Generic graphs

This module implements the base class for graphs and digraphs, and methods that can be applied on both. Here is what it can do:

Basic Graph operations:

networkx_graph() Creates a new NetworkX graph from the Sage graph
to_dictionary() Creates a dictionary encoding the graph.
adjacency_matrix() Returns the adjacency matrix of the (di)graph.
incidence_matrix() Returns an incidence matrix of the (di)graph
distance_matrix() Returns the distance matrix of the (strongly) connected (di)graph
weighted_adjacency_matrix() Returns the weighted adjacency matrix of the graph
kirchhoff_matrix() Returns the Kirchhoff matrix (a.k.a. the Laplacian) of the graph.
get_boundary() Returns the boundary of the (di)graph.
set_boundary() Sets the boundary of the (di)graph.
has_loops() Returns whether there are loops in the (di)graph.
allows_loops() Returns whether loops are permitted in the (di)graph.
allow_loops() Changes whether loops are permitted in the (di)graph.
loops() Returns any loops in the (di)graph.
has_multiple_edges() Returns whether there are multiple edges in the (di)graph.
allows_multiple_edges() Returns whether multiple edges are permitted in the (di)graph.
allow_multiple_edges() Changes whether multiple edges are permitted in the (di)graph.
multiple_edges() Returns any multiple edges in the (di)graph.
name() Returns or sets the graph’s name.
weighted() Whether the (di)graph is to be considered as a weighted (di)graph.
antisymmetric() Tests whether the graph is antisymmetric
density() Returns the density
order() Returns the number of vertices.
size() Returns the number of edges.
add_vertex() Creates an isolated vertex.
add_vertices() Add vertices to the (di)graph from an iterable container
delete_vertex() Deletes a vertex, removing all incident edges.
delete_vertices() Remove vertices from the (di)graph taken from an iterable container of vertices.
has_vertex() Return True if vertex is one of the vertices of this graph.
random_vertex() Returns a random vertex of self.
random_edge() Returns a random edge of self.
vertex_boundary() Returns a list of all vertices in the external boundary of vertices1, intersected with vertices2.
set_vertices() Associate arbitrary objects with each vertex
set_vertex() Associate an arbitrary object with a vertex.
get_vertex() Retrieve the object associated with a given vertex.
get_vertices() Return a dictionary of the objects associated to each vertex.
loop_vertices() Returns a list of vertices with loops.
vertex_iterator() Returns an iterator over the vertices.
neighbor_iterator() Return an iterator over neighbors of vertex.
vertices() Return a list of the vertices.
neighbors() Return a list of neighbors (in and out if directed) of vertex.
merge_vertices() Merge vertices.
add_edge() Adds an edge from u and v.
add_edges() Add edges from an iterable container.
subdivide_edge() Subdivides an edge k times.
subdivide_edges() Subdivides k times edges from an iterable container.
delete_edge() Delete the edge from u to v
delete_edges() Delete edges from an iterable container.
delete_multiedge() Deletes all edges from u and v.
set_edge_label() Set the edge label of a given edge.
has_edge() Returns True if (u, v) is an edge, False otherwise.
edges() Return a list of edges.
edge_boundary() Returns a list of edges (u,v,l) with u in vertices1
edge_iterator() Returns an iterator over edges.
edges_incident() Returns incident edges to some vertices.
edge_label() Returns the label of an edge.
edge_labels() Returns a list of edge labels.
remove_multiple_edges() Removes all multiple edges, retaining one edge for each.
remove_loops() Removes loops on vertices in vertices. If vertices is None, removes all loops.
loop_edges() Returns a list of all loops in the graph.
number_of_loops() Returns the number of edges that are loops.
clear() Empties the graph of vertices and edges and removes name, boundary, associated objects, and position information.
degree() Gives the degree (in + out for digraphs) of a vertex or of vertices.
average_degree() Returns the average degree of the graph.
degree_histogram() Returns a list, whose ith entry is the frequency of degree i.
degree_iterator() Returns an iterator over the degrees of the (di)graph.
degree_sequence() Return the degree sequence of this (di)graph.
random_subgraph() Return a random subgraph that contains each vertex with prob. p.
add_cycle() Adds a cycle to the graph with the given vertices.
add_path() Adds a cycle to the graph with the given vertices.
complement() Returns the complement of the (di)graph.
line_graph() Returns the line graph of the (di)graph.
to_simple() Returns a simple version of itself (i.e., undirected and loops and multiple edges are removed).
disjoint_union() Returns the disjoint union of self and other.
union() Returns the union of self and other.
relabel() Relabels the vertices of self
degree_to_cell() Returns the number of edges from vertex to an edge in cell.
subgraph() Returns the subgraph containing the given vertices and edges.
is_subgraph() Tests whether self is a subgraph of other.

Graph products:

cartesian_product() Returns the Cartesian product of self and other.
tensor_product() Returns the tensor product, also called the categorical product, of self and other.
lexicographic_product() Returns the lexicographic product of self and other.
strong_product() Returns the strong product of self and other.
disjunctive_product() Returns the disjunctive product of self and other.

Paths and cycles:

eulerian_orientation() Returns a DiGraph which is an Eulerian orientation of the current graph.
eulerian_circuit() Return a list of edges forming an eulerian circuit if one exists.
cycle_basis() Returns a list of cycles which form a basis of the cycle space of self.
interior_paths() Returns an exhaustive list of paths (also lists) through only interior vertices from vertex start to vertex end in the (di)graph.
all_paths() Returns a list of all paths (also lists) between a pair of vertices in the (di)graph.
triangles_count() Returns the number of triangles in the (di)graph.

Linear algebra:

spectrum() Returns a list of the eigenvalues of the adjacency matrix.
eigenvectors() Returns the right eigenvectors of the adjacency matrix of the graph.
eigenspaces() Returns the right eigenspaces of the adjacency matrix of the graph.

Some metrics:

cluster_triangles() Returns the number of triangles for nbunch of vertices as a dictionary keyed by vertex.
clustering_average() Returns the average clustering coefficient.
clustering_coeff() Returns the clustering coefficient for each vertex in nbunch
cluster_transitivity() Returns the transitivity (fraction of transitive triangles) of the graph.
szeged_index() Returns the Szeged index of the graph.

Automorphism group:

coarsest_equitable_refinement() Returns the coarsest partition which is finer than the input partition, and equitable with respect to self.
automorphism_group() Returns the largest subgroup of the automorphism group of the (di)graph whose orbit partition is finer than the partition given.
is_vertex_transitive() Returns whether the automorphism group of self is transitive within the partition provided
is_isomorphic() Tests for isomorphism between self and other.
canonical_label() Returns the unique graph on \{0,1,...,n-1\} ( n = self.order() ) which 1) is isomorphic to self 2) is invariant in the isomorphism class.

Graph properties:

is_eulerian() Return true if the graph has a (closed) tour that visits each edge exactly once.
is_planar() Tests whether the graph is planar.
is_circular_planar() Tests whether the graph is circular planar (outerplanar)
is_regular() Return True if this graph is (k-)regular.
is_chordal() Tests whether the given graph is chordal.
is_circulant() Tests whether the graph is a circulant graph.
is_interval() Check whether self is an interval graph
is_gallai_tree() Returns whether the current graph is a Gallai tree.
is_clique() Tests whether a set of vertices is a clique
is_independent_set() Tests whether a set of vertices is an independent set
is_transitively_reduced() Tests whether the digraph is transitively reduced.
is_equitable() Checks whether the given partition is equitable with respect to self.

Traversals:

breadth_first_search() Returns an iterator over the vertices in a breadth-first ordering.
depth_first_search() Returns an iterator over the vertices in a depth-first ordering.
lex_BFS() Performs a Lex BFS on the graph.

Distances:

distance() Returns the (directed) distance from u to v in the (di)graph
distance_all_pairs() Returns the distances between all pairs of vertices.
distances_distribution() Returns the distances distribution of the (di)graph in a dictionary.
eccentricity() Return the eccentricity of vertex (or vertices) v.
radius() Returns the radius of the (di)graph.
center() Returns the set of vertices in the center of the graph
diameter() Returns the largest distance between any two vertices.
distance_graph() Returns the graph on the same vertex set as the original graph but vertices are adjacent in the returned graph if and only if they are at specified distances in the original graph.
girth() Computes the girth of the graph.
periphery() Returns the set of vertices in the periphery
shortest_path() Returns a list of vertices representing some shortest path from u to v
shortest_path_length() Returns the minimal length of paths from u to v
shortest_paths() Returns a dictionary associating to each vertex v a shortest path from u to v, if it exists.
shortest_path_lengths() Returns a dictionary of shortest path lengths keyed by targets that are connected by a path from u.
shortest_path_all_pairs() Computes a shortest path between each pair of vertices.
wiener_index() Returns the Wiener index of the graph.
average_distance() Returns the average distance between vertices of the graph.

Flows, connectivity, trees:

is_connected() Tests whether the (di)graph is connected.
connected_components() Returns the list of connected components
connected_components_number() Returns the number of connected components.
connected_components_subgraphs() Returns a list of connected components as graph objects.
connected_component_containing_vertex() Returns a list of the vertices connected to vertex.
blocks_and_cut_vertices() Computes the blocks and cut vertices of the graph.
blocks_and_cuts_tree() Computes the blocks-and-cuts tree of the graph.
is_cut_edge() Returns True if the input edge is a cut-edge or a bridge.
is_cut_vertex() Returns True if the input vertex is a cut-vertex.
edge_cut() Returns a minimum edge cut between vertices s and t
vertex_cut() Returns a minimum vertex cut between non-adjacent vertices s and t
flow() Returns a maximum flow in the graph from x to y
edge_disjoint_paths() Returns a list of edge-disjoint paths between two vertices
vertex_disjoint_paths() Returns a list of vertex-disjoint paths between two vertices as given by Menger’s theorem.
edge_connectivity() Returns the edge connectivity of the graph.
vertex_connectivity() Returns the vertex connectivity of the graph.
transitive_closure() Computes the transitive closure of a graph and returns it.
transitive_reduction() Returns a transitive reduction of a graph.
min_spanning_tree() Returns the edges of a minimum spanning tree.
spanning_trees_count() Returns the number of spanning trees in a graph.

Plot/embedding-related methods:

set_embedding() Sets a combinatorial embedding dictionary to _embedding attribute.
get_embedding() Returns the attribute _embedding if it exists.
check_embedding_validity() Checks whether an _embedding attribute is well defined
get_pos() Returns the position dictionary
check_pos_validity() Checks whether pos specifies two (resp. 3) coordinates for every vertex (and no more vertices).
set_pos() Sets the position dictionary.
set_planar_positions() Compute a planar layout for self using Schnyder’s algorithm
layout_planar() Uses Schnyder’s algorithm to compute a planar layout for self.
is_drawn_free_of_edge_crossings() Tests whether the position dictionary gives a planar embedding.
latex_options() Returns an instance of GraphLatex for the graph.
set_latex_options() Sets multiple options for rendering a graph with LaTeX.
layout() Returns a layout for the vertices of this graph.
layout_spring() Computes a spring layout for this graph
layout_ranked() Computes a ranked layout for this graph
layout_extend_randomly() Extends randomly a partial layout
layout_circular() Computes a circular layout for this graph
layout_tree() Computes an ordered tree layout for this graph, which should be a tree (no non-oriented cycles).
layout_graphviz() Calls graphviz to compute a layout of the vertices of this graph.
graphplot() Returns a GraphPlot object.
plot() Returns a graphics object representing the (di)graph.
show() Shows the (di)graph.
plot3d() Plot a graph in three dimensions.
show3d() Plots the graph using Tachyon, and shows the resulting plot.
graphviz_string() Returns a representation in the dot language.
graphviz_to_file_named() Write a representation in the dot in a file.

Algorithmically hard stuff:

steiner_tree() Returns a tree of minimum weight connecting the given set of vertices.
edge_disjoint_spanning_trees() Returns the desired number of edge-disjoint spanning trees/arborescences.
feedback_vertex_set() Computes the minimum feedback vertex set of a (di)graph.
multiway_cut() Returns a minimum edge multiway cut
max_cut() Returns a maximum edge cut of the graph.
longest_path() Returns a longest path of self.
traveling_salesman_problem() Solves the traveling salesman problem (TSP)
is_hamiltonian() Tests whether the current graph is Hamiltonian.
hamiltonian_cycle() Returns a Hamiltonian cycle/circuit of the current graph/digraph
multicommodity_flow() Solves a multicommodity flow problem.
disjoint_routed_paths() Returns a set of disjoint routed paths.
dominating_set() Returns a minimum dominating set of the graph
subgraph_search() Returns a copy of G in self.
subgraph_search_count() Returns the number of labelled occurences of G in self.
subgraph_search_iterator() Returns an iterator over the labelled copies of G in self.
characteristic_polynomial() Returns the characteristic polynomial of the adjacency matrix of the (di)graph.
genus() Returns the minimal genus of the graph.
trace_faces() A helper function for finding the genus of a graph.

Methods

class sage.graphs.generic_graph.GenericGraph

Bases: sage.graphs.generic_graph_pyx.GenericGraph_pyx

Base class for graphs and digraphs.

add_cycle(vertices)

Adds a cycle to the graph with the given vertices. If the vertices are already present, only the edges are added.

For digraphs, adds the directed cycle, whose orientation is determined by the list. Adds edges (vertices[u], vertices[u+1]) and (vertices[-1], vertices[0]).

INPUT:

  • vertices – a list of indices for the vertices of the cycle to be added.

EXAMPLES:

sage: G = Graph()
sage: G.add_vertices(range(10)); G
Graph on 10 vertices
sage: show(G)
sage: G.add_cycle(range(20)[10:20])
sage: show(G)
sage: G.add_cycle(range(10))
sage: show(G)
sage: D = DiGraph()
sage: D.add_cycle(range(4))
sage: D.edges()
[(0, 1, None), (1, 2, None), (2, 3, None), (3, 0, None)]
add_edge(u, v=None, label=None)

Adds an edge from u and v.

INPUT: The following forms are all accepted:

  • G.add_edge( 1, 2 )
  • G.add_edge( (1, 2) )
  • G.add_edges( [ (1, 2) ])
  • G.add_edge( 1, 2, ‘label’ )
  • G.add_edge( (1, 2, ‘label’) )
  • G.add_edges( [ (1, 2, ‘label’) ] )

WARNING: The following intuitive input results in nonintuitive output:

sage: G = Graph()
sage: G.add_edge((1,2), 'label')
sage: G.networkx_graph().adj           # random output order
{'label': {(1, 2): None}, (1, 2): {'label': None}}

Use one of these instead:

sage: G = Graph()
sage: G.add_edge((1,2), label="label")
sage: G.networkx_graph().adj           # random output order
{1: {2: 'label'}, 2: {1: 'label'}}
sage: G = Graph()
sage: G.add_edge(1,2,'label')
sage: G.networkx_graph().adj           # random output order
{1: {2: 'label'}, 2: {1: 'label'}}

The following syntax is supported, but note that you must use the label keyword:

sage: G = Graph()
sage: G.add_edge((1,2), label='label')
sage: G.edges()
[(1, 2, 'label')]
sage: G = Graph()
sage: G.add_edge((1,2), 'label')
sage: G.edges()
[('label', (1, 2), None)]

Vertex name cannot be None, so:

sage: G = Graph()
sage: G.add_edge(None, 4)
sage: G.vertices()
[0, 4]
add_edges(edges)

Add edges from an iterable container.

EXAMPLES:

sage: G = graphs.DodecahedralGraph()
sage: H = Graph()
sage: H.add_edges( G.edge_iterator() ); H
Graph on 20 vertices
sage: G = graphs.DodecahedralGraph().to_directed()
sage: H = DiGraph()
sage: H.add_edges( G.edge_iterator() ); H
Digraph on 20 vertices
add_path(vertices)

Adds a cycle to the graph with the given vertices. If the vertices are already present, only the edges are added.

For digraphs, adds the directed path vertices[0], ..., vertices[-1].

INPUT:

  • vertices - a list of indices for the vertices of the cycle to be added.

EXAMPLES:

sage: G = Graph()
sage: G.add_vertices(range(10)); G
Graph on 10 vertices
sage: show(G)
sage: G.add_path(range(20)[10:20])
sage: show(G)
sage: G.add_path(range(10))
sage: show(G)
sage: D = DiGraph()
sage: D.add_path(range(4))
sage: D.edges()
[(0, 1, None), (1, 2, None), (2, 3, None)]
add_vertex(name=None)

Creates an isolated vertex. If the vertex already exists, then nothing is done.

INPUT:

  • name - Name of the new vertex. If no name is specified, then the vertex will be represented by the least integer not already representing a vertex. Name must be an immutable object, and cannot be None.

As it is implemented now, if a graph G has a large number of vertices with numeric labels, then G.add_vertex() could potentially be slow, if name is None.

OUTPUT:

If name``=``None, the new vertex name is returned. None otherwise.

EXAMPLES:

sage: G = Graph(); G.add_vertex(); G
0
Graph on 1 vertex
sage: D = DiGraph(); D.add_vertex(); D
0
Digraph on 1 vertex
add_vertices(vertices)

Add vertices to the (di)graph from an iterable container of vertices. Vertices that already exist in the graph will not be added again.

