Our great sponsors
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
I believe a tool that can hold all dependencies (such as Nix) is a true solution for this kind of packages.
pytorch and tensorflow are nightmare packages with tons of non-python implicit requirements: https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/python-modules/torch/default.nix
It's still conda, but mamba-forge with conda-forge as the source has solved almost all my problems. I even wrote a quick "how-to": https://github.com/Paradoxdruid/mamba-how-to
Poetry won’t install dependencies that are already present in the env. However, one issue is that build artifacts are typically included in the version for packages installed from conda-forge, which poetry doesn’t recognise as the same version. The workaround is to run ‘find $CONDA_PREFIX -name “direct_url.json” -delete’. Note that this corrupts the conda env so you might not be able to use conda to make changes to the environment anymore, so definitely make sure you don’t run this while base is activated!
As detailed at the top of its wiki page, it was spun out and now has its own governance model.
that said this export plugin for requirements.txt files made life a lot easier