I’ve been hosting the MoleculeNet data on my personal AWS S3 for several years. Charges have been low for a long time (around $10/month) but recently have spiked to ~$60/month. I am trying to understand why. My theory is that perhaps a DeepChem user or company put MoleculeNet onto their CI build? There’s about 600 GB of data being downloaded per month which suggests a heavy volume of automated repeat downloads possibly.
If you are using MoleculeNet in your CI, would you please chime in on this thread so I understand your use case better? If charges continue rising, we may need to throttle downloads somehow to prevent automated repeat downloads