Data File Testing Code Optimization

[boerzia]

The current code for testing data and labels files has issues. Lacks proper exception handling, has duplicate code for reading TOML and JSON files, suboptimal unittest assertions, and limited code readability/documentation. Suggest refactoring for robustness and maintainability.
当前用于测试数据和标签文件的代码存在问题。缺乏适当的异常处理，读取 TOML 和 JSON 文件的代码重复，代码可读性/文档较差。因此为了健壮性和可维护性，建议进行重构。

[Lemeri123]

Assign this to me @boerzia