Recog An/Sn unknown degree: conder, holt #265

ssiccha · 2021-04-09T15:25:27Z

A PR based on #176.

We are going to implement the extensions by Conder and Derek Holt's
GuessAltSymDegree.

ssiccha · 2021-04-12T12:12:12Z

Rebased the last commit onto the current version of #176.

codecov · 2021-04-12T12:21:57Z

Codecov Report

❌ Patch coverage is 75.32468% with 95 lines in your changes missing coverage. Please review.
✅ Project coverage is 70.48%. Comparing base (c309d8a) to head (16bace0).

Files with missing lines	Patch %	Lines
gap/generic/SnAnUnknownDegree.gi	75.13%	95 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #265      +/-   ##
==========================================
+ Coverage   69.12%   70.48%   +1.35%     
==========================================
  Files          43       43              
  Lines       18321    18637     +316     
==========================================
+ Hits        12665    13136     +471     
+ Misses       5656     5501     -155

Files with missing lines	Coverage Δ
gap/matrix.gi	`92.97% <100.00%> (+0.01%)`	⬆️
gap/perm.gi	`79.76% <100.00%> (+0.07%)`	⬆️
gap/projective.gi	`83.17% <100.00%> (+0.08%)`	⬆️
gap/generic/SnAnUnknownDegree.gi	`85.07% <75.13%> (-5.02%)`	⬇️

... and 12 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

FriedrichRober · 2022-03-28T19:53:53Z

Link to our Hackmd

ssiccha · 2022-03-29T23:35:23Z

We should be almost done. We think there are two TODOs left: clean up the tests and figure out what to do with small classical groups like: Omega(-1,4,3). It has element orders <= 5, thus slips through our heuristics to weed out groups which are not isomorphic to An or Sn. Hence the SnAnUnknownDegree algorithm takes forever until it realizes, that the group in question is not an An or Sn. We think we should be able to alleviate this by simply saying that we're NeverApplicable if the vector space our group acts has less than e.g. 100 elements. And then we'll post a challenge on slack or the GAP forum whether somebody can find a small group, which makes the algorithm run slow.

And get the tests to pass again. :)

And in Make CallMethods directly write into the RecogNode update the docs of CallMethods.

And, @FriedrichRober is writing a short text, how all the parts fit together, that is how our implementation differs from a "vanilla" implementation of the JLNP algorithm according to the paper.

fingolfin · 2022-03-30T13:45:31Z

Awesome, thank you!

fingolfin · 2022-04-06T22:29:37Z

@FriedrichRober do you think you can finish this on your own, now that @ssiccha is gone? Perhaps I can help a bit?

FriedrichRober · 2022-04-07T06:47:22Z

Yes, I think I can finish this on my own. The code is finished (except for handling small classical groups). What is missing now is to finish the documentation (comments) and looking at the runtime of the test suite.

FriedrichRober · 2022-11-07T03:42:42Z

@fingolfin, ich habe nun die Änderungen implementiert, wie wir es besprochen haben. Wir geben früher auf, und speichern uns die Iteratoren im recog node ab, um sie später weiter zu verwenden. Trotzdem dauert die Test-Suite auf meinem Rechner 3-4 Minuten länger als auf dem master. Meiner Meinung nach ist dieser Zustand inakzeptabel, um das einfach so zu mergen :(

Ich muss dann nochmal genau schauen, an welchen Stellen wir zu viel Zeit verschwenden, und ob sich da was machen lässt.

Compute ThreeCycleCandidates in a "lazy" way, as done in Magma Code GetNextThreeCycle.

FriedrichRober · 2025-12-08T15:33:58Z

Getting a lot of errors now of the sort:

# But found:
#I  Increasing tolerance to 0
#I  Increasing tolerance to 1
#I  Increasing tolerance to 2
#I  Increasing tolerance to 3
#I  Increasing tolerance to 4
#I  Increasing tolerance to 5
#I  Increasing tolerance to 6
#I  Increasing tolerance to 7
#I  Increasing tolerance to 8
#I  Increasing tolerance to 9
#I  Increasing tolerance to 10
#I  Giving up!
Error, the recognition described by this recognition node has failed!

