Fix bug with comparison to 1 in IndexedFreeModuleElement

[trevorkarn]

There is a monomial basis comparison to 1 in IndexedFreeModule element. This ticket expands it to a comparison with .one_basis().

CC: @tscrim @trevorkarn

Component: linear algebra

Keywords: gsoc2022, one basis, indexed free module

Author: Trevor K. Karn

Branch/Commit: 2269b23

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/34084

[tscrim]

Reviewer: Travis Scrimshaw

[tscrim]

comment:2

This is dangerous as a module does not have to implement a one_basis() (in fact, some of them do not because it is not an element of the basis). I agree with the principle of this change, but you should pull out getting the self.parent().one_basis() outside of the loops and into a try-except AttributeError: block. I would set the stored one_basis variable you'll need to None for the exception branch.

The same change should be done in _unicode_art_ as well.

[trevorkarn]

comment:3

Sounds good. Is there any examples of modules that don’t implement .one_basis that come to mind? I want to make sure I can add an example in the doctest where the fix I had breaks.

[tscrim]

comment:4

I think one of the bases for the DescentAlgebra doesn’t have a one_basis() since one() is a sum of basis elements. There is also a basis of one of the combinatorial Hopf algebras called the orbit basis that also has this IIRC.

[sagetrac-git]

Changed commit from c767729 to 382cb46

[sagetrac-git]

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

382cb46

Fix to work with algebras without a one_basis

[sagetrac-git]

Changed commit from 382cb46 to 2269b23

[sagetrac-git]

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

2269b23

Change one in _unicode_art_

[tscrim]

comment:8

Thanks. LGTM.

[tscrim]

Changed keywords from gsoc2022 one-basis indexed-free-module to gsoc2022, one basis, indexed free module

[vbraun]

Changed branch from u/tkarn/indexed-element-one-basis to 2269b23