Remove tox support

[Saransh-cpp]

Resolves #1211

As title!

TODO:

• Add unit tests and/or doctests in docstrings
• Add docstrings and API docs for any new/modified user-facing classes and functions
• New/modified features documented in docs/tutorial.rst
• Changes documented in docs/release.rst
• GitHub Actions have all passed
• Test coverage is 100% (Codecov passes)

[jakirkham]

Thanks Saransh! 🙏

Should we also remove these instances too?

zarr-python/release.txt

Line 1 in daad2fa

tox

zarr-python/requirements_dev_optional.txt

Line 16 in f361631

tox==3.26.0

zarr-python/.gitignore

Line 38 in daad2fa

.tox/

(not sure if the last one might be worth keeping)

[GbotemiB]

Hi @Saransh-cpp and @jakirkham. since support tox has been removed. Can I go ahead to update the docs to build docs with 'make html'

[jakirkham]

Hi @GbotemiB. Thanks for the offer. Think that is already in Saransh's PR, but if you see somewhere else that needs to be updated please let us know and we can fix that here.

[joshmoore]

jakirkham requested a review from joshmoore 13 hours ago

I'm a bit torn, to be honest. The tox.ini is definitely not in an ideal state, especially now with everything being so nice and clean thanks to @Saransh-cpp! However, I do like the idea of having one place that encapsulates the various commands that are needed. I could see preparing a second PR which does the opposite, namely to go all-in on tox and then we could compare the two.

[Saransh-cpp]

May I also suggest looking into nox? We use nox in most of the Scikit-HEP packages (for example, vector's noxfile), and I personally like how one can specify instructions in pure python, instead of having a .ini file.

I'll be happy to work on either of them (tox.ini or noxfile.py)!

[jakirkham]

I'm a bit torn, to be honest. The tox.ini is definitely not in an ideal state, especially now with everything being so nice and clean thanks to @Saransh-cpp! However, I do like the idea of having one place that encapsulates the various commands that are needed. I could see preparing a second PR which does the opposite, namely to go all-in on tox and then we could compare the two.

Given how valuable the conda-incubator/setup-miniconda GHA workflow has been in cutting down maintenance, am not eager to replace that with tox (where we have to maintain even more stuff)

The fact that new contributors are stumbling over tox here just makes it that much clear in my mind that it is time to drop it

[joshmoore]

Definitely agreed on not reproducing the environmental logic within tox. I rarely if ever use tox for running multiple environments. That type of testing I defer to GHA. For me, it's more a matter of having the individual commands that are hard-coded in the workflows available to be run locally. At the moment, we have different sets of commands spread between:

• tox.ini
• the docs/ files
• the .github workflows

If there were a way to reduce that down to 1️⃣, I'd be pretty happy.

[Saransh-cpp]

@joshmoore @jakirkham, could you please have a look at the last commit? I have added nox support to address @joshmoore's comments!

I feel like maintaining a noxfile would be much easier than maintaining a tox.ini file. Plus it is all written in Python! I have also added documentation for the same!

[codecov]

Codecov Report

Merging #1219 (c9a0dac) into main (519f7ef) will not change coverage.
The diff coverage is n/a.

❗ Current head c9a0dac differs from pull request most recent head 422e720. Consider uploading reports for the commit 422e720 to get more accurate results

@@           Coverage Diff           @@
##             main    #1219   +/-   ##
=======================================
  Coverage   99.99%   99.99%           
=======================================
  Files          35       35           
  Lines       14136    14136           
=======================================
  Hits        14135    14135           
  Misses          1        1           

Impacted Files	Coverage Δ
zarr/hierarchy.py	99.78% <0.00%> (-0.01%)	⬇️
zarr/util.py	100.00% <0.00%> (ø)
zarr/storage.py	100.00% <0.00%> (ø)
zarr/creation.py	100.00% <0.00%> (ø)
zarr/convenience.py	100.00% <0.00%> (ø)

[jakirkham]

One could run GitHub Actions locally

https://github.com/nektos/act

[joshmoore]

Thanks for drafting the nox support, @Saransh-cpp. I agree that the Python syntax is a nice improvement, but I'm not sure it warrants the addition of a new tool. Open to other opinions, but I'd err on the side of reverting those changes. I can understand if you'd like to then just get this merged as is, otherwise happy to discuss other options for the unification.

[Saransh-cpp]

Completely understandable! Maybe we can then link or document the usage of https://github.com/nektos/act for Zarr?

[jakirkham]

Documenting how to use act to run Zarr's tests locally sounds reasonable

[joshmoore]

Thanks all. @Saransh-cpp, if you can revert, then let's go ahead and get this merged. Someone can work on documenting akt, and I might look into unifying the docs & the GHA commands.

[Saransh-cpp]

Reverted!

[joshmoore]

Awesome. Thank you!

[jakirkham]

Just noticed that we lost the flake8 configuration here. Adding back in PR ( #1249 ).