owned this note changed 3 years ago
Linked with GitHub

Notepad for GAP Days Summer 2022

Beware that the contents of this notepad can be read and modified by anyone with access to the URL, so you should not share any sensitive information here.

Table of contents:

Practical information

Main focus: Matrix Object

List of functionality we would like for Matrix Objects:

Other project descriptions and reports

For each project, please maintain a short description of the project and its progress.

Writing up MatrixObj specification

Revising how matrices and vectors are implemented in GAP has been in the works for over a decade now. We had two GAP meetings on this in the past few year

People:

New Matrix Object Implementations

  • SparseMatrixRep
  • DiagonalMatrixRep
  • (upper | lower) UnitriangularMatrixRep → Daniel
  • PermutationMatrixRep
  • MonomialMatrixRep
  • BlockMatrixRep
  • WreathProductElementMatrixRep (Permuted Diagonal-Block Matrix)
  • Compressed
  • CVec
  • Generic pListRep
  • \(\mathbb{Z}/n\mathbb{Z}\)
  • almost simple?
  • minimal MatrixObject → Lucas

People: Daniel, Lucas

Generic Test Suite

Adapt Library Codes, but progress has been mostly stalled since then. A major blocker for this is that while we discussed how to go about things, we never finished writing up a specification for others to work with. The plan is to get that done this week to enable other people to contribute to this.

People: Max Horn, Lucas, Anna

Use MatrixObjects and Document/Fix problems

People: Dominika

GAP discussion topics and outcomes

Nothing so far.

Programme

Day 1: Monday 17th

Schedule

10:00 - 10:20 Opening
10:20 - 10:50 Setup Session
11:00 - 11:30 Talk: Clique and design finding in parallel using GAP, GRAPE, DESIGN, and C; Leonard Soicher
11:30 - 17:40 Questions and Short break
12:00 - 13:00 Lunch break
13:30 - 14:00 Talk: Introduction to MatrixObj; Max Horn
14:30 - 16:30 Working session
16:30 - 16:45 Stand-up

Day 2: Tuesday 18th

Schedule

09:00 - 09:10 Stand-up
09:20 - 10:50 Package Session Part 1
10:50 - 12:00 Working session
12:00 - 13:00 Lunch break
13:20 - 13:30 Group photo (please be there)
13:00 - 13:30 Working session
13:30 - 15:00 Package Session Part 2
15:00 - 16:30 Working session
16:30 - 16:50 Stand-up

Morning stand-up

  • Max
    • work on MatrixObj specification, tests, implementation
    • help people
  • Ruth
    • automata package updates
    • graphs
  • Mun-See
    • write my talk
    • work on normalizers in vole
    • help Ruth with graph stuff
  • Lucas
    • start work on generic test suit
    • help people with new matrix objects
  • Friedrich
    • work on SnAnRecognition code for the recog package
    • construct some graphs with Tom and Reymond
  • Daniel
    • Start a new type for MatrixObj (UpperTriangularMatrices) -> Constructor already working
  • Marvin
    • help Daniel
  • ChrisJ
    • Work on Vole
    • Clean up emscripten code?
  • Tom
    • work on SimplicialSurface Package
    • create some funny graphs related to quimp groups (with Friedrich and Reymond)
  • Anna
    • work on MatrixObj tests
    • enjoy the package session :)
  • other Anna
    • work on the clique graph operator on triangulations of surfaces
  • Leonard
    • Install nauty and traces version 2.7r4 in GRAPE development version
  • Iryna
    • extend the library of local nearrings and publish on GitHub and Zenodo
  • Maryna
    • extend the library of local nearrings and publish on GitHub and Zenodo
  • Eileen
    • work on the SOTGrps package
    • do some last-minute changes to my talk

Afternoon stand-up / Results and accomplishments

  • Lucas
    • worked on the Generic Test Suite
    • helped people getting started with their work on MatObj
    • orga stuff
  • Anna
    • worked on the Generic Test Suite
    • orga stuff
  • Max
    • discussions on Generic Test Suite and other MatrixObj stuff
    • fixed bugs in GAP adn packages
    • help Eileen with SOTGroup package
    • work towards GAP 4.12.1 release (due this week)
  • Mun See
    • decided on recycling my old talk
    • talked to Chris about normalisers in vole, but haven't started implementation
    • worked on some graph stuff with Ruth
  • Dominika
    • went through MatrixObj documentation and functions
  • Ruth
    • Got the triangle merge in graphs to work
    • (had the talk ready before the morning stand-up)
  • Leonard
    • nauty and traces: version 2.7 r4 seems to work with the development version of GRAPE
  • Eileen
    • gave a talk on SOTGrps package and continued working to improve the package (infrastructure and maths)
  • Iryna
    • prepared for published the libraries on GitHub and Zenodo
    • restarted our tasks on the server
  • Maryna
    • prepared for published the libraries on GitHub and Zenodo
    • restarted our tasks on the server
  • Fatemeh
    • worked on the MatrixObj functions
  • other Anna
    • stupid implementation of subgraph isomorphism testing -> too slow

