Add SatImp and WGA methods

[Puning97]

What does this PR do?

This PR introduces the integration of the SatImp (ICML2025) and WGA (ICLR2025) methods.

The original papers are:

SatImp: arxiv, Github
WGA: arxiv, Github

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Have you gone through the contributions guide?
• Are your changes documented? Read documentation guidelines here.

[molereddy]

Can you link to the papers you are implementing?

[Puning97]

Can you link to the papers you are implementing?

Of course! In fact, I’ve already added the relevant links in the README.md and in docs/links.

| SatImp | Paper📄, Code 🐙 |
| WGA (G-effect) | Paper📄, Code 🐙 |

[molereddy]

Sorry I meant in the PR description. It's hard to navigate to these.

[Puning97]

Sorry I meant in the PR description. It's hard to navigate to these.

Sorry for the misunderstanding. I guess you mean that I need to modify the top PR description, adding these links so that others can easily find them? Am I correct?

Now the PR is:
We add the SatImp (ICML2025) method for open-unlearning framework. 5 files are changed.

I need to modify to:

What does this PR do?

This PR introduces the integration of the SatImp and WGA methods.
The original papers are:

SatImp: arxiv, Github
WGA: arxiv, Github

Have you gone through the contributions guide? Yes

Are your changes documented? Read documentation guidelines here. Yes

[molereddy]

It seems the ruff check tests we have (for code formatting) are failing. Please follow the instructions in docs/contributing.md and re-format the code.
Also, just to re-confirm, since we will not be testing before merging,

1. this exact version of the codebase
2. in an enviroment with the package versions described in the readme
3. is working without any errors

Thank you a lot for contributing!

[molereddy]

Aside from code formatting, LGTM.

[Puning97]

Aside from code formatting, LGTM.

Sorry for the multiple attempts! The reason is the difference between ruff==0.12.0 and ruff==0.6.6, which makes the different format in utils.py!

[Dornavineeth]

@Puning97, thank you for adding your methods with clean implementations. Just added minor comments, which shouldn't take much time but can help with implementation consistency throughout the repo.

[Dornavineeth]

@Puning97 Congratulations on your two new method additions to Open-Unlearning 🎉🎉🎉! We look forward to seeing more of your advanced work and contributions to OU.

[Puning97]

@Puning97 Congratulations on your two new method additions to Open-Unlearning 🎉🎉🎉! We look forward to seeing more of your advanced work and contributions to OU.

Thank you for your valuable contributions too. I look forward to continued collaboration on advanced research in the future.