INPUT:

  • vertices: iterator of vertex labels. A new label is created, used and returned in the output list for all None values in vertices.

OUTPUT:

Generated names of new vertices if there is at least one None value present in vertices. None otherwise.

EXAMPLES:

sage: d = {0: [1,4,5], 1: [2,6], 2: [3,7], 3: [4,8], 4: [9], 5: [7,8], 6: [8,9], 7: [9]}
sage: G = Graph(d)
sage: G.add_vertices([10,11,12])
sage: G.vertices()
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
sage: G.add_vertices(graphs.CycleGraph(25).vertices())
sage: G.vertices()
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]
sage: G = Graph()
sage: G.add_vertices([1,2,3])
sage: G.add_vertices([4,None,None,5])
[0, 6]
adjacency_matrix(sparse=None, boundary_first=False)

Returns the adjacency matrix of the (di)graph.

Each vertex is represented by its position in the list returned by the vertices() function.

The matrix returned is over the integers. If a different ring is desired, use either the change_ring function or the matrix function.

INPUT:

  • sparse - whether to represent with a sparse matrix
  • boundary_first - whether to represent the boundary vertices in the upper left block

EXAMPLES:

sage: G = graphs.CubeGraph(4)
sage: G.adjacency_matrix()
[0 1 1 0 1 0 0 0 1 0 0 0 0 0 0 0]
[1 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0]
[1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0]
[0 1 1 0 0 0 0 1 0 0 0 1 0 0 0 0]
[1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0]
[0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 0]
[0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0]
[0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 1]
[1 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0]
[0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0]
[0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0]
[0 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1]
[0 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0]
[0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 1]
[0 0 0 0 0 0 1 0 0 0 1 0 1 0 0 1]
[0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0]
sage: matrix(GF(2),G) # matrix over GF(2)
[0 1 1 0 1 0 0 0 1 0 0 0 0 0 0 0]
[1 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0]
[1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0]
[0 1 1 0 0 0 0 1 0 0 0 1 0 0 0 0]
[1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0]
[0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 0]
[0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0]
[0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 1]
[1 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0]
[0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0]
[0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0]
[0 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1]
[0 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0]
[0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 1]
[0 0 0 0 0 0 1 0 0 0 1 0 1 0 0 1]
[0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0]
sage: D = DiGraph( { 0: [1,2,3], 1: [0,2], 2: [3], 3: [4], 4: [0,5], 5: [1] } )
sage: D.adjacency_matrix()
[0 1 1 1 0 0]
[1 0 1 0 0 0]
[0 0 0 1 0 0]
[0 0 0 0 1 0]
[1 0 0 0 0 1]
[0 1 0 0 0 0]

TESTS:

sage: graphs.CubeGraph(8).adjacency_matrix().parent()
Full MatrixSpace of 256 by 256 dense matrices over Integer Ring
sage: graphs.CubeGraph(9).adjacency_matrix().parent()
Full MatrixSpace of 512 by 512 sparse matrices over Integer Ring
all_paths(start, end)

Returns a list of all paths (also lists) between a pair of vertices (start, end) in the (di)graph. If start is the same vertex as end, then [[start]] is returned – a list containing the 1-vertex, 0-edge path “start”.

EXAMPLES:

sage: eg1 = Graph({0:[1,2], 1:[4], 2:[3,4], 4:[5], 5:[6]})
sage: eg1.all_paths(0,6)
[[0, 1, 4, 5, 6], [0, 2, 4, 5, 6]]
sage: eg2 = graphs.PetersenGraph()
sage: sorted(eg2.all_paths(1,4))
[[1, 0, 4],
 [1, 0, 5, 7, 2, 3, 4],
 [1, 0, 5, 7, 2, 3, 8, 6, 9, 4],
 [1, 0, 5, 7, 9, 4],
 [1, 0, 5, 7, 9, 6, 8, 3, 4],
 [1, 0, 5, 8, 3, 2, 7, 9, 4],
 [1, 0, 5, 8, 3, 4],
 [1, 0, 5, 8, 6, 9, 4],
 [1, 0, 5, 8, 6, 9, 7, 2, 3, 4],
 [1, 2, 3, 4],
 [1, 2, 3, 8, 5, 0, 4],
 [1, 2, 3, 8, 5, 7, 9, 4],
 [1, 2, 3, 8, 6, 9, 4],
 [1, 2, 3, 8, 6, 9, 7, 5, 0, 4],
 [1, 2, 7, 5, 0, 4],
 [1, 2, 7, 5, 8, 3, 4],
 [1, 2, 7, 5, 8, 6, 9, 4],
 [1, 2, 7, 9, 4],
 [1, 2, 7, 9, 6, 8, 3, 4],
 [1, 2, 7, 9, 6, 8, 5, 0, 4],
 [1, 6, 8, 3, 2, 7, 5, 0, 4],
 [1, 6, 8, 3, 2, 7, 9, 4],
 [1, 6, 8, 3, 4],
 [1, 6, 8, 5, 0, 4],
 [1, 6, 8, 5, 7, 2, 3, 4],
 [1, 6, 8, 5, 7, 9, 4],
 [1, 6, 9, 4],
 [1, 6, 9, 7, 2, 3, 4],
 [1, 6, 9, 7, 2, 3, 8, 5, 0, 4],
 [1, 6, 9, 7, 5, 0, 4],
 [1, 6, 9, 7, 5, 8, 3, 4]]
sage: dg = DiGraph({0:[1,3], 1:[3], 2:[0,3]})
sage: sorted(dg.all_paths(0,3))
[[0, 1, 3], [0, 3]]
sage: ug = dg.to_undirected()
sage: sorted(ug.all_paths(0,3))
[[0, 1, 3], [0, 2, 3], [0, 3]]

Starting and ending at the same vertex (see trac ticket #13006):

sage: graphs.CompleteGraph(4).all_paths(2,2)
[[2]]
allow_loops(new, check=True)

Changes whether loops are permitted in the (di)graph.

INPUT:

  • new - boolean.

EXAMPLES:

sage: G = Graph(loops=True); G
Looped graph on 0 vertices
sage: G.has_loops()
False
sage: G.allows_loops()
True
sage: G.add_edge((0,0))
sage: G.has_loops()
True
sage: G.loops()
[(0, 0, None)]
sage: G.allow_loops(False); G
Graph on 1 vertex
sage: G.has_loops()
False
sage: G.edges()
[]

sage: D = DiGraph(loops=True); D
Looped digraph on 0 vertices
sage: D.has_loops()
False
sage: D.allows_loops()
True
sage: D.add_edge((0,0))
sage: D.has_loops()
True
sage: D.loops()
[(0, 0, None)]
sage: D.allow_loops(False); D
Digraph on 1 vertex
sage: D.has_loops()
False
sage: D.edges()
[]
allow_multiple_edges(new, check=True)

Changes whether multiple edges are permitted in the (di)graph.

INPUT:

  • new - boolean.

EXAMPLES:

sage: G = Graph(multiedges=True,sparse=True); G
Multi-graph on 0 vertices
sage: G.has_multiple_edges()
False
sage: G.allows_multiple_edges()
True
sage: G.add_edges([(0,1)]*3)
sage: G.has_multiple_edges()
True
sage: G.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: G.allow_multiple_edges(False); G
Graph on 2 vertices
sage: G.has_multiple_edges()
False
sage: G.edges()
[(0, 1, None)]

sage: D = DiGraph(multiedges=True,sparse=True); D
Multi-digraph on 0 vertices
sage: D.has_multiple_edges()
False
sage: D.allows_multiple_edges()
True
sage: D.add_edges([(0,1)]*3)
sage: D.has_multiple_edges()
True
sage: D.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: D.allow_multiple_edges(False); D
Digraph on 2 vertices
sage: D.has_multiple_edges()
False
sage: D.edges()
[(0, 1, None)]
allows_loops()

Returns whether loops are permitted in the (di)graph.

EXAMPLES:

sage: G = Graph(loops=True); G
Looped graph on 0 vertices
sage: G.has_loops()
False
sage: G.allows_loops()
True
sage: G.add_edge((0,0))
sage: G.has_loops()
True
sage: G.loops()
[(0, 0, None)]
sage: G.allow_loops(False); G
Graph on 1 vertex
sage: G.has_loops()
False
sage: G.edges()
[]

sage: D = DiGraph(loops=True); D
Looped digraph on 0 vertices
sage: D.has_loops()
False
sage: D.allows_loops()
True
sage: D.add_edge((0,0))
sage: D.has_loops()
True
sage: D.loops()
[(0, 0, None)]
sage: D.allow_loops(False); D
Digraph on 1 vertex
sage: D.has_loops()
False
sage: D.edges()
[]
allows_multiple_edges()

Returns whether multiple edges are permitted in the (di)graph.

EXAMPLES:

sage: G = Graph(multiedges=True,sparse=True); G
Multi-graph on 0 vertices
sage: G.has_multiple_edges()
False
sage: G.allows_multiple_edges()
True
sage: G.add_edges([(0,1)]*3)
sage: G.has_multiple_edges()
True
sage: G.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: G.allow_multiple_edges(False); G
Graph on 2 vertices
sage: G.has_multiple_edges()
False
sage: G.edges()
[(0, 1, None)]

sage: D = DiGraph(multiedges=True,sparse=True); D
Multi-digraph on 0 vertices
sage: D.has_multiple_edges()
False
sage: D.allows_multiple_edges()
True
sage: D.add_edges([(0,1)]*3)
sage: D.has_multiple_edges()
True
sage: D.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: D.allow_multiple_edges(False); D
Digraph on 2 vertices
sage: D.has_multiple_edges()
False
sage: D.edges()
[(0, 1, None)]
am(sparse=None, boundary_first=False)

Returns the adjacency matrix of the (di)graph.

Each vertex is represented by its position in the list returned by the vertices() function.

The matrix returned is over the integers. If a different ring is desired, use either the change_ring function or the matrix function.

INPUT:

  • sparse - whether to represent with a sparse matrix
  • boundary_first - whether to represent the boundary vertices in the upper left block

EXAMPLES:

sage: G = graphs.CubeGraph(4)
sage: G.adjacency_matrix()
[0 1 1 0 1 0 0 0 1 0 0 0 0 0 0 0]
[1 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0]
[1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0]
[0 1 1 0 0 0 0 1 0 0 0 1 0 0 0 0]
[1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0]
[0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 0]
[0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0]
[0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 1]
[1 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0]
[0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0]
[0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0]
[0 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1]
[0 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0]
[0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 1]
[0 0 0 0 0 0 1 0 0 0 1 0 1 0 0 1]
[0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0]
sage: matrix(GF(2),G) # matrix over GF(2)
[0 1 1 0 1 0 0 0 1 0 0 0 0 0 0 0]
[1 0 0 1 0 1 0 0 0 1 0 0 0 0 0 0]
[1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0]
[0 1 1 0 0 0 0 1 0 0 0 1 0 0 0 0]
[1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0]
[0 1 0 0 1 0 0 1 0 0 0 0 0 1 0 0]
[0 0 1 0 1 0 0 1 0 0 0 0 0 0 1 0]
[0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 1]
[1 0 0 0 0 0 0 0 0 1 1 0 1 0 0 0]
[0 1 0 0 0 0 0 0 1 0 0 1 0 1 0 0]
[0 0 1 0 0 0 0 0 1 0 0 1 0 0 1 0]
[0 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1]
[0 0 0 0 1 0 0 0 1 0 0 0 0 1 1 0]
[0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 1]
[0 0 0 0 0 0 1 0 0 0 1 0 1 0 0 1]
[0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 0]
sage: D = DiGraph( { 0: [1,2,3], 1: [0,2], 2: [3], 3: [4], 4: [0,5], 5: [1] } )
sage: D.adjacency_matrix()
[0 1 1 1 0 0]
[1 0 1 0 0 0]
[0 0 0 1 0 0]
[0 0 0 0 1 0]
[1 0 0 0 0 1]
[0 1 0 0 0 0]

TESTS:

sage: graphs.CubeGraph(8).adjacency_matrix().parent()
Full MatrixSpace of 256 by 256 dense matrices over Integer Ring
sage: graphs.CubeGraph(9).adjacency_matrix().parent()
Full MatrixSpace of 512 by 512 sparse matrices over Integer Ring
antisymmetric()

Tests whether the graph is antisymmetric.

A graph represents an antisymmetric relation if there being a path from a vertex x to a vertex y implies that there is not a path from y to x unless x=y.

A directed acyclic graph is antisymmetric. An undirected graph is never antisymmetric unless it is just a union of isolated vertices.

sage: graphs.RandomGNP(20,0.5).antisymmetric()
False
sage: digraphs.RandomDirectedGNR(20,0.5).antisymmetric()
True
automorphism_group(partition=None, verbosity=0, edge_labels=False, order=False, return_group=True, orbits=False)

Returns the largest subgroup of the automorphism group of the (di)graph whose orbit partition is finer than the partition given. If no partition is given, the unit partition is used and the entire automorphism group is given.

INPUT:

  • partition - default is the unit partition, otherwise computes the subgroup of the full automorphism group respecting the partition.
  • edge_labels - default False, otherwise allows only permutations respecting edge labels.
  • order - (default False) if True, compute the order of the automorphism group
  • return_group - default True
  • orbits - returns the orbits of the group acting on the vertices of the graph

Warning

Since trac ticket #14319 the domain of the automorphism group is equal to the graph’s vertex set, and the translation argument has become useless.

OUTPUT: The order of the output is group, order, orbits. However, there are options to turn each of these on or off.

EXAMPLES:

Graphs:

sage: graphs_query = GraphQuery(display_cols=['graph6'],num_vertices=4)
sage: L = graphs_query.get_graphs_list()
sage: graphs_list.show_graphs(L)
sage: for g in L:
...    G = g.automorphism_group()
...    G.order(), G.gens()
(24, [(2,3), (1,2), (0,1)])
(4, [(2,3), (0,1)])
(2, [(1,2)])
(6, [(1,2), (0,1)])
(6, [(2,3), (1,2)])
(8, [(1,2), (0,1)(2,3)])
(2, [(0,1)(2,3)])
(2, [(1,2)])
(8, [(2,3), (0,1), (0,2)(1,3)])
(4, [(2,3), (0,1)])
(24, [(2,3), (1,2), (0,1)])
sage: C = graphs.CubeGraph(4)
sage: G = C.automorphism_group()
sage: M = G.character_table() # random order of rows, thus abs() below
sage: QQ(M.determinant()).abs()
712483534798848
sage: G.order()
384
sage: D = graphs.DodecahedralGraph()
sage: G = D.automorphism_group()
sage: A5 = AlternatingGroup(5)
sage: Z2 = CyclicPermutationGroup(2)
sage: H = A5.direct_product(Z2)[0] #see documentation for direct_product to explain the [0]
sage: G.is_isomorphic(H)
True

Multigraphs:

sage: G = Graph(multiedges=True,sparse=True)
sage: G.add_edge(('a', 'b'))
sage: G.add_edge(('a', 'b'))
sage: G.add_edge(('a', 'b'))
sage: G.automorphism_group()
Permutation Group with generators [('a','b')]

Digraphs:

sage: D = DiGraph( { 0:[1], 1:[2], 2:[3], 3:[4], 4:[0] } )
sage: D.automorphism_group()
Permutation Group with generators [(0,1,2,3,4)]

Edge labeled graphs:

sage: G = Graph(sparse=True)
sage: G.add_edges( [(0,1,'a'),(1,2,'b'),(2,3,'c'),(3,4,'b'),(4,0,'a')] )
sage: G.automorphism_group(edge_labels=True)
Permutation Group with generators [(1,4)(2,3)]
sage: G = Graph({0 : {1 : 7}})
sage: G.automorphism_group(edge_labels=True)
Permutation Group with generators [(0,1)]

sage: foo = Graph(sparse=True)
sage: bar = Graph(implementation='c_graph',sparse=True)
sage: foo.add_edges([(0,1,1),(1,2,2), (2,3,3)])
sage: bar.add_edges([(0,1,1),(1,2,2), (2,3,3)])
sage: foo.automorphism_group(edge_labels=True)
Permutation Group with generators [()]
sage: foo.automorphism_group()
Permutation Group with generators [(0,3)(1,2)]
sage: bar.automorphism_group(edge_labels=True)
Permutation Group with generators [()]

You can also ask for just the order of the group:

sage: G = graphs.PetersenGraph()
sage: G.automorphism_group(return_group=False, order=True)
120

Or, just the orbits (note that each graph here is vertex transitive)

sage: G = graphs.PetersenGraph()
sage: G.automorphism_group(return_group=False, orbits=True)
[[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]]
sage: G.automorphism_group(partition=[[0],range(1,10)], return_group=False, orbits=True)
[[0], [2, 3, 6, 7, 8, 9], [1, 4, 5]]
sage: C = graphs.CubeGraph(3)
sage: C.automorphism_group(orbits=True, return_group=False)
[['000', '001', '010', '011', '100', '101', '110', '111']]

TESTS:

We get a KeyError when given an invalid partition (trac #6087):

sage: g=graphs.CubeGraph(3)
sage: g.relabel()
sage: g.automorphism_group(partition=[[0,1,2],[3,4,5]])
Traceback (most recent call last):
...
KeyError: 6

Labeled automorphism group:

sage: digraphs.DeBruijn(3,2).automorphism_group()
Permutation Group with generators [('01','02')('10','20')('11','22')('12','21'), ('00','11')('01','10')('02','12')('20','21')]
sage: d = digraphs.DeBruijn(3,2)
sage: d.allow_multiple_edges(True)
sage: d.add_edge(d.edges()[0])
sage: d.automorphism_group()
Permutation Group with generators [('01','02')('10','20')('11','22')('12','21')]

The labeling is correct:

sage: g = graphs.PetersenGraph()
sage: ag = g.automorphism_group()
sage: for u,v in g.edges(labels = False):
...       if len(ag.orbit((u,v),action="OnPairs")) != 30:
...           print "ARggggggggggggg !!!"

