feat: improve model rate limit tracking

apps/api/internal/runner/limits_test.go

@@ -0,0 +1,29 @@
+package runner
+
+import (
+	"testing"
+
+	"github.com/easyai/easyai-ai-gateway/apps/api/internal/clients"
+)
+
+func TestTokenUsageAmountsUsesActualUsageForTPM(t *testing.T) {
+	got := tokenUsageAmounts(clients.Usage{InputTokens: 12, OutputTokens: 8, TotalTokens: 21})
+
+	if got["tpm_input"] != 12 {
+		t.Fatalf("expected input token amount 12, got %v", got["tpm_input"])
+	}
+	if got["tpm_output"] != 8 {
+		t.Fatalf("expected output token amount 8, got %v", got["tpm_output"])
+	}
+	if got["tpm_total"] != 21 {
+		t.Fatalf("expected total token amount 21, got %v", got["tpm_total"])
+	}
+}
+
+func TestTokenUsageAmountsFallsBackToInputOutputTotal(t *testing.T) {
+	got := tokenUsageAmounts(clients.Usage{InputTokens: 3, OutputTokens: 5})
+
+	if got["tpm_total"] != 8 {
+		t.Fatalf("expected total token fallback 8, got %v", got["tpm_total"])
+	}
+}