Any help is appreciated! I guess we are missing some requirement for the recog tree because the method SnAnUnknownDegree seems to work correctly when testing it manually on the erroring groups. However, instead of

<recognition node SnAnUnknownDegree AlmostSimple Size=5040>

we often get in the (projective) matrix group case

<failed recognition node  AlmostSimple Size=5040 Dim=7 Field=3>

So this is probably a technical error of inserting our method into the database.

FriedrichRober · 2025-12-08T15:36:47Z

Oh, I messed up the rebase. I will fix this, and this should solve the error

FriedrichRober · 2025-12-08T15:45:48Z

The quick test files seem to work as fast as on master on my machine. I am testing now the slow testsuite, since this is where we encoutered problems with Sergio

FriedrichRober · 2025-12-08T16:11:54Z

The slow test suite takes on master almost 5 minutes and on this branch almost 7 minutes

…ithm, and highlight the adjustments made to the original algorithm.

FriedrichRober · 2025-12-09T11:49:41Z

@fingolfin , I think I fixed the performance problem for most non-AnSn examples now. I changed the iterator for the threecyclecandidates to work in batches and give up very quickly. It remembers in its cache where it stopped, and continues from there if called again. However, the numbers for the batches I chose arbitrarily now. If the specific numbers are a good choice or not, I do not know. So the pull request is close to being finished now

FriedrichRober · 2025-12-09T11:50:14Z

We can see which tests fail after everything has run. I think I remember that some small classical groups were recognised now as alternating or symmetric and therefore destroy the output of some tests.

FriedrichRober · 2025-12-09T11:56:29Z

Exactly, for example GL(4,2) is now recognized as A8, which is correct but not what the test file expects.

Edit: this seems to be the only error in the tests

- deactivate guessdegree code - insert SnAnUnknownDegree after deletedpermutationmodule code in projective.gi

FriedrichRober · 2025-12-09T15:57:05Z

@fingolfin, the pull request is ready now. The test suite on my laptop was as quick as on master, however some unrelated errors occures. Two where the stamp is not what was expected in the test file. One where the group was wrongly recognized which would require further investigation. I assume that this is a random error which simply occured from inserting the recognition method into the database.

fingolfin · 2025-12-11T19:09:07Z

Fantastic, @FriedrichRober, thank you so much!

fingolfin · 2025-12-11T21:22:30Z

doc/recog.bib

+    AUTHOR = {Conder, Jonathan},
+     TITLE = {Algorithms for Permutation Groups},
+      YEAR = {2012},
+      NOTE = {TODO},


Could this be filled in by someone? At which university was this thesis written; what kind of thesis was it?

I talked to @aniemeyer and learned that this was a thesis in Auckland, but I didn't find something initially. But now it seems this was indeed a B.Sc. thesis there, and it was written by @jonathan-conder . I have now pinged them, and will also write an email. Perhaps we get permissions to post the thesis somewhere, or perhaps Jonathan is willing to put it on e.g. the arXiv (I am not sure if they accept such things, but it seems to contain original research, and they definitely have at least phd theses?)

Jonathan replied to my email and generously gave permission to share copies of his thesis, and he also made it available at https://jonathan-conder.github.io/afpg.pdf -- this should keep working for the foreseeable future, and I also made sure archive.org has a snapshot. Still, we should probably see if we can also get this onto e.g. Zenodo or so.

Anyway, that gives us enough info to complete this bib record. I'll do so eventually (if nobody beats me to it).

fingolfin

Thank you! I've pushed a bunch of changes (see commit log). I'll try to understand the errors better tomorrow

fingolfin · 2025-12-11T21:24:45Z

gap/generic/SnAnUnknownDegree.gi

+    # This is usually much smaller than RECOG.SnAnUpperBoundForDegree.
+    # The number to compare N with was chosen arbitrarily as a "large" degree.
+    # if N > 20 then
+    #     degreeData := RECOG.GuessSnAnDegree(ri);
+    #     if degreeData = fail then
+    #         cache.N := TemporaryFailure;
+    #         return;
+    #     fi;
+    #     N := Minimum(N, degreeData.degree);
+    #     cache.N := N;
+    # fi;


Is this the only call to RECOG.GuessSnAnDegree?? If so... do we need it at all?

