Cluster algebras

This file constructs cluster algebras using the Parent-Element framework. The implementation mainly utilizes structural theorems from [FZ2007].

The key points being used here are these:

  • cluster variables are parametrized by their g-vectors;
  • g-vectors (together with c-vectors) provide a self-standing model for the combinatorics behind any cluster algebra;
  • each cluster variable in any cluster algebra can be computed, by the separation of additions formula, from its g-vector and F-polynomial.

Accordingly this file provides three classes:

ClusterAlgebra, constructed as a subobject of sage.rings.polynomial.laurent_polynomial_ring.LaurentPolynomialRing_generic, is the frontend of this implementation. It provides all the algebraic features (like ring morphisms), it computes cluster variables, it is responsible for controlling the exploration of the exchange graph and serves as the repository for all the data recursively computed so far. In particular, all g-vectors and all F-polynomials of known cluster variables as well as a mutation path by which they can be obtained are recorded. In the optic of efficiency, this implementation does not store directly the exchange graph nor the exchange relations. Both of these could be added to ClusterAlgebra with minimal effort.

ClusterAlgebraSeed provides the combinatorial backbone for ClusterAlgebra. It is an auxiliary class and therefore its instances should not be directly created by the user. Rather it should be accessed via ClusterAlgebra.current_seed() and ClusterAlgebra.initial_seed(). The task of performing current seed mutations is delegated to this class. Seeds are considered equal if they have the same parent cluster algebra and they can be obtained from each other by a permutation of their data (i.e. if they coincide as unlabelled seeds). Cluster algebras whose initial seeds are equal in the above sense are not considered equal but are endowed with coercion maps to each other. More generally, a cluster algebra is endowed with coercion maps from any cluster algebra which is obtained by freezing a collection of initial cluster variables and/or permuting both cluster variables and coefficients.

ClusterAlgebraElement is a thin wrapper around sage.rings.polynomial.laurent_polynomial.LaurentPolynomial providing all the functions specific to cluster variables. Elements of a cluster algebra with principal coefficients have special methods and these are grouped in the subclass PrincipalClusterAlgebraElement.

One more remark about this implementation. Instances of ClusterAlgebra are built by identifying the initial cluster variables with the generators of ClusterAlgebra.ambient(). In particular, this forces a specific embedding into the ambient field of rational expressions. In view of this, although cluster algebras themselves are independent of the choice of initial seed, ClusterAlgebra.mutate_initial() is forced to return a different instance of ClusterAlgebra. At the moment there is no coercion implemented among the two instances but this could in principle be added to ClusterAlgebra.mutate_initial().

REFERENCES:

AUTHORS:

  • Dylan Rupel (2015-06-15): initial version
  • Salvatore Stella (2015-06-15): initial version

EXAMPLES:

We begin by creating a simple cluster algebra and printing its initial exchange matrix:

sage: A = ClusterAlgebra(['A', 2]); A
A Cluster Algebra with cluster variables x0, x1 and no coefficients over Integer Ring
sage: A.b_matrix()
[ 0  1]
[-1  0]

A is of finite type so we can explore all its exchange graph:

sage: A.explore_to_depth(infinity)

and get all its g-vectors, F-polynomials, and cluster variables:

sage: sorted(A.g_vectors_so_far())
[(-1, 0), (-1, 1), (0, -1), (0, 1), (1, 0)]
sage: sorted(A.F_polynomials_so_far(), key=str)
[1, 1, u0 + 1, u0*u1 + u0 + 1, u1 + 1]
sage: sorted(A.cluster_variables_so_far(), key=str)
[(x0 + 1)/x1, (x0 + x1 + 1)/(x0*x1), (x1 + 1)/x0, x0, x1]

Simple operations among cluster variables behave as expected:

sage: s = A.cluster_variable((0, -1)); s
(x0 + 1)/x1
sage: t = A.cluster_variable((-1, 1)); t
(x1 + 1)/x0
sage: t + s
(x0^2 + x1^2 + x0 + x1)/(x0*x1)
sage: _.parent() == A
True
sage: t - s
(-x0^2 + x1^2 - x0 + x1)/(x0*x1)
sage: _.parent() == A
True
sage: t*s
(x0*x1 + x0 + x1 + 1)/(x0*x1)
sage: _.parent() == A
True
sage: t/s
(x1^2 + x1)/(x0^2 + x0)
sage: _.parent() == A
False

