Robyn

#Conda is still driving me crazy.🙈
The confusion is actually perfect...
Conda is actually a package manager.
But then there is also #Anaconda and #Miniconda as package managers.

Both Miniconda and Anaconda install conda.
Then there is #mamba A reimplementation of Conda in C++. There is also #Micromamba...but no #Minimamba...

#bioinformatics

🥴 🐍

Sep 21, 2024 at 1:38pmWeb
Show threaddwardoricSep 21, 2024
@necrosis IMHO the python people messed dependency stuff up even more than the js folks. 🤷🏽‍♀️
Show threadBenjamin S-B Sep 21, 2024
@necrosis Nowadays In my lab, I tell every new student this hard-earned truth: Conda is never the solution, it's always the problem...
Show threadMahesh Binzer-PanchalSep 21, 2024
@necrosis You missed the new kid on the block. https://pixi.sh/latest/
Pixi by prefix.dev

Pixi Documentation — Next-gen package manager for reproducible development setups

Show threadPeter CockSep 21, 2024
@mbp @necrosis or is it uv? https://github.com/astral-sh/uv (I jest, but serious there are as usual too many ways to do Python package management…)
GitHub - astral-sh/uv: An extremely fast Python package and project manager, written in Rust.

An extremely fast Python package and project manager, written in Rust. - astral-sh/uv

GitHub
Show threadMahesh Binzer-PanchalSep 21, 2024
@pjacock @necrosis Pixi uses uv for the Python package part. I actually really like Pixi because it's much better for reproducibility. Conda/Mamba environments will not be created with the same libraries, generally, 6 months later. Pixi includes a lock.file so it's all tracked. You can make a lock file with conda too but people don't generate those by default.
Show threadPeter CockSep 21, 2024
@mbp @necrosis I did read up a little on uv, but less so into Pixi. Perhaps I should, conda has its flaws and frustrations, but is better than everything I’d used previously.
Show threadMahesh Binzer-PanchalSep 21, 2024
@pjacock @necrosis I have to admit I mostly use Conda/Pixi for making an environment for Nextflow, or Quarto, or developing an exploratory chunk of code. Then I'll switch to making a container for anything exploratory if a container doesn't already exist (and connecting it in a Nextflow workflow). Between Docker, Singularity, and Podman, people can then reuse the analysis reproducibly. Nextflow/Quarto stay in Conda/Pixi environment to freely use the users choice of container and/or job submitter
Show threadMahesh Binzer-PanchalSep 21, 2024
@pjacock @necrosis Biocontainers has been great at making containers for individual tools, but making multi-tool packages has been not so nice. Seqera containers filled a niche here and really made making multi-tool containers easy to make.