Day 3: Wednesday 19th

Schedule

09:00 - 09:10 Stand-up
09:20 - 09:50 Talk: Mun See
09:50 - 10:00 Questions and short break
10:00 - 10:30 Talk: GAP for Permutation Pattern Classes; Ruth Hoffmann
10:39 - 10:40 Questions and short break
10:40 - 12:00 Working session
12:00 - 13:00 Lunch break
13:00 - 16:30 Working session
16:30 - 16:50 Stand-up

Morning Stand-up

  • Anna
    • continue to work on generic test suite for MatrixObj
  • Max
    • work on MatrixObj (for real today)
    • work towards GAP 4.12.1 release
  • Lucas
    • work on generic test suite for MatrixObj
    • help others
  • Daniel and Marvin
    • Continue work on new MatObject Type triangular Matrices
  • Mun See
    • Work on normalisers in Vole
  • Ruth
    • Give a talk
    • Work on automata PR
    • Maybe some more graphs
  • other Anna
    • Learn about induced subraph isomorphism testing in the digraphs package
    • Apply to the clique graphs of triangulations
  • Tom and Meike
    • Work on Simplicial Surface Package
  • Fatemeh
    • continue to work on checking MatrixObj functions and find problems
  • ChrisJ
    • Clean up orbital graphs related code in BacktrackKit
  • Iryna
    • extend the library of local nearrings and publish on GitHub and Zenodo
  • Maryna
    • extend the library of local nearrings and publish on GitHub and Zenodo

Results and accomplishments

  • other Anna
    • found out how to apply the digraphs package and the glasgow subgraph solver to iterated clique graphs
  • Anna
    • work on generic test suite for MatrixObj
  • Max
    • helped with SimplicialSurfaces package, submitted a PR and a few issues
    • helped people submit issues and pull requests related to MatrixObj and how to find what causes a bug
    • wrote a CI test for GAP's make install feature and fixed the bug this revealed
    • discussions about GAP and math problems with a bunch of people
    • no MatrixObj work on my own
  • Friedrich
    • work on RecogniseSnAnUnknownDegree
  • Ruth
    • Gave talk
    • Finished fixing automata PR (the first)
    • Abandonned own graphs code for Mun See's
  • ChrisJ
    • Improve printing and creation code for HashMap/HashSet
    • PR for utils package
  • Lucas
    • work on generic test suite for MatrixObj
  • Fatemeh
    • work on MatrixObj and working with functions in it
    • disscution with Max and getting help on Matrixobj
  • Iryna
    • published the libraries of local nearrings of order 128 on GitHub and Zenodo
    • prepared our report
    • noted several questions with Sonata
  • Maryna
    • published the libraries of local nearrings of order 128 on GitHub and Zenodo
    • prepared our report
    • noted several questions with Sonata
  • Daniel and Marvin
    • Finished the code for new matrix obj type
  • Tom
    • worked on the SimplicialSurface Package
    • constructing graphs
  • Leonard
    • Working on an algorithm and program to determine whether a given simple GRAPE graph gamma is gamma.group-undirected-edge-transitive, without forming an explicit orbit of edges

Day 4: Thursday 20th

Schedule

09:00 - 09:10 Stand-up
09:20 - 09:35 Talk: An algorithm for constructing all supercharacter theories of a finite group; Fatemeh Koorepazan-Moftakhar
09:45 - 10:15 Talk: Finite local nearrings and their classification; Iryna Raievska and Maryna Raievska:
10:15 - 10:30 Talk: Classification of Gyrogroups of Small Order; Kuroš Mavaddat Nežâd
10:30 - 10:40 Questions and short break
10:40 - 12:00 Working session
12:00 - 13:00 Lunch break
13:00 - 16:30 Working session
16:30 - 16:50 Stand-up

Morning Stand-up

  • Leonard
    • finish a GAP function to determine whether a given simple GRAPE graph gamma is gamma.group-undirected-edge-transitive, without forming an explicit orbit of edges
    • and check and test this function with some examples from Alice Niemeyer
    • maybe start on a GAP function to determine whether a given simple GRAPE graph is locally gamma.group-t-arc-transitive
  • Marvin
    • adapt the previous days' work on upper triangular matrices to lower ones
  • Lucas
    • continue with the MatrixObj tests
    • make a minimal MatrixObj example
    • review some PRs
  • Max
    • review various PRs folks here made yesterday
    • release GAP 4.12.1
    • work with MatrixObj
    • help people
  • Anna
    • continue with the MatrixObj tests
    • a little teaching
  • Tom
    • work on the SimplicialSurface package:
    • (DrawSurfaceToSVG + DrawSurfaceToSTL)
  • Daniel
    • Finish PR for new MatObj type (except Max has too many complaints)
  • other Anna (Annegret)
    • Construct examples to test second clique graph inclusions
  • Fatemeh
    • Give a talk
    • work with MatrixObj test
  • Iryna
    • Give a talk
    • extend (migrate) the library of local nearrings and publish on GitHub and Zenodo
  • Maryna
    • Give a talk
    • extend (migrate) the library of local nearrings and publish on GitHub and Zenodo

