Environment not working on platform osx-arm64

[darioflute]

What is the problem?

Currently, the instruction to create the AstronomicalData environment does not work with the platform osx-arm64 since there is no compiled version of "gala" for this distribution.
The most recent version on conda-forge for mac-osx is the: osx-64 1.9.1

I found two possible ways out:
a) Create the environment by forcing the osx-64 platform files to be installed:
CONDA_SUBDIR=osx-64 conda env create -f environment.yml
b) Change the environment file by adding at the end instead of the line with gala, the two lines:

• pip
• pip:
  • gala

In order to use pip to install the latest version of gala. Note that in this case a more recent version is available (1.10.1).
In this second case, at least if running the latest version of the operative system, one has to first accept the license agreement of Xcode (which can be done launching the Xcode.app).

Location of problem (optional)

https://datacarpentry.github.io/astronomy-python/ Summary and Setup

[darioflute]

I noticed that this problem has been already found in the issue #220, but no solution was given.

[ShebMichel]

@darioflute can you share the few line of installations steps you've taken so that I can try to test it on my mac, and if it works then I can add it to the course. Cheers!

[abostroem]

I think we actually came up with a plan to avoid having to install gala all together. I'll have time to look into this later this week. In the meantime, it would be great to hear any work around @darioflute found.

[darioflute]

The environment can be checked with the command:

uname -m

If the result is:

arm64

there are two ways to proceed:

1st method

In this case we install the available conda packages created for osx-64.

Use the environment.yml found in the "Summary and Setup" of https://datacarpentry.github.io/astronomy-python/
and create the environment with the command:

CONDA_SUBDIR=osx-64 conda env create -f environment.yml

2nd method

In this case, we install gala directly from pip. The advantage is that we can use a more recent version not available yet as anaconda package.

In this case, the environment file (environment.yml) should be modified to:

====================
name: AstronomicalData
channels:

• conda-forge
• astropy
  dependencies:
• python>=3.6
• jupyter
• numpy
• matplotlib
• pandas
• pytables
• h5py
• scipy
• astropy
• astroquery
• pip
• pip:
  • gala
    ====================

and the normal command can be used:

conda env create -f environment.yml

Let me know if these instructions are clear and thanks for your help,
Dario