So far, the BCSS dataset needs to be downloaded separately to run the demo notebook.
A detailed guide is available in docs/datasets/BCSS.md, but it would be great to automate the process in a script (starting with "git clone" as a subprocess) to make the quick start smoother.