabled *bool `json:"cache_ttl_override_enabled,omitempty"` CacheTTLOverrideTarget *string `json:"cache_ttl_override_target,omitempty"` - // 客户端亲和调度（仅 Anthropic/Antigravity 平台有效） + // 客户端亲和调度（仅 Anthropic OAuth/SetupToken 账号有效） // 启用后新会话会优先调度到客户端之前使用过的账号 ClientAffinityEnabled *bool `json:"client_affinity_enabled,omitempty"` diff --git a/backend/internal/service/account.go b/backend/internal/service/account.go index 390790a325..d36f55978c 100644 --- a/backend/internal/service/account.go +++ b/backend/internal/service/account.go @@ -1100,10 +1100,10 @@ func (a *Account) IsSessionIDMaskingEnabled() bool { } // IsClientAffinityEnabled 检查是否启用客户端亲和调度 -// 仅适用于 Anthropic 平台账号（包括 anthropic 和 antigravity 平台） +// 仅适用于 Anthropic OAuth/SetupToken 账号（Claude Code） // 启用后，新会话会优先调度到之前使用过的账号 func (a *Account) IsClientAffinityEnabled() bool { - if a.Platform != PlatformAnthropic && a.Platform != PlatformAntigravity { + if !a.IsAnthropicOAuthOrSetupToken() { return false } if a.Extra == nil { diff --git a/frontend/src/components/account/EditAccountModal.vue b/frontend/src/components/account/EditAccountModal.vue index 6026eb9570..0bc5f479cb 100644 --- a/frontend/src/components/account/EditAccountModal.vue +++ b/frontend/src/components/account/EditAccountModal.vue @@ -645,6 +645,36 @@
+ +
+
+
+ +

+ {{ t('admin.accounts.quotaControl.clientAffinity.hint') }} +

+
+ +
+
+
@@ -1165,38 +1195,6 @@
- -
-
-
-
- -

- {{ t('admin.accounts.quotaControl.clientAffinity.hint') }} -

-
- -
-
-
-
@@ -1907,16 +1905,16 @@ function loadQuotaControlSettings(account: Account) { cacheTTLOverrideTarget.value = '5m' clientAffinityEnabled.value = false - // Client affinity applies to both Anthropic and Antigravity platforms - if (account.client_affinity_enabled === true) { - clientAffinityEnabled.value = true - } - // Only applies to Anthropic OAuth/SetupToken accounts if (account.platform !== 'anthropic' || (account.type !== 'oauth' && account.type !== 'setup-token')) { return } + // Client affinity (Anthropic OAuth/SetupToken only) + if (account.client_affinity_enabled === true) { + clientAffinityEnabled.value = true + } + // Load from extra field (via backend DTO fields) if (account.window_cost_limit != null && account.window_cost_limit > 0) { windowCostEnabled.value = true @@ -2235,12 +2233,6 @@ const handleSubmit = async () => { } else { delete newExtra.mixed_scheduling } - // Client affinity setting - if (clientAffinityEnabled.value) { - newExtra.client_affinity_enabled = true - } else { - delete newExtra.client_affinity_enabled - } updatePayload.extra = newExtra } From 60f050d184466c8c143bcb3c7d0064c90a9a3322 Mon Sep 17 00:00:00 2001 From: erio Date: Fri, 6 Mar 2026 15:48:58 +0800 Subject: [PATCH 35/53] feat: add account affinity client count indicator Add reverse Redis index (account_affinity:{accountID}) to track which clients are bound to each account via affinity scheduling. Display a blue badge with client count in the admin account list, with a hover tooltip showing client details (groupID:clientID + TTL). Backend: reverse index in Lua scripts, batch query methods, new GET /accounts/:id/affinity-clients endpoint. Frontend: blue badge + lazy-loaded tooltip in AccountStatusIndicator. --- .../admin/account_data_handler_test.go | 1 + .../internal/handler/admin/account_handler.go | 54 ++++++- .../account_handler_mixed_channel_test.go | 2 +- .../admin/account_handler_passthrough_test.go | 1 + .../admin/batch_update_credentials_test.go | 2 +- backend/internal/repository/gateway_cache.go | 141 +++++++++++++++++- backend/internal/server/api_contract_test.go | 2 +- backend/internal/server/routes/admin.go | 1 + .../service/antigravity_smart_retry_test.go | 9 ++ .../gateway_hotpath_optimization_test.go | 9 ++ .../service/gateway_multiplatform_test.go | 9 ++ backend/internal/service/gateway_service.go | 19 +++ .../service/gemini_multiplatform_test.go | 9 ++ .../service/openai_gateway_service_test.go | 9 ++ .../service/openai_ws_state_store_test.go | 9 ++ backend/internal/testutil/stubs.go | 9 ++ .../account/AccountStatusIndicator.vue | 66 +++++++- frontend/src/types/index.ts | 1 + 18 files changed, 339 insertions(+), 14 deletions(-) diff --git a/backend/internal/handler/admin/account_data_handler_test.go b/backend/internal/handler/admin/account_data_handler_test.go index 285033a17d..e516b275f2 100644 --- a/backend/internal/handler/admin/account_data_handler_test.go +++ b/backend/internal/handler/admin/account_data_handler_test.go @@ -65,6 +65,7 @@ func setupAccountDataRouter() (*gin.Engine, *stubAdminService) { nil, nil, nil, + nil, ) router.GET("/api/v1/admin/accounts/data", h.ExportData) diff --git a/backend/internal/handler/admin/account_handler.go b/backend/internal/handler/admin/account_handler.go index 843d51c516..2fa0e20c4c 100644 --- a/backend/internal/handler/admin/account_handler.go +++ b/backend/internal/handler/admin/account_handler.go @@ -55,6 +55,7 @@ type AccountHandler struct { sessionLimitCache service.SessionLimitCache rpmCache service.RPMCache tokenCacheInvalidator service.TokenCacheInvalidator + gatewayCache service.GatewayCache } // NewAccountHandler creates a new admin account handler @@ -72,6 +73,7 @@ func NewAccountHandler( sessionLimitCache service.SessionLimitCache, rpmCache service.RPMCache, tokenCacheInvalidator service.TokenCacheInvalidator, + gatewayCache service.GatewayCache, ) *AccountHandler { return &AccountHandler{ adminService: adminService, @@ -87,6 +89,7 @@ func NewAccountHandler( sessionLimitCache: sessionLimitCache, rpmCache: rpmCache, tokenCacheInvalidator: tokenCacheInvalidator, + gatewayCache: gatewayCache, } } @@ -158,9 +161,10 @@ type AccountWithConcurrency struct { *dto.Account CurrentConcurrency int `json:"current_concurrency"` // 以下字段仅对 Anthropic OAuth/SetupToken 账号有效，且仅在启用相应功能时返回 - CurrentWindowCost *float64 `json:"current_window_cost,omitempty"` // 当前窗口费用 - ActiveSessions *int `json:"active_sessions,omitempty"` // 当前活跃会话数 - CurrentRPM *int `json:"current_rpm,omitempty"` // 当前分钟 RPM 计数 + CurrentWindowCost *float64 `json:"current_window_cost,omitempty"` // 当前窗口费用 + ActiveSessions *int `json:"active_sessions,omitempty"` // 当前活跃会话数 + CurrentRPM *int `json:"current_rpm,omitempty"` // 当前分钟 RPM 计数 + AffinityClientCount *int `json:"affinity_client_count,omitempty"` // 亲和客户端数量 } func (h *AccountHandler) buildAccountResponseWithRuntime(ctx context.Context, account *service.Account) AccountWithConcurrency { @@ -243,6 +247,7 @@ func (h *AccountHandler) List(c *gin.Context) { var windowCosts map[int64]float64 var activeSessions map[int64]int var rpmCounts map[int64]int + var affinityCounts map[int64]int // 始终获取并发数（Redis ZCARD，极低开销） if h.concurrencyService != nil { @@ -251,10 +256,11 @@ func (h *AccountHandler) List(c *gin.Context) { } } - // 识别需要查询窗口费用、会话数和 RPM 的账号（Anthropic OAuth/SetupToken 且启用了相应功能） + // 识别需要查询窗口费用、会话数、RPM 和亲和客户端的账号（Anthropic OAuth/SetupToken 且启用了相应功能） windowCostAccountIDs := make([]int64, 0) sessionLimitAccountIDs := make([]int64, 0) rpmAccountIDs := make([]int64, 0) + affinityAccountIDs := make([]int64, 0) sessionIdleTimeouts := make(map[int64]time.Duration) // 各账号的会话空闲超时配置 for i := range accounts { acc := &accounts[i] @@ -269,6 +275,9 @@ func (h *AccountHandler) List(c *gin.Context) { if acc.GetBaseRPM() > 0 { rpmAccountIDs = append(rpmAccountIDs, acc.ID) } + if acc.IsClientAffinityEnabled() { + affinityAccountIDs = append(affinityAccountIDs, acc.ID) + } } } @@ -288,6 +297,11 @@ func (h *AccountHandler) List(c *gin.Context) { } } + // 获取亲和客户端数量（Redis pipeline，低开销） + if len(affinityAccountIDs) > 