Attributes of polyhedra are exposed

[kliem]

Currently the f_vector is exposed.

sage: P = polytopes.simplex()
sage: P.f_vector()[0] = 2
sage: P.f_vector()
(2, 4, 6, 4, 1)

Same applies for

• incidence matrix,
• vertex-facet graph,
• vertices matrix,
• vertex adjacency matrix,
• facet adjacency matrix,
• gale transform.

Some of the above are probably more relevant than others.

CC: @jplab @LaisRast

Component: geometry

Keywords: polyhedron, exposed attributes

Author: Jonathan Kliem

Branch/Commit: a7eeece

Reviewer: Laith Rastanawi

Issue created by migration from https://trac.sagemath.org/ticket/28828

[kliem]

Changed author from Jonathan Kliem to none

[kliem]

Changed keywords from polyhedron, f_vector to polyhedron, exposed attributes

[sagetrac-nailuj]

comment:2

An older open ticket might be worth a look in this context: #25509. It proposes to simplify the creation of immutable vectors/matrices, e.g. by adding an immutable=True/False switch to the vector and matrix constructors.

[kliem]

comment:3

Thanks for pointing this out.

As for vectors and matrices there is set_immutable, which should work fine here.
I just didn't get around to doing it.

[kliem]

Branch: public/28828

[kliem]

Commit: c113802

[kliem]

New commits:

c113802

make exposed invariants of polyhedron immutable

[kliem]

Author: Jonathan Kliem

[LaisRast]

comment:6

I am not sure we need to change everything to be immutable.
But since you are doing this, you may also want to consider the following:

sage: P = polytopes.cube()
....: P.restricted_automorphism_group(output='matrixlist')[0][0,0] = 1000
....: P.restricted_automorphism_group(output='matrixlist')[0]
....: 
[1000    0    0    0]
[   0    1    0    0]
[   0    0    1    0]
[   0    0    0    1]

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

50a4975	make exposed invariants of polyhedron immutable
a7eeece	make restricted automorphism group as matrixlist immutable

[sagetrac-git]

Changed commit from c113802 to a7eeece

[kliem]

comment:10

As mentioned in the description of the ticket, some methods are worse than others.

E.g. it can easily happen, that one does something with the graph. There is even an example in the doctests (the one I modified in ticket), where the original graph was reversed.

Replying to @LaisRast:

I am not sure we need to change everything to be immutable.
But since you are doing this, you may also want to consider the following:

sage: P = polytopes.cube()
....: P.restricted_automorphism_group(output='matrixlist')[0][0,0] = 1000
....: P.restricted_automorphism_group(output='matrixlist')[0]
....: 
[1000    0    0    0]
[   0    1    0    0]
[   0    0    1    0]
[   0    0    0    1]

[LaisRast]

Reviewer: Laith Rastanawi

[LaisRast]

comment:11

I believe now it is good to go.

[vbraun]

Changed branch from public/28828 to a7eeece