Bug in Poset.is_slender(?)

[jdemeyer]

CC: @mezzarobba @jm58660

Component: combinatorics

Author: Marc Mezzarobba, Jori Mäntysalo, Frédéric Chapoton

Branch/Commit: 7608202

Reviewer: Jori Mäntysalo, Frédéric Chapoton

Issue created by migration from https://trac.sagemath.org/ticket/22373

[jdemeyer]

Branch: u/jdemeyer/bug_in_poset_is_slender___

[fchapoton]

comment:2

iteritems is not python3-compatible

you should rather use

from six import iteritems
for xx in iteritems(dd):
    pass

---

New commits:

90a31a1

Bug in Poset.is_slender(?)

[fchapoton]

Commit: 90a31a1

[mezzarobba]

Changed branch from u/jdemeyer/bug_in_poset_is_slender___ to u/mmezzarobba/22373-is_slender

[mezzarobba]

Changed commit from 90a31a1 to 4796945

[mezzarobba]

comment:4

Replying to @fchapoton:

iteritems is not python3-compatible

Fixed, thanks!

Without this patch, #22029 made one of the doctests of is_slender() fail.
Jori, can you tell if there was indeed a bug and perhaps how to add a test for the fix that would work even without #22029? Thanks!

---

New commits:

4796945

Bug in Poset.is_slender(?)

[jm58660]

comment:5

Replying to @mezzarobba:

Without this patch, #22029 made one of the doctests of is_slender() fail.
Jori, can you tell if there was indeed a bug and perhaps how to add a test for the fix that would work even without #22029?

What was the failing example?

I would guess that the other patch shows some hidden bug in an example poset. That should be easy to spot, just make all the example posets with and without the patch and try

L.hasse_diagram().canonical_label().dig6_string()

[mezzarobba]

comment:6

Replying to @jm58660:

What was the failing example?

It must have been the one with certificate=True, since my patch only touches the certificate. But I don't remember exactly which changes elsewhere made it fail.

[jm58660]

comment:7

Yes, I have been an idiot when doing the certificate-code. Test code:

set_random_seed(332211)
D = Posets.DiamondPoset(5)._hasse_diagram

for i in range(100):
    P = Posets.RandomPoset(randint(10, 20), 0.2)
    if P.is_slender():
        if P._hasse_diagram.subgraph_search(D) is not None:
            print("Bug 1")
            break
    else:
        if P._hasse_diagram.subgraph_search(D) is None:
            print("Bug 2")
            break
        _, cert = P.is_slender(certificate=True)
        P_ = P.subposet(P.interval(cert[0], cert[1]))
        if P_.height() != 3:
            print("Bug 3")
            break
        if P_.cardinality() < 5:
            print("Bug 4")
            break
else:
    print("All OK")

[jm58660]

comment:8

How should we define this to be on

LatticePoset({0: [1, 2, 3, 4], 1: [6], 2: [6], 3: [6], 4: [5], 5: [6]})

? I think that my test code has error, when I compare it to the definition in the docstring.

[fchapoton]

Changed branch from u/mmezzarobba/22373-is_slender to public/22373

[fchapoton]

Changed commit from 4796945 to 7608202

[fchapoton]

comment:9

I added a doctest, currently return the wrong answer 5.

I also changed six.iteritems to iteritems in the full file.

---

New commits:

6feb62f	Merge branch 'u/mmezzarobba/22373-is_slender' in 7.6.rc0
7608202	trac 22373 adding a doctest

[fchapoton]

comment:10

ok, green bot. Looks good to me. Jori, if you agree, you can set to positive.

[jm58660]

comment:11

Replying to @fchapoton:

ok, green bot. Looks good to me. Jori, if you agree, you can set to positive.

Yes, seems to be OK.

[mezzarobba]

Changed author from Marc Mezzarobba to Marc Mezzarobba, Jori Mäntysalo, Frédéric Chapoton

[mezzarobba]

Reviewer: Jori Mäntysalo, Frédéric Chapoton

[vbraun]

Changed branch from public/22373 to 7608202