refactor: update repository interfaces and improve dependency injection

- Refactored repository structures to use interfaces instead of concrete types, enhancing flexibility and testability.
- Updated various repository methods to accept interfaces, allowing for better dependency management.
- Modified wire generation to accommodate new repository interfaces, ensuring proper service injection throughout the application.
- Enhanced handler methods to utilize DTOs for filtering and data handling, improving code clarity and maintainability.

internal/pkg/openai/client.go

@@ -69,14 +69,46 @@ func (c *clientImpl) ModeratePost(ctx context.Context, title, content string, im
 	if !c.IsEnabled() {
 		return true, "", nil
 	}
-	return c.moderateContentInBatches(ctx, fmt.Sprintf("帖子标题：%s\n帖子内容：%s", title, content), images)
+
+	// 构建帖子提示词，处理标题或内容为空的情况
+	var prompt strings.Builder
+	if strings.TrimSpace(title) != "" {
+		prompt.WriteString("帖子标题：" + title)
+	}
+	if strings.TrimSpace(content) != "" {
+		if prompt.Len() > 0 {
+			prompt.WriteString("\n")
+		}
+		prompt.WriteString("帖子内容：" + content)
+	}
+	// 如果标题和内容都为空但有图片，说明是纯图片帖子
+	if prompt.Len() == 0 && len(normalizeImageURLs(images)) > 0 {
+		prompt.WriteString("帖子内容：[仅包含图片]")
+	}
+	if prompt.Len() == 0 {
+		prompt.WriteString("帖子内容：[空]")
+	}
+
+	return c.moderateContentInBatches(ctx, prompt.String(), images)
 }
 
 func (c *clientImpl) ModerateComment(ctx context.Context, content string, images []string) (bool, string, error) {
 	if !c.IsEnabled() {
 		return true, "", nil
 	}
-	return c.moderateContentInBatches(ctx, fmt.Sprintf("评论内容：%s", content), images)
+
+	// 构建评论提示词，处理纯图片评论的情况
+	var prompt strings.Builder
+	if strings.TrimSpace(content) != "" {
+		prompt.WriteString("评论内容：" + content)
+	} else if len(normalizeImageURLs(images)) > 0 {
+		// 评论只有图片没有文字，说明是纯图片评论
+		prompt.WriteString("评论内容：[仅包含图片]")
+	} else {
+		prompt.WriteString("评论内容：[空]")
+	}
+
+	return c.moderateContentInBatches(ctx, prompt.String(), images)
 }
 
 func (c *clientImpl) moderateContentInBatches(ctx context.Context, contentPrompt string, images []string) (bool, string, error) {