Updated Dataset

[GriffinYoung]

Hi,
The original PLINDER dataset was a really rigorous way of assessing leakage and constructing benchmarks. Due to the inevitable march of time it is now out of date and cannot be used to assess eg the latest Protenix model which has been trained up to mid-2025. Are there plans to release an updated version of the dataset? Or in lieu of that can someone direct me as to how I would use the code in this repo to generate an updated version myself?

[Ninjani]

Hi, thanks for the comment! We're currently in the middle of a major refactor and have a tentative timeline of updating the dataset in July. For something more immediate, there's some documentation on running the entire pipeline - just note that it requires significant storage and compute resources, which will also be reduced as part of the refactor.