I think we want to keep it just for the sake of Issue #348. But right now the code is not used by SnAnUnknownDegree, and we can surely move it somewhere else.

fingolfin · 2025-12-11T21:26:34Z

gap/projective.gi

+# AddMethod(FindHomDbProjective, FindHomMethodsGeneric.SnAnSmallUnknownDegree, 1075);
+


This method does not seem to exist?

Suggested change

# AddMethod(FindHomDbProjective, FindHomMethodsGeneric.SnAnSmallUnknownDegree, 1075);

Yes, might be an artifact of some old code we wrote. You can safely delete this.

fingolfin · 2025-12-11T22:17:32Z

gap/generic/SnAnUnknownDegree.gi

+        yTo5 := y ^ 5;
+        if not isone(ri)(yTo5) and not isone(ri)(yTo5 * y) then


Computing $y^5 = (y^2)^2\cdot y$ require three multiplications, and then $y^5\cdot y$ a fourth one. So either 3 or 4 multiplications

Wouldn't it then be even better to do something like this, which requires either 1, 2 or 3 multiplications ?

Suggested change

yTo5 := y ^ 5;

if not isone(ri)(yTo5) and not isone(ri)(yTo5 * y) then

y2 := y^2;

if not isone(ri)(y2) then

y3 := y2 * y;

if not isone(ri)(y3) and not isone(ri)(y2 * y3) then

yes, your suggestion is definitely better

fingolfin · 2025-12-11T22:21:26Z

gap/generic/SnAnUnknownDegree.gi

-    constants := RECOG.ThreeCycleCandidatesConstants(1. / 4., N);
-    for i in [1 .. T] do
-        iterator := RECOG.ThreeCycleCandidatesIterator(ri, constants);
+# This is the main method used by RECOG.RecogniseSnAn and RECOG.RecogniseSnAnLazy.


As far as I can tell, RECOG.RecogniseSnAn is not used anymore. Can we just remove it then, and perhaps rename RECOG.RecogniseSnAnLazy to RECOG.RecogniseSnAn ?

So, in theory RECOG.RecogniseSnAn is the algorithm that you can run with an arbitrary epsilon and it guarantees you to recognise SnAn with this error bound, so it is running all iterations as often as computed in the respective paper. However, in practice it is too expensive to run this method for the recog tree, so we wrote the lazy variant that exits the iterations way sooner. Whether we want to keep it or not depends on whether or not it is interesting to have this version of the algorithm to play around with, which I honestly do not have a strong opinion on. We can definitely rename the functions to RECOG.RecogniseSnAnExact or something like this and then drop the lazy from the other :)

I think we should keep the code, but rename as Friedrich suggested.

FriedrichRober · 2025-12-12T19:52:17Z

Thank you! I've pushed a bunch of changes (see commit log). I'll try to understand the errors better tomorrow

The rest of the changes look good to me

ssiccha marked this pull request as draft April 9, 2021 15:26

ssiccha mentioned this pull request Apr 12, 2021

Implement "recognize An/Sn of unknown degree" algorithm by Jambor et al. #176

Merged

ssiccha force-pushed the frss/AnSnRecogUnknownDegree-2 branch from 5ae5367 to d40a988 Compare April 12, 2021 12:11

ssiccha force-pushed the frss/AnSnRecogUnknownDegree-2 branch from d40a988 to dc4a895 Compare April 12, 2021 12:32

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch from dc4a895 to f162f36 Compare June 23, 2021 14:20

ssiccha force-pushed the frss/AnSnRecogUnknownDegree-2 branch from cca25fd to 35a39de Compare June 23, 2021 16:37

ssiccha mentioned this pull request Oct 5, 2021

Wrong size for finite deleted permutation module #296

Open

ssiccha force-pushed the frss/AnSnRecogUnknownDegree-2 branch from 96f9a25 to 705504c Compare March 28, 2022 22:01

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch 3 times, most recently from e63f4a8 to 956884d Compare March 29, 2022 22:51

ssiccha force-pushed the frss/AnSnRecogUnknownDegree-2 branch 2 times, most recently from 3b12114 to daceec3 Compare March 29, 2022 23:52

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch 2 times, most recently from d45adbb to 7ba5ec6 Compare June 1, 2022 13:32

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch 4 times, most recently from 85ca56f to deb2974 Compare November 7, 2022 03:36

fingolfin added the recog days 2025 label Dec 8, 2025

ssiccha added 2 commits December 8, 2025 15:23

Add option infoLevel to RECOG.TestGroup

f381cc7

Fix for howtowritearecogmethod.autodoc

2aed139

FriedrichRober added 2 commits December 8, 2025 15:23

Improvement of SnAnUnknownDegree

65dc977

Compute ThreeCycleCandidates in a "lazy" way, as done in Magma Code GetNextThreeCycle.