Division is not guaranteed to yield an element of A so it returns an element of A.ambient().fraction_field() instead:

sage: (t/s).parent() == A.ambient().fraction_field()
True

We can compute denominator vectors of any element of A:

sage: (t*s).d_vector()
(1, 1)

Since we are in rank 2 and we do not have coefficients we can compute the greedy element associated to any denominator vector:

sage: A.rank() == 2 and A.coefficients() == ()
True
sage: A.greedy_element((1, 1))
(x0 + x1 + 1)/(x0*x1)
sage: _ == t*s
False

not surprising since there is no cluster in A containing both t and s:

sage: seeds = A.seeds(mutating_F=false)
sage: [ S for S in seeds if (0, -1) in S and (-1, 1) in S ]
[]

indeed:

sage: A.greedy_element((1, 1)) == A.cluster_variable((-1, 0))
True

Disabling F-polynomials in the computation just done was redundant because we already explored the whole exchange graph before. Though in different circumstances it could have saved us considerable time.

g-vectors and F-polynomials can be computed from elements of A only if A has principal coefficients at the initial seed:

sage: (t*s).g_vector()
Traceback (most recent call last):
...
AttributeError: 'ClusterAlgebra_with_category.element_class' object has no attribute 'g_vector'
sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.explore_to_depth(infinity)
sage: s = A.cluster_variable((0, -1)); s
(x0*y1 + 1)/x1
sage: t = A.cluster_variable((-1, 1)); t
(x1 + y0)/x0
sage: (t*s).g_vector()
(-1, 0)
sage: (t*s).F_polynomial()
u0*u1 + u0 + u1 + 1
sage: (t*s).is_homogeneous()
True
sage: (t+s).is_homogeneous()
False
sage: (t+s).homogeneous_components()
{(-1, 1): (x1 + y0)/x0, (0, -1): (x0*y1 + 1)/x1}

Each cluster algebra is endowed with a reference to a current seed; it could be useful to assign a name to it:

sage: A = ClusterAlgebra(['F', 4])
sage: len(A.g_vectors_so_far())
4
sage: A.current_seed()
The initial seed of a Cluster Algebra with cluster variables x0, x1, x2, x3
 and no coefficients over Integer Ring
sage: A.current_seed() == A.initial_seed()
True
sage: S = A.current_seed()
sage: S.b_matrix()
[ 0  1  0  0]
[-1  0 -1  0]
[ 0  2  0  1]
[ 0  0 -1  0]
sage: S.g_matrix()
[1 0 0 0]
[0 1 0 0]
[0 0 1 0]
[0 0 0 1]
sage: S.cluster_variables()
[x0, x1, x2, x3]

and use S to walk around the exchange graph of A:

sage: S.mutate(0); S
The seed of a Cluster Algebra with cluster variables x0, x1, x2, x3
 and no coefficients over Integer Ring obtained from the initial
 by mutating in direction 0
sage: S.b_matrix()
[ 0 -1  0  0]
[ 1  0 -1  0]
[ 0  2  0  1]
[ 0  0 -1  0]
sage: S.g_matrix()
[-1  0  0  0]
[ 1  1  0  0]
[ 0  0  1  0]
[ 0  0  0  1]
sage: S.cluster_variables()
[(x1 + 1)/x0, x1, x2, x3]
sage: S.mutate('sinks'); S
The seed of a Cluster Algebra with cluster variables x0, x1, x2, x3
 and no coefficients over Integer Ring obtained from the initial
 by mutating along the sequence [0, 2]
sage: S.mutate([2, 3, 2, 1, 0]); S
The seed of a Cluster Algebra with cluster variables x0, x1, x2, x3
 and no coefficients over Integer Ring obtained from the initial
 by mutating along the sequence [0, 3, 2, 1, 0]
sage: S.g_vectors()
[(0, 1, -2, 0), (-1, 2, -2, 0), (0, 1, -1, 0), (0, 0, 0, -1)]
sage: S.cluster_variable(3)
(x2 + 1)/x3

Walking around by mutating S updates the informations stored in A:

sage: len(A.g_vectors_so_far())
10
sage: A.current_seed().path_from_initial_seed()
[0, 3, 2, 1, 0]
sage: A.current_seed() == S
True

Starting from A.initial_seed() still records data in A but does not update A.current_seed():

sage: S1 = A.initial_seed()
sage: S1.mutate([2, 1, 3])
sage: len(A.g_vectors_so_far())
11
sage: S1 == A.current_seed()
False

