Suzuki-related strongly regular graphs

[ncohen]

This ticket adds several Suzuki-related groups and graphs.

[ncohen]

New commits:

​006fa91	trac #19184: HigmanSims design
​a380d6d	trac #19184: SRG on 105 vertices for L(3,4)
​6c56dda	trac #19216: Suzuki-related strongly regular graphs

[dimpase]

this seems to be the place where using GAP instead of libGAP makes things much slower...

[dimpase]

also, if we care about being fast, the faster way to build the graph on 416 vertices is from its automorphism group, which is available as Atlas group:

​http://brauer.maths.qmul.ac.uk/Atlas/v3/exc/G24/

[git]

Branch pushed to git repo; I updated commit sha1. Last 10 new commits:

​848a972	trac #19184: the SRG at last !
​fbe7bb9	trac #19184: Broken doctest
​2ceb7ee	Merge branch 'public/19184' of git://trac.sagemath.org/sage into hs
​a0eabd4	cite [BvL84]
​c07f2bc	no need to take complement
​994aa33	SRG on 105 vertices for L(3,4)
​8d615c2	less crude code in SRG_176_49_12_14
​717e356	Merge branch 'public/19184' of git://trac.sagemath.org/sage into hs
​e8c2668	Merge remote-tracking branch 'trac/public/19216' into suz
​9fac029	3.5-fold speedup of SRG_560 using libgap

[git]

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

​4d8c733

optimised constuction of SRG_416_100_36_20 and added SRG_144_39_6_12

[ncohen]

Merging this branch with the NO/NU branch was unnecessary (as was probably proved by the fact that there was no conflict) and results in a very unclean history, that I find very hard to read if I have to review the commits that you added.

[ncohen]

I pushed back my original branch, on top of which I cherry-picked (git terminology) what I believe to be the only two commits that you added that are related to this branch. I also updated their message so that it contains the ticket number (please do this in the future).

I did not overwrite public/19216 but pushed everything to u/ncohen/19216, you can thus find your old branch if you need it. In the future, however, please ask me whenever you want to see one of my tickets merged with something else, as by doing it yourself you give me much more work.

I agree with your two commits, and I thank you for the improvements and addition. You also told me often to use libgap, and I often forget to. Sorry for that.

Please check that I did not miss any of the commits you wanted to see here, and set the ticket to positive_review if you agree with it. Please fill your name in the 'reviewer' field, or Volker will invalidate the ticket.

Nathann

---

New commits:

​b049232	trac #19216: 3.5-fold speedup of SRG_560 using libgap
​262998b	trac #19216: optimised constuction of SRG_416_100_36_20 and added SRG_144_39_6_12

[dimpase]

Replying to ncohen:

Merging this branch with the NO/NU branch was unnecessary (as was probably proved by the fact that there was no conflict) and results in a very unclean history, that I find very hard to read if I have to review the commits that you added.

Apologies. Probably my develop branch is messed up, and messing up anything on top of it. Anyhow, I did not overwrite anything this time.

[dimpase]

by the way, how about reorganising the source of strongly_regular_db: namely, moving all the graph constructions out of there, say, to generators/srg.py(x) ?

[ncohen]

Hmmmmm... Everything in "generators/" is usually exposed in graphs.<tab>, and in this case I am not sure that we want that.

To me having this stuff in this file means that it is not very important of its own. What about only reorganizing the file by moving all the graph constructors at the end of it? It does not bother me much personally.

Nathann

[ncohen]

Thank you for this review, by the way.

Nathann