Command "mamba" is not found
stargaser opened this issue · 3 comments
At least three users reported that the mamba info
step in the instructions just brought up an error that the mamba command is not found.
The solution is to specify the full path to mamba, e.g. $HOME/bin/mamba info
, or $HOME/bin/mamba init
to initialize the user's shell.
Now that the libmamba solver is included in conda by default, we could switch the instructions back to conda...
I don't know the history of how we got here – do we want that?
My preferred approach is to have users:
- install miniconda
- make an astropy workshop environment with a recent python
- use
pip install -r requirements.txt
to install every dependency
That's been the fastest way to get people up and running in my recent experience. Is there any reason that isn't sufficient for this workshop?
This history of how we got here is that several years ago, users were just told to install Anaconda since it included astropy and Jupyter notebooks. Then for the AAS 235 workshop in January 2020, we switched to a full-blown conda install, IIRC.
I like your preferred approach. It would make everything a lot simpler. Let's do it!