Since ClusterAlgebra inherits from UniqueRepresentation, computed data is shared across instances:

sage: A1 = ClusterAlgebra(['F', 4])
sage: A1 is A
True
sage: len(A1.g_vectors_so_far())
11

It can be useful, at times to forget all computed data. Because of UniqueRepresentation this cannot be achieved by simply creating a new instance; instead it has to be manually triggered by:

sage: A.clear_computed_data()
sage: len(A.g_vectors_so_far())
4

Given a cluster algebra A we may be looking for a specific cluster variable:

sage: A = ClusterAlgebra(['E', 8, 1])
sage: v = (-1, 1, -1, 1, -1, 1, 0, 0, 1)
sage: A.find_g_vector(v, depth=2)
sage: seq = A.find_g_vector(v); seq  # random
[0, 1, 2, 4, 3]
sage: v in A.initial_seed().mutate(seq, inplace=False).g_vectors()
True

This also performs mutations of F-polynomials:

sage: A.F_polynomial((-1, 1, -1, 1, -1, 1, 0, 0, 1))
u0*u1*u2*u3*u4 + u0*u1*u2*u4 + u0*u2*u3*u4 + u0*u1*u2 + u0*u2*u4
 + u2*u3*u4 + u0*u2 + u0*u4 + u2*u4 + u0 + u2 + u4 + 1

which might not be a good idea in algebras that are too big. One workaround is to first disable F-polynomials and then recompute only the desired mutations:

sage: A.reset_exploring_iterator(mutating_F=False)  # long time
sage: v = (-1, 1, -2, 2, -1, 1, -1, 1, 1)  # long time
sage: seq = A.find_g_vector(v); seq  # long time random
[1, 0, 2, 6, 5, 4, 3, 8, 1]
sage: S = A.initial_seed().mutate(seq, inplace=False)   # long time
sage: v in S.g_vectors()   # long time
True
sage: A.current_seed().mutate(seq)    # long time
sage: A.F_polynomial((-1, 1, -2, 2, -1, 1, -1, 1, 1))   # long time
u0*u1^2*u2^2*u3*u4*u5*u6*u8 +
...
2*u2 + u4 + u6 + 1

We can manually freeze cluster variables and get coercions in between the two algebras:

sage: A = ClusterAlgebra(['F', 4]); A
A Cluster Algebra with cluster variables x0, x1, x2, x3 and no coefficients
 over Integer Ring
sage: A1 = ClusterAlgebra(A.b_matrix().matrix_from_columns([0, 1, 2]), coefficient_prefix='x'); A1
A Cluster Algebra with cluster variables x0, x1, x2 and coefficient x3
 over Integer Ring
sage: A.has_coerce_map_from(A1)
True

and we also have an immersion of A.base() into A and of A into A.ambient():

sage: A.has_coerce_map_from(A.base())
True
sage: A.ambient().has_coerce_map_from(A)
True

but there is currently no coercion in between algebras obtained by mutating at the initial seed:

sage: A1 = A.mutate_initial(0); A1
A Cluster Algebra with cluster variables x0, x1, x2, x3 and no coefficients
 over Integer Ring
sage: A.b_matrix() == A1.b_matrix()
False
sage: [X.has_coerce_map_from(Y) for X, Y in [(A, A1), (A1, A)]]
[False, False]
class sage.algebras.cluster_algebra.ClusterAlgebra(Q, **kwargs)

Bases: sage.structure.parent.Parent, sage.structure.unique_representation.UniqueRepresentation

A Cluster Algebra.

INPUT:

  • data – some data defining a cluster algebra; it can be anything that can be parsed by ClusterQuiver
  • scalars – a ring (default \(\ZZ\)); the scalars over which the cluster algebra is defined
  • cluster_variable_prefix – string (default 'x'); it needs to be a valid variable name
  • cluster_variable_names – a list of strings; each element needs to be a valid variable name; supersedes cluster_variable_prefix
  • coefficient_prefix – string (default 'y'); it needs to be a valid variable name.
  • coefficient_names – a list of strings; each element needs to be a valid variable name; supersedes cluster_variable_prefix
  • principal_coefficients – bool (default False); supersedes any coefficient defined by data

ALGORITHM:

The implementation is mainly based on [FZ2007] and [NZ2012].

