Put preprocessed data on an appropriate server

[albangossard]

Data are now preprocessed in Python using https://github.com/ODINN-SciML/Gungnir/
This means that everything is prepared for Julia and Sleipnir just has to read the data. This was implemented in a89c40e.

For the moment the data are stored in a google drive folder https://docs.google.com/uc?export=download&id=1d070a_YqN5aPAONpnzL9hfInv1DA8z3p but we need to put them on a dedicated server.

[JordiBolibar]

We should start having a proper discussion on where to store the preprocessed files.

I would like to have @fmaussion's opinion on this. Do you think we could store the pre-processed gdirs in the Bremen server? The way this works right now, is that we have Gungnir, a Python package acting as an OGGM wrapper, handle all the generation of gdirs and initial files. And then we read directly the files stored in a server from ODINN.jl.

Since with OGGM you are already doing this, it could be quite a straightforward solution. The only thing would be to know how scalable that would be. The goal would be to progressively add files for different regions. Since for now we are not targeting anything global, it would mostly be focused around Central Europe, Scandinavia and the Andes. If this was not a possibility, we could find a way to store it in a server in Grenoble.