refactor: replace scope-level rate limiting with model-level rate limiting

Merge functional changes from develop branch:
- Remove AntigravityQuotaScope system (claude/gemini_text/gemini_image)
- Replace with per-model rate limiting using resolveAntigravityModelKey
- Remove model load statistics (IncrModelCallCount/GetModelLoadBatch)
- Simplify account selection to unified priority→load→LRU algorithm
- Remove SetAntigravityQuotaScopeLimit from AccountRepository
- Clean up scope-related UI indicators and API fields

backend/internal/service/openai_gateway_service_test.go

@@ -204,14 +204,6 @@ func (c *stubGatewayCache) DeleteSessionAccountID(ctx context.Context, groupID i
 	return nil
 }
 
-func (c *stubGatewayCache) IncrModelCallCount(ctx context.Context, accountID int64, model string) (int64, error) {
-	return 0, nil
-}
-
-func (c *stubGatewayCache) GetModelLoadBatch(ctx context.Context, accountIDs []int64, model string) (map[int64]*ModelLoadInfo, error) {
-	return nil, nil
-}
-
 func TestOpenAISelectAccountWithLoadAwareness_FiltersUnschedulable(t *testing.T) {
 	now := time.Now()
 	resetAt := now.Add(10 * time.Minute)