EXAMPLES:

sage: B = matrix([(0, 1, 0, 0), (-1, 0, -1, 0), (0, 1, 0, 1), (0, 0, -2, 0), (-1, 0, 0, 0), (0, -1, 0, 0)])
sage: A = ClusterAlgebra(B); A
A Cluster Algebra with cluster variables x0, x1, x2, x3
 and coefficients y0, y1 over Integer Ring
sage: A.gens()
(x0, x1, x2, x3, y0, y1)
sage: A = ClusterAlgebra(['A', 2]); A
A Cluster Algebra with cluster variables x0, x1 and no coefficients
 over Integer Ring
sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True); A.gens()
(x0, x1, y0, y1)
sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True, coefficient_prefix='x'); A.gens()
(x0, x1, x2, x3)
sage: A = ClusterAlgebra(['A', 3], principal_coefficients=True, cluster_variable_names=['a', 'b', 'c']); A.gens()
(a, b, c, y0, y1, y2)
sage: A = ClusterAlgebra(['A', 3], principal_coefficients=True, cluster_variable_names=['a', 'b'])
Traceback (most recent call last):
...
ValueError: cluster_variable_names should be a list of 3 valid variable names
sage: A = ClusterAlgebra(['A', 3], principal_coefficients=True, coefficient_names=['a', 'b', 'c']); A.gens()
(x0, x1, x2, a, b, c)
sage: A = ClusterAlgebra(['A', 3], principal_coefficients=True, coefficient_names=['a', 'b'])
Traceback (most recent call last):
...
ValueError: coefficient_names should be a list of 3 valid variable names
F_polynomial(g_vector)

Return the F-polynomial with g-vector g_vector if it has been found.

INPUT:

  • g_vector – tuple; the g-vector of the F-polynomial to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.F_polynomial((-1, 1))
Traceback (most recent call last):
...
KeyError: 'the g-vector (-1, 1) has not been found yet'
sage: A.initial_seed().mutate(0, mutating_F=False)
sage: A.F_polynomial((-1, 1))
Traceback (most recent call last):
...
KeyError: 'the F-polynomial with g-vector (-1, 1) has not been computed yet;
 you can compute it by mutating from the initial seed along the sequence [0]'
sage: A.initial_seed().mutate(0)
sage: A.F_polynomial((-1, 1))
u0 + 1
F_polynomials()

Return an iterator producing all the F_polynomials of self.

ALGORITHM:

This method does not use the caching framework provided by self, but recomputes all the F-polynomials from scratch. On the other hand it stores the results so that other methods like F_polynomials_so_far() can access them afterwards.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: len(list(A.F_polynomials()))
9
F_polynomials_so_far()

Return a list of the F-polynomials encountered so far.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.current_seed().mutate(0)
sage: sorted(A.F_polynomials_so_far(), key=str)
[1, 1, u0 + 1]
ambient()

Return the Laurent polynomial ring containing self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.ambient()
Multivariate Laurent Polynomial Ring in x0, x1, y0, y1 over Integer Ring
b_matrix()

Return the initial exchange matrix of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.b_matrix()
[ 0  1]
[-1  0]
clear_computed_data()

Clear the cache of computed g-vectors and F-polynomials and reset both the current seed and the exploring iterator.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: sorted(A.g_vectors_so_far())
[(0, 1), (1, 0)]
sage: A.current_seed().mutate([1, 0])
sage: sorted(A.g_vectors_so_far())
[(-1, 0), (0, -1), (0, 1), (1, 0)]
sage: A.clear_computed_data()
sage: sorted(A.g_vectors_so_far())
[(0, 1), (1, 0)]
cluster_fan(depth=+Infinity)

Return the cluster fan (the fan of g-vectors) of self.

INPUT:

  • depth – a positive integer or infinity (default infinity); the maximum depth at which to compute

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.cluster_fan()
Rational polyhedral fan in 2-d lattice N
cluster_variable(g_vector)

Return the cluster variable with g-vector g_vector if it has been found.

INPUT:

  • g_vector – tuple; the g-vector of the cluster variable to return

ALGORITHM:

This function computes cluster variables from their g-vectors and F-polynomials using the “separation of additions” formula of Theorem 3.7 in [FZ2007].

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.initial_seed().mutate(0)
sage: A.cluster_variable((-1, 1))
(x1 + 1)/x0
cluster_variables()

Return an iterator producing all the cluster variables of self.

