status of `scikit-optimize`

[fkiraly]

May I kindly ask what exactly is the status of scikit-optimize?
It looks confusing:

• GitHub archived and seemingly defunct
• this repository and you, @holgern, committed to maintaining

If you are continuing maintenance, why not in the original repo?

I'd also be available to help, although I'd like to understand the current situation better.

[holgern]

Hi @fkiraly ,
I was a maintainer for scikit-optimize in 2021 and during this time I was added to the pypi-repository as maintainer.

Then my work was moving away from classification and others had taken my maintenance role, until they stopped being active.

Then the owner of scikit-optimize decided to archive scikit-optimize, as no one was doing something. This was the moment when I was again aware of the scikit-optimize project.

I was willing to do the maintaince again and I asked the project owner to allow me to do this. They did not response.

As I'm still maintainer of the pypi-repo. I decided to move the code to this new location and I worked through all open PR and integrated them into the codebase.

I'm would be more than happy to contribute to the original git location, but I do not have access.

Hope this helps to understand the situation better.

[fkiraly]

Ah, I see. That makes sense, summarizing so I just understand:

1. you were a maintainer back some years ago and drifted off the project but never officially left
2. other maintainers decided to close down the project but did not contact all other maintainers (2023-2024)
3. they archived it in Feb 2024 on their own initiative, and that's when you heard of it again being closed down
4. you thought, you'd be happy to take a more active role to prevent the project from closing and keep it maintained
5. you contacted the maintainers about this, and also proved that you were willing and able to continue maintaining the project by handling the PR and variou supdates (March 2024) but they did not respond since then

Is the above a correct summary of the project's current history?

If so, I could offer to help out as well, at sktime we are maintaining multiple sklearn-like packages under a streamlined maintenance framework, so one more would be only a fractional addition to the workload, compared to maintaining packages separately. After a bit of upfront refactor of release pipelines and package structure (non-breaking), things should be easier.

What do you think?

[fkiraly]

?

[holgern]

Yes, that is a correct summary.

It seems that adding scikit-optimize to sktime would be a good thing. So I generally agree with integrating scikit-optimize into the sktime organization (assuming I would still be a maintainer and would be able to release new versions of scikit-optimize).