Closed costrouc closed 6 years ago
Did you succeed convincing your friends? I'm currently packaging mayavi. How would you prioritise other packages?
I would prioritize the packages that you need. Nix already covers a huge amount of python packages in python-packages.nix
2,356 as of today (and more if you include applications). Personally I have found that both packages managers can coexist. I still think that nix needs more momentum for true widespread adoption. I still use conda heavily but every time I use a package I check if it is in nix and add it if not.
Got genson
ready to be merged here: https://github.com/NixOS/nixpkgs/pull/237775
I want to convince my friends that nix is the future way to do reproducible data science. Anaconda is the current way that most people develop and run their workloads for data science with python and R. This list will never be 100% complete since some are windows specific packages and are derivations of the same package. But for linux and OSX users, nix can be the tool. I want to be able to say that nix would be an easy switch. I will spend my time trying to complete as much of this list handling gpu support will be tricky since I do not have much experience in this area. I welcome others to help as well! This list was constructed from https://anaconda.org/anaconda/repo . There are currently 1530 packages.
I have not checked beyond the last check mark