ALGORITHM:

This method does not use the caching framework provided by self, but recomputes all the cluster variables from scratch. On the other hand it stores the results so that other methods like cluster_variables_so_far() can access them afterwards.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: len(list(A.cluster_variables()))
9
cluster_variables_so_far()

Return a list of the cluster variables encountered so far.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.current_seed().mutate(0)
sage: sorted(A.cluster_variables_so_far(), key=str)
[(x1 + 1)/x0, x0, x1]
coefficient(j)

Return the j-th coefficient of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the coefficient to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.coefficient(0)
y0
coefficient_names()

Return the list of coefficient names.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: A.coefficient_names()
()
sage: A1 = ClusterAlgebra(['B', 2], principal_coefficients=True)
sage: A1.coefficient_names()
('y0', 'y1')
sage: A2 = ClusterAlgebra(['C', 3], principal_coefficients=True, coefficient_prefix='x')
sage: A2.coefficient_names()
('x3', 'x4', 'x5')
coefficients()

Return the list of coefficients of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.coefficients()
(y0, y1)
sage: A1 = ClusterAlgebra(['B', 2])
sage: A1.coefficients()
()
contains_seed(seed)

Test if seed is a seed of self.

INPUT:

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True); A
A Cluster Algebra with cluster variables x0, x1 and coefficients y0, y1 over Integer Ring
sage: S = copy(A.current_seed())
sage: A.contains_seed(S)
True
current_seed()

Return the current seed of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.current_seed()
The initial seed of a Cluster Algebra with cluster variables x0, x1
 and no coefficients over Integer Ring
explore_to_depth(depth)

Explore the exchange graph of self up to distance depth from the initial seed.

INPUT:

  • depth – a positive integer or infinity; the maximum depth at which to stop searching

EXAMPLES:

sage: A = ClusterAlgebra(['A', 4])
sage: A.explore_to_depth(infinity)
sage: len(A.g_vectors_so_far())
14
find_g_vector(g_vector, depth=+Infinity)

Return a mutation sequence to obtain a seed containing the g-vector g_vector from the initial seed.

INPUT:

  • g_vector – a tuple: the g-vector to find
  • depth – a positive integer or infinity (default infinity); the maximum distance from self.current_seed to reach

OUTPUT:

This function returns a list of integers if it can find g_vector, otherwise it returns None. If the exploring iterator stops, it means that the algebra is of finite type and g_vector is not the g-vector of any cluster variable. In this case the function resets the iterator and raises an error.

EXAMPLES:

sage: A = ClusterAlgebra(['G', 2], principal_coefficients=True)
sage: A.clear_computed_data()
sage: A.find_g_vector((-2, 3), depth=2)
sage: A.find_g_vector((-2, 3), depth=3)
[0, 1, 0]
sage: A.find_g_vector((1, 1), depth=3)
sage: A.find_g_vector((1, 1), depth=4)
Traceback (most recent call last):
...
ValueError: (1, 1) is not the g-vector of any cluster variable of a
 Cluster Algebra with cluster variables x0, x1 and coefficients y0, y1
 over Integer Ring
g_vectors(mutating_F=True)

Return an iterator producing all the g-vectors of self.

INPUT:

  • mutating_F – bool (default True); whether to compute F-polynomials; disable this for speed considerations

ALGORITHM:

This method does not use the caching framework provided by self, but recomputes all the g-vectors from scratch. On the other hand it stores the results so that other methods like g_vectors_so_far() can access them afterwards.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: len(list(A.g_vectors()))
9
g_vectors_so_far()

Return a list of the g-vectors of cluster variables encountered so far.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.current_seed().mutate(0)
sage: sorted(A.g_vectors_so_far())
[(-1, 1), (0, 1), (1, 0)]
gens()

Return the list of initial cluster variables and coefficients of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.gens()
(x0, x1, y0, y1)
sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True, coefficient_prefix='x')
sage: A.gens()
(x0, x1, x2, x3)
greedy_element(d_vector)

Return the greedy element with denominator vector d_vector.

INPUT:

  • d_vector – tuple of 2 integers; the denominator vector of the element to compute

ALGORITHM:

This implements greedy elements of a rank 2 cluster algebra using Equation (1.5) from [LLZ2014].