Empty group, correct domain:

sage: Graph({'a':['a'], 'b':[]}).automorphism_group()
Permutation Group with generators [()]
sage: Graph({'a':['a'], 'b':[]}).automorphism_group().domain()
{'a', 'b'}
average_degree()

Returns the average degree of the graph.

The average degree of a graph G=(V,E) is equal to \frac {2|E|}{|V|}.

EXAMPLES:

The average degree of a regular graph is equal to the degree of any vertex:

sage: g = graphs.CompleteGraph(5)
sage: g.average_degree() == 4
True

The average degree of a tree is always strictly less than 2:

sage: g = graphs.RandomGNP(20,.5)
sage: tree = Graph()
sage: tree.add_edges(g.min_spanning_tree())
sage: tree.average_degree() < 2
True

For any graph, it is equal to \frac {2|E|}{|V|}:

sage: g = graphs.RandomGNP(50,.8)
sage: g.average_degree() == 2*g.size()/g.order()
True
average_distance()

Returns the average distance between vertices of the graph.

Formally, for a graph G this value is equal to \frac 1 {n(n-1)} \sum_{u,v\in G} d(u,v) where d(u,v) denotes the distance between vertices u and v and n is the number of vertices in G.

EXAMPLE:

From [GYLL93]:

sage: g=graphs.PathGraph(10)
sage: w=lambda x: (x*(x*x -1)/6)/(x*(x-1)/2)
sage: g.average_distance()==w(10)
True

REFERENCE:

[GYLL93]I. Gutman, Y.-N. Yeh, S.-L. Lee, and Y.-L. Luo. Some recent results in the theory of the Wiener number. Indian Journal of Chemistry, 32A:651–661, 1993.

TEST:

sage: g = Graph()
sage: g.average_distance()
Traceback (most recent call last):
...
ValueError: The graph must have at least two vertices for this value to be defined
blocks_and_cut_vertices()

Computes the blocks and cut vertices of the graph.

In the case of a digraph, this computation is done on the underlying graph.

A cut vertex is one whose deletion increases the number of connected components. A block is a maximal induced subgraph which itself has no cut vertices. Two distinct blocks cannot overlap in more than a single cut vertex.

OUTPUT: ( B, C ), where B is a list of blocks- each is a list of vertices and the blocks are the corresponding induced subgraphs-and C is a list of cut vertices.

ALGORITHM:

We implement the algorithm proposed by Tarjan in [Tarjan72]. The original version is recursive. We emulate the recursion using a stack.

EXAMPLES:

sage: graphs.PetersenGraph().blocks_and_cut_vertices()
([[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]], [])
sage: graphs.PathGraph(6).blocks_and_cut_vertices()
([[4, 5], [3, 4], [2, 3], [1, 2], [0, 1]], [1, 2, 3, 4])
sage: graphs.CycleGraph(7).blocks_and_cut_vertices()
([[0, 1, 2, 3, 4, 5, 6]], [])
sage: graphs.KrackhardtKiteGraph().blocks_and_cut_vertices()
([[8, 9], [7, 8], [0, 1, 2, 3, 4, 5, 6, 7]], [7, 8])
sage: G=Graph()  # make a bowtie graph where 0 is a cut vertex
sage: G.add_vertices(range(5))
sage: G.add_edges([(0,1),(0,2),(0,3),(0,4),(1,2),(3,4)])
sage: G.blocks_and_cut_vertices()
([[0, 1, 2], [0, 3, 4]], [0])
sage: graphs.StarGraph(3).blocks_and_cut_vertices()
([[0, 1], [0, 2], [0, 3]], [0])

TESTS:

sage: Graph(0).blocks_and_cut_vertices()
([], [])
sage: Graph(1).blocks_and_cut_vertices()
([[0]], [])
sage: Graph(2).blocks_and_cut_vertices()
Traceback (most recent call last):
...
NotImplementedError: ...

REFERENCE:

[Tarjan72]R.E. Tarjan. Depth-First Search and Linear Graph Algorithms. SIAM J. Comput. 1(2): 146-160 (1972).
blocks_and_cuts_tree()

Returns the blocks-and-cuts tree of self.

This new graph has two different kinds of vertices, some representing the blocks (type B) and some other the cut vertices of the graph self (type C).

There is an edge between a vertex u of type B and a vertex v of type C if the cut-vertex corresponding to v is in the block corresponding to u.

The resulting graph is a tree, with the additional characteristic property that the distance between two leaves is even.

EXAMPLES:

sage: T = graphs.KrackhardtKiteGraph().blocks_and_cuts_tree(); T
Graph on 5 vertices
sage: T.is_isomorphic(graphs.PathGraph(5))
True

sage: T = graphs.RandomTree(40).blocks_and_cuts_tree()
sage: T.is_tree()
True
sage: leaves = [v for v in T if T.degree(v) == 1]
sage: all(T.distance(u,v) % 2 == 0 for u in leaves for v in leaves)
True

REFERENCES:

[HarPri]F. Harary and G. Prins. The block-cutpoint-tree of a graph. Publ. Math. Debrecen 13 1966 103-107.
[Gallai]T. Gallai, Elementare Relationen bezueglich der Glieder und trennenden Punkte von Graphen, Magyar Tud. Akad. Mat. Kutato Int. Kozl. 9 (1964) 235-236

Returns an iterator over the vertices in a breadth-first ordering.

INPUT:

  • start - vertex or list of vertices from which to start the traversal
  • ignore_direction - (default False) only applies to directed graphs. If True, searches across edges in either direction.
  • distance - the maximum distance from the start nodes to traverse. The start nodes are distance zero from themselves.
  • neighbors - a function giving the neighbors of a vertex. The function should take a vertex and return a list of vertices. For a graph, neighbors is by default the neighbors() function of the graph. For a digraph, the neighbors function defaults to the successors() function of the graph.

See also

EXAMPLES:

sage: G = Graph( { 0: [1], 1: [2], 2: [3], 3: [4], 4: [0]} )
sage: list(G.breadth_first_search(0))
[0, 1, 4, 2, 3]

By default, the edge direction of a digraph is respected, but this can be overridden by the ignore_direction parameter:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.breadth_first_search(0))
[0, 1, 2, 3, 4, 5, 6, 7]
sage: list(D.breadth_first_search(0, ignore_direction=True))
[0, 1, 2, 3, 7, 4, 5, 6]

You can specify a maximum distance in which to search. A distance of zero returns the start vertices:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.breadth_first_search(0,distance=0))
[0]
sage: list(D.breadth_first_search(0,distance=1))
[0, 1, 2, 3]

Multiple starting vertices can be specified in a list:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.breadth_first_search([0]))
[0, 1, 2, 3, 4, 5, 6, 7]
sage: list(D.breadth_first_search([0,6]))
[0, 6, 1, 2, 3, 7, 4, 5]
sage: list(D.breadth_first_search([0,6],distance=0))
[0, 6]
sage: list(D.breadth_first_search([0,6],distance=1))
[0, 6, 1, 2, 3, 7]
sage: list(D.breadth_first_search(6,ignore_direction=True,distance=2))
[6, 3, 7, 0, 5]

More generally, you can specify a neighbors function. For example, you can traverse the graph backwards by setting neighbors to be the neighbors_in() function of the graph:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.breadth_first_search(5,neighbors=D.neighbors_in, distance=2))
[5, 1, 2, 0]
sage: list(D.breadth_first_search(5,neighbors=D.neighbors_out, distance=2))
[5, 7, 0]
sage: list(D.breadth_first_search(5,neighbors=D.neighbors, distance=2))
[5, 1, 2, 7, 0, 4, 6]

TESTS:

sage: D = DiGraph({1:[0], 2:[0]})
sage: list(D.breadth_first_search(0))
[0]
sage: list(D.breadth_first_search(0, ignore_direction=True))
[0, 1, 2]
canonical_label(partition=None, certify=False, verbosity=0, edge_labels=False)

Returns the unique graph on \{0,1,...,n-1\} ( n = self.order() ) which

  • is isomorphic to self,
  • is invariant in the isomorphism class.

In other words, given two graphs G and H which are isomorphic, suppose G_c and H_c are the graphs returned by canonical_label. Then the following hold:

  • G_c == H_c
  • G_c.adjacency_matrix() == H_c.adjacency_matrix()
  • G_c.graph6_string() == H_c.graph6_string()

INPUT:

  • partition - if given, the canonical label with respect to this set partition will be computed. The default is the unit set partition.
  • certify - if True, a dictionary mapping from the (di)graph to its canonical label will be given.
  • verbosity - gets passed to nice: prints helpful output.
  • edge_labels - default False, otherwise allows only permutations respecting edge labels.

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: E = D.canonical_label(); E
Dodecahedron: Graph on 20 vertices
sage: D.canonical_label(certify=True)
(Dodecahedron: Graph on 20 vertices, {0: 0, 1: 19, 2: 16, 3: 15, 4: 9, 5: 1, 6: 10, 7: 8, 8: 14, 9: 12, 10: 17, 11: 11, 12: 5, 13: 6, 14: 2, 15: 4, 16: 3, 17: 7, 18: 13, 19: 18})
sage: D.is_isomorphic(E)
True

Multigraphs:

sage: G = Graph(multiedges=True,sparse=True)
sage: G.add_edge((0,1))
sage: G.add_edge((0,1))
sage: G.add_edge((0,1))
sage: G.canonical_label()
Multi-graph on 2 vertices
sage: Graph('A?', implementation='c_graph').canonical_label()
Graph on 2 vertices

Digraphs:

sage: P = graphs.PetersenGraph()
sage: DP = P.to_directed()
sage: DP.canonical_label().adjacency_matrix()
[0 0 0 0 0 0 0 1 1 1]
[0 0 0 0 1 0 1 0 0 1]
[0 0 0 1 0 0 1 0 1 0]
[0 0 1 0 0 1 0 0 0 1]
[0 1 0 0 0 1 0 0 1 0]
[0 0 0 1 1 0 0 1 0 0]
[0 1 1 0 0 0 0 1 0 0]
[1 0 0 0 0 1 1 0 0 0]
[1 0 1 0 1 0 0 0 0 0]
[1 1 0 1 0 0 0 0 0 0]

Edge labeled graphs:

sage: G = Graph(sparse=True)
sage: G.add_edges( [(0,1,'a'),(1,2,'b'),(2,3,'c'),(3,4,'b'),(4,0,'a')] )
sage: G.canonical_label(edge_labels=True)
Graph on 5 vertices
sage: G.canonical_label(edge_labels=True,certify=True)
(Graph on 5 vertices, {0: 4, 1: 3, 2: 0, 3: 1, 4: 2})
cartesian_product(other)

Returns the Cartesian product of self and other.

The Cartesian product of G and H is the graph L with vertex set V(L) equal to the Cartesian product of the vertices V(G) and V(H), and ((u,v), (w,x)) is an edge iff either - (u, w) is an edge of self and v = x, or - (v, x) is an edge of other and u = w.

See also

TESTS:

Cartesian product of graphs:

sage: G = Graph([(0,1),(1,2)])
sage: H = Graph([('a','b')])
sage: C1 = G.cartesian_product(H)
sage: C1.edges(labels=None)
[((0, 'a'), (0, 'b')), ((0, 'a'), (1, 'a')), ((0, 'b'), (1, 'b')), ((1, 'a'), (1, 'b')), ((1, 'a'), (2, 'a')), ((1, 'b'), (2, 'b')), ((2, 'a'), (2, 'b'))]
sage: C2 = H.cartesian_product(G)
sage: C1.is_isomorphic(C2)
True

Construction of a Toroidal grid:

sage: A = graphs.CycleGraph(3)
sage: B = graphs.CycleGraph(4)
sage: T = A.cartesian_product(B)
sage: T.is_isomorphic( graphs.ToroidalGrid2dGraph(3,4) )
True

Cartesian product of digraphs:

sage: P = DiGraph([(0,1)])
sage: B = digraphs.DeBruijn( ['a','b'], 2 )
sage: Q = P.cartesian_product(B)
sage: Q.edges(labels=None)
[((0, 'aa'), (0, 'aa')), ((0, 'aa'), (0, 'ab')), ((0, 'aa'), (1, 'aa')), ((0, 'ab'), (0, 'ba')), ((0, 'ab'), (0, 'bb')), ((0, 'ab'), (1, 'ab')), ((0, 'ba'), (0, 'aa')), ((0, 'ba'), (0, 'ab')), ((0, 'ba'), (1, 'ba')), ((0, 'bb'), (0, 'ba')), ((0, 'bb'), (0, 'bb')), ((0, 'bb'), (1, 'bb')), ((1, 'aa'), (1, 'aa')), ((1, 'aa'), (1, 'ab')), ((1, 'ab'), (1, 'ba')), ((1, 'ab'), (1, 'bb')), ((1, 'ba'), (1, 'aa')), ((1, 'ba'), (1, 'ab')), ((1, 'bb'), (1, 'ba')), ((1, 'bb'), (1, 'bb'))]
sage: Q.strongly_connected_components_digraph().num_verts()
2
sage: V = Q.strongly_connected_component_containing_vertex( (0, 'aa') )
sage: B.is_isomorphic( Q.subgraph(V) )
True
categorical_product(other)

Returns the tensor product of self and other.

The tensor product of G and H is the graph L with vertex set V(L) equal to the Cartesian product of the vertices V(G) and V(H), and ((u,v), (w,x)) is an edge iff - (u, w) is an edge of self, and - (v, x) is an edge of other.

The tensor product is also known as the categorical product and the kronecker product (refering to the kronecker matrix product). See Wikipedia article on the Kronecker product.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: C = graphs.CycleGraph(5)
sage: T = C.tensor_product(Z); T
Graph on 10 vertices
sage: T.size()
10
sage: T.plot() # long time
sage: D = graphs.DodecahedralGraph()
sage: P = graphs.PetersenGraph()
sage: T = D.tensor_product(P); T
Graph on 200 vertices
sage: T.size()
900
sage: T.plot() # long time

TESTS:

Tensor product of graphs:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.tensor_product(H)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((0, 'b'), (1, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a'))]
sage: T.is_isomorphic( H.tensor_product(G) )
True

Tensor product of digraphs:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.tensor_product(J)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((1, 'a'), (2, 'b'))]
sage: T.is_isomorphic( J.tensor_product(I) )
True

The tensor product of two DeBruijn digraphs of same diameter is a DeBruijn digraph:

sage: B1 = digraphs.DeBruijn(2, 3)
sage: B2 = digraphs.DeBruijn(3, 3)
sage: T = B1.tensor_product( B2 )
sage: T.is_isomorphic( digraphs.DeBruijn( 2*3, 3) )
True
center()

Returns the set of vertices in the center, i.e. whose eccentricity is equal to the radius of the (di)graph.

In other words, the center is the set of vertices achieving the minimum eccentricity.

EXAMPLES:

sage: G = graphs.DiamondGraph()
sage: G.center()
[1, 2]
sage: P = graphs.PetersenGraph()
sage: P.subgraph(P.center()) == P
True
sage: S = graphs.StarGraph(19)
sage: S.center()
[0]
sage: G = Graph()
sage: G.center()
[]
sage: G.add_vertex()
0
sage: G.center()
[0]
characteristic_polynomial(var='x', laplacian=False)

Returns the characteristic polynomial of the adjacency matrix of the (di)graph.

Let G be a (simple) graph with adjacency matrix A. Let I be the identity matrix of dimensions the same as A. The characteristic polynomial of G is defined as the determinant \det(xI - A).

Note

characteristic_polynomial and charpoly are aliases and thus provide exactly the same method.

INPUT:

  • x – (default: 'x') the variable of the characteristic polynomial.
  • laplacian – (default: False) if True, use the Laplacian matrix.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.characteristic_polynomial()
x^10 - 15*x^8 + 75*x^6 - 24*x^5 - 165*x^4 + 120*x^3 + 120*x^2 - 160*x + 48
sage: P.charpoly()
x^10 - 15*x^8 + 75*x^6 - 24*x^5 - 165*x^4 + 120*x^3 + 120*x^2 - 160*x + 48
sage: P.characteristic_polynomial(laplacian=True)
x^10 - 30*x^9 + 390*x^8 - 2880*x^7 + 13305*x^6 - 
39882*x^5 + 77640*x^4 - 94800*x^3 + 66000*x^2 - 20000*x
charpoly(var='x', laplacian=False)

Returns the characteristic polynomial of the adjacency matrix of the (di)graph.

Let G be a (simple) graph with adjacency matrix A. Let I be the identity matrix of dimensions the same as A. The characteristic polynomial of G is defined as the determinant \det(xI - A).

