Hi,
Thank you for releasing this tool/dataset! The splitting algorithm presented in the paper looks quite rigorous and well-designed. However, as far as I understand, the current implementation seems to only work for the Plinder dataset. Given that the same algorithm can also be very useful for other use cases, I was wondering if it would be possible to use the existing codebase to process other datasets. If not, are there any modifications you would recommend to make this feasible?