EXAMPLES:

sage: A = ClusterAlgebra(['A', [1, 1], 1])
sage: A.greedy_element((1, 1))
(x0^2 + x1^2 + 1)/(x0*x1)
initial_cluster_variable(j)

Return the j-th initial cluster variable of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the cluster variable to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.initial_cluster_variable(0)
x0
initial_cluster_variable_names()

Return the list of initial cluster variable names.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.initial_cluster_variable_names()
('x0', 'x1')
sage: A1 = ClusterAlgebra(['B', 2], cluster_variable_prefix='a')
sage: A1.initial_cluster_variable_names()
('a0', 'a1')
initial_cluster_variables()

Return the list of initial cluster variables of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: A.initial_cluster_variables()
(x0, x1)
initial_seed()

Return the initial seed of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.initial_seed()
The initial seed of a Cluster Algebra with cluster variables x0, x1 and no coefficients over Integer Ring
lift(x)

Return x as an element of ambient().

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: x = A.cluster_variable((1, 0))
sage: A.lift(x).parent()
Multivariate Laurent Polynomial Ring in x0, x1, y0, y1 over Integer Ring
lower_bound()

Return the lower bound associated to self.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4])
sage: A.lower_bound()
Traceback (most recent call last):
...
NotImplementedError: not implemented yet
mutate_initial(direction)

Return the cluster algebra obtained by mutating self at the initial seed.

INPUT:

  • direction – in which direction(s) to mutate, it can be:
    • an integer in range(self.rank()) to mutate in one direction only
    • an iterable of such integers to mutate along a sequence
    • a string “sinks” or “sources” to mutate at all sinks or sources simultaneously

ALGORITHM:

This function computes data for the new algebra from known data for the old algebra using Equation (4.2) from [NZ2012] for g-vectors, and Equation (6.21) from [FZ2007] for F-polynomials. The exponent \(h\) in the formula for F-polynomials is -min(0, old_g_vect[k]) due to [NZ2012] Proposition 4.2.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4])
sage: A.explore_to_depth(infinity)
sage: B = A.b_matrix()
sage: B.mutate(0)
sage: A1 = ClusterAlgebra(B)
sage: A1.explore_to_depth(infinity)
sage: A2 = A1.mutate_initial(0)
sage: A2._F_poly_dict == A._F_poly_dict
True

Check that we did not mess up the original algebra because of UniqueRepresentation:

sage: A = ClusterAlgebra(['A',2])
sage: A.mutate_initial(0) is A
False
rank()

Return the rank of self, i.e. the number of cluster variables in any seed.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True); A
A Cluster Algebra with cluster variables x0, x1
 and coefficients y0, y1 over Integer Ring
sage: A.rank()
2
reset_current_seed()

Reset the value reported by current_seed() to initial_seed().

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: A.current_seed().mutate([1, 0])
sage: A.current_seed() == A.initial_seed()
False
sage: A.reset_current_seed()
sage: A.current_seed() == A.initial_seed()
True
reset_exploring_iterator(mutating_F=True)

Reset the iterator used to explore self.

INPUT:

  • mutating_F – bool (default True); whether to also compute F-polynomials; disable this for speed considerations

EXAMPLES:

sage: A = ClusterAlgebra(['A', 4])
sage: A.clear_computed_data()
sage: A.reset_exploring_iterator(mutating_F=False)
sage: A.explore_to_depth(infinity)
sage: len(A.g_vectors_so_far())
14
sage: len(A.F_polynomials_so_far())
4
retract(x)

Return x as an element of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2], principal_coefficients=True)
sage: L = A.ambient()
sage: x = L.gen(0)
sage: A.retract(x).parent()
A Cluster Algebra with cluster variables x0, x1 and coefficients y0, y1 over Integer Ring
scalars()

Return the ring of scalars over which self is defined.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.scalars()
Integer Ring
seeds(**kwargs)

Return an iterator running over seeds of self.