Note

characteristic_polynomial and charpoly are aliases and thus provide exactly the same method.

INPUT:

  • x – (default: 'x') the variable of the characteristic polynomial.
  • laplacian – (default: False) if True, use the Laplacian matrix.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.characteristic_polynomial()
x^10 - 15*x^8 + 75*x^6 - 24*x^5 - 165*x^4 + 120*x^3 + 120*x^2 - 160*x + 48
sage: P.charpoly()
x^10 - 15*x^8 + 75*x^6 - 24*x^5 - 165*x^4 + 120*x^3 + 120*x^2 - 160*x + 48
sage: P.characteristic_polynomial(laplacian=True)
x^10 - 30*x^9 + 390*x^8 - 2880*x^7 + 13305*x^6 - 
39882*x^5 + 77640*x^4 - 94800*x^3 + 66000*x^2 - 20000*x
check_embedding_validity(embedding=None)

Checks whether an _embedding attribute is well defined.

If the _embedding attribute exists, it is checked for accuracy. Returns True if everything is okay, False otherwise.

If embedding=None will test the attribute _embedding.

EXAMPLES:

sage: d = {0: [1, 5, 4], 1: [0, 2, 6], 2: [1, 3, 7], 3: [8, 2, 4], 4: [0, 9, 3], 5: [0, 8, 7], 6: [8, 1, 9], 7: [9, 2, 5], 8: [3, 5, 6], 9: [4, 6, 7]}
sage: G = graphs.PetersenGraph()
sage: G.check_embedding_validity(d)
True
check_pos_validity(pos=None, dim=2)

Checks whether pos specifies two (resp. 3) coordinates for every vertex (and no more vertices).

INPUT:

  • pos - a position dictionary for a set of vertices
  • dim - 2 or 3 (default: 3

OUTPUT:

If pos is None then the position dictionary of self is investigated, otherwise the position dictionary provided in pos is investigated. The function returns True if the dictionary is of the correct form for self.

EXAMPLES:

sage: p = {0: [1, 5], 1: [0, 2], 2: [1, 3], 3: [8, 2], 4: [0, 9], 5: [0, 8], 6: [8, 1], 7: [9, 5], 8: [3, 5], 9: [6, 7]}
sage: G = graphs.PetersenGraph()
sage: G.check_pos_validity(p)
True
clear()

Empties the graph of vertices and edges and removes name, boundary, associated objects, and position information.

EXAMPLES:

sage: G=graphs.CycleGraph(4); G.set_vertices({0:'vertex0'})
sage: G.order(); G.size()
4
4
sage: len(G._pos)
4
sage: G.name()
'Cycle graph'
sage: G.get_vertex(0)
'vertex0'
sage: H = G.copy(implementation='c_graph', sparse=True)
sage: H.clear()
sage: H.order(); H.size()
0
0
sage: len(H._pos)
0
sage: H.name()
''
sage: H.get_vertex(0)
sage: H = G.copy(implementation='c_graph', sparse=False)
sage: H.clear()
sage: H.order(); H.size()
0
0
sage: len(H._pos)
0
sage: H.name()
''
sage: H.get_vertex(0)
sage: H = G.copy(implementation='networkx')
sage: H.clear()
sage: H.order(); H.size()
0
0
sage: len(H._pos)
0
sage: H.name()
''
sage: H.get_vertex(0)
cluster_transitivity()

Returns the transitivity (fraction of transitive triangles) of the graph.

Transitivity is the fraction of all existing triangles and all connected triples (triads), T = 3\times\text{triangles}
/ \text{triads}.

See also section “Clustering” in chapter “Algorithms” of [HSSNX].

EXAMPLES:

sage: (graphs.FruchtGraph()).cluster_transitivity()
0.25
cluster_triangles(nbunch=None, with_labels=False)

Returns the number of triangles for the set nbunch of vertices as a dictionary keyed by vertex.

See also section “Clustering” in chapter “Algorithms” of [HSSNX].

INPUT:

  • nbunch - The vertices to inspect. If nbunch=None, returns data for all vertices in the graph.

REFERENCE:

[HSSNX](1, 2, 3, 4) Aric Hagberg, Dan Schult and Pieter Swart. NetworkX documentation. [Online] Available: http://networkx.github.io/documentation/latest/reference/index.html

EXAMPLES:

sage: (graphs.FruchtGraph()).cluster_triangles().values()
[1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0]
sage: (graphs.FruchtGraph()).cluster_triangles()
{0: 1, 1: 1, 2: 0, 3: 1, 4: 1, 5: 1, 6: 1, 7: 1, 8: 0, 9: 1, 10: 1, 11: 0}
sage: (graphs.FruchtGraph()).cluster_triangles(nbunch=[0,1,2])
{0: 1, 1: 1, 2: 0}
clustering_average()

Returns the average clustering coefficient.

The clustering coefficient of a node i is the fraction of existing triangles containing node i and all possible triangles containing i: c_i = T(i) / \binom {k_i} 2 where T(i) is the number of existing triangles through i, and k_i is the degree of vertex i.

A coefficient for the whole graph is the average of the c_i.

See also section “Clustering” in chapter “Algorithms” of [HSSNX].

EXAMPLES:

sage: (graphs.FruchtGraph()).clustering_average()
0.25
clustering_coeff(nodes=None, weight=False, return_vertex_weights=True)

Returns the clustering coefficient for each vertex in nodes as a dictionary keyed by vertex.

For an unweighted graph, the clustering coefficient of a node i is the fraction of existing triangles containing node i and all possible triangles containing i: c_i = T(i) / \binom {k_i} 2 where T(i) is the number of existing triangles through i, and k_i is the degree of vertex i.

For weighted graphs the clustering is defined as the geometric average of the subgraph edge weights, normalized by the maximum weight in the network.

The value of c_i is assigned 0 if k_i < 2.

See also section “Clustering” in chapter “Algorithms” of [HSSNX].

INPUT:

  • nodes - the vertices to inspect (default None, returns data on all vertices in graph)
  • weight - string or boolean (default is False). If it is a string it used the indicated edge property as weight. weight = True is equivalent to weight = 'weight'
  • return_vertex_weights is a boolean ensuring backwards compatibility with deprecated features of NetworkX 1.2. It should be set to False for all production code.

EXAMPLES:

sage: (graphs.FruchtGraph()).clustering_coeff().values()
[0.3333333333333333, 0.3333333333333333, 0.0, 0.3333333333333333,
 0.3333333333333333, 0.3333333333333333, 0.3333333333333333,
 0.3333333333333333, 0.0, 0.3333333333333333, 0.3333333333333333,
 0.0]
sage: (graphs.FruchtGraph()).clustering_coeff()
{0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0,
 3: 0.3333333333333333, 4: 0.3333333333333333,
 5: 0.3333333333333333, 6: 0.3333333333333333,
 7: 0.3333333333333333, 8: 0.0, 9: 0.3333333333333333,
 10: 0.3333333333333333, 11: 0.0}

sage: (graphs.FruchtGraph()).clustering_coeff(weight=True,
...     return_vertex_weights=False)
{0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0,
3: 0.3333333333333333, 4: 0.3333333333333333,
5: 0.3333333333333333, 6: 0.3333333333333333,
7: 0.3333333333333333, 8: 0.0, 9: 0.3333333333333333,
10: 0.3333333333333333, 11: 0.0}
sage: (graphs.FruchtGraph()).clustering_coeff(nodes=[0,1,2])
{0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0}

sage: (graphs.FruchtGraph()).clustering_coeff(nodes=[0,1,2],
...     weight=True, return_vertex_weights=False)
{0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0}

TESTS:

Doctests that demonstrate the deprecation of the two-dictionary return value due to the NetworkX API change after 1.2. The return_vertex_weights keyword is provided with a default value of True for backwards compatibility with older versions of Sage. When the deprecation period has expired and the keyword is removed, these doctests should be removed as well.

sage: (graphs.FruchtGraph()).clustering_coeff(weight=True,
...     return_vertex_weights=True)
doctest:...: DeprecationWarning: The option 'return_vertex_weights'
has been deprecated. Only offered for backwards compatibility with
NetworkX 1.2.
See http://trac.sagemath.org/12806 for details.
({0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0,
  3: 0.3333333333333333, 4: 0.3333333333333333,
  5: 0.3333333333333333, 6: 0.3333333333333333,
  7: 0.3333333333333333, 8: 0.0, 9: 0.3333333333333333,
  10: 0.3333333333333333, 11: 0.0}, {0: 0.08333333333333333,
  1: 0.08333333333333333, 2: 0.08333333333333333,
  3: 0.08333333333333333, 4: 0.08333333333333333,
  5: 0.08333333333333333, 6: 0.08333333333333333,
  7: 0.08333333333333333, 8: 0.08333333333333333,
  9: 0.08333333333333333, 10: 0.08333333333333333,
  11: 0.08333333333333333})

sage: (graphs.FruchtGraph()).clustering_coeff(nodes=[0, 1, 2],
...     weight=True, return_vertex_weights=True)
({0: 0.3333333333333333, 1: 0.3333333333333333, 2: 0.0},
 {0: 0.3333333333333333, 1: 0.3333333333333333,
 2: 0.3333333333333333})
coarsest_equitable_refinement(partition, sparse=True)

Returns the coarsest partition which is finer than the input partition, and equitable with respect to self.

A partition is equitable with respect to a graph if for every pair of cells C1, C2 of the partition, the number of edges from a vertex of C1 to C2 is the same, over all vertices in C1.

A partition P1 is finer than P2 (P2 is coarser than P1) if every cell of P1 is a subset of a cell of P2.

INPUT:

  • partition - a list of lists
  • sparse - (default False) whether to use sparse or dense representation- for small graphs, use dense for speed

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.coarsest_equitable_refinement([[0],range(1,10)])
[[0], [2, 3, 6, 7, 8, 9], [1, 4, 5]]
sage: G = graphs.CubeGraph(3)
sage: verts = G.vertices()
sage: Pi = [verts[:1], verts[1:]]
sage: Pi
[['000'], ['001', '010', '011', '100', '101', '110', '111']]
sage: G.coarsest_equitable_refinement(Pi)
[['000'], ['011', '101', '110'], ['111'], ['001', '010', '100']]

Note that given an equitable partition, this function returns that partition:

sage: P = graphs.PetersenGraph()
sage: prt = [[0], [1, 4, 5], [2, 3, 6, 7, 8, 9]]
sage: P.coarsest_equitable_refinement(prt)
[[0], [1, 4, 5], [2, 3, 6, 7, 8, 9]]
sage: ss = (graphs.WheelGraph(6)).line_graph(labels=False)
sage: prt = [[(0, 1)], [(0, 2), (0, 3), (0, 4), (1, 2), (1, 4)], [(2, 3), (3, 4)]]
sage: ss.coarsest_equitable_refinement(prt)
Traceback (most recent call last):
...
TypeError: Partition ([[(0, 1)], [(0, 2), (0, 3), (0, 4), (1, 2), (1, 4)], [(2, 3), (3, 4)]]) is not valid for this graph: vertices are incorrect.
sage: ss = (graphs.WheelGraph(5)).line_graph(labels=False)
sage: ss.coarsest_equitable_refinement(prt)
[[(0, 1)], [(1, 2), (1, 4)], [(0, 3)], [(0, 2), (0, 4)], [(2, 3), (3, 4)]]

ALGORITHM: Brendan D. McKay’s Master’s Thesis, University of Melbourne, 1976.

complement()

Returns the complement of the (di)graph.

The complement of a graph has the same vertices, but exactly those edges that are not in the original graph. This is not well defined for graphs with multiple edges.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.plot() # long time
sage: PC = P.complement()
sage: PC.plot() # long time
sage: graphs.TetrahedralGraph().complement().size()
0
sage: graphs.CycleGraph(4).complement().edges()
[(0, 2, None), (1, 3, None)]
sage: graphs.CycleGraph(4).complement()
complement(Cycle graph): Graph on 4 vertices
sage: G = Graph(multiedges=True, sparse=True)
sage: G.add_edges([(0,1)]*3)
sage: G.complement()
Traceback (most recent call last):
...
TypeError: complement not well defined for (di)graphs with multiple edges
connected_component_containing_vertex(vertex)

Returns a list of the vertices connected to vertex.

EXAMPLES:

sage: G = Graph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: G.connected_component_containing_vertex(0)
[0, 1, 2, 3]
sage: D = DiGraph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: D.connected_component_containing_vertex(0)
[0, 1, 2, 3]
connected_components()

Returns the list of connected components.

Returns a list of lists of vertices, each list representing a connected component. The list is ordered from largest to smallest component.

EXAMPLES:

sage: G = Graph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: G.connected_components()
[[0, 1, 2, 3], [4, 5, 6]]
sage: D = DiGraph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: D.connected_components()
[[0, 1, 2, 3], [4, 5, 6]]
connected_components_number()

Returns the number of connected components.

EXAMPLES:

sage: G = Graph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: G.connected_components_number()
2
sage: D = DiGraph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: D.connected_components_number()
2
connected_components_subgraphs()

Returns a list of connected components as graph objects.

EXAMPLES:

sage: G = Graph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: L = G.connected_components_subgraphs()
sage: graphs_list.show_graphs(L)
sage: D = DiGraph( { 0 : [1, 3], 1 : [2], 2 : [3], 4 : [5, 6], 5 : [6] } )
sage: L = D.connected_components_subgraphs()
sage: graphs_list.show_graphs(L)
copy(implementation='c_graph', data_structure=None, sparse=None)

Creates a copy of the graph.

INPUT:

  • implementation - string (default: ‘networkx’) the implementation goes here. Current options are only ‘networkx’ or ‘c_graph’.
  • sparse (boolean) – sparse=True is an alias for data_structure="sparse", and sparse=False is an alias for data_structure="dense".
  • data_structure – one of "sparse", "static_sparse", or "dense". See the documentation of Graph or DiGraph.

OUTPUT:

A Graph object.

Warning

Please use this method only if you need to copy but change the underlying implementation. Otherwise simply do copy(g) instead of doing g.copy().

EXAMPLES:

sage: g=Graph({0:[0,1,1,2]},loops=True,multiedges=True,sparse=True)
sage: g==copy(g)
True
sage: g=DiGraph({0:[0,1,1,2],1:[0,1]},loops=True,multiedges=True,sparse=True)
sage: g==copy(g)
True

Note that vertex associations are also kept:

sage: d = {0 : graphs.DodecahedralGraph(), 1 : graphs.FlowerSnark(), 2 : graphs.MoebiusKantorGraph(), 3 : graphs.PetersenGraph() }
sage: T = graphs.TetrahedralGraph()
sage: T.set_vertices(d)
sage: T2 = copy(T)
sage: T2.get_vertex(0)
Dodecahedron: Graph on 20 vertices

Notice that the copy is at least as deep as the objects:

sage: T2.get_vertex(0) is T.get_vertex(0)
False

Examples of the keywords in use:

sage: G = graphs.CompleteGraph(19)
sage: H = G.copy(implementation='c_graph')
sage: H == G; H is G
True
False
sage: G1 = G.copy(sparse=True)
sage: G1==G
True
sage: G1 is G
False
sage: G2 = copy(G)
sage: G2 is G
False

TESTS: We make copies of the _pos and _boundary attributes.

sage: g = graphs.PathGraph(3)
sage: h = copy(g)
sage: h._pos is g._pos
False
sage: h._boundary is g._boundary
False
cycle_basis()

Returns a list of cycles which form a basis of the cycle space of self.

A basis of cycles of a graph is a minimal collection of cycles (considered as sets of edges) such that the edge set of any cycle in the graph can be written as a Z/2Z sum of the cycles in the basis.

OUTPUT:

A list of lists, each of them representing the vertices of a cycle in a basis.

ALGORITHM:

Uses the NetworkX library.

EXAMPLE:

A cycle basis in Petersen’s Graph

sage: g = graphs.PetersenGraph()
sage: g.cycle_basis()
[[1, 2, 7, 5, 0], [8, 3, 2, 7, 5], [4, 3, 2, 7, 5, 0], [4, 9, 7, 5, 0], [8, 6, 9, 7, 5], [1, 6, 9, 7, 5, 0]]

Checking the given cycles are algebraically free:

sage: g = graphs.RandomGNP(30,.4)
sage: basis = g.cycle_basis()

Building the space of (directed) edges over Z/2Z. On the way, building a dictionary associating an unique vector to each undirected edge:

sage: m = g.size()
sage: edge_space = VectorSpace(FiniteField(2),m)
sage: edge_vector = dict( zip( g.edges(labels = False), edge_space.basis() ) )
sage: for (u,v),vec in edge_vector.items():
...      edge_vector[(v,u)] = vec

Defining a lambda function associating a vector to the vertices of a cycle:

sage: vertices_to_edges = lambda x : zip( x, x[1:] + [x[0]] )
sage: cycle_to_vector = lambda x : sum( edge_vector[e] for e in vertices_to_edges(x) )

Finally checking the cycles are a free set:

sage: basis_as_vectors = map( cycle_to_vector, basis )
sage: edge_space.span(basis_as_vectors).rank() == len(basis)
True
degree(vertices=None, labels=False)

Gives the degree (in + out for digraphs) of a vertex or of vertices.

