Implement a catalog for algebras

[tscrim]

Like groups, crystals, etc.

[tscrim]

New commits:

​e1373c9	Initial catalog of algebras.
​936590d	Merge branch 'develop' into public/algebras/catalog-16219
​fe1850c	Merge branch 'develop' into public/algebras/catalog-16219
​c951590	Added some more algebras.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​61d6bd1

Added quaternion algebra to doc and some cleanup of the file.

[nthiery]

I have just been through the diff, and overall it looks good.

It seems a bit redundant to both import the algebras and link to their doc. We should find a better way to handle this automatically at some point. But that's for a later ticket.

The documentation improvements elsewhere in the algebra module is nice. Yet's it's orthogonal to the main purpose of the ticket. Next time, please make it a separate ticket.

Why not lazy importing everything?

Up to this last point, I believe it's good to go (assuming all tests pass).

[nthiery]

Ah, also: at some point, we will want to point to all the Hopf algebras from combinat. But it's probably easier to do this once #16256 is done.

[tscrim]

Replying to nthiery:

I have just been through the diff, and overall it looks good.

Thanks for taking a look at this.

It seems a bit redundant to both import the algebras and link to their doc. We should find a better way to handle this automatically at some point. But that's for a later ticket.

Concur.

The documentation improvements elsewhere in the algebra module is nice. Yet's it's orthogonal to the main purpose of the ticket. Next time, please make it a separate ticket.

When I added the quaternion_algebra.py to the doc, it causes docbuild failures. The cleanup was just as I was going through reading the file in order for it to build.

Why not lazy importing everything?

I lazy imported things that were lazily imported into the global namespace, so it wouldn't trigger an import when doing tab completion on algebras. (as what happens for groups).

There's a more trickier question of what algebras could we remove from the global namespace (it's not nearly as clear as it was for crystals to me), but I think that can be done on a followup.

Up to this last point, I believe it's good to go (assuming all tests pass).

Was this suppose to be set as needs_info? I don't quite understand why this was set to needs_work.

Ah, also: at some point, we will want to point to all the Hopf algebras from combinat. But it's probably easier to do this once #16256 is done.

Yep, and it should probably be a subcatalog called combinatorial_hopf (or maybe just .hopf and include all Hopf algebras). Anyways, stuff for later and I'll take a look at #16256 soon.

[nthiery]

Replying to tscrim:

Why not lazy importing everything?

I lazy imported things that were lazily imported into the global namespace, so it wouldn't trigger an import when doing tab completion on algebras. (as what happens for groups). Was this suppose to be set as needs_info? I don't quite understand why this was set to needs_work.

I was thinking of lazy importing everything here right away. Maybe using a flag "at_startup" to make the point that those should really be lazy imported elsewhere as well. What do you think?

There's a more trickier question of what algebras could we remove from the global namespace (it's not nearly as clear as it was for crystals to me), but I think that can be done on a followup.

+1

Yep, and it should probably be a subcatalog called combinatorial_hopf (or maybe just .hopf and include all Hopf algebras). Anyways, stuff for later and I'll take a look at #16256 soon.

And at some point we will want to automatize the building of all those catalogs. E.g. following the prototype Florent wrote during the Sage days in Edinburgh that instruments TestSuite to build a database of all parents with their category, from which we can get all parents in a category by reverse lookup. I just created #17219 for this.

Cheers,

Nicolas

[tscrim]

Replying to nthiery:

I was thinking of lazy importing everything here right away. Maybe using a flag "at_startup" to make the point that those should really be lazy imported elsewhere as well. What do you think?

So do you mean you want everything to be lazily imported or do you want to lazy import the catalog? If you want the latter, this causes problems because then the imports are triggered on tab completion (you can see this occur with the groups.<tab>).

[nthiery]

Replying to tscrim:

So do you mean you want everything to be lazily imported or do you want to lazy import the catalog? If you want the latter, this causes problems because then the imports are triggered on tab completion (you can see this occur with the groups.<tab>).

The former: lazy importing everything that is in the catalog.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​850bebb

Merge branch 'public/algebras/catalog-16219' of trac.sagemath.org:sage into public/algebras/catalog-16219

[tscrim]

IMO, we should not lazy import things which are (currently) not lazily imported into the global namespace as it is more coherent. I think we should have a (separate) discussion about which algebras should be lazily imported and handle it there.

[nthiery]

I am not convinced, but I don't have a strong opinion either, and it's minor. I won't hold up this ticket just for this. Please proceed if you are convinced this is the way to go.

[tscrim]

I believe it is, and I'm interpreting that as a positive review. I've created #17271 for the lazy importing. Thanks Nicolas.

[git]

Branch pushed to git repo; I updated commit sha1 and set ticket back to needs_review. Last 10 new commits:

​a4d5b24	Cached methods should return immutable matrices
​aae8f5b	Merge branch 'u/jhpalmieri/DGA_new' of trac.sagemath.org:sage into public/algebras/cdga-16508
​7b6f118	Fixed doctests.
​fb16080	Merge branch 'develop' into public/algebras/cdga-16508
​b8328e6	Second pass of reworking things around by adding functionality.
​d732dbe	Commutative DGAs: rename some classes
​497d9d6	Merge branch 'public/algebras/cdga-16508' of trac.sagemath.org:sage into public/algebras/cdga-16508
​3900c66	Renamed CDGAlgebra -> cdg_algebra.
​73ee329	Merge branch 'public/algebras/catalog-16219' of trac.sagemath.org:sage into public/algebras/cdga-16508
​1a8a288	Added GradedCommutative to the algebras catalog.

[tscrim]

I've added #16508 to the catalog and moved DifferentialWeyl to it's correct spot in the catalog (it changed its name from Weyl). Very quick check needed.

[nthiery]

Check done. Ok for me!

Just for the record: the comment for 73ee329 looks backward ...

[tscrim]

Replying to nthiery:

Just for the record: the comment for 73ee329 looks backward ...

I was going to push this to #16508 until I noticed the out-of-order DifferentialWeyl, so in a sense it is backwards. ^^;;