fix: extract and log Claude output_config.effort in usage records

Claude's output_config.effort parameter (low/medium/high/max) was not
being extracted from requests or logged in the reasoning_effort column
of usage logs. Only the OpenAI path populated this field.

Changes:
- Extract output_config.effort in ParseGatewayRequest
- Add ReasoningEffort field to ForwardResult
- Populate reasoning_effort in both RecordUsage and RecordUsageWithLongContext
- Guard against overwriting service-set effort values in handler
- Update stale comments that described reasoning_effort as OpenAI-only
- Add unit tests for extraction, normalization, and persistence

backend/internal/service/gateway_service.go

@@ -492,6 +492,7 @@ type ForwardResult struct {
 	Duration         time.Duration
 	FirstTokenMs     *int // 首字时间（流式请求）
 	ClientDisconnect bool // 客户端是否在流式传输过程中断开
+	ReasoningEffort  *string
 
 	// 图片生成计费字段（图片生成模型使用）
 	ImageCount int    // 生成的图片数量
@@ -7523,6 +7524,7 @@ func (s *GatewayService) RecordUsage(ctx context.Context, input *RecordUsageInpu
 		AccountID:             account.ID,
 		RequestID:             requestID,
 		Model:                 result.Model,
+		ReasoningEffort:       result.ReasoningEffort,
 		InputTokens:           result.Usage.InputTokens,
 		OutputTokens:          result.Usage.OutputTokens,
 		CacheCreationTokens:   result.Usage.CacheCreationInputTokens,
@@ -7699,6 +7701,7 @@ func (s *GatewayService) RecordUsageWithLongContext(ctx context.Context, input *
 		AccountID:             account.ID,
 		RequestID:             requestID,
 		Model:                 result.Model,
+		ReasoningEffort:       result.ReasoningEffort,
 		InputTokens:           result.Usage.InputTokens,
 		OutputTokens:          result.Usage.OutputTokens,
 		CacheCreationTokens:   result.Usage.CacheCreationInputTokens,