0 && h.gatewayCache != nil { + affinityCounts, _ = h.gatewayCache.GetAccountAffinityCountBatch(c.Request.Context(), affinityAccountIDs, service.ClientAffinityTTL()) + } + // 窗口费用获取：lite 模式从快照缓存读取，非 lite 模式执行 PostgreSQL 查询后写入缓存 if len(windowCostAccountIDs) > 0 { if lite { @@ -362,6 +376,13 @@ func (h *AccountHandler) List(c *gin.Context) { } } + // 添加亲和客户端数量（仅当启用时） + if affinityCounts != nil { + if count, ok := affinityCounts[acc.ID]; ok { + item.AffinityClientCount = &count + } + } + result[i] = item } @@ -1852,3 +1873,28 @@ func sanitizeExtraBaseRPM(extra map[string]any) { } extra["base_rpm"] = v } + +// GetAffinityClients 返回指定账号的亲和客户端列表 +// GET /api/v1/admin/accounts/:id/affinity-clients +func (h *AccountHandler) GetAffinityClients(c *gin.Context) { + id, err := strconv.ParseInt(c.Param("id"), 10, 64) + if err != nil { + response.Error(c, http.StatusBadRequest, "invalid id") + return + } + + if h.gatewayCache == nil { + response.Success(c, []service.AffinityClientInfo{}) + return + } + + clients, err := h.gatewayCache.GetAccountAffinityClients(c.Request.Context(), id, service.ClientAffinityTTL()) + if err != nil { + response.ErrorFrom(c, err) + return + } + if clients == nil { + clients = []service.AffinityClientInfo{} + } + response.Success(c, clients) +} diff --git a/backend/internal/handler/admin/account_handler_mixed_channel_test.go b/backend/internal/handler/admin/account_handler_mixed_channel_test.go index 5b81db2aad..5e45619064 100644 --- a/backend/internal/handler/admin/account_handler_mixed_channel_test.go +++ b/backend/internal/handler/admin/account_handler_mixed_channel_test.go @@ -15,7 +15,7 @@ import ( func setupAccountMixedChannelRouter(adminSvc *stubAdminService) *gin.Engine { gin.SetMode(gin.TestMode) router := gin.New() - accountHandler := NewAccountHandler(adminSvc, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) + accountHandler := NewAccountHandler(adminSvc, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) router.POST("/api/v1/admin/accounts/check-mixed-channel", accountHandler.CheckMixedChannel) router.POST("/api/v1/admin/accounts", accountHandler.Create) router.PUT("/api/v1/admin/accounts/:id", accountHandler.Update) diff --git a/backend/internal/handler/admin/account_handler_passthrough_test.go b/backend/internal/handler/admin/account_handler_passthrough_test.go index d86501c047..a11f2c5195 100644 --- a/backend/internal/handler/admin/account_handler_passthrough_test.go +++ b/backend/internal/handler/admin/account_handler_passthrough_test.go @@ -29,6 +29,7 @@ func TestAccountHandler_Create_AnthropicAPIKeyPassthroughExtraForwarded(t *testi nil, nil, nil, + nil, ) router := gin.New() diff --git a/backend/internal/handler/admin/batch_update_credentials_test.go b/backend/internal/handler/admin/batch_update_credentials_test.go index 0b1b669174..f2cd1e3a1c 100644 --- a/backend/internal/handler/admin/batch_update_credentials_test.go +++ b/backend/internal/handler/admin/batch_update_credentials_test.go @@ -36,7 +36,7 @@ func (f *failingAdminService) UpdateAccount(ctx context.Context, id int64, input func setupAccountHandlerWithService(adminSvc service.AdminService) (*gin.Engine, *AccountHandler) { gin.SetMode(gin.TestMode) router := gin.New() - handler := NewAccountHandler(adminSvc, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) + handler := NewAccountHandler(adminSvc, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) router.POST("/api/v1/admin/accounts/batch-update-credentials", handler.BatchUpdateCredentials) return router, handler } diff --git a/backend/internal/repository/gateway_cache.go b/backend/internal/repository/gateway_cache.go index 8fa6d0c401..6fed5b6874 100644 --- a/backend/internal/repository/gateway_cache.go +++ b/backend/internal/repository/gateway_cache.go @@ -4,6 +4,7 @@ import ( "context" "fmt" "strconv" + "strings" "time" "github.com/Wei-Shaw/sub2api/internal/service" @@ -11,8 +12,9 @@ import ( ) const ( - stickySessionPrefix = "sticky_session:" - clientAffinityPrefix = "client_affinity:" + stickySessionPrefix = "sticky_session:" + clientAffinityPrefix = "client_affinity:" + accountAffinityPrefix = "account_affinity:" ) type gatewayCache struct { @@ -70,16 +72,21 @@ redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[1]) return redis.call('ZREVRANGE', KEYS[1], 0, -1) `) -// updateAffinityScript: 清理过期成员、添加/更新亲和关系、刷新 TTL +// updateAffinityScript: 清理过期成员、添加/更新亲和关系、刷新 TTL， +// 同时维护反向索引 account_affinity:{accountID} // KEYS[1] = client_affinity:{groupID}:{clientID} +// KEYS[2] = account_affinity:{accountID} // ARGV[1] = 当前时间戳 // ARGV[2] = TTL 秒数 // ARGV[3] = accountID // ARGV[4] = 过期阈值时间戳 (now - ttl) +// ARGV[5] = 反向索引成员 "{groupID}:{clientID}" var updateAffinityScript = redis.NewScript(` redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[4]) redis.call('ZADD', KEYS[1], ARGV[1], ARGV[3]) redis.call('EXPIRE', KEYS[1], ARGV[2]) +redis.call('ZADD', KEYS[2], ARGV[1], ARGV[5]) +redis.call('EXPIRE', KEYS[2], ARGV[2]) return 1 `) @@ -109,16 +116,138 @@ func (c *gatewayCache) GetClientAffinityAccounts(ctx context.Context, groupID in func (c *gatewayCache) UpdateClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64, ttl time.Duration) error { key := buildAffinityKey(groupID, clientID) + reverseKey := buildAccountAffinityKey(accountID) + reverseMember := fmt.Sprintf("%d:%s", groupID, clientID) now := time.Now().Unix() ttlSeconds := int64(ttl.Seconds()) expireThreshold := now - ttlSeconds - return updateAffinityScript.Run(ctx, c.rdb, []string{key}, - now, ttlSeconds, accountID, expireThreshold, + return updateAffinityScript.Run(ctx, c.rdb, []string{key, reverseKey}, + now, ttlSeconds, accountID, expireThreshold, reverseMember, ).Err() } func (c *gatewayCache) RemoveClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64) error { key := buildAffinityKey(groupID, clientID) - return c.rdb.ZRem(ctx, key, accountID).Err() + reverseKey := buildAccountAffinityKey(accountID) + reverseMember := fmt.Sprintf("%d:%s", groupID, clientID) + pipe := c.rdb.Pipeline() + pipe.ZRem(ctx, key, accountID) + pipe.ZRem(ctx, reverseKey, reverseMember) + _, err := pipe.Exec(ctx) + return err +} + +// buildAccountAffinityKey 构建账号亲和反向索引 key +// 格式: account_affinity:{accountID} +func buildAccountAffinityKey(accountID int64) string { + return fmt.Sprintf("%s%d", accountAffinityPrefix, accountID) +} + +// getAccountAffinityScript: 清理过期成员后返回 ZCARD +// KEYS[1] = account_affinity:{accountID} +// ARGV[1] = 过期阈值时间戳 (now - ttl) +var getAccountAffinityScript = redis.NewScript(` +redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[1]) +return redis.call('ZCARD', KEYS[1]) +`) + +func (c *gatewayCache) GetAccountAffinityCount(ctx context.Context, accountID int64, ttl time.Duration) (int, error) { + key := buildAccountAffinityKey(accountID) + expireThreshold := time.Now().Unix() - int64(ttl.Seconds()) + + count, err := getAccountAffinityScript.Run(ctx, c.rdb, []string{key}, expireThreshold).Int() + if err != nil && err != redis.Nil { + return 0, err + } + return count, nil +} + +func (c *gatewayCache) GetAccountAffinityCountBatch(ctx context.Context, accountIDs []int64, ttl time.Duration) (map[int64]int, error) { + if len(accountIDs) == 0 { + return nil, nil + } + + expireThreshold := time.Now().Unix() - int64(ttl.Seconds()) + pipe := c.rdb.Pipeline() + cmds := make([]*redis.Cmd, len(accountIDs)) + + for i, id := range accountIDs { + key := buildAccountAffinityKey(id) + cmds[i] = getAccountAffinityScript.Run(ctx, pipe, []string{key}, expireThreshold) + } + + _, err := pipe.Exec(ctx) + if err != nil && err != redis.Nil { + return nil, err + } + + result := make(map[int64]int, len(accountIDs)) + for i, id := range accountIDs { + count, err := cmds[i].Int() + if err != nil && err != redis.Nil { + continue + } + if count > 0 { + result[id] = count + } + } + return result, nil +} + +// getAccountAffinityClientsScript: 清理过期成员后返回成员列表及 score +// KEYS[1] = account_affinity:{accountID} +// ARGV[1] = 过期阈值时间戳 (now - ttl) +var getAccountAffinityClientsScript = redis.NewScript(` +redis.call('ZREMRANGEBYSCORE', KEYS[1], '-inf', ARGV[1]) +return redis.call('ZRANGE', KEYS[1], 0, -1, 'WITHSCORES') +`) + +func (c *gatewayCache) GetAccountAffinityClients(ctx context.Context, accountID int64, ttl time.Duration) ([]service.AffinityClientInfo, error) { + key := buildAccountAffinityKey(accountID) + now := time.Now().Unix() + expireThreshold := now - int64(ttl.Seconds()) + + raw, err := getAccountAffinityClientsScript.Run(ctx, c.rdb, []string{key}, expireThreshold).StringSlice() + if err != nil { + if err == redis.Nil { + return nil, nil + } + return nil, err + } + + // ZRANGE WITHSCORES returns [member1, score1, member2, score2, ...] + clients := make([]service.AffinityClientInfo, 0, len(raw)/2) + for i := 0; i+1 < len(raw); i += 2 { + member := raw[i] + scoreStr := raw[i+1] + + score, parseErr := strconv.ParseInt(scoreStr, 10, 64) + if parseErr != nil { + continue + } + + // member format: "{groupID}:{clientID}" + sepIdx := strings.IndexByte(member, ':') + if sepIdx < 0 { + continue + } + groupID, parseErr := strconv.ParseInt(member[:sepIdx], 10, 64) + if parseErr != nil { + continue + } + clientID := member[sepIdx+1:] + + remaining := score + int64(ttl.Seconds()) - now + if remaining < 0 { + remaining = 0 + } + + clients = append(clients, service.AffinityClientInfo{ + GroupID: groupID, + ClientID: clientID, + TTL: remaining, + }) + } + return clients, nil } diff --git a/backend/internal/server/api_contract_test.go b/backend/internal/server/api_contract_test.go index aafbbe21cb..537e64b4ea 100644 --- a/backend/internal/server/api_contract_test.go +++ b/backend/internal/server/api_contract_test.go @@ -648,7 +648,7 @@ func newContractDeps(t *testing.T) *contractDeps { apiKeyHandler := handler.NewAPIKeyHandler(apiKeyService) usageHandler := handler.NewUsageHandler(usageService, apiKeyService) adminSettingHandler := adminhandler.NewSettingHandler(settingService, nil, nil, nil, nil) - adminAccountHandler := adminhandler.NewAccountHandler(adminService, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) + adminAccountHandler := adminhandler.NewAccountHandler(adminService, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil, nil) jwtAuth := func(c *gin.Context) { c.Set(string(middleware.ContextKeyUser), middleware.AuthSubject{ diff --git a/backend/internal/server/routes/admin.go b/backend/internal/server/routes/admin.go index 2e53feb36f..46366a03b5 100644 --- a/backend/internal/server/routes/admin.go +++ b/backend/internal/server/routes/admin.go @@ -257,6 +257,7 @@ func registerAccountRoutes(admin *gin.RouterGroup, h *handler.Handlers) { accounts.DELETE("/:id/temp-unschedulable", h.Admin.Account.ClearTempUnschedulable) accounts.POST("/:id/schedulable", h.Admin.Account.SetSchedulable) accounts.GET("/:id/models", h.Admin.Account.GetAvailableModels) + accounts.GET("/:id/affinity-clients", h.Admin.Account.GetAffinityClients) accounts.POST("/batch", h.Admin.Account.BatchCreate) accounts.GET("/data", h.Admin.Account.ExportData) accounts.POST("/data", h.Admin.Account.ImportData) diff --git a/backend/internal/service/antigravity_smart_retry_test.go b/backend/internal/service/antigravity_smart_retry_test.go index 7f836bf47c..a516805758 100644 --- a/backend/internal/service/antigravity_smart_retry_test.go +++ b/backend/internal/service/antigravity_smart_retry_test.go @@ -39,6 +39,15 @@ func (c *stubSmartRetryCache) UpdateClientAffinity(_ context.Context, _ int64, _ func (c *stubSmartRetryCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (c *stubSmartRetryCache) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (c *stubSmartRetryCache) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (c *stubSmartRetryCache) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} // mockSmartRetryUpstream 用于 handleSmartRetry 测试的 mock upstream type mockSmartRetryUpstream struct { diff --git a/backend/internal/service/gateway_hotpath_optimization_test.go b/backend/internal/service/gateway_hotpath_optimization_test.go index 137490e8db..5eb97de993 100644 --- a/backend/internal/service/gateway_hotpath_optimization_test.go +++ b/backend/internal/service/gateway_hotpath_optimization_test.go @@ -152,6 +152,15 @@ func (s *stickyGatewayCacheHotpathStub) UpdateClientAffinity(_ context.Context, func (s *stickyGatewayCacheHotpathStub) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (s *stickyGatewayCacheHotpathStub) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (s *stickyGatewayCacheHotpathStub) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (s *stickyGatewayCacheHotpathStub) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} func (s *modelsListAccountRepoStub) ListSchedulableByGroupID(ctx context.Context, groupID int64) ([]Account, error) { s.listByGroupCalls.Add(1) diff --git a/backend/internal/service/gateway_multiplatform_test.go b/backend/internal/service/gateway_multiplatform_test.go index 466ff294bb..c513d6b179 100644 --- a/backend/internal/service/gateway_multiplatform_test.go +++ b/backend/internal/service/gateway_multiplatform_test.go @@ -244,6 +244,15 @@ func (m *mockGatewayCacheForPlatform) UpdateClientAffinity(_ context.Context, _ func (m *mockGatewayCacheForPlatform) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (m *mockGatewayCacheForPlatform) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (m *mockGatewayCacheForPlatform) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (m *mockGatewayCacheForPlatform) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} type mockGroupRepoForGateway struct { groups map[int64]*Group diff --git a/backend/internal/service/gateway_service.go b/backend/internal/service/gateway_service.go index 0fe2545ec0..964e1a0348 100644 --- a/backend/internal/service/gateway_service.go +++ b/backend/internal/service/gateway_service.go @@ -375,6 +375,25 @@ type GatewayCache interface { UpdateClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64, ttl time.Duration) error // RemoveClientAffinity 删除单个客户端亲和关系 RemoveClientAffinity(ctx context.Context, groupID int64, clientID string, accountID int64) error + + // GetAccountAffinityCount 获取账号的亲和客户端数量（清理过期后） + GetAccountAffinityCount(ctx context.Context, accountID int64, ttl time.Duration) (int, error) + // GetAccountAffinityCountBatch 批量获取多个账号的亲和客户端数量 + GetAccountAffinityCountBatch(ctx context.Context, accountIDs []int64, ttl time.Duration) (map[int64]int, error) + // GetAccountAffinityClients 获取账号的亲和客户端详情列表 + GetAccountAffinityClients(ctx context.Context, accountID int64, ttl time.Duration) ([]AffinityClientInfo, error) +} + +// AffinityClientInfo 表示一个亲和客户端的详情 +type AffinityClientInfo struct { + GroupID int64 `json:"group_id"` + ClientID string `json:"client_id"` + TTL int64 `json:"ttl"` // 剩余秒数 +} + +// ClientAffinityTTL 返回客户端亲和 TTL（与粘性会话 TTL 相同） +func ClientAffinityTTL() time.Duration { + return stickySessionTTL } // derefGroupID safely dereferences *int64 to int64, returning 0 if nil diff --git a/backend/internal/service/gemini_multiplatform_test.go b/backend/internal/service/gemini_multiplatform_test.go index 