plural wrapper

[burcin]

During SD10 in Nancy, Michael Brickenstein and Burcin Erocal worked on making Plural (the non-commutative extension of Singular) accessible from Sage. Burcin and Michael also worked at the Plural wrapper on SD 23.5 in Kaiserslautern. Oleksandr Motsak and Alexander Dreyer continued this at SD 24 in Linz.

The patches that resulted from this work are attached.

Newest functionality:

• non-commutative rings/polynomials/ideals are fully featured classes now (no deriving from commutative ones)!
• coercion from basering/Integer types (still needs tests)
• flag to check degeneracy conditions on init
• relations for non-commutative rings
• most relevant functions for rings/polynomials/ideals (mostly adopted from MPolynomialRing_libsingular/MPolynomialRing_libsingular/...) e.g. std/twostd/syzygy_module/lc/lm/lt/monomial operations
• RingWrap and TermOrder were extended
• quick and dirty conversion of RingWrap to Sage rings (needs some care as the resulting rings may not be unique and therefore may confuse coercion)
• quotient of a non-commutative ring by a two-sided Groebner basis
• shortcut to create graded commutative algebras: SCA

Possible topics that need work are:

• put the files in sage/algebra/ ???
• make sure element does not export functions it doesn't support (e.g. gcd)
• predefined structures from the library

Apply

• attachment: trac4539_libplural_rel11761.patch
• attachment: trac4539_pickling_rel10903.patch
• attachment: trac4539_normal_forms_rel10903.patch
• attachment: trac4539_fix_docs_rel10903.patch

Depends on #11316
Depends on #11856
Depends on #10903
Depends on #9138
Depends on #11900
Depends on #11115
Depends on #11068
Depends on #11761

CC: @saliola @mwhansen @alexanderdreyer @sagetrac-OleksandrMotsak @sagetrac-PolyBoRi @malb @simon-king-jena

Component: algebra

Keywords: libsingular plural wrapper sd10 sd23.5 sd24 sd34

Author: Michael Brickenstein, Burcin Erocal, Oleksandr Motsak, Alexander Dreyer, Simon King

Reviewer: Simon King, Alexander Dreyer

Merged: sage-5.0.beta1

Issue created by migration from https://trac.sagemath.org/ticket/4539

[burcin]

initial wrapper for plural

[burcin]

Attachment: plural_1.patch.gz

better user interface for plural rings

[sagetrac-PolyBoRi]

Attachment: plural_2.patch.gz

[aghitza]

comment:1

Attachment: letterplace.py.gz

[burcin]

comment:3

The letterplace interface in attachment: letterplace.py is now at #7797.

[burcin]

Attachment: plural_1.sage-4.4.4.patch.gz

rebased to 4.4.4

[burcin]

rebased to 4.4.4

[burcin]

Attachment: plural_2.sage-4.4.4.patch.gz

Attachment: plural_3.patch.gz

appy on top of 1 and 2, new classes for plural objects which don't inherit from the commutative ones

[sagetrac-PolyBoRi]

Attachment: plural_functions.2.patch.gz

[sagetrac-PolyBoRi]

comment:4

sorry, for multiple files. Apply patches in this order:

plural_1.sage-4.4.4.patch

plural_2.sage-4.4.4.patch

plural_3.patch Download

plural_functions.patch

[mwhansen]

comment:5

Attachment: plural_functions.patch.gz

[sagetrac-OleksandrMotsak]

i have just folded all the previous patches by Michael & Burcin into plural_folded-4.4.4.patch (should be applied before anything else)

[alexanderdreyer]

Attachment: plural_folded-4.4.4.patch.gz

Part one Olekandr's work in Linz

[alexanderdreyer]

Attachment: extending_plural_folded-4.4.4.patch.gz

Attachment: extplural-more.gz

Doctest fixes by Alexander

[alexanderdreyer]

Attachment: extplural-more.patch.gz

Doctest fixes by Alexander

[alexanderdreyer]

Attachment: extplural-more2.patch.gz

noncommunative ring functionality

[alexanderdreyer]

Attachment: extplural-more3.patch.gz

[alexanderdreyer]

Attachment: plural-wrapper-2010-07-22.patch.gz

This folds of the following patches, a crucial subset of the noncommutation fucntionality as well as extensive documentation and doctests

[alexanderdreyer]

comment:7

We have an first release ready for reviewing!

Regards, Oleksandr and Alexander

[simon-king-jena]

comment:117

Attachment: trac4539_docbuild_reviewer.patch.gz

Hi Alexander,

Note that there is a new patch at #10903 (where the docbuild crash was introduced), and it fixes the problem in a more satisfying way. The problem was that under certain circumstances the name of a deprecated Cython method could not be determined - but with the new patch from #10903 (actually there are TWO new patches) the problem is fixed.

So, the docbuild reviewer patch is not needed.

[alexanderdreyer]

comment:118

Building, installing and testing succeeded on SuSE 11 Enterprise amd64 and OS X 10.5 ppc. So we can switch back to positive review.

[jdemeyer]

Changed dependencies from #11068 #11316 #11856 #10903 to #11316, #11856, #10903, #9138, #11900, #11115, #11068

[jdemeyer]

Changed dependencies from #11316, #11856, #10903, #9138, #11900, #11115, #11068 to #11316, #11856, #10903, #9138, #11900, #11115, #11068, #11761

[jdemeyer]

comment:121

Attachment: trac4539_libplural_rel11761.patch.gz

[simon-king-jena]

comment:122

It could be that we need some work here. The first patch does not apply when we start with sage-4.8.alpha0. Namely, in sage/libs/singular/function.pyx, it expects the line

       ring2 = None

but this line has been removed. I don't know in which ticket that has happened. By consequence, a rather complicated chunk of the patch does not apply.

What shall we do? In order to avoid premature work, it would be an option to wait until finally all the dependencies got a positive review.

[alexanderdreyer]

comment:123

That line comes from https://github.com/sagemath/sage-prod/files/10653584/11761-cython-0.15.patch.gz (Cython is more strict here.) That patch was not merged in alpha0.

[simon-king-jena]

comment:124

Replying to @alexanderdreyer:

That line comes from https://github.com/sagemath/sage-prod/files/10653584/11761-cython-0.15.patch.gz (Cython is more strict here.)

Yes, now I see it: #11761 is a dependency! So, sorry for the noise, and back to a positive review - which needs two steps. One...

[simon-king-jena]

comment:125

... and two.

[jdemeyer]

Merged: sage-5.0.beta1