INPUT:

  • from_current_seed – bool (default False); whether to start the iterator from current_seed() or initial_seed()
  • mutating_F – bool (default True); whether to compute F-polynomials also; disable this for speed considerations
  • allowed_directions – iterable of integers (default range(self.rank())); the directions in which to mutate
  • depth – a positive integer or infinity (default infinity); the maximum depth at which to stop searching
  • catch_KeyboardInterrupt – bool (default False); whether to catch KeyboardInterrupt and return it rather then raising an exception – this allows the iterator returned by this method to be resumed after being interrupted

ALGORITHM:

This function traverses the exchange graph in a breadth-first search.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 4])
sage: A.clear_computed_data()
sage: seeds = A.seeds(allowed_directions=[3, 0, 1])
sage: _ = list(seeds)
sage: sorted(A.g_vectors_so_far())
[(-1, 0, 0, 0),
 (-1, 1, 0, 0),
 (0, -1, 0, 0),
 (0, 0, 0, -1),
 (0, 0, 0, 1),
 (0, 0, 1, 0),
 (0, 1, 0, 0),
 (1, 0, 0, 0)]
set_current_seed(seed)

Set the value reported by current_seed() to seed, if it makes sense.

INPUT:

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: A.clear_computed_data()
sage: S = copy(A.current_seed())
sage: S.mutate([0, 1, 0])
sage: A.current_seed() == S
False
sage: A.set_current_seed(S)
sage: A.current_seed() == S
True
sage: A1 = ClusterAlgebra(['B', 2])
sage: A.set_current_seed(A1.initial_seed())
Traceback (most recent call last):
...
ValueError: This is not a seed in this cluster algebra
theta_basis_element(g_vector)

Return the element of the theta basis with g-vector g_vector.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4])
sage: A.theta_basis_element((1, 0, 0, 0))
Traceback (most recent call last):
...
NotImplementedError: not implemented yet
upper_bound()

Return the upper bound associated to self.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4])
sage: A.upper_bound()
Traceback (most recent call last):
...
NotImplementedError: not implemented yet
upper_cluster_algebra()

Return the upper cluster algebra associated to self.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4])
sage: A.upper_cluster_algebra()
Traceback (most recent call last):
...
NotImplementedError: not implemented yet
class sage.algebras.cluster_algebra.ClusterAlgebraElement

Bases: sage.structure.element_wrapper.ElementWrapper

An element of a cluster algebra.

d_vector()

Return the denominator vector of self as a tuple of integers.

EXAMPLES:

sage: A = ClusterAlgebra(['F', 4], principal_coefficients=True)
sage: A.current_seed().mutate([0, 2, 1])
sage: x = A.cluster_variable((-1, 2, -2, 2)) * A.cluster_variable((0, 0, 0, 1))**2
sage: x.d_vector()
(1, 1, 2, -2)
class sage.algebras.cluster_algebra.ClusterAlgebraSeed(B, C, G, parent, **kwargs)

Bases: sage.structure.sage_object.SageObject

A seed in a Cluster Algebra.

INPUT:

  • B – a skew-symmetrizable integer matrix
  • C – the matrix of c-vectors of self
  • G – the matrix of g-vectors of self
  • parentClusterAlgebra; the algebra to which the seed belongs
  • path – list (default []); the mutation sequence from the initial seed of parent to self

Warning

Seeds should not be created manually: no test is performed to assert that they are built from consistent data nor that they really are seeds of parent. If you create seeds with inconsistent data all sort of things can go wrong, even __eq__() is no longer guaranteed to give correct answers. Use at your own risk.

F_polynomial(j)

Return the j-th F-polynomial of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the F-polynomial to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.F_polynomial(0)
1
F_polynomials()

Return all the F-polynomials of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.F_polynomials()
[1, 1, 1]
b_matrix()

Return the exchange matrix of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.b_matrix()
[ 0  1  0]
[-1  0 -1]
[ 0  1  0]
c_matrix()

Return the matrix whose columns are the c-vectors of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.c_matrix()
[1 0 0]
[0 1 0]
[0 0 1]
c_vector(j)

Return the j-th c-vector of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the c-vector to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.c_vector(0)
(1, 0, 0)
sage: S.mutate(0)
sage: S.c_vector(0)
(-1, 0, 0)
sage: S.c_vector(1)
(1, 1, 0)
c_vectors()

Return all the c-vectors of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.c_vectors()
[(1, 0, 0), (0, 1, 0), (0, 0, 1)]
cluster_variable(j)

Return the j-th cluster variable of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the cluster variable to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.cluster_variable(0)
x0
sage: S.mutate(0)
sage: S.cluster_variable(0)
(x1 + 1)/x0
cluster_variables()

Return all the cluster variables of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.cluster_variables()
[x0, x1, x2]
depth()
Return the length of a mutation sequence from the initial seed
of parent() to self.