Integrate SnAn code into library

af0aba4

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch from 7925157 to af0aba4 Compare December 8, 2025 14:24

Remove hack from test file

6539468

fixed wrong rebase

18b7a18

FriedrichRober force-pushed the frss/AnSnRecogUnknownDegree-2 branch from e6f5836 to 18b7a18 Compare December 8, 2025 15:41

FriedrichRober added 2 commits December 9, 2025 10:25

Rename counters to better reflect what they are counting in the algor…

7fb6ed2

…ithm, and highlight the adjustments made to the original algorithm.

Change iterator to work in small batches

6089439

Update changes text

adb2fce

FriedrichRober mentioned this pull request Dec 9, 2025

SnAn naming algorithm #348

Open

Performance fixes:

458b1b1

- deactivate guessdegree code - insert SnAnUnknownDegree after deletedpermutationmodule code in projective.gi

FriedrichRober marked this pull request as ready for review December 9, 2025 15:46

fingolfin reviewed Dec 11, 2025

View reviewed changes

fingolfin added 6 commits December 11, 2025 22:34

Merge branch 'master' into frss/AnSnRecogUnknownDegree-2

2959e1a

Removing duplicate of RECOG.GuessSnAnDegree

01f39bf

Turn RECOG_SnAn into a code comment

0a3e5a9

Get rid of SnAnRepeatImmediately

a6d26f9

trivial tweaks

7f2e750

restore bug fix

16bace0

fingolfin reviewed Dec 11, 2025

View reviewed changes

		# AddMethod(FindHomDbProjective, FindHomMethodsGeneric.SnAnSmallUnknownDegree, 1075);

		yTo5 := y ^ 5;
		if not isone(ri)(yTo5) and not isone(ri)(yTo5 * y) then

-        yTo5 := y ^ 5;
-        if not isone(ri)(yTo5) and not isone(ri)(yTo5 * y) then
+        y2 := y^2;
+        if not isone(ri)(y2) then
+          y3 := y2 * y;
+          if not isone(ri)(y3) and not isone(ri)(y2 * y3) then

Recog An/Sn unknown degree: conder, holt #265

Are you sure you want to change the base?

Recog An/Sn unknown degree: conder, holt #265

Uh oh!

Conversation

ssiccha commented Apr 9, 2021

Uh oh!

ssiccha commented Apr 12, 2021

Uh oh!

codecov bot commented Apr 12, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

FriedrichRober commented Mar 28, 2022

Uh oh!

ssiccha commented Mar 29, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

fingolfin commented Mar 30, 2022

Uh oh!

fingolfin commented Apr 6, 2022

Uh oh!

FriedrichRober commented Apr 7, 2022

Uh oh!

FriedrichRober commented Nov 7, 2022

Uh oh!

FriedrichRober commented Dec 8, 2025

Uh oh!

FriedrichRober commented Dec 8, 2025

Uh oh!

FriedrichRober commented Dec 8, 2025

Uh oh!

FriedrichRober commented Dec 8, 2025

Uh oh!

FriedrichRober commented Dec 9, 2025

Uh oh!

FriedrichRober commented Dec 9, 2025

Uh oh!

FriedrichRober commented Dec 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

FriedrichRober commented Dec 9, 2025

Uh oh!

fingolfin commented Dec 11, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fingolfin left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fingolfin Dec 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

FriedrichRober commented Dec 12, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

codecov bot commented Apr 12, 2021 •

edited

Loading

ssiccha commented Mar 29, 2022 •

edited

Loading

FriedrichRober commented Dec 9, 2025 •

edited

Loading

fingolfin Dec 11, 2025 •

edited

Loading