Sage implementation of arithmetic matroids and toric arrangements

# Arithmat

Sage implementation of arithmetic matroids and toric arrangements.

Authors: Roberto Pagaria and Giovanni Paolini

## Installation

### From PyPI

The easiest way to start using Arithmat is to install the latest release from PyPI. Type the following command:

sage -pip install --upgrade arithmat


### From source

git clone https://github.com/giove91/arithmat.git


Change to the root directory and run:

sage -pip install --upgrade --no-index -v .


## Overview

Arithmat is a Sage package that implements arithmetic matroids and toric arrangements.

At its core there is the class ArithmeticMatroidMixin, which is intended to be used in combination with any existing matroid class of Sage (e.g. RankMatroid, BasisMatroid, LinearMatroid) via multiple inheritance. The most common combinations are already defined: ArithmeticMatroid (deriving from RankMatroid), BasisArithmeticMatroid (deriving from BasisMatroid), and LinearArithmeticMatroid (deriving from LinearMatroid).

An additional class ToricArithmeticMatroid is implemented, for arithmetic matroids constructed from a fixed given representation.

## Citing Arithmat

To cite Arithmat, you can use (some variant of) the following BibTeX code:

@Misc{arithmat,
author = {Pagaria, Roberto and Paolini, Giovanni},
title = {{Arithmat}: {Sage} implementation of arithmetic matroids and toric arrangements},
year = {2019},
url = "https://github.com/giove91/arithmat",
}


## Documentation

### Import

All defined classes and functions can be imported at once:

from arithmat import *


Alternatively, it is possible to import only specific classes and/or functions. For example:

from arithmat import ArithmeticMatroid, ToricArithmeticMatroid


### Available classes for arithmetic matroids

All classes for arithmetic matroids derive from ArithmeticMatroidMixin and from some subclass of Sage's Matroid. The class ArithmeticMatroidMixin is not intended to be used by itself, but it is possible to subclass it in order to create new classes for arithmetic matroids.

The classes which are already provided in arithmat are the following.

• ArithmeticMatroid(groundset, rank_function, multiplicity_function)

This is the simplest arithmetic matroid class, and derives from ArithmeticMatroidMixin and RankMatroid.

E = [1,2,3,4,5]

def rk(X):
return min(2, len(X))

def m(X):
if len(X) == 2 and all(x in [3,4,5] for x in X):
return 2
else:
return 1

M = ArithmeticMatroid(E, rk, m)
print M
# Arithmetic matroid of rank 2 on 5 elements

print M.arithmetic_tutte_polynomial()
# y^3 + x^2 + 2*y^2 + 3*x + 3*y + 3

print M.representation()
# None (this arithmetic matroid is not representable)

• ToricArithmeticMatroid(arrangement_matrix, torus_matrix=None, ordered_groundset=None)

Arithmetic matroid associated with a given toric arrangement. This class derives from ArithmeticMatroidMixin and Matroid.

The constructor requires an integer matrix arrangement_matrix representing the toric arrangement. Optionally it accepts another integer matrix torus_matrix (whose cokernel describes the ambient torus, and defaults to matrix(ZZ, arrangement_matrix.nrows(), 0)) and/or an ordered copy ordered_groundset of the groundset (defaults to range(matrix.ncols())). The number of rows of arrangement_matrix must be equal to the numer of rows of torus_matrix.

The two matrices are not guaranteed to remain unchanged: internally,torus_matrix is kept in Smith normal form (this also affects arrangement_matrix).

A = matrix(ZZ, [[-1, 1, 0, 2], [3, 1, -1, -2]])
M = ToricArithmeticMatroid(A)

print M
# Toric arithmetic matroid of rank 2 on 4 elements

print M.arrangement_matrix()
# [-1  1  0  2]
# [ 3  1 -1 -2]

print M.torus_matrix()
# []

P = M.poset_of_layers()
print P
# Finite poset containing 11 elements

P.show(label_elements=False)

A = matrix(ZZ, [[-1, 1, 0, 2], [3, 1, -1, -2]])
Q = matrix(ZZ, [[5], [1]])
N = ToricArithmeticMatroid(A, Q)

print N
# Toric arithmetic matroid of rank 1 on 4 elements

print N.arrangement_matrix()
# [-16  -4   5  12]

print N.torus_matrix()
# []


The classes BasisArithmeticMatroid and LinearArithmeticMatroid derive from the Sage classes BasisMatroid and LinearMatroid, respectively. An instance of XxxArithmeticMatroid can be constructed with XxxArithmeticMatroid(..., multiplicity_function=m), where ... should be replaced by arguments to construct an instance of XxxMatroid, and m is the multiplicity function. The multiplicity function needs to be passed as a keyword argument (and not as a positional argument).