Warning

This is the length of the mutation sequence returned by path_from_initial_seed(), which need not be the shortest possible.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: S1 = A.initial_seed()
sage: S1.mutate([0, 1, 0, 1])
sage: S1.depth()
4
sage: S2 = A.initial_seed()
sage: S2.mutate(1)
sage: S2.depth()
1
sage: S1 == S2
True
g_matrix()

Return the matrix whose columns are the g-vectors of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.g_matrix()
[1 0 0]
[0 1 0]
[0 0 1]
g_vector(j)

Return the j-th g-vector of self.

INPUT:

  • j – an integer in range(self.parent().rank()); the index of the g-vector to return

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.g_vector(0)
(1, 0, 0)
g_vectors()

Return all the g-vectors of self.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 3])
sage: S = A.initial_seed()
sage: S.g_vectors()
[(1, 0, 0), (0, 1, 0), (0, 0, 1)]
mutate(direction, **kwargs)

Mutate self.

INPUT:

  • direction – in which direction(s) to mutate, it can be:
    • an integer in range(self.rank()) to mutate in one direction only
    • an iterable of such integers to mutate along a sequence
    • a string “sinks” or “sources” to mutate at all sinks or sources simultaneously
  • inplace – bool (default True); whether to mutate in place or to return a new object
  • mutating_F – bool (default True); whether to compute F-polynomials while mutating

Note

While knowing F-polynomials is essential to computing cluster variables, the process of mutating them is quite slow. If you care only about combinatorial data like g-vectors and c-vectors, setting mutating_F=False yields significant benefits in terms of speed.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: S = A.initial_seed()
sage: S.mutate(0); S
The seed of a Cluster Algebra with cluster variables x0, x1
 and no coefficients over Integer Ring obtained from the initial
 by mutating in direction 0
sage: S.mutate(5)
Traceback (most recent call last):
...
ValueError: cannot mutate in direction 5
parent()

Return the parent of self.

EXAMPLES:

sage: A = ClusterAlgebra(['B', 3])
sage: A.current_seed().parent() == A
True
path_from_initial_seed()

Return a mutation sequence from the initial seed of parent() to self.

Warning

This is the path used to compute self and it does not have to be the shortest possible.

EXAMPLES:

sage: A = ClusterAlgebra(['A', 2])
sage: S1 = A.initial_seed()
sage: S1.mutate([0, 1, 0, 1])
sage: S1.path_from_initial_seed()
[0, 1, 0, 1]
sage: S2 = A.initial_seed()
sage: S2.mutate(1)
sage: S2.path_from_initial_seed()
[1]
sage: S1 == S2
True
class sage.algebras.cluster_algebra.PrincipalClusterAlgebraElement

Bases: sage.algebras.cluster_algebra.ClusterAlgebraElement

An element in a cluster algebra with principle coefficients.

F_polynomial()

Return the F-polynomial of self.

EXAMPLES:

sage: A = ClusterAlgebra(['B', 2], principal_coefficients=True)
sage: S = A.initial_seed()
sage: S.mutate([0, 1, 0])
sage: S.cluster_variable(0).F_polynomial() == S.F_polynomial(0)
True
sage: sum(A.initial_cluster_variables()).F_polynomial()
Traceback (most recent call last):
...
ValueError: this element is not homogeneous
g_vector()

Return the g-vector of self.

EXAMPLES:

sage: A = ClusterAlgebra(['B', 2], principal_coefficients=True)
sage: A.cluster_variable((1, 0)).g_vector() == (1, 0)
True
sage: sum(A.initial_cluster_variables()).g_vector()
Traceback (most recent call last):
...
ValueError: this element is not homogeneous
homogeneous_components()

Return a dictionary of the homogeneous components of self.

OUTPUT:

A dictionary whose keys are homogeneous degrees and whose values are the summands of self of the given degree.

EXAMPLES:

sage: A = ClusterAlgebra(['B', 2], principal_coefficients=True)
sage: x = A.cluster_variable((1, 0)) + A.cluster_variable((0, 1))
sage: x.homogeneous_components()
{(0, 1): x1, (1, 0): x0}
is_homogeneous()

Return True if self is a homogeneous element of self.parent().

EXAMPLES:

sage: A = ClusterAlgebra(['B', 2], principal_coefficients=True)
sage: A.cluster_variable((1, 0)).is_homogeneous()
True
sage: x = A.cluster_variable((1, 0)) + A.cluster_variable((0, 1))
sage: x.is_homogeneous()
False