I am a long-term user of the mambaforge/miniforge distributions for various tasks in experimental physical science.
In the past, I always installed miniforge via the github release page for my personal machines.
However, I was recently tasked to create a workflow for less python/programming-savvy users to install a predefined set of packages (numpy, scipy, pandas, jupyterlab + dependencies) and environments (of course with their own packages).
I want to avoid a lengthy tutorial and one-by-one installing of packages and or importing of environment YAMLs, as I would usually do for my own installations.
My basic idea was to build/create a custom exe-Installer based off the miniforge installer, but with the list of packages and environments included. I am also open to completely different approaches (also with pip/vanilla python).
The packages can be included in an installer and/or downloaded at runtime based on a list of requirements. Ideally, the installer should be easily adaptable by me or another more experienced user, while the userbase would only be exposed to a simple installer exe/script.
I found this but am unsure where to start and how to proceed or wether this is a good starting point.
As this seems like a problem that was solved many times, is there a tutorial/doc/wiki outlining the necessary steps?
You can edit https://github.com/conda-forge/miniforge/blob/main/Miniforge3/construct.yaml and create your own installer that is built and released in github releases.