affine hull of one point polyhedron

[mo271]

When using the orthogonal/orthonormal feature of the affine hull function of a polyhedron (introduced in #22804), there is a problem with the one point polyhedron:

sage: P = Polyhedron([[7]])
sage: P
A 0-dimensional polyhedron in ZZ^1 defined as the convex hull of 1 vertex
sage: P.affine_hull()
A 0-dimensional polyhedron in ZZ^0 defined as the convex hull of 1 verte
sage: P.affine_hull(orthogonal='True')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
...
TypeError: unsupported operand parent(s) for *: 'Full MatrixSpace of 0 by 0 dense matrices over Integer Ring' and 'Ambient free module of rank 1 over the principal ideal domain Integer Ring'
sage: P.affine_hull(orthonormal='True')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
...
TypeError: unsupported operand parent(s) for *: 'Full MatrixSpace of 0 by 0 dense matrices over Integer Ring' and 'Ambient free module of rank 1 over the principal ideal domain Integer Ring'

This is how it should be:

sage: P = Polyhedron([[7]]); P
A 0-dimensional polyhedron in ZZ^1 defined as the convex hull of 1 vertex
sage: P.affine_hull()
A 0-dimensional polyhedron in ZZ^0 defined as the convex hull of 1 vertex
sage: P.affine_hull(orthonormal='True')
A 0-dimensional polyhedron in QQ^0 defined as the convex hull of 1 vertex
sage: P.affine_hull(orthogonal='True')
A 0-dimensional polyhedron in QQ^0 defined as the convex hull of 1 vertex

Depends on #22804

CC: @videlec @jplab

Component: geometry

Keywords: polyhedron

Author: Moritz Firsching

Branch/Commit: 9fccf4e

Reviewer: Jean-Philippe Labbé

Issue created by migration from https://trac.sagemath.org/ticket/23355

[mo271]

Branch: u/moritz/fix_one_point_affine_hull

[mo271]

comment:2

The reason why this didn't work properly is the fact that if a Matrix is generated from a list it cannot correctly guess its dimensions if the list is empty. The whole fix is simply to exchange the line

M = matrix(self.base_ring(), [list(w) for w in itertools.islice(v.neighbors(), self.dim())])

by

M = matrix(self.base_ring(), self.dim(), self.ambient_dim(), [list(w) for w in itertools.islice(v.neighbors(), self.dim())])

giving the dimensions explicitly. I also added a doctest (under the section TESTS) documenting that the bug is fixed.

---

New commits:

029246d	prescribe size of matrix explicitely
08d951f	added hidden doctest

[mo271]

Commit: 08d951f

[jplab]

comment:3

Hi Moritz,

Could you add a reference to this ticket just before the tests you added?

Apart from that it looks good to me! You can set it to positive review on my behalf once this is done!

[jplab]

Reviewer: Jean-Philippe Labbé

[sagetrac-git]

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

9fccf4e

added trac number in test

[sagetrac-git]

Changed commit from 08d951f to 9fccf4e

[mo271]

comment:6

thanks for the review, JP!

I added the desired reference and will put it on 'positive review' as you suggested