56f94d5f1f..a08ef690b2 100644 --- a/backend/internal/service/gemini_multiplatform_test.go +++ b/backend/internal/service/gemini_multiplatform_test.go @@ -297,6 +297,15 @@ func (m *mockGatewayCacheForGemini) UpdateClientAffinity(_ context.Context, _ in func (m *mockGatewayCacheForGemini) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (m *mockGatewayCacheForGemini) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (m *mockGatewayCacheForGemini) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (m *mockGatewayCacheForGemini) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} // TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_GeminiPlatform 测试 Gemini 单平台选择 func TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_GeminiPlatform(t *testing.T) { diff --git a/backend/internal/service/openai_gateway_service_test.go b/backend/internal/service/openai_gateway_service_test.go index 7e53c8996e..7b7b6da415 100644 --- a/backend/internal/service/openai_gateway_service_test.go +++ b/backend/internal/service/openai_gateway_service_test.go @@ -274,6 +274,15 @@ func (c *stubGatewayCache) UpdateClientAffinity(_ context.Context, _ int64, _ st func (c *stubGatewayCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (c *stubGatewayCache) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (c *stubGatewayCache) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (c *stubGatewayCache) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} func TestOpenAISelectAccountWithLoadAwareness_FiltersUnschedulable(t *testing.T) { now := time.Now() diff --git a/backend/internal/service/openai_ws_state_store_test.go b/backend/internal/service/openai_ws_state_store_test.go index 6f77f64356..273bfadddb 100644 --- a/backend/internal/service/openai_ws_state_store_test.go +++ b/backend/internal/service/openai_ws_state_store_test.go @@ -202,6 +202,15 @@ func (c *openAIWSStateStoreTimeoutProbeCache) UpdateClientAffinity(_ context.Con func (c *openAIWSStateStoreTimeoutProbeCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (c *openAIWSStateStoreTimeoutProbeCache) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (c *openAIWSStateStoreTimeoutProbeCache) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (c *openAIWSStateStoreTimeoutProbeCache) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]AffinityClientInfo, error) { + return nil, nil +} func TestOpenAIWSStateStore_RedisOpsUseShortTimeout(t *testing.T) { probe := &openAIWSStateStoreTimeoutProbeCache{} diff --git a/backend/internal/testutil/stubs.go b/backend/internal/testutil/stubs.go index 669372ef51..8c6f54dd07 100644 --- a/backend/internal/testutil/stubs.go +++ b/backend/internal/testutil/stubs.go @@ -106,6 +106,15 @@ func (c StubGatewayCache) UpdateClientAffinity(_ context.Context, _ int64, _ str func (c StubGatewayCache) RemoveClientAffinity(_ context.Context, _ int64, _ string, _ int64) error { return nil } +func (c StubGatewayCache) GetAccountAffinityCount(_ context.Context, _ int64, _ time.Duration) (int, error) { + return 0, nil +} +func (c StubGatewayCache) GetAccountAffinityCountBatch(_ context.Context, _ []int64, _ time.Duration) (map[int64]int, error) { + return nil, nil +} +func (c StubGatewayCache) GetAccountAffinityClients(_ context.Context, _ int64, _ time.Duration) ([]service.AffinityClientInfo, error) { + return nil, nil +} // ============================================================ // StubSessionLimitCache — service.SessionLimitCache 的空实现 diff --git a/frontend/src/components/account/AccountStatusIndicator.vue b/frontend/src/components/account/AccountStatusIndicator.vue index 728b310ccc..3a8d632f42 100644 --- a/frontend/src/components/account/AccountStatusIndicator.vue +++ b/frontend/src/components/account/AccountStatusIndicator.vue @@ -1,5 +1,29 @@