Fusion ring refactoring

[willieab]

📚 Description

Re-factoring the FusionRing code to expand its functionality from dealing exclusively with Verlinde algebras. Unifies VerlindeAlgebra and FusionDouble class.

📝 Checklist

• The title is concise, informative, and self-explanatory.
• The description explains in detail what this PR is about.
• I have linked a relevant issue or discussion.
• I have created tests covering the changes.
• [] I have updated the documentation accordingly.

⌛ Dependencies

• The Fusion Ring of the Drinfeld Double of a Finite Group #35387

[willieab]

@tscrim @dwbump

[willieab]

@dimpase @mkoeppe Could you please add me to the Sage triage team? I've contributed to Sage before, mostly to the fusion_rings folder: #30423 #31284 #30257 #29615, amongst others. @tscrim said I should contact you. Many thanks in advance!

[mkoeppe]

I've sent the invite to org and team

[willieab]

@mkoeppe Thanks!

[dwbump]

@mkoeppe wrote:

I've sent the invite to org and team

I still can't add Willie as a reviewer on #35387

[tscrim]

Some quick comments on this for @willieab

1. I pushed some changes to The Fusion Ring of the Drinfeld Double of a Finite Group #35387 that I would like to see pulled here and analogously made to the generic doc.
2. We should do a deprecated lazy import of FusionRing in fusion_ring.py (for an example of this, see sage/rings/all.py).
3. Don't make _repr_ an @abstract_method; it isn't (and shouldn't be) required.
4. @abstract_method with a docstring doesn't need a pass (IMO, that actually makes it slightly more confusing).
5. Remove _element_constructor; that is something for the internal's of Sage Parents.
6. For VerlindeAlgebra, it is a bit more to code, but it seems natural to me to have it take a ``WeylCharacterRingring as input and have aclasscall_private` that parses everything to a WCR object to pass along to the `init` (with the appropriate additional options).

[dwbump]

Also the tests show a problem with set_order. I think the test failures in verlinde_algebra.py need to be debugged.

[willieab]

Thanks @dwbump @tscrim! I'll work on these changes. @tscrim Why do we need the deprecated lazy_import? We are neither changing the location nor the functionality of FusionRing.

[dwbump]

Thanks @dwbump @tscrim! I'll work on these changes. @tscrim Why do we need the deprecated lazy_import? We are neither changing the location nor the functionality of FusionRing.

I started a PR that would import Travis' changes to #35387 and can probably finish it today or early tomorrow. I think doctest errors need investigating.

[tscrim]

Why do we need the deprecated lazy_import? We are neither changing the location nor the functionality of FusionRing.

@willieab Because there is no longer a file fusion_ring.py. So if someone was directly importing from that file (say, in their own .py files/library), then their code will break. So it needs a lazy import redirect to the new location. (This might also be needed for old pickles too, but that is less important I think.)

[tscrim]

Also, another general docstring comment. Remember that docstrings should have a short one-sentence description first, then followed by the more detailed paragraph explanation (if necessary).

[dwbump]

The PR #35569 brings in some of the edits to #35387, namely those to the main docstring in fusion_double.py.

[tscrim]

@dwbump It would be better for reviewer changes to have PRs done to the author’s individual fork rather than in the general Sage repo. Otherwise this creates separate threads/PRs that need further manual intervention and pollutes the search feature here.

[dwbump]

@dwbump It would be better for reviewer changes to have PRs done to the author’s individual fork rather than in the general Sage repo. Otherwise this creates separate threads/PRs that need further manual intervention and pollutes the search feature here.

The only github repository I can push to is my own. Willie` can merge this PR into his. Eventually we can try to merge Willie's branch into sage. at which point there will be a (minor) problem, because the two branches are forks and the merge will not be a clean one. However that won't be too hard to take care of and I don't see a better way to proceed.

Am I missing something?

[mkoeppe]

From your repository you can open a PR against any fork of the Sage repo, including Willie's, not just the main repo.

[tscrim]

@dwbump It is as @mkoeppe said. We did this before during our Zoom call (and it is what I did with your branch). It also makes it easier for @willieab to merge it into his branch (and mark it as having done so).