INPUT:

  • vertices - If vertices is a single vertex, returns the number of neighbors of vertex. If vertices is an iterable container of vertices, returns a list of degrees. If vertices is None, same as listing all vertices.
  • labels - see OUTPUT

OUTPUT: Single vertex- an integer. Multiple vertices- a list of integers. If labels is True, then returns a dictionary mapping each vertex to its degree.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.degree(5)
3
sage: K = graphs.CompleteGraph(9)
sage: K.degree()
[8, 8, 8, 8, 8, 8, 8, 8, 8]
sage: D = DiGraph( { 0: [1,2,3], 1: [0,2], 2: [3], 3: [4], 4: [0,5], 5: [1] } )
sage: D.degree(vertices = [0,1,2], labels=True)
{0: 5, 1: 4, 2: 3}
sage: D.degree()
[5, 4, 3, 3, 3, 2]
degree_histogram()

Returns a list, whose ith entry is the frequency of degree i.

EXAMPLES:

sage: G = graphs.Grid2dGraph(9,12)
sage: G.degree_histogram()
[0, 0, 4, 34, 70]
sage: G = graphs.Grid2dGraph(9,12).to_directed()
sage: G.degree_histogram()
[0, 0, 0, 0, 4, 0, 34, 0, 70]
degree_iterator(vertices=None, labels=False)

Returns an iterator over the degrees of the (di)graph.

In the case of a digraph, the degree is defined as the sum of the in-degree and the out-degree, i.e. the total number of edges incident to a given vertex.

INPUT:

  • labels (boolean) – if set to False (default) the method returns an iterator over degrees. Otherwise it returns an iterator over tuples (vertex, degree).
  • vertices - if specified, restrict to this subset.

EXAMPLES:

sage: G = graphs.Grid2dGraph(3,4)
sage: for i in G.degree_iterator():
...    print i
3
4
2
...
2
4
sage: for i in G.degree_iterator(labels=True):
...    print i
((0, 1), 3)
((1, 2), 4)
((0, 0), 2)
...
((0, 3), 2)
((1, 1), 4)
sage: D = graphs.Grid2dGraph(2,4).to_directed()
sage: for i in D.degree_iterator():
...    print i
6
6
...
4
6
sage: for i in D.degree_iterator(labels=True):
...    print i
((0, 1), 6)
((1, 2), 6)
...
((0, 3), 4)
((1, 1), 6)
degree_sequence()

Return the degree sequence of this (di)graph.

EXAMPLES:

The degree sequence of an undirected graph:

sage: g = Graph({1: [2, 5], 2: [1, 5, 3, 4], 3: [2, 5], 4: [3], 5: [2, 3]})
sage: g.degree_sequence()
[4, 3, 3, 2, 2]

The degree sequence of a digraph:

sage: g = DiGraph({1: [2, 5, 6], 2: [3, 6], 3: [4, 6], 4: [6], 5: [4, 6]})
sage: g.degree_sequence()
[5, 3, 3, 3, 3, 3]

Degree sequences of some common graphs:

sage: graphs.PetersenGraph().degree_sequence()
[3, 3, 3, 3, 3, 3, 3, 3, 3, 3]
sage: graphs.HouseGraph().degree_sequence()
[3, 3, 2, 2, 2]
sage: graphs.FlowerSnark().degree_sequence()
[3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3]
degree_to_cell(vertex, cell)

Returns the number of edges from vertex to an edge in cell. In the case of a digraph, returns a tuple (in_degree, out_degree).

EXAMPLES:

sage: G = graphs.CubeGraph(3)
sage: cell = G.vertices()[:3]
sage: G.degree_to_cell('011', cell)
2
sage: G.degree_to_cell('111', cell)
0
sage: D = DiGraph({ 0:[1,2,3], 1:[3,4], 3:[4,5]})
sage: cell = [0,1,2]
sage: D.degree_to_cell(5, cell)
(0, 0)
sage: D.degree_to_cell(3, cell)
(2, 0)
sage: D.degree_to_cell(0, cell)
(0, 2)
delete_edge(u, v=None, label=None)

Delete the edge from u to v, returning silently if vertices or edge does not exist.

INPUT: The following forms are all accepted:

  • G.delete_edge( 1, 2 )
  • G.delete_edge( (1, 2) )
  • G.delete_edges( [ (1, 2) ] )
  • G.delete_edge( 1, 2, ‘label’ )
  • G.delete_edge( (1, 2, ‘label’) )
  • G.delete_edges( [ (1, 2, ‘label’) ] )

EXAMPLES:

sage: G = graphs.CompleteGraph(19).copy(implementation='c_graph')
sage: G.size()
171
sage: G.delete_edge( 1, 2 )
sage: G.delete_edge( (3, 4) )
sage: G.delete_edges( [ (5, 6), (7, 8) ] )
sage: G.size()
167

Note that NetworkX accidentally deletes these edges, even though the labels do not match up:

sage: N = graphs.CompleteGraph(19).copy(implementation='networkx')
sage: N.size()
171
sage: N.delete_edge( 1, 2 )
sage: N.delete_edge( (3, 4) )
sage: N.delete_edges( [ (5, 6), (7, 8) ] )
sage: N.size()
167
sage: N.delete_edge( 9, 10, 'label' )
sage: N.delete_edge( (11, 12, 'label') )
sage: N.delete_edges( [ (13, 14, 'label') ] )
sage: N.size()
167
sage: N.has_edge( (11, 12) )
True

However, CGraph backends handle things properly:

sage: G.delete_edge( 9, 10, 'label' )
sage: G.delete_edge( (11, 12, 'label') )
sage: G.delete_edges( [ (13, 14, 'label') ] )
sage: G.size()
167            
sage: C = graphs.CompleteGraph(19).to_directed(sparse=True)
sage: C.size()
342
sage: C.delete_edge( 1, 2 )
sage: C.delete_edge( (3, 4) )
sage: C.delete_edges( [ (5, 6), (7, 8) ] )

sage: D = graphs.CompleteGraph(19).to_directed(sparse=True, implementation='networkx')
sage: D.size()
342
sage: D.delete_edge( 1, 2 )
sage: D.delete_edge( (3, 4) )
sage: D.delete_edges( [ (5, 6), (7, 8) ] )
sage: D.delete_edge( 9, 10, 'label' )
sage: D.delete_edge( (11, 12, 'label') )
sage: D.delete_edges( [ (13, 14, 'label') ] )
sage: D.size()
338
sage: D.has_edge( (11, 12) )
True
sage: C.delete_edge( 9, 10, 'label' )
sage: C.delete_edge( (11, 12, 'label') )
sage: C.delete_edges( [ (13, 14, 'label') ] )
sage: C.size() # correct!
338
sage: C.has_edge( (11, 12) ) # correct!
True
delete_edges(edges)

Delete edges from an iterable container.

EXAMPLES:

sage: K12 = graphs.CompleteGraph(12)
sage: K4 = graphs.CompleteGraph(4)
sage: K12.size()
66
sage: K12.delete_edges(K4.edge_iterator())
sage: K12.size()
60
sage: K12 = graphs.CompleteGraph(12).to_directed()
sage: K4 = graphs.CompleteGraph(4).to_directed()
sage: K12.size()
132
sage: K12.delete_edges(K4.edge_iterator())
sage: K12.size()
120
delete_multiedge(u, v)

Deletes all edges from u and v.

EXAMPLES:

sage: G = Graph(multiedges=True,sparse=True)
sage: G.add_edges([(0,1), (0,1), (0,1), (1,2), (2,3)])
sage: G.edges()
[(0, 1, None), (0, 1, None), (0, 1, None), (1, 2, None), (2, 3, None)]
sage: G.delete_multiedge( 0, 1 )
sage: G.edges()
[(1, 2, None), (2, 3, None)]
sage: D = DiGraph(multiedges=True,sparse=True)
sage: D.add_edges([(0,1,1), (0,1,2), (0,1,3), (1,0), (1,2), (2,3)])
sage: D.edges()
[(0, 1, 1), (0, 1, 2), (0, 1, 3), (1, 0, None), (1, 2, None), (2, 3, None)]
sage: D.delete_multiedge( 0, 1 )
sage: D.edges()
[(1, 0, None), (1, 2, None), (2, 3, None)]
delete_vertex(vertex, in_order=False)

Deletes vertex, removing all incident edges. Deleting a non-existent vertex will raise an exception.

INPUT:

  • in_order - (default False) If True, this deletes the ith vertex in the sorted list of vertices, i.e. G.vertices()[i]

EXAMPLES:

sage: G = Graph(graphs.WheelGraph(9))
sage: G.delete_vertex(0); G.show()
sage: D = DiGraph({0:[1,2,3,4,5],1:[2],2:[3],3:[4],4:[5],5:[1]})
sage: D.delete_vertex(0); D
Digraph on 5 vertices
sage: D.vertices()
[1, 2, 3, 4, 5]
sage: D.delete_vertex(0)
Traceback (most recent call last):
...
RuntimeError: Vertex (0) not in the graph.
sage: G = graphs.CompleteGraph(4).line_graph(labels=False)
sage: G.vertices()
[(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]
sage: G.delete_vertex(0, in_order=True)
sage: G.vertices()
[(0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]
sage: G = graphs.PathGraph(5)
sage: G.set_vertices({0: 'no delete', 1: 'delete'})
sage: G.set_boundary([1,2])
sage: G.delete_vertex(1)
sage: G.get_vertices()
{0: 'no delete', 2: None, 3: None, 4: None}
sage: G.get_boundary()
[2]
sage: G.get_pos()
{0: (0, 0), 2: (2, 0), 3: (3, 0), 4: (4, 0)}
delete_vertices(vertices)

Remove vertices from the (di)graph taken from an iterable container of vertices. Deleting a non-existent vertex will raise an exception.

EXAMPLES:

sage: D = DiGraph({0:[1,2,3,4,5],1:[2],2:[3],3:[4],4:[5],5:[1]})
sage: D.delete_vertices([1,2,3,4,5]); D
Digraph on 1 vertex
sage: D.vertices()
[0]
sage: D.delete_vertices([1])
Traceback (most recent call last):
...
RuntimeError: Vertex (1) not in the graph.
density()

Returns the density (number of edges divided by number of possible edges).

In the case of a multigraph, raises an error, since there is an infinite number of possible edges.

EXAMPLES:

sage: d = {0: [1,4,5], 1: [2,6], 2: [3,7], 3: [4,8], 4: [9], 5: [7, 8], 6: [8,9], 7: [9]}
sage: G = Graph(d); G.density()
1/3
sage: G = Graph({0:[1,2], 1:[0] }); G.density()
2/3
sage: G = DiGraph({0:[1,2], 1:[0] }); G.density()
1/2

Note that there are more possible edges on a looped graph:

sage: G.allow_loops(True)
sage: G.density()
1/3

Returns an iterator over the vertices in a depth-first ordering.

INPUT:

  • start - vertex or list of vertices from which to start the traversal
  • ignore_direction - (default False) only applies to directed graphs. If True, searches across edges in either direction.
  • distance - the maximum distance from the start nodes to traverse. The start nodes are distance zero from themselves.
  • neighbors - a function giving the neighbors of a vertex. The function should take a vertex and return a list of vertices. For a graph, neighbors is by default the neighbors() function of the graph. For a digraph, the neighbors function defaults to the successors() function of the graph.

See also

EXAMPLES:

sage: G = Graph( { 0: [1], 1: [2], 2: [3], 3: [4], 4: [0]} )
sage: list(G.depth_first_search(0))
[0, 4, 3, 2, 1]

By default, the edge direction of a digraph is respected, but this can be overridden by the ignore_direction parameter:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.depth_first_search(0))
[0, 3, 6, 7, 2, 5, 1, 4]
sage: list(D.depth_first_search(0, ignore_direction=True))
[0, 7, 6, 3, 5, 2, 1, 4]

You can specify a maximum distance in which to search. A distance of zero returns the start vertices:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.depth_first_search(0,distance=0))
[0]
sage: list(D.depth_first_search(0,distance=1))
[0, 3, 2, 1]

Multiple starting vertices can be specified in a list:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.depth_first_search([0]))                                    
[0, 3, 6, 7, 2, 5, 1, 4]
sage: list(D.depth_first_search([0,6]))
[0, 3, 6, 7, 2, 5, 1, 4]
sage: list(D.depth_first_search([0,6],distance=0))
[0, 6]
sage: list(D.depth_first_search([0,6],distance=1))
[0, 3, 2, 1, 6, 7]
sage: list(D.depth_first_search(6,ignore_direction=True,distance=2))
[6, 7, 5, 0, 3]

More generally, you can specify a neighbors function. For example, you can traverse the graph backwards by setting neighbors to be the neighbors_in() function of the graph:

sage: D = DiGraph( { 0: [1,2,3], 1: [4,5], 2: [5], 3: [6], 5: [7], 6: [7], 7: [0]})
sage: list(D.depth_first_search(5,neighbors=D.neighbors_in, distance=2))
[5, 2, 0, 1]
sage: list(D.depth_first_search(5,neighbors=D.neighbors_out, distance=2))  
[5, 7, 0]
sage: list(D.depth_first_search(5,neighbors=D.neighbors, distance=2)) 
[5, 7, 6, 0, 2, 1, 4]

TESTS:

sage: D = DiGraph({1:[0], 2:[0]})
sage: list(D.depth_first_search(0))
[0]
sage: list(D.depth_first_search(0, ignore_direction=True))
[0, 2, 1]
diameter()

Returns the largest distance between any two vertices. Returns Infinity if the (di)graph is not connected.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.diameter()
2
sage: G = Graph( { 0 : [], 1 : [], 2 : [1] } )
sage: G.diameter()
+Infinity

Although max( ) is usually defined as -Infinity, since the diameter will never be negative, we define it to be zero:

sage: G = graphs.EmptyGraph()
sage: G.diameter()
0
disjoint_routed_paths(pairs, solver=None, verbose=0)

Returns a set of disjoint routed paths.

Given a set of pairs (s_i,t_i), a set of disjoint routed paths is a set of s_i-t_i paths which can interset at their endpoints and are vertex-disjoint otherwise.

INPUT:

  • pairs – list of pairs of vertices
  • solver – Specify a Linear Program solver to be used. If set to None, the default one is used. function of MixedIntegerLinearProgram. See the documentation of MixedIntegerLinearProgram.solve for more informations.
  • verbose (integer) – sets the level of verbosity. Set to 0 by default (quiet).

EXAMPLE:

Given a grid, finding two vertex-disjoint paths, the first one from the top-left corner to the bottom-left corner, and the second from the top-right corner to the bottom-right corner is easy

sage: g = graphs.GridGraph([5,5])
sage: p1,p2 = g.disjoint_routed_paths( [((0,0), (0,4)), ((4,4), (4,0))])

Though there is obviously no solution to the problem in which each corner is sending information to the opposite one:

sage: g = graphs.GridGraph([5,5])
sage: p1,p2 = g.disjoint_routed_paths( [((0,0), (4,4)), ((0,4), (4,0))])
Traceback (most recent call last):
...
ValueError: The disjoint routed paths do not exist.
disjoint_union(other, verbose_relabel=True)

Returns the disjoint union of self and other.

INPUT:

  • verbose_relabel - (defaults to True) If True, each vertex v in the first graph will be named ‘0,v’ and each vertex u in the second graph will be named ‘1,u’ in the final graph. If False, the vertices of the first graph and the second graph will be relabeled with consecutive integers.

See also

EXAMPLES:

sage: G = graphs.CycleGraph(3)
sage: H = graphs.CycleGraph(4)
sage: J = G.disjoint_union(H); J
Cycle graph disjoint_union Cycle graph: Graph on 7 vertices
sage: J.vertices()
[(0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (1, 3)]
sage: J = G.disjoint_union(H, verbose_relabel=False); J
Cycle graph disjoint_union Cycle graph: Graph on 7 vertices
sage: J.vertices()
[0, 1, 2, 3, 4, 5, 6]
sage: G=Graph({'a': ['b']})
sage: G.name("Custom path")
sage: G.name()
'Custom path'
sage: H=graphs.CycleGraph(3)
sage: J=G.disjoint_union(H); J
Custom path disjoint_union Cycle graph: Graph on 5 vertices
sage: J.vertices()
[(0, 'a'), (0, 'b'), (1, 0), (1, 1), (1, 2)]
disjunctive_product(other)

Returns the disjunctive product of self and other.