Results and accomplishments

  • Lucas
    • finished PR
    • implemented minimal matrix obj (still needs testing)
    • worked on generic test suite (PR imminent)
  • Anna
    • finished concept, implemented and tested generic test suite
  • Friedrich
    • further work on graph constructions with Reymond and Tom
  • Daniel
    • Pull request for infinite loop
    • Draft Pull request for upper triangular matrices (waiting for the test suite)
    • Started new issue for new matrix group categories
  • Iryna
  • Maryna
  • Max
    • finished (?) 4.12.1 preparations, and submitted PR with propose release notes
    • helped folks with creating PRs, using GitHub
    • discussed MatrixObj details with various people
  • Leonard
    • finished and tested first versions of the functions described above
  • Fatemeh
    • Gave a talk
    • Creating PR with helping Max, using GitHub
    • work on my package using PackageMaker and with Friedrich's help

Day 5: Friday 21th

Morning Stand-up

  • Lucas
    • make PR for MatObj test suite
    • review Daniel & Marvins PR
    • incorporate Max suggestions into implementation of IsFlistMatrixRep
    • maybe talk to Chris about Kernel debugging
  • Max
    • recover from GAP 4.12.1 release
    • review PRs
    • work on some generic helper for MatrixObj
    • look into fixes for various GAP packages sent to me by someone who packages GAP for a Linux distribution
    • look with Friedrich & Alice into An/Sn recognition issues???
  • Ruth
    • Integrate the Glasgow SIP solver into Mun See's distillation code
  • Anna
    • test suite for MatObj
  • Daniel
    • Work on Issue #3568
  • Marvin
    • help Daniel
  • Leonard
    • install GAP 4.12.1 on my laptop
    • perform further checks and tests of the functions written during the workshop
  • other Anna (Annegret)
    • implement the inverse Clique Operator
  • Iryna
  • Maryna

Schedule

09:00 - 09:10 Stand-up
09:10 - 11:40 Working session
11:40 - 12:00 Time for Feedback
12:00 - 13:00 Lunch break
13:00 - 15:00 Working session
15:00 - 15:30 Closing

Results and accomplishments

  • other Anna (Annegret)
    • finished implementing the reversed clique operator

Scratchpad

Put your random notes here, minutes for discussions etc.

MatrixObj

Some random MatrixObj TODOs

documentation

  • perhaps move ChangedBaseDomain documentation into the "Constructing Vector and Matrix Objects" section?
  • "Implementing New Vector and Matrix Objects Types" used to be last sectionin chapter 26, now this is "Basic operations for row/column reductions" swap them again?
  • Matrix( <M1>, <M2> ) documentation is not helpful; we need to explain what M1 and M2 mean; better argument names would help (e.g. Matrix( <dataMat>, <exampleMat> ) is not great but arguably at least a bit better)

Returning errors

  • When operations in general return 'fail', or return an error?

  • Example, for a non-invertible IsPlistMatrixRep, M^(-1) returns fail, but M^(-2) enters the break loop.

  • Should all matrices (including ones with different fields / sizes / representations) be comparable with =, and/or < ?

constructors

  • clarify difference between NewMatrix and Matrix (and likewise NewVector and Vector)
    • Max thinks that at some point there idea here was that the "New*" constructors are internal, and the other ones (wittout New) are implemented generically in terms of the New... ones. The idea being that there are many variants for Matrix and we don't want every implementation to have to provide all of them
    • if this is to be still the plan, let's make it very explicit in the documentation
    • perhaps it would be even better to rename NewMatrix/NewVector to say NewMatrixInternal or _NewMatrix or so, to make accidental use less likely?
    • though I am not sure accidental use is a problem at all?
    • see also https://github.com/gap-system/gap/issues/4549 and https://github.com/gap-system/gap/pull/4584
  • CompatibleVector(M) returns a vector of length NrRows(M). Shouldn't it be possible to get one with length NrCols(M), too? Perhaps an optional length argument would be a good idea?
  • should user facing constructors perform input validation, e.g. check that all entries are really integers etc. (see https://github.com/gap-system/gap/issues/4884)
    • Max thinks YES these days but have NC variants?? (Thomas works on this)
  • inverting an Integer matrix: raise an error if result is not defined over integers? or return a matrix over Q? or what?
    • Max favors the error
Select a repo