feat: implement tiered billing expression evaluation and related functionality

- Added support for tiered billing expressions in the billing system.
- Introduced new types and functions for handling billing expressions, including caching and execution.
- Updated existing billing logic to accommodate tiered billing scenarios.
- Enhanced request handling to support incoming billing expression requests.
- Added tests for tiered billing functionality to ensure correctness.

service/tiered_settle.go

@@ -0,0 +1,36 @@
+package service
+
+import (
+	"github.com/QuantumNous/new-api/pkg/billingexpr"
+	relaycommon "github.com/QuantumNous/new-api/relay/common"
+)
+
+// TieredResultWrapper wraps billingexpr.TieredResult for use at the service layer.
+type TieredResultWrapper = billingexpr.TieredResult
+
+// TryTieredSettle checks if the request uses tiered_expr billing and, if so,
+// computes the actual quota using the frozen BillingSnapshot. Returns:
+//   - ok=true, quota, result  when tiered billing applies
+//   - ok=false, 0, nil        when it doesn't (caller should fall through to existing logic)
+func TryTieredSettle(relayInfo *relaycommon.RelayInfo, params billingexpr.TokenParams) (ok bool, quota int, result *billingexpr.TieredResult) {
+	snap := relayInfo.TieredBillingSnapshot
+	if snap == nil || snap.BillingMode != "tiered_expr" {
+		return false, 0, nil
+	}
+
+	requestInput := billingexpr.RequestInput{}
+	if relayInfo.BillingRequestInput != nil {
+		requestInput = *relayInfo.BillingRequestInput
+	}
+
+	tr, err := billingexpr.ComputeTieredQuotaWithRequest(snap, params, requestInput)
+	if err != nil {
+		quota = relayInfo.FinalPreConsumedQuota
+		if quota <= 0 {
+			quota = snap.EstimatedQuotaAfterGroup
+		}
+		return true, quota, nil
+	}
+
+	return true, tr.ActualQuotaAfterGroup, &tr
+}