The disjunctive product of G and H is the graph L with vertex set V(L)=V(G)\times V(H), and ((u,v), (w,x)) is an edge iff either :

  • (u, w) is an edge of G, or
  • (v, x) is an edge of H.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: D = Z.disjunctive_product(Z); D
Graph on 4 vertices
sage: D.plot() # long time
sage: C = graphs.CycleGraph(5)
sage: D = C.disjunctive_product(Z); D
Graph on 10 vertices
sage: D.plot() # long time

TESTS:

Disjunctive product of graphs:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.disjunctive_product(H)
sage: T.edges(labels=None)
[((0, 'a'), (0, 'b')), ((0, 'a'), (1, 'a')), ((0, 'a'), (1, 'b')), ((0, 'a'), (2, 'b')), ((0, 'b'), (1, 'a')), ((0, 'b'), (1, 'b')), ((0, 'b'), (2, 'a')), ((1, 'a'), (1, 'b')), ((1, 'a'), (2, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a')), ((1, 'b'), (2, 'b')), ((2, 'a'), (2, 'b'))]
sage: T.is_isomorphic( H.disjunctive_product(G) )
True

Disjunctive product of digraphs:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.disjunctive_product(J)
sage: T.edges(labels=None)
[((0, 'a'), (0, 'b')), ((0, 'a'), (1, 'a')), ((0, 'a'), (1, 'b')), ((0, 'a'), (2, 'b')), ((0, 'b'), (1, 'a')), ((0, 'b'), (1, 'b')), ((1, 'a'), (0, 'b')), ((1, 'a'), (1, 'b')), ((1, 'a'), (2, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a')), ((1, 'b'), (2, 'b')), ((2, 'a'), (0, 'b')), ((2, 'a'), (1, 'b')), ((2, 'a'), (2, 'b'))]
sage: T.is_isomorphic( J.disjunctive_product(I) )
True
distance(u, v, by_weight=False)

Returns the (directed) distance from u to v in the (di)graph, i.e. the length of the shortest path from u to v.

INPUT:

  • by_weight - if False, uses a breadth first search. If True, takes edge weightings into account, using Dijkstra’s algorithm.

EXAMPLES:

sage: G = graphs.CycleGraph(9)
sage: G.distance(0,1)
1
sage: G.distance(0,4)
4
sage: G.distance(0,5)
4
sage: G = Graph( {0:[], 1:[]} )
sage: G.distance(0,1)
+Infinity
sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse = True)
sage: G.plot(edge_labels=True).show() # long time
sage: G.distance(0, 3)
2
sage: G.distance(0, 3, by_weight=True)
3
distance_all_pairs(algorithm='auto')

Returns the distances between all pairs of vertices.

INPUT:

  • "algorithm" (string) – two algorithms are available

    • algorithm = "BFS" in which case the distances are computed through n different breadth-first-search.
    • algorithm = "Floyd-Warshall", in which case the Floyd-Warshall algorithm is used.
    • algorithm = "auto", in which case the Floyd-Warshall algorithm is used for graphs on less than 20 vertices, and BFS otherwise.

    The default is algorithm = "BFS".

OUTPUT:

A doubly indexed dictionary

Note

There is a Cython version of this method that is usually much faster for large graphs, as most of the time is actually spent building the final double dictionary. Everything on the subject is to be found in the distances_all_pairs module.

EXAMPLE:

The Petersen Graph:

sage: g = graphs.PetersenGraph()
sage: print g.distance_all_pairs()
{0: {0: 0, 1: 1, 2: 2, 3: 2, 4: 1, 5: 1, 6: 2, 7: 2, 8: 2, 9: 2}, 1: {0: 1, 1: 0, 2: 1, 3: 2, 4: 2, 5: 2, 6: 1, 7: 2, 8: 2, 9: 2}, 2: {0: 2, 1: 1, 2: 0, 3: 1, 4: 2, 5: 2, 6: 2, 7: 1, 8: 2, 9: 2}, 3: {0: 2, 1: 2, 2: 1, 3: 0, 4: 1, 5: 2, 6: 2, 7: 2, 8: 1, 9: 2}, 4: {0: 1, 1: 2, 2: 2, 3: 1, 4: 0, 5: 2, 6: 2, 7: 2, 8: 2, 9: 1}, 5: {0: 1, 1: 2, 2: 2, 3: 2, 4: 2, 5: 0, 6: 2, 7: 1, 8: 1, 9: 2}, 6: {0: 2, 1: 1, 2: 2, 3: 2, 4: 2, 5: 2, 6: 0, 7: 2, 8: 1, 9: 1}, 7: {0: 2, 1: 2, 2: 1, 3: 2, 4: 2, 5: 1, 6: 2, 7: 0, 8: 2, 9: 1}, 8: {0: 2, 1: 2, 2: 2, 3: 1, 4: 2, 5: 1, 6: 1, 7: 2, 8: 0, 9: 2}, 9: {0: 2, 1: 2, 2: 2, 3: 2, 4: 1, 5: 2, 6: 1, 7: 1, 8: 2, 9: 0}}

Testing on Random Graphs:

sage: g = graphs.RandomGNP(20,.3)
sage: distances = g.distance_all_pairs()
sage: all([g.distance(0,v) == distances[0][v] for v in g])
True

See also

  • distance_matrix()
distance_graph(dist)

Returns the graph on the same vertex set as the original graph but vertices are adjacent in the returned graph if and only if they are at specified distances in the original graph.

INPUT:

  • dist is a nonnegative integer or a list of nonnegative integers. Infinity may be used here to describe vertex pairs in separate components.

OUTPUT:

The returned value is an undirected graph. The vertex set is identical to the calling graph, but edges of the returned graph join vertices whose distance in the calling graph are present in the input dist. Loops will only be present if distance 0 is included. If the original graph has a position dictionary specifying locations of vertices for plotting, then this information is copied over to the distance graph. In some instances this layout may not be the best, and might even be confusing when edges run on top of each other due to symmetries chosen for the layout.

EXAMPLES:

sage: G = graphs.CompleteGraph(3)
sage: H = G.cartesian_product(graphs.CompleteGraph(2))
sage: K = H.distance_graph(2)
sage: K.am()
[0 0 0 1 0 1]
[0 0 1 0 1 0]
[0 1 0 0 0 1]
[1 0 0 0 1 0]
[0 1 0 1 0 0]
[1 0 1 0 0 0]

To obtain the graph where vertices are adjacent if their distance apart is d or less use a range() command to create the input, using d+1 as the input to range. Notice that this will include distance 0 and hence place a loop at each vertex. To avoid this, use range(1,d+1).

sage: G = graphs.OddGraph(4)
sage: d = G.diameter()
sage: n = G.num_verts()
sage: H = G.distance_graph(range(d+1))
sage: H.is_isomorphic(graphs.CompleteGraph(n))
False
sage: H = G.distance_graph(range(1,d+1))
sage: H.is_isomorphic(graphs.CompleteGraph(n))
True

A complete collection of distance graphs will have adjacency matrices that sum to the matrix of all ones.

sage: P = graphs.PathGraph(20)
sage: all_ones = sum([P.distance_graph(i).am() for i in range(20)])
sage: all_ones == matrix(ZZ, 20, 20, [1]*400)
True

Four-bit strings differing in one bit is the same as four-bit strings differing in three bits.

sage: G = graphs.CubeGraph(4)
sage: H = G.distance_graph(3)
sage: G.is_isomorphic(H)
True

The graph of eight-bit strings, adjacent if different in an odd number of bits.

sage: G = graphs.CubeGraph(8) # long time
sage: H = G.distance_graph([1,3,5,7]) # long time
sage: degrees = [0]*sum([binomial(8,j) for j in [1,3,5,7]]) # long time
sage: degrees.append(2^8) # long time
sage: degrees == H.degree_histogram() # long time
True

An example of using Infinity as the distance in a graph that is not connected.

sage: G = graphs.CompleteGraph(3)
sage: H = G.disjoint_union(graphs.CompleteGraph(2))
sage: L = H.distance_graph(Infinity)
sage: L.am()
[0 0 0 1 1]
[0 0 0 1 1]
[0 0 0 1 1]
[1 1 1 0 0]
[1 1 1 0 0]

TESTS:

Empty input, or unachievable distances silently yield empty graphs.

sage: G = graphs.CompleteGraph(5)
sage: G.distance_graph([]).num_edges()
0
sage: G = graphs.CompleteGraph(5)
sage: G.distance_graph(23).num_edges()
0

It is an error to provide a distance that is not an integer type.

sage: G = graphs.CompleteGraph(5)
sage: G.distance_graph('junk')
Traceback (most recent call last):
...
TypeError: unable to convert x (=junk) to an integer

It is an error to provide a negative distance.

sage: G = graphs.CompleteGraph(5)
sage: G.distance_graph(-3)
Traceback (most recent call last):
...
ValueError: Distance graph for a negative distance (d=-3) is not defined

AUTHOR:

Rob Beezer, 2009-11-25

distance_matrix()

Returns the distance matrix of the (strongly) connected (di)graph.

The distance matrix of a (strongly) connected (di)graph is a matrix whose rows and columns are indexed with the vertices of the (di) graph. The intersection of a row and column contains the respective distance between the vertices indexed at these position.

Warning

The ordering of vertices in the matrix has no reason to correspond to the order of vertices in vertices(). In particular, if two integers i,j are vertices of a graph G with distance matrix M, then M[i][i] is not necessarily the distance between vertices i and j.

EXAMPLES:

sage: G = graphs.CubeGraph(3)
sage: G.distance_matrix()
[0 1 1 2 1 2 2 3]
[1 0 2 1 2 1 3 2]
[1 2 0 1 2 3 1 2]
[2 1 1 0 3 2 2 1]
[1 2 2 3 0 1 1 2]
[2 1 3 2 1 0 2 1]
[2 3 1 2 1 2 0 1]
[3 2 2 1 2 1 1 0]

The well known result of Graham and Pollak states that the determinant of the distance matrix of any tree of order n is (-1)^{n-1}(n-1)2^{n-2}

sage: all(T.distance_matrix().det() == (-1)^9*(9)*2^8 for T in graphs.trees(10))
True

See also

  • distance_all_pairs() – computes the distance between any two vertices.
distances_distribution(G)

Returns the distances distribution of the (di)graph in a dictionary.

This method ignores all edge labels, so that the distance considered is the topological distance.

OUTPUT:

A dictionary d such that the number of pairs of vertices at distance k (if any) is equal to d[k] \cdot |V(G)| \cdot (|V(G)|-1).

Note

We consider that two vertices that do not belong to the same connected component are at infinite distance, and we do not take the trivial pairs of vertices (v, v) at distance 0 into account. Empty (di)graphs and (di)graphs of order 1 have no paths and so we return the empty dictionary {}.

EXAMPLES:

An empty Graph:

sage: g = Graph()
sage: g.distances_distribution()
{}

A Graph of order 1:

sage: g = Graph()
sage: g.add_vertex(1)
sage: g.distances_distribution()
{}

A Graph of order 2 without edge:

sage: g = Graph()
sage: g.add_vertices([1,2])
sage: g.distances_distribution()
{+Infinity: 1}

The Petersen Graph:

sage: g = graphs.PetersenGraph()
sage: g.distances_distribution()
{1: 1/3, 2: 2/3}

A graph with multiple disconnected components:

sage: g = graphs.PetersenGraph()
sage: g.add_edge('good','wine')
sage: g.distances_distribution()
{1: 8/33, 2: 5/11, +Infinity: 10/33}

The de Bruijn digraph dB(2,3):

sage: D = digraphs.DeBruijn(2,3)
sage: D.distances_distribution()
{1: 1/4, 2: 11/28, 3: 5/14}
dominating_set(independent=False, value_only=False, solver=None, verbose=0)

Returns a minimum dominating set of the graph represented by the list of its vertices. For more information, see the Wikipedia article on dominating sets.

A minimum dominating set S of a graph G is a set of its vertices of minimal cardinality such that any vertex of G is in S or has one of its neighbors in S.

As an optimization problem, it can be expressed as:

\mbox{Minimize : }&\sum_{v\in G} b_v\\
\mbox{Such that : }&\forall v \in G, b_v+\sum_{(u,v)\in G.edges()} b_u\geq 1\\
&\forall x\in G, b_x\mbox{ is a binary variable}

INPUT:

  • independent – boolean (default: False). If independent=True, computes a minimum independent dominating set.
  • value_only – boolean (default: False)
    • If True, only the cardinality of a minimum dominating set is returned.
    • If False (default), a minimum dominating set is returned as the list of its vertices.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

EXAMPLES:

A basic illustration on a PappusGraph:

sage: g=graphs.PappusGraph()
sage: g.dominating_set(value_only=True)
5

If we build a graph from two disjoint stars, then link their centers we will find a difference between the cardinality of an independent set and a stable independent set:

sage: g = 2 * graphs.StarGraph(5)
sage: g.add_edge(0,6)
sage: len(g.dominating_set())
2
sage: len(g.dominating_set(independent=True))
6
eccentricity(v=None, dist_dict=None, with_labels=False)

Return the eccentricity of vertex (or vertices) v.

The eccentricity of a vertex is the maximum distance to any other vertex.

INPUT:

  • v - either a single vertex or a list of vertices. If it is not specified, then it is taken to be all vertices.
  • dist_dict - optional, a dict of dicts of distance.
  • with_labels - Whether to return a list or a dict.

EXAMPLES:

sage: G = graphs.KrackhardtKiteGraph()
sage: G.eccentricity()
[4, 4, 4, 4, 4, 3, 3, 2, 3, 4]
sage: G.vertices()
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
sage: G.eccentricity(7)
2
sage: G.eccentricity([7,8,9])
[3, 4, 2]
sage: G.eccentricity([7,8,9], with_labels=True) == {8: 3, 9: 4, 7: 2}
True
sage: G = Graph( { 0 : [], 1 : [], 2 : [1] } )
sage: G.eccentricity()
[+Infinity, +Infinity, +Infinity]
sage: G = Graph({0:[]})
sage: G.eccentricity(with_labels=True)
{0: 0}
sage: G = Graph({0:[], 1:[]})
sage: G.eccentricity(with_labels=True)
{0: +Infinity, 1: +Infinity}
edge_boundary(vertices1, vertices2=None, labels=True, sort=True)

Returns a list of edges (u,v,l) with u in vertices1 and v in vertices2. If vertices2 is None, then it is set to the complement of vertices1.

In a digraph, the external boundary of a vertex v are those vertices u with an arc (v, u).

INPUT:

  • labels - if False, each edge is a tuple (u,v) of vertices.

EXAMPLES:

sage: K = graphs.CompleteBipartiteGraph(9,3)
sage: len(K.edge_boundary( [0,1,2,3,4,5,6,7,8], [9,10,11] ))
27
sage: K.size()
27

Note that the edge boundary preserves direction:

sage: K = graphs.CompleteBipartiteGraph(9,3).to_directed()
sage: len(K.edge_boundary( [0,1,2,3,4,5,6,7,8], [9,10,11] ))
27
sage: K.size()
54
sage: D = DiGraph({0:[1,2], 3:[0]})
sage: D.edge_boundary([0])
[(0, 1, None), (0, 2, None)]
sage: D.edge_boundary([0], labels=False)
[(0, 1), (0, 2)]

TESTS:

sage: G = graphs.DiamondGraph()
sage: G.edge_boundary([0,1])
[(0, 2, None), (1, 2, None), (1, 3, None)]
sage: G.edge_boundary([0], [0])
[]
sage: G.edge_boundary([2], [0])
[(0, 2, None)]
edge_connectivity(value_only=True, use_edge_labels=False, vertices=False, solver=None, verbose=0)

Returns the edge connectivity of the graph. For more information, see the Wikipedia article on connectivity.

Note

When the graph is a directed graph, this method actually computes the strong connectivity, (i.e. a directed graph is strongly k-connected if there are k disjoint paths between any two vertices u, v). If you do not want to consider strong connectivity, the best is probably to convert your DiGraph object to a Graph object, and compute the connectivity of this other graph.

INPUT:

  • value_only – boolean (default: True)
    • When set to True (default), only the value is returned.
    • When set to False, both the value and a minimum edge cut are returned.
  • use_edge_labels – boolean (default: False)
    • When set to True, computes a weighted minimum cut where each edge has a weight defined by its label. (If an edge has no label, 1 is assumed.)
    • When set to False, each edge has weight 1.
  • vertices – boolean (default: False)
    • When set to True, also returns the two sets of vertices that are disconnected by the cut. Implies value_only=False.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

EXAMPLES:

A basic application on the PappusGraph:

sage: g = graphs.PappusGraph()
sage: g.edge_connectivity()
3

The edge connectivity of a complete graph ( and of a random graph ) is its minimum degree, and one of the two parts of the bipartition is reduced to only one vertex. The cutedges isomorphic to a Star graph:

sage: g = graphs.CompleteGraph(5)
sage: [ value, edges, [ setA, setB ]] = g.edge_connectivity(vertices=True)
sage: print value
4
sage: len(setA) == 1 or len(setB) == 1
True
sage: cut = Graph()
sage: cut.add_edges(edges)
sage: cut.is_isomorphic(graphs.StarGraph(4))
True

Even if obviously in any graph we know that the edge connectivity is less than the minimum degree of the graph:

sage: g = graphs.RandomGNP(10,.3)
sage: min(g.degree()) >= g.edge_connectivity()
True

If we build a tree then assign to its edges a random value, the minimum cut will be the edge with minimum value:

sage: g = graphs.RandomGNP(15,.5)
sage: tree = Graph()
sage: tree.add_edges(g.min_spanning_tree())
sage: for u,v in tree.edge_iterator(labels=None):
...        tree.set_edge_label(u,v,random())
sage: minimum = min([l for u,v,l in tree.edge_iterator()])
sage: [value, [(u,v,l)]] = tree.edge_connectivity(value_only=False, use_edge_labels=True)
sage: l == minimum
True

When value_only = True, this function is optimized for small connectivity values and does not need to build a linear program.