• BasisArithmeticMatroid(M=None, groundset=None, bases=None, ..., multiplicity_function)

def m(X):
if len(X) == 2 and all(x in ['b','c','d'] for x in X):
return 2
else:
return 1

M = BasisArithmeticMatroid(groundset='abcd', bases=['ab', 'ac', 'ad', 'bc', 'bd', 'cd'], multiplicity_function=m)

print M
# Basis arithmetic matroid of rank 2 on 4 elements

print M.is_valid()
# True


It is possible to cast any arithmetic matroid to a BasisArithmeticMatroid:

M1 = ToricArithmeticMatroid(matrix(ZZ, [[-1, 1, 0, 7], [6, 1, -1, -2]]))
M2 = BasisArithmeticMatroid(M1)

print M1
# Toric arithmetic matroid of rank 2 on 4 elements

print M2
# Basis arithmetic matroid of rank 2 on 4 elements

print M2.full_multiplicity() == M1.full_multiplicity()
# True

• LinearArithmeticMatroid(matrix=None, groundset=None, ..., multiplicity_function)

A = matrix(GF(2), [[1, 0, 0, 1, 1], [0, 1, 0, 1, 0], [0, 0, 1, 0, 1]])

def m(X):
return 1

M = LinearArithmeticMatroid(A, multiplicity_function=m)

print M
# Linear arithmetic matroid of rank 3 on 5 elements

print M.is_valid()
# True


Finally, MinorArithmeticMatroid and DualArithmeticMatroid are the analogs of MinorMatroid and DualMatroid. They are used when calling contract(X), delete(X), dual() (see below).

### Available methods

All classes for arithmetic matroids must also derive from some subclass of Sage's Matroid. In particular, Matroid methods are still available. For example:

M = ToricArithmeticMatroid(matrix(ZZ, [[1,2,3], [0,1, 1]]))
print list(M.bases())
# [frozenset([0, 1]), frozenset([0, 2]), frozenset([1, 2])]


All subclasses of ArithmeticMatroidMixin also (re-)implement the following methods.

• multiplicity(X) Return the multiplicity of the set X.

• full_multiplicity() Return the multiplicity of the groundset.

• is_valid() Check if the arithmetic matroid axioms are satisfied. This method overwrites Matroid.is_valid.

• is_torsion_free() Check if the matroid is torsion-free, i.e. the multiplicity of the empty set is equal to 1.

• is_surjective() Check if the matroid is surjective, i.e. the multiplicity of the groundset is equal to 1.

• is_gcd() Check if the matroid satisfies the gcd property, defined in [DM13, Section 3].

• is_strong_gcd() Check if the matroid satisfies the strong gcd property, defined in [PP19, Section 3].

• is_isomorphic(other, morphism) Check if the two arithmetic matroids are isomorphic. This method overwrites Matroid.is_isomorphic.

• is_isomorphism(other, morphism) Check if the given morphism of groundsets is an isomorphism of arithmetic matroids. It works also when comparing instances of different subclasses of ArithmeticMatroid. This method overwrites Matroid.is_isomorphism.

• contract(X) Contract elements. This method overwrites Matroid.contract.

• delete(X) Delete elements. This method overwrites Matroid.delete.

• minor(contractions=None, deletions=None) Contract some elements and delete some other elements. This method overwrites Matroid.minor.

• dual() Return the dual of the matroid. This method overwrites Matroid.dual.

• arithmetic_tutte_polynomial(x=None, y=None) Return the arithmetic Tutte polynomial of the matroid.

• reduction() Return the reduction of the matroid, defined in [PP19, Section 4].

• check_representation(A, ordered_groundset=None) Check if the given integer matrix A is a representation of the matroid. The optional parameter ordered_groundset specifies the bijection between the columns of the matrix and the groundset.

• all_representations(ordered_groundset=None) Generator of all non-equivalent essential representations of the matroid, computed using the algorithm of [PP19, Section 5].

• num_representations() Return the number of non-equivalent essential representations of the matroid. This calls all_representations().

• representation(ordered_groundset=None) Return any essential representation of the matroid, or None if the matroid is not representable.

• is_representable() Check if the matroid is representable. This calls representation().

• is_orientable() Check if the matroid is orientable as an arithmetic matroid, as defined in [Pag18].

In addition, ToricArithmeticMatroid has the following methods.

• ordered_groundset() Return the groundset as a list, in the same order as the columns of the matrix of the defining toric arrangement.

• arrangement_matrix() Return the matrix of the defining toric arrangement.

• torus_matrix() Return the matrix describing the ambient torus.

• poset_of_layers() Return the poset of layers of the toric arrangement, computed using Lenz's algorithm [Len17a].

• arithmetic_independence_poset() Return the arithmetic independence poset of the toric arrangement, defined in [Len17b, Definition 5], [Mar18, Definitions 2.1 and 2.2], [DD18, Section 7]. Notice that it is not the same as the independence poset of the underlying matroid.

• decomposition() Return the partition of the groundset corresponding to the decomposition of the matroid as a direct sum of indecomposable matroids. Uses the algorithm of [PP19, Section 7].

• is_decomposable() Check if the matroid is decomposable.

• is_indecomposable() Check if the matroid is not decomposable.

• is_equivalent(other, morphism=None) Check if the two ToricArithmeticMatroids are equivalent (i.e. the defining representations are equivalent; see [PP19, Section 2]). If the morphism is not given, the two ordered groundsets are assumed to be the same.

The following function is available outside of arithmetic matroid classes.

• signed_hermite_normal_form(A) Return the signed Hermite normal form of the integer matrix A, defined in [PP19, Section 6].
from arithmat import signed_hermite_normal_form

print signed_hermite_normal_form(matrix([[3, 2, 1], [-1, 1, 3]]))
# [  1   1  -3]
# [  0   5 -10]


More examples can be found in the test file.

## Bibliography

[BM14] P. Brändén and L. Moci, The multivariate arithmetic Tutte polynomial, Transactions of the American Mathematical Society 366 (2014), no. 10, 5523-5540.

[DM13] M. D'Adderio and L. Moci, Arithmetic matroids, the Tutte polynomial and toric arrangements, Advances in Mathematics 232 (2013), 335-367.

[DD18] A. D'Alì and E. Delucchi, Stanley-Reisner rings for symmetric simplicial complexes, G-semimatroids and Abelian arrangements, ArXiv preprint 1804.07366 (2018).

[DGP17] E. Delucchi, N. Girard, and G. Paolini, Shellability of posets of labeled partitions and arrangements defined by root systems, ArXiv preprint 1706.06360 (2017).

[Len17a] M. Lenz, Computing the poset of layers of a toric arrangement, ArXiv preprint 1708.06646 (2017).

[Len17b] M. Lenz, Stanley-Reisner rings for quasi-arithmetic matroids, ArXiv preprint 1709.03834 (2017).

[Len19] M. Lenz, Representations of weakly multiplicative arithmetic matroids are unique, Annals of Combinatorics 23 (2019), no. 2, 335-346.

[Mar18] I. Martino, Face module for realizable Z-matroids, Contributions to Discrete Mathematics 13 (2018).

[Pag17] R. Pagaria, Combinatorics of Toric Arrangements, ArXiv preprint 1710.00409 (2017).

[Pag18] R. Pagaria, Orientable arithmetic matroids, ArXiv preprint 1805.11888 (2018).

[Pag19] R. Pagaria, Two Examples of Toric Arrangements, Journal of Combinatorial Theory, Series A 167 (2019), 389-402.

[PP19] R. Pagaria and G. Paolini, Representations of torsion-free arithmetic matroids, ArXiv preprint 1908.04137 (2019).

## Project details

Uploaded source
Uploaded py2