fix: include Gemini thoughtsTokenCount in output token billing

Gemini 2.5 Pro/Flash thinking models return thoughtsTokenCount separately
from candidatesTokenCount in usageMetadata, but this field was not parsed
or included in billing calculations, causing thinking tokens to be
unbilled.

- Add ThoughtsTokenCount field to GeminiUsageMetadata struct
- Include thoughtsTokenCount in OutputTokens across all 3 Gemini usage
  parsing paths (non-streaming, streaming, compat layer)
- Add tests covering thinking token scenarios

Closes #554

backend/internal/service/gemini_messages_compat_service.go

@@ -2663,11 +2663,12 @@ func extractGeminiUsage(geminiResp map[string]any) *ClaudeUsage {
 	prompt, _ := asInt(usageMeta["promptTokenCount"])
 	cand, _ := asInt(usageMeta["candidatesTokenCount"])
 	cached, _ := asInt(usageMeta["cachedContentTokenCount"])
+	thoughts, _ := asInt(usageMeta["thoughtsTokenCount"])
 	// 注意：Gemini 的 promptTokenCount 包含 cachedContentTokenCount，
 	// 但 Claude 的 input_tokens 不包含 cache_read_input_tokens，需要减去
 	return &ClaudeUsage{
 		InputTokens:          prompt - cached,
-		OutputTokens:         cand,
+		OutputTokens:         cand + thoughts,
 		CacheReadInputTokens: cached,
 	}
 }