Posets in OSCAR #3610
Posets in OSCAR #3610
Conversation
| parent::Poset | ||
| end | ||
|
|
||
| function index(x::PosetElem) |
There was a problem hiding this comment.
The already existing index methods do something completely different. Does somebody know of other occurrences in the Oscar world of something like this and how we call it there?
There was a problem hiding this comment.
its mostly called data to access the defining datum...
|
|
||
| # add k to the stack and continue from k | ||
| push!(q, k) | ||
| @goto outer |
There was a problem hiding this comment.
I think you can just replace this goto by a break and then get rid of the @label as well.
There was a problem hiding this comment.
If I replace this with break, I will have to introduce a variable and an if statement after for loop. In the end I felt that goto is easier and the flow is clear too.
Co-authored-by: Lars Göttgens <lars.goettgens@gmail.com>
|
This is not an initial implementation of posets in OSCAR, there is one in |
|
I didn't know there was an existing implementation, since it isn't exported; thanks for letting me know. The goals pursued in GaloisGrp are different from my implementation. For my application I am mostly interested in maximal chains, so I need covering relations and the edges have bonds/weights. Currently the constructor uses (and only needs) the adjacency matrix of the Hasse diagram. @lgoettgens suggested to bring this up in the next OSCAR meeting, so other people (who need posets) can voice their needs, and then plan from there. |
|
There is a new polymake release in the making. One of its main features is substantial new functionality concerning posets. For instance, a fairly fast implementation of all maximal chains etc. We need to align our efforts. |
|
@micjoswig is there already some kind of preliminary documentation on this somewhere, or a person to talk to about it, so that @felix-roehrich can coordinate with them? |
|
Let's discuss posets on Friday (26 April) in the regular zoom. |
|
We will see the release of poylmake 4.12 soon (say, by May 3). That features some additions to (and renaming of) polymake's posets (which are still called lattice in 4.11 and before). It will trickle down through Polymake.jl to OSCAR. In case you want to see this in action, just as an example, here is the polymake 4.11 code for computing an arbitrary order polytope. One nontrivial ingredient is how the maximal anti-chains are computed; this goes through a max clique computation in the comparability graph. See also Section 5 of our preprint. |
|
Polymake 4.12 has been released now; but integration into OSCAR is still work in progress, see PR #3819. |
This is an initial implementation for posets in OSCAR, feedback is welcome.