switch from docopt to docopt-ng

[n-gao]

Since docopt is deprecated and hasn't been updated in 10 years, its regex is throwing errors in Python >3.12. This PR changes the dependency from docopt to docopt-ng, which is a maintained fork of docopt.

I also changed some unittests to work again. However, the unittests do not properly support numpy 2.0. To get them to run successfully, one has to install numpy<2.0.

In general, sacred could use some cleaning as it heavily relies on deprecated packages like pkgutil and pkg_resources. Another example is datetime.datetime.utcnow() which is now datetime.datetime.now(datetime.UTC).

[thequilo]

Hi @n-gao! I wasn't aware of this, it's a good idea to switch to a maintained version of docopt. And you also fixed the AWS mock.

I'm just wondering why the tests didn't get triggered. I don't like to merge this without making sure that it doesn't break the tests.

[thequilo]

In general, sacred could use some cleaning as it heavily relies on deprecated packages like pkgutil and pkg_resources. Another example is datetime.datetime.utcnow() which is now datetime.datetime.now(datetime.UTC).

I agree. There is a whole lot of old baggage that sacred carries around. There's more than just deprecated packages, but fixing all that would be a lot of work. I currently don't have much time to work on these things, so I would be grateful if you would provide PRs for this.

To get them to run successfully, one has to install numpy<2.0.

I'm also aware of this, and I think I know what has to be changed for it to work with numpy 2.0. I can prepare a PR for this.

[n-gao]

So, the CI is not triggering here because it would run on your budget and with your secrets, you would have to first include pull_request as trigger to the CI. However, I enabled the tests on my fork:
https://github.com/n-gao/sacred/actions/runs/10521691384

The tests that are failing are due to numpy 2.0, with nump<2 they work.

I'm also aware of this, and I think I know what has to be changed for it to work with numpy 2.0. I can prepare a PR for this.

That would be amazing! This seems to be an actual breaking change and not just a deprecation warning.

[thequilo]

So, the CI is not triggering here because it would run on your budget and with your secrets, you would have to first include pull_request as trigger to the CI.

Hmm, then I understood the triggers wrong. I'll add that as well.

This seems to be an actual breaking change and not just a deprecation warning.

Yes, numpy2 has a few breaking changes. It also broke a lot of our internal code base. But the impact on sacred is not super large.

[n-gao]

Tests are now passing! :)

[n-gao]

Can we merge this and make a new release?

[thequilo]

Yes, looks good to me!