It is the case for connected graphs which are not connected

sage: g = 2 * graphs.PetersenGraph()
sage: g.edge_connectivity()
0.0

Or if they are just 1-connected

sage: g = graphs.PathGraph(10)
sage: g.edge_connectivity()
1.0

For directed graphs, the strong connectivity is tested through the dedicated function

sage: g = digraphs.ButterflyGraph(3)
sage: g.edge_connectivity()
0.0
edge_cut(s, t, value_only=True, use_edge_labels=False, vertices=False, method='FF', solver=None, verbose=0)

Returns a minimum edge cut between vertices s and t represented by a list of edges.

A minimum edge cut between two vertices s and t of self is a set A of edges of minimum weight such that the graph obtained by removing A from self is disconnected. For more information, see the Wikipedia article on cuts.

INPUT:

  • s – source vertex

  • t – sink vertex

  • value_only – boolean (default: True). When set to True, only the weight of a minimum cut is returned. Otherwise, a list of edges of a minimum cut is also returned.

  • use_edge_labels – boolean (default: False). When set to True, computes a weighted minimum cut where each edge has a weight defined by its label (if an edge has no label, 1 is assumed). Otherwise, each edge has weight 1.

  • vertices – boolean (default: False). When set to True, returns a list of edges in the edge cut and the two sets of vertices that are disconnected by the cut.

    Note: vertices=True implies value_only=False.

  • method – There are currently two different implementations of this method :

    • If method = "FF" (default), a Python implementation of the Ford-Fulkerson algorithm is used.
    • If method = "LP", the flow problem is solved using Linear Programming.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.

  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

Note

The use of Linear Programming for non-integer problems may possibly mean the presence of a (slight) numerical noise.

OUTPUT:

Real number or tuple, depending on the given arguments (examples are given below).

EXAMPLES:

A basic application in the Pappus graph:

sage: g = graphs.PappusGraph()
sage: g.edge_cut(1, 2, value_only=True)
3

Or on Petersen’s graph, with the corresponding bipartition of the vertex set:

sage: g = graphs.PetersenGraph()
sage: g.edge_cut(0, 3, vertices=True)
[3, [(0, 1, None), (0, 4, None), (0, 5, None)], [[0], [1, 2, 3, 4, 5, 6, 7, 8, 9]]]

If the graph is a path with randomly weighted edges:

sage: g = graphs.PathGraph(15)
sage: for (u,v) in g.edge_iterator(labels=None):
...      g.set_edge_label(u,v,random())

The edge cut between the two ends is the edge of minimum weight:

sage: minimum = min([l for u,v,l in g.edge_iterator()])
sage: minimum == g.edge_cut(0, 14, use_edge_labels=True)
True
sage: [value,[e]] = g.edge_cut(0, 14, use_edge_labels=True, value_only=False)
sage: g.edge_label(e[0],e[1]) == minimum
True

The two sides of the edge cut are obviously shorter paths:

sage: value,edges,[set1,set2] = g.edge_cut(0, 14, use_edge_labels=True, vertices=True)
sage: g.subgraph(set1).is_isomorphic(graphs.PathGraph(len(set1)))
True
sage: g.subgraph(set2).is_isomorphic(graphs.PathGraph(len(set2)))
True
sage: len(set1) + len(set2) == g.order()
True

TESTS:

If method is set to an exotic value:

sage: g = graphs.PetersenGraph()
sage: g.edge_cut(0,1, method="Divination")
Traceback (most recent call last):
...
ValueError: The method argument has to be equal to either "FF" or "LP"

Same result for both methods:

sage: g = graphs.RandomGNP(20,.3)
sage: for u,v in g.edges(labels=False):
...      g.set_edge_label(u,v,round(random(),5))
sage: g.edge_cut(0,1, method="FF") == g.edge_cut(0,1,method="LP")
True

Rounded return value when using the LP method:

sage: g = graphs.PappusGraph()
sage: g.edge_cut(1, 2, value_only=True, method = "LP")
3
edge_disjoint_paths(s, t, method='FF')

Returns a list of edge-disjoint paths between two vertices as given by Menger’s theorem.

The edge version of Menger’s theorem asserts that the size of the minimum edge cut between two vertices s and`t` (the minimum number of edges whose removal disconnects s and t) is equal to the maximum number of pairwise edge-independent paths from s to t.

This function returns a list of such paths.

INPUT:

  • method – There are currently two different implementations of this method :

    • If method = "FF" (default), a Python implementation of the Ford-Fulkerson algorithm is used.
    • If method = "LP", the flow problem is solved using Linear Programming.

Note

This function is topological : it does not take the eventual weights of the edges into account.

EXAMPLE:

In a complete bipartite graph

sage: g = graphs.CompleteBipartiteGraph(2,3)
sage: g.edge_disjoint_paths(0,1)
[[0, 2, 1], [0, 3, 1], [0, 4, 1]]
edge_disjoint_spanning_trees(k, root=None, solver=None, verbose=0)

Returns the desired number of edge-disjoint spanning trees/arborescences.

INPUT:

  • k (integer) – the required number of edge-disjoint spanning trees/arborescences
  • root (vertex) – root of the disjoint arborescences when the graph is directed. If set to None, the first vertex in the graph is picked.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

ALGORITHM:

Mixed Integer Linear Program. The formulation can be found in [LPForm].

There are at least two possible rewritings of this method which do not use Linear Programming:

  • The algorithm presented in the paper entitled “A short proof of the tree-packing theorem”, by Thomas Kaiser [KaisPacking].
  • The implementation of a Matroid class and of the Matroid Union Theorem (see section 42.3 of [SchrijverCombOpt]), applied to the cycle Matroid (see chapter 51 of [SchrijverCombOpt]).

EXAMPLES:

The Petersen Graph does have a spanning tree (it is connected):

sage: g = graphs.PetersenGraph()
sage: [T] = g.edge_disjoint_spanning_trees(1)
sage: T.is_tree()
True

Though, it does not have 2 edge-disjoint trees (as it has less than 2(|V|-1) edges):

sage: g.edge_disjoint_spanning_trees(2)
Traceback (most recent call last):
...
ValueError: This graph does not contain the required number of trees/arborescences !

By Edmond’s theorem, a graph which is k-connected always has k edge-disjoint arborescences, regardless of the root we pick:

sage: g = digraphs.RandomDirectedGNP(28,.3) # reduced from 30 to 28, cf. #9584
sage: k = Integer(g.edge_connectivity())
sage: arborescences = g.edge_disjoint_spanning_trees(k)  # long time (up to 15s on sage.math, 2011)
sage: all([a.is_directed_acyclic() for a in arborescences])  # long time
True
sage: all([a.is_connected() for a in arborescences])  # long time
True

In the undirected case, we can only ensure half of it:

sage: g = graphs.RandomGNP(30,.3)
sage: k = floor(Integer(g.edge_connectivity())/2)
sage: trees = g.edge_disjoint_spanning_trees(k)
sage: all([t.is_tree() for t in trees])
True

REFERENCES:

[LPForm]Nathann Cohen, Several Graph problems and their Linear Program formulations, http://hal.archives-ouvertes.fr/inria-00504914/en
[KaisPacking]Thomas Kaiser A short proof of the tree-packing theorem Arxiv 0911.2809
[SchrijverCombOpt](1, 2) Alexander Schrijver Combinatorial optimization: polyhedra and efficiency 2003
edge_iterator(vertices=None, labels=True, ignore_direction=False)

Returns an iterator over edges.

The iterator returned is over the edges incident with any vertex given in the parameter vertices. If the graph is directed, iterates over edges going out only. If vertices is None, then returns an iterator over all edges. If self is directed, returns outgoing edges only.

INPUT:

  • vertices - (default: None) a vertex, a list of vertices or None
  • labels - if False, each edge is a tuple (u,v) of vertices.
  • ignore_direction - bool (default: False) - only applies to directed graphs. If True, searches across edges in either direction.

EXAMPLES:

sage: for i in graphs.PetersenGraph().edge_iterator([0]):
...    print i
(0, 1, None)
(0, 4, None)
(0, 5, None)
sage: D = DiGraph( { 0 : [1,2], 1: [0] } )
sage: for i in D.edge_iterator([0]):
...    print i
(0, 1, None)
(0, 2, None)
sage: G = graphs.TetrahedralGraph()
sage: list(G.edge_iterator(labels=False))
[(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]
sage: D = DiGraph({1:[0], 2:[0]})
sage: list(D.edge_iterator(0))
[]
sage: list(D.edge_iterator(0, ignore_direction=True))
[(1, 0, None), (2, 0, None)]
edge_label(u, v=None)

Returns the label of an edge. Note that if the graph allows multiple edges, then a list of labels on the edge is returned.

EXAMPLES:

sage: G = Graph({0 : {1 : 'edgelabel'}}, sparse=True)
sage: G.edges(labels=False)
[(0, 1)]
sage: G.edge_label( 0, 1 )
'edgelabel'
sage: D = DiGraph({0 : {1 : 'edgelabel'}}, sparse=True)
sage: D.edges(labels=False)
[(0, 1)]
sage: D.edge_label( 0, 1 )
'edgelabel'
sage: G = Graph(multiedges=True, sparse=True)
sage: [G.add_edge(0,1,i) for i in range(1,6)]
[None, None, None, None, None]
sage: sorted(G.edge_label(0,1))
[1, 2, 3, 4, 5]

TESTS:

sage: G = Graph()
sage: G.add_edge(0,1,[7])
sage: G.add_edge(0,2,[7])
sage: G.edge_label(0,1)[0] += 1
sage: G.edges()
[(0, 1, [8]), (0, 2, [7])]
edge_labels()

Returns a list of edge labels.

EXAMPLES:

sage: G = Graph({0:{1:'x',2:'z',3:'a'}, 2:{5:'out'}}, sparse=True)
sage: G.edge_labels()
['x', 'z', 'a', 'out']
sage: G = DiGraph({0:{1:'x',2:'z',3:'a'}, 2:{5:'out'}}, sparse=True)
sage: G.edge_labels()
['x', 'z', 'a', 'out']
edges(labels=True, sort=True, key=None)

Return a list of edges.

Each edge is a triple (u,v,l) where u and v are vertices and l is a label. If the parameter labels is False then a list of couple (u,v) is returned where u and v are vertices.

INPUT:

  • labels - default: True - if False, each edge is simply a pair (u,v) of vertices.
  • sort - default: True - if True, edges are sorted according to the default ordering.
  • key - default: None - a function takes an edge (a pair or a triple, according to the labels keyword) as its one argument and returns a value that can be used for comparisons in the sorting algorithm.

OUTPUT: A list of tuples. It is safe to change the returned list.

Warning

Since any object may be a vertex, there is no guarantee that any two vertices will be comparable, and thus no guarantee how two edges may compare. With default objects for vertices (all integers), or when all the vertices are of the same simple type, then there should not be a problem with how the vertices will be sorted. However, if you need to guarantee a total order for the sorting of the edges, use the key argument, as illustrated in the examples below.

EXAMPLES:

sage: graphs.DodecahedralGraph().edges()
[(0, 1, None), (0, 10, None), (0, 19, None), (1, 2, None), (1, 8, None), (2, 3, None), (2, 6, None), (3, 4, None), (3, 19, None), (4, 5, None), (4, 17, None), (5, 6, None), (5, 15, None), (6, 7, None), (7, 8, None), (7, 14, None), (8, 9, None), (9, 10, None), (9, 13, None), (10, 11, None), (11, 12, None), (11, 18, None), (12, 13, None), (12, 16, None), (13, 14, None), (14, 15, None), (15, 16, None), (16, 17, None), (17, 18, None), (18, 19, None)]
sage: graphs.DodecahedralGraph().edges(labels=False)
[(0, 1), (0, 10), (0, 19), (1, 2), (1, 8), (2, 3), (2, 6), (3, 4), (3, 19), (4, 5), (4, 17), (5, 6), (5, 15), (6, 7), (7, 8), (7, 14), (8, 9), (9, 10), (9, 13), (10, 11), (11, 12), (11, 18), (12, 13), (12, 16), (13, 14), (14, 15), (15, 16), (16, 17), (17, 18), (18, 19)]
sage: D = graphs.DodecahedralGraph().to_directed()
sage: D.edges()
[(0, 1, None), (0, 10, None), (0, 19, None), (1, 0, None), (1, 2, None), (1, 8, None), (2, 1, None), (2, 3, None), (2, 6, None), (3, 2, None), (3, 4, None), (3, 19, None), (4, 3, None), (4, 5, None), (4, 17, None), (5, 4, None), (5, 6, None), (5, 15, None), (6, 2, None), (6, 5, None), (6, 7, None), (7, 6, None), (7, 8, None), (7, 14, None), (8, 1, None), (8, 7, None), (8, 9, None), (9, 8, None), (9, 10, None), (9, 13, None), (10, 0, None), (10, 9, None), (10, 11, None), (11, 10, None), (11, 12, None), (11, 18, None), (12, 11, None), (12, 13, None), (12, 16, None), (13, 9, None), (13, 12, None), (13, 14, None), (14, 7, None), (14, 13, None), (14, 15, None), (15, 5, None), (15, 14, None), (15, 16, None), (16, 12, None), (16, 15, None), (16, 17, None), (17, 4, None), (17, 16, None), (17, 18, None), (18, 11, None), (18, 17, None), (18, 19, None), (19, 0, None), (19, 3, None), (19, 18, None)]
sage: D.edges(labels = False)
[(0, 1), (0, 10), (0, 19), (1, 0), (1, 2), (1, 8), (2, 1), (2, 3), (2, 6), (3, 2), (3, 4), (3, 19), (4, 3), (4, 5), (4, 17), (5, 4), (5, 6), (5, 15), (6, 2), (6, 5), (6, 7), (7, 6), (7, 8), (7, 14), (8, 1), (8, 7), (8, 9), (9, 8), (9, 10), (9, 13), (10, 0), (10, 9), (10, 11), (11, 10), (11, 12), (11, 18), (12, 11), (12, 13), (12, 16), (13, 9), (13, 12), (13, 14), (14, 7), (14, 13), (14, 15), (15, 5), (15, 14), (15, 16), (16, 12), (16, 15), (16, 17), (17, 4), (17, 16), (17, 18), (18, 11), (18, 17), (18, 19), (19, 0), (19, 3), (19, 18)]

The default is to sort the returned list in the default fashion, as in the above examples. this can be overridden by specifying a key function. This first example just ignores the labels in the third component of the triple.

sage: G=graphs.CycleGraph(5)
sage: G.edges(key = lambda x: (x[1],-x[0]))
[(0, 1, None), (1, 2, None), (2, 3, None), (3, 4, None), (0, 4, None)]

We set the labels to characters and then perform a default sort followed by a sort according to the labels.

sage: G=graphs.CycleGraph(5)
sage: for e in G.edges():
...     G.set_edge_label(e[0], e[1], chr(ord('A')+e[0]+5*e[1]))
sage: G.edges(sort=True)
[(0, 1, 'F'), (0, 4, 'U'), (1, 2, 'L'), (2, 3, 'R'), (3, 4, 'X')]
sage: G.edges(key=lambda x: x[2])
[(0, 1, 'F'), (1, 2, 'L'), (2, 3, 'R'), (0, 4, 'U'), (3, 4, 'X')]

TESTS:

It is an error to turn off sorting while providing a key function for sorting.

sage: P=graphs.PetersenGraph()
sage: P.edges(sort=False, key=lambda x: x)
Traceback (most recent call last):
...
ValueError: sort keyword is False, yet a key function is given
edges_incident(vertices=None, labels=True, sort=True)

Returns incident edges to some vertices.

If vertices` is a vertex, then it returns the list of edges incident to that vertex. If ``vertices is a list of vertices then it returns the list of all edges adjacent to those vertices. If vertices is None, returns a list of all edges in graph. For digraphs, only lists outward edges.

INPUT:

  • vertices - object (default: None) - a vertex, a list of vertices or None.
  • labels - bool (default: True) - if False, each edge is a tuple (u,v) of vertices.
  • sort - bool (default: True) - if True the returned list is sorted.

EXAMPLES:

sage: graphs.PetersenGraph().edges_incident([0,9], labels=False)
[(0, 1), (0, 4), (0, 5), (4, 9), (6, 9), (7, 9)]
sage: D = DiGraph({0:[1]})
sage: D.edges_incident([0])
[(0, 1, None)]
sage: D.edges_incident([1])
[]

TESTS:

sage: G = Graph({0:[0]}, loops=True)  # ticket 9581
sage: G.edges_incident(0)
[(0, 0, None)]
eigenspaces(laplacian=False)

Returns the right eigenspaces of the adjacency matrix of the graph.

INPUT:

OUTPUT:

A list of pairs. Each pair is an eigenvalue of the adjacency matrix of the graph, followed by the vector space that is the eigenspace for that eigenvalue, when the eigenvectors are placed on the right of the matrix.

For some graphs, some of the the eigenspaces are described exactly by vector spaces over a NumberField. For numerical eigenvectors use eigenvectors().

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.eigenspaces()
[
(3, Vector space of degree 10 and dimension 1 over Rational Field
User basis matrix:
[1 1 1 1 1 1 1 1 1 1]),
(-2, Vector space of degree 10 and dimension 4 over Rational Field
User basis matrix:
[ 1  0  0  0 -1 -1 -1  0  1  1]
[ 0  1  0  0 -1  0 -2 -1  1  2]
[ 0  0  1  0 -1  1 -1 -2  0  2]
[ 0  0  0  1 -1  1  0 -1 -1  1]),
(1, Vector space of degree 10 and dimension 5 over Rational Field
User basis matrix:
[ 1  0  0  0  0  1 -1  0  0 -1]
[ 0  1  0  0  0 -1  1 -1  0  0]
[ 0  0  1  0  0  0 -1  1 -1  0]
[ 0  0  0  1  0  0  0 -1  1 -1]
[ 0  0  0  0  1 -1  0  0 -1  1])
]

Eigenspaces for the Laplacian should be identical since the Petersen graph is regular. However, since the output also contains the eigenvalues, the two outputs are slightly different.

sage: P.eigenspaces(laplacian=True)
[
(0, Vector space of degree 10 and dimension 1 over Rational Field
User basis matrix:
[1 1 1 1 1 1 1 1 1 1]),
(5, Vector space of degree 10 and dimension 4 over Rational Field
User basis matrix:
[ 1  0  0  0 -1 -1 -1  0  1  1]
[ 0  1  0  0 -1  0 -2 -1  1  2]
[ 0  0  1  0 -1  1 -1 -2  0  2]
[ 0  0  0  1 -1  1  0 -1 -1  1]),
(2, Vector space of degree 10 and dimension 5 over Rational Field
User basis matrix:
[ 1  0  0  0  0  1 -1  0  0 -1]
[ 0  1  0  0  0 -1  1 -1  0  0]
[ 0  0  1  0  0  0 -1  1 -1  0]
[ 0  0  0  1  0  0  0 -1  1 -1]
[ 0  0  0  0  1 -1  0  0 -1  1])
]

Notice how one eigenspace below is described with a square root of 2. For the two possible values (positive and negative) there is a corresponding eigenspace.

sage: C = graphs.CycleGraph(8)
sage: C.eigenspaces()
[
(2, Vector space of degree 8 and dimension 1 over Rational Field
User basis matrix:
[1 1 1 1 1 1 1 1]),
(-2, Vector space of degree 8 and dimension 1 over Rational Field
User basis matrix:
[ 1 -1  1 -1  1 -1  1 -1]),
(0, Vector space of degree 8 and dimension 2 over Rational Field
User basis matrix:
[ 1  0 -1  0  1  0 -1  0]
[ 0  1  0 -1  0  1  0 -1]),
(a3, Vector space of degree 8 and dimension 2 over Number Field in a3 with defining polynomial x^2 - 2
User basis matrix:
[  1   0  -1 -a3  -1   0   1  a3]
[  0   1  a3   1   0  -1 -a3  -1])
]

A digraph may have complex eigenvalues and eigenvectors. For a 3-cycle, we have:

sage: T = DiGraph({0:[1], 1:[2], 2:[0]})
sage: T.eigenspaces()
[
(1, Vector space of degree 3 and dimension 1 over Rational Field
User basis matrix:
[1 1 1]),
(a1, Vector space of degree 3 and dimension 1 over Number Field in a1 with defining polynomial x^2 + x + 1
User basis matrix:
[      1      a1 -a1 - 1])
]
eigenvectors(laplacian=False)

Returns the right eigenvectors of the adjacency matrix of the graph.

INPUT:

OUTPUT:

A list of triples. Each triple begins with an eigenvalue of the adjacency matrix of the graph. This is followed by a list of eigenvectors for the eigenvalue, when the eigenvectors are placed on the right side of the matrix. Together, the eigenvectors form a basis for the eigenspace. The triple concludes with the algebraic multiplicity of the eigenvalue.

For some graphs, the exact eigenspaces provided by eigenspaces() provide additional insight into the structure of the eigenspaces.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.eigenvectors()
[(3, [
(1, 1, 1, 1, 1, 1, 1, 1, 1, 1)
], 1), (-2, [
(1, 0, 0, 0, -1, -1, -1, 0, 1, 1),
(0, 1, 0, 0, -1, 0, -2, -1, 1, 2),
(0, 0, 1, 0, -1, 1, -1, -2, 0, 2),
(0, 0, 0, 1, -1, 1, 0, -1, -1, 1)
], 4), (1, [
(1, 0, 0, 0, 0, 1, -1, 0, 0, -1),
(0, 1, 0, 0, 0, -1, 1, -1, 0, 0),
(0, 0, 1, 0, 0, 0, -1, 1, -1, 0),
(0, 0, 0, 1, 0, 0, 0, -1, 1, -1),
(0, 0, 0, 0, 1, -1, 0, 0, -1, 1)
], 5)]

Eigenspaces for the Laplacian should be identical since the Petersen graph is regular. However, since the output also contains the eigenvalues, the two outputs are slightly different.

sage: P.eigenvectors(laplacian=True)
[(0, [
(1, 1, 1, 1, 1, 1, 1, 1, 1, 1)
], 1), (5, [
(1, 0, 0, 0, -1, -1, -1, 0, 1, 1),
(0, 1, 0, 0, -1, 0, -2, -1, 1, 2),
(0, 0, 1, 0, -1, 1, -1, -2, 0, 2),
(0, 0, 0, 1, -1, 1, 0, -1, -1, 1)
], 4), (2, [
(1, 0, 0, 0, 0, 1, -1, 0, 0, -1),
(0, 1, 0, 0, 0, -1, 1, -1, 0, 0),
(0, 0, 1, 0, 0, 0, -1, 1, -1, 0),
(0, 0, 0, 1, 0, 0, 0, -1, 1, -1),
(0, 0, 0, 0, 1, -1, 0, 0, -1, 1)
], 5)]
sage: C = graphs.CycleGraph(8)
sage: C.eigenvectors()
[(2, [
(1, 1, 1, 1, 1, 1, 1, 1)
], 1), (-2, [
(1, -1, 1, -1, 1, -1, 1, -1)
], 1), (0, [
(1, 0, -1, 0, 1, 0, -1, 0),
(0, 1, 0, -1, 0, 1, 0, -1)
], 2), (-1.4142135623..., [(1, 0, -1, 1.4142135623..., -1, 0, 1, -1.4142135623...), (0, 1, -1.4142135623..., 1, 0, -1, 1.4142135623..., -1)], 2), (1.4142135623..., [(1, 0, -1, -1.4142135623..., -1, 0, 1, 1.4142135623...), (0, 1, 1.4142135623..., 1, 0, -1, -1.4142135623..., -1)], 2)]

A digraph may have complex eigenvalues. Previously, the complex parts of graph eigenvalues were being dropped. For a 3-cycle, we have:

sage: T = DiGraph({0:[1], 1:[2], 2:[0]})
sage: T.eigenvectors()
[(1, [
(1, 1, 1)
], 1), (-0.5000000000... - 0.8660254037...*I, [(1, -0.5000000000... - 0.8660254037...*I, -0.5000000000... + 0.8660254037...*I)], 1), (-0.5000000000... + 0.8660254037...*I, [(1, -0.5000000000... + 0.8660254037...*I, -0.5000000000... - 0.8660254037...*I)], 1)]
eulerian_circuit(return_vertices=False, labels=True, path=False)

Return a list of edges forming an eulerian circuit if one exists. Otherwise return False.

This is implemented using Hierholzer’s algorithm.

INPUT:

  • return_vertices – (default: False) optionally provide a list of vertices for the path
  • labels – (default: True) whether to return edges with labels (3-tuples)
  • path – (default: False) find an eulerian path instead

OUTPUT:

either ([edges], [vertices]) or [edges] of an Eulerian circuit (or path)

EXAMPLES:

sage: g=graphs.CycleGraph(5);
sage: g.eulerian_circuit()
[(0, 4, None), (4, 3, None), (3, 2, None), (2, 1, None), (1, 0, None)]
sage: g.eulerian_circuit(labels=False)
[(0, 4), (4, 3), (3, 2), (2, 1), (1, 0)]
sage: g = graphs.CompleteGraph(7)
sage: edges, vertices = g.eulerian_circuit(return_vertices=True)
sage: vertices
[0, 6, 5, 4, 6, 3, 5, 2, 4, 3, 2, 6, 1, 5, 0, 4, 1, 3, 0, 2, 1, 0]
sage: graphs.CompleteGraph(4).eulerian_circuit()
False

A disconnected graph can be eulerian:

sage: g = Graph({0: [], 1: [2], 2: [3], 3: [1], 4: []})
sage: g.eulerian_circuit(labels=False)
[(1, 3), (3, 2), (2, 1)]
sage: g = DiGraph({0: [1], 1: [2, 4], 2:[3], 3:[1]})
sage: g.eulerian_circuit(labels=False, path=True)
[(0, 1), (1, 2), (2, 3), (3, 1), (1, 4)]
sage: g = Graph({0:[1,2,3], 1:[2,3], 2:[3,4], 3:[4]})
sage: g.is_eulerian(path=True)
(0, 1)
sage: g.eulerian_circuit(labels=False, path=True)
[(1, 3), (3, 4), (4, 2), (2, 3), (3, 0), (0, 2), (2, 1), (1, 0)]

TESTS:

sage: Graph({'H': ['G','L','L','D'], 'L': ['G','D']}).eulerian_circuit(labels=False)
[('H', 'D'), ('D', 'L'), ('L', 'G'), ('G', 'H'), ('H', 'L'), ('L', 'H')]
sage: Graph({0: [0, 1, 1, 1, 1]}).eulerian_circuit(labels=False)
[(0, 1), (1, 0), (0, 1), (1, 0), (0, 0)]
eulerian_orientation()

Returns a DiGraph which is an Eulerian orientation of the current graph.

An Eulerian graph being a graph such that any vertex has an even degree, an Eulerian orientation of a graph is an orientation of its edges such that each vertex v verifies d^+(v)=d^-(v)=d(v)/2, where d^+ and d^- respectively represent the out-degree and the in-degree of a vertex.

If the graph is not Eulerian, the orientation verifies for any vertex v that | d^+(v)-d^-(v) | \leq 1.

ALGORITHM:

This algorithm is a random walk through the edges of the graph, which orients the edges according to the walk. When a vertex is reached which has no non-oriented edge ( this vertex must have odd degree ), the walk resumes at another vertex of odd degree, if any.

This algorithm has complexity O(m), where m is the number of edges in the graph.

EXAMPLES:

The CubeGraph with parameter 4, which is regular of even degree, has an Eulerian orientation such that d^+=d^-:

sage: g=graphs.CubeGraph(4)
sage: g.degree()
[4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4]
sage: o=g.eulerian_orientation()
sage: o.in_degree()
[2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]
sage: o.out_degree()
[2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]

Secondly, the Petersen Graph, which is 3 regular has an orientation such that the difference between d^+ and d^- is at most 1:

sage: g=graphs.PetersenGraph()
sage: o=g.eulerian_orientation()
sage: o.in_degree()
[2, 2, 2, 2, 2, 1, 1, 1, 1, 1]
sage: o.out_degree()
[1, 1, 1, 1, 1, 2, 2, 2, 2, 2]
feedback_vertex_set(value_only=False, solver=None, verbose=0, constraint_generation=True)

Computes the minimum feedback vertex set of a (di)graph.

The minimum feedback vertex set of a (di)graph is a set of vertices that intersect all of its cycles. Equivalently, a minimum feedback vertex set of a (di)graph is a set S of vertices such that the digraph G-S is acyclic. For more information, see Wikipedia article Feedback_vertex_set.

INPUT:

  • value_only – boolean (default: False)
    • When set to True, only the minimum cardinal of a minimum vertex set is returned.
    • When set to False, the Set of vertices of a minimal feedback vertex set is returned.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.
  • constraint_generation (boolean) – whether to use constraint generation when solving the Mixed Integer Linear Program (default: True).

ALGORITHMS:

(Constraints generation)

When the parameter constraint_generation is enabled (default) the following MILP formulation is used to solve the problem:

\mbox{Minimize : }&\sum_{v\in G} b_{v}\\
\mbox{Such that : }&\\
&\forall C\text{ circuits }\subseteq G, \sum_{v\in C}b_{v}\geq 1\\

As the number of circuits contained in a graph is exponential, this LP is solved through constraint generation. This means that the solver is sequentially asked to solve the problem, knowing only a portion of the circuits contained in G, each time adding to the list of its constraints the circuit which its last answer had left intact.

(Another formulation based on an ordering of the vertices)

When the graph is directed, a second (and very slow) formulation is available, which should only be used to check the result of the first implementation in case of doubt.

\mbox{Minimize : }&\sum_{v\in G} b_v\\
\mbox{Such that : }&\\
&\forall (u,v)\in G, d_u-d_v+nb_u+nb_v\geq 0\\
&\forall u\in G, 0\leq d_u\leq |G|\\

A brief explanation:

An acyclic digraph can be seen as a poset, and every poset has a linear extension. This means that in any acyclic digraph the vertices can be ordered with a total order < in such a way that if (u,v)\in G, then u<v. Thus, this linear program is built in order to assign to each vertex v a number d_v\in [0,\dots,n-1] such that if there exists an edge (u,v)\in G then either d_v<d_u or one of u or v is removed. The number of vertices removed is then minimized, which is the objective.

EXAMPLES:

The necessary example:

sage: g = graphs.PetersenGraph()
sage: fvs = g.feedback_vertex_set()
sage: len(fvs)
3
sage: g.delete_vertices(fvs)
sage: g.is_forest()
True

In a digraph built from a graph, any edge is replaced by arcs going in the two opposite directions, thus creating a cycle of length two. Hence, to remove all the cycles from the graph, each edge must see one of its neighbors removed: a feedback vertex set is in this situation a vertex cover:

sage: cycle = graphs.CycleGraph(5)
sage: dcycle = DiGraph(cycle)
sage: cycle.vertex_cover(value_only=True)
3
sage: feedback = dcycle.feedback_vertex_set()
sage: len(feedback)
3
sage: (u,v,l) = cycle.edge_iterator().next()
sage: u in feedback or v in feedback
True

For a circuit, the minimum feedback arc set is clearly 1:

sage: circuit = digraphs.Circuit(5)
sage: circuit.feedback_vertex_set(value_only=True) == 1
True

TESTS:

Comparing with/without constraint generation:

sage: g = digraphs.RandomDirectedGNP(10,.3)
sage: x = g.feedback_vertex_set(value_only = True)
sage: y = g.feedback_vertex_set(value_only = True,
....:          constraint_generation = False)
sage: x == y
True

Bad algorithm:

sage: g = graphs.PetersenGraph()
sage: g.feedback_vertex_set(constraint_generation = False)
Traceback (most recent call last):
...
ValueError: The only implementation available for undirected graphs is with constraint_generation set to True.
flow(x, y, value_only=True, integer=False, use_edge_labels=True, vertex_bound=False, method=None, solver=None, verbose=0)

Returns a maximum flow in the graph from x to y represented by an optimal valuation of the edges. For more information, see the Wikipedia article on maximum flow.

As an optimization problem, is can be expressed this way :

\mbox{Maximize : }&\sum_{e\in G.edges()} w_e b_e\\
\mbox{Such that : }&\forall v \in G, \sum_{(u,v)\in G.edges()} b_{(u,v)}\leq 1\\
&\forall x\in G, b_x\mbox{ is a binary variable}

INPUT:

  • x – Source vertex

  • y – Sink vertex

  • value_only – boolean (default: True)

    • When set to True, only the value of a maximal flow is returned.
    • When set to False, is returned a pair whose first element is the value of the maximum flow, and whose second value is a flow graph (a copy of the current graph, such that each edge has the flow using it as a label, the edges without flow being omitted).
  • integer – boolean (default: False)

    • When set to True, computes an optimal solution under the constraint that the flow going through an edge has to be an integer.
  • use_edge_labels – boolean (default: True)

    • When set to True, computes a maximum flow where each edge has a capacity defined by its label. (If an edge has no label, 1 is assumed.)
    • When set to False, each edge has capacity 1.
  • vertex_bound – boolean (default: False)

    • When set to True, sets the maximum flow leaving a vertex different from x to 1 (useful for vertex connectivity parameters).
  • method – There are currently two different implementations of this method :

    • If method = "FF", a Python implementation of the Ford-Fulkerson algorithm is used (only available when vertex_bound = False)
    • If method = "LP", the flow problem is solved using Linear Programming.
    • If method = None (default), the Ford-Fulkerson implementation is used iif vertex_bound = False.
  • solver – Specify a Linear Program solver to be used. If set to None, the default one is used. function of MixedIntegerLinearProgram. See the documentation of MixedIntegerLinearProgram.solve for more information.

    Only useful when LP is used to solve the flow problem.

  • verbose (integer) – sets the level of verbosity. Set to 0 by default (quiet).

    Only useful when LP is used to solve the flow problem.

Note

Even though the two different implementations are meant to return the same Flow values, they can not be expected to return the same Flow graphs.

Besides, the use of Linear Programming may possibly mean a (slight) numerical noise.

EXAMPLES:

Two basic applications of the flow method for the PappusGraph and the ButterflyGraph with parameter 2

sage: g=graphs.PappusGraph()
sage: g.flow(1,2)
3
sage: b=digraphs.ButterflyGraph(2)
sage: b.flow(('00',1),('00',2))
1

The flow method can be used to compute a matching in a bipartite graph by linking a source s to all the vertices of the first set and linking a sink t to all the vