perf(网关): 复用分组上下文减少热路径查询

新增 GetByIDLite 并在网关与 Gemini 选择流程复用上下文 group，避免 COUNT 触发
更新 API key 中间件注入 group 上下文，减少重复查库
补充 gateway/gemini 中间件与仓库层回归测试

测试: make test

backend/internal/service/gemini_multiplatform_test.go

@@ -8,6 +8,7 @@ import (
 	"testing"
 	"time"
 
+	"github.com/Wei-Shaw/sub2api/internal/pkg/ctxkey"
 	"github.com/Wei-Shaw/sub2api/internal/pkg/pagination"
 	"github.com/stretchr/testify/require"
 )
@@ -146,10 +147,21 @@ var _ AccountRepository = (*mockAccountRepoForGemini)(nil)
 
 // mockGroupRepoForGemini Gemini 测试用的 group repo mock
 type mockGroupRepoForGemini struct {
-	groups map[int64]*Group
+	groups           map[int64]*Group
+	getByIDCalls     int
+	getByIDLiteCalls int
 }
 
 func (m *mockGroupRepoForGemini) GetByID(ctx context.Context, id int64) (*Group, error) {
+	m.getByIDCalls++
+	if g, ok := m.groups[id]; ok {
+		return g, nil
+	}
+	return nil, errors.New("group not found")
+}
+
+func (m *mockGroupRepoForGemini) GetByIDLite(ctx context.Context, id int64) (*Group, error) {
+	m.getByIDLiteCalls++
 	if g, ok := m.groups[id]; ok {
 		return g, nil
 	}
@@ -242,6 +254,76 @@ func TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_GeminiP
 	require.Equal(t, PlatformGemini, acc.Platform, "无分组时应只返回 gemini 平台账户")
 }
 
+func TestGeminiMessagesCompatService_GroupResolution_ReusesContextGroup(t *testing.T) {
+	ctx := context.Background()
+	groupID := int64(7)
+	group := &Group{
+		ID:       groupID,
+		Platform: PlatformGemini,
+		Status:   StatusActive,
+	}
+	ctx = context.WithValue(ctx, ctxkey.Group, group)
+
+	repo := &mockAccountRepoForGemini{
+		accounts: []Account{
+			{ID: 1, Platform: PlatformGemini, Priority: 1, Status: StatusActive, Schedulable: true},
+		},
+		accountsByID: map[int64]*Account{},
+	}
+	for i := range repo.accounts {
+		repo.accountsByID[repo.accounts[i].ID] = &repo.accounts[i]
+	}
+
+	cache := &mockGatewayCacheForGemini{}
+	groupRepo := &mockGroupRepoForGemini{groups: map[int64]*Group{}}
+
+	svc := &GeminiMessagesCompatService{
+		accountRepo: repo,
+		groupRepo:   groupRepo,
+		cache:       cache,
+	}
+
+	acc, err := svc.SelectAccountForModelWithExclusions(ctx, &groupID, "", "gemini-2.5-flash", nil)
+	require.NoError(t, err)
+	require.NotNil(t, acc)
+	require.Equal(t, 0, groupRepo.getByIDCalls)
+	require.Equal(t, 0, groupRepo.getByIDLiteCalls)
+}
+
+func TestGeminiMessagesCompatService_GroupResolution_UsesLiteFetch(t *testing.T) {
+	ctx := context.Background()
+	groupID := int64(7)
+
+	repo := &mockAccountRepoForGemini{
+		accounts: []Account{
+			{ID: 1, Platform: PlatformGemini, Priority: 1, Status: StatusActive, Schedulable: true},
+		},
+		accountsByID: map[int64]*Account{},
+	}
+	for i := range repo.accounts {
+		repo.accountsByID[repo.accounts[i].ID] = &repo.accounts[i]
+	}
+
+	cache := &mockGatewayCacheForGemini{}
+	groupRepo := &mockGroupRepoForGemini{
+		groups: map[int64]*Group{
+			groupID: {ID: groupID, Platform: PlatformGemini},
+		},
+	}
+
+	svc := &GeminiMessagesCompatService{
+		accountRepo: repo,
+		groupRepo:   groupRepo,
+		cache:       cache,
+	}
+
+	acc, err := svc.SelectAccountForModelWithExclusions(ctx, &groupID, "", "gemini-2.5-flash", nil)
+	require.NoError(t, err)
+	require.NotNil(t, acc)
+	require.Equal(t, 0, groupRepo.getByIDCalls)
+	require.Equal(t, 1, groupRepo.getByIDLiteCalls)
+}
+
 // TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_AntigravityGroup 测试 antigravity 分组
 func TestGeminiMessagesCompatService_SelectAccountForModelWithExclusions_AntigravityGroup(t *testing.T) {
 	ctx := context.Background()