robinhood VS hts

https://github.com/cea-hpc/robinhood maybe this could work for you?

Assuming this was a "scratch" HPC filesystem, as I'd guess, "scratch" is used advisedly -- users should be prepared to lose anything on it, not that it should happen with finger trouble. However, if I understand correctly from the comments, I'm surprised at the tools, and that the vendor was managing the filesystem. I'd expect to use https://github.com/cea-hpc/robinhood/wiki with Lustre, though I thought I'd seen a Cray presentation about tools of their own.

sadly a lot of statistics research is done with R and is unavailable with Python, hopefully this kind of work will also motivate new libraries for Python. I am particularly interested in hierarchical forecasting. Are there Python alternatives to the hts library?(https://github.com/earowang/hts)

Additionally, you could use one of the more complex methods from the aforementioned hts package. This will allow you to make forecasts on all levels of the hierarchy, and use the bootstrapped errors to make adjustments to all forecasts in the hierarchy using a constrained least-squares approach, in order to make all forecasts sum-consistent (make the aggregates of the forecasts equal the forecasts of the aggregates). This allows you to model cannibalisation effects between different products, for example. However for this to work, you'd need quite good models, as the bootstrapped errors are taken as the 'wiggle room' for the adjustments, which means that if you have a badly fitting model, the adjustments might be quite large and no longer make sense (eg. be negative for a sales forecast).