feat: support cached input token billing

apps/api/internal/runner/limits_test.go

@@ -1,6 +1,7 @@
 package runner
 
 import (
+	"context"
 	"testing"
 
 	"github.com/easyai/easyai-ai-gateway/apps/api/internal/clients"
@@ -29,6 +30,44 @@ func TestTokenUsageAmountsFallsBackToInputOutputTotal(t *testing.T) {
 	}
 }
 
+func TestBillingsSplitsCachedInputTokens(t *testing.T) {
+	service := &Service{}
+	candidate := store.RuntimeModelCandidate{
+		ModelName: "test-model",
+		BillingConfig: map[string]any{
+			"textInputPer1k":       1.0,
+			"textCachedInputPer1k": 0.25,
+			"textOutputPer1k":      2.0,
+		},
+	}
+	response := clients.Response{
+		Usage: clients.Usage{
+			InputTokens:       1000,
+			CachedInputTokens: 400,
+			OutputTokens:      500,
+			TotalTokens:       1500,
+		},
+	}
+
+	lines := service.billings(context.Background(), nil, "chat.completions", nil, candidate, response, false)
+
+	if len(lines) != 3 {
+		t.Fatalf("expected uncached input, cached input, and output lines, got %+v", lines)
+	}
+	uncached, _ := lines[0].(map[string]any)
+	cached, _ := lines[1].(map[string]any)
+	output, _ := lines[2].(map[string]any)
+	if uncached["resourceType"] != "text_input" || uncached["quantity"] != 600 || uncached["amount"] != 0.6 {
+		t.Fatalf("unexpected uncached input billing line: %+v", uncached)
+	}
+	if cached["resourceType"] != "text_cached_input" || cached["quantity"] != 400 || cached["amount"] != 0.1 {
+		t.Fatalf("unexpected cached input billing line: %+v", cached)
+	}
+	if output["resourceType"] != "text_output" || output["quantity"] != 500 || output["amount"] != 1.0 {
+		t.Fatalf("unexpected output billing line: %+v", output)
+	}
+}
+
 func TestEffectiveRateLimitPolicyTreatsEmptyRuntimePolicyAsUnlimited(t *testing.T) {
 	policy := effectiveRateLimitPolicy(store.RuntimeModelCandidate{
 		PlatformRateLimitPolicy: map[string]any{"rules": []any{