conda
No Change
assess
conda. (Anaconda/Miniconda) manages Python and non-Python binary dependencies in isolated environments.
Summary
What it is: environment.yml, channel-based packages, conda-forge ecosystem, cross-platform binaries.
When to use: ML/data notebooks needing CUDA, BLAS, or scientific stacks that pip alone struggles with.
When to skip: Standard web services on Python with wheels available on PyPI.
Key features: conda create, exportable env files, mamba faster solver (community).
Details
| Topic | Notes |
|---|---|
| Fit | Data science laptops, not typical GitHub Actions service CI |
| Contrast | pyenv + venv for app dev; Nix for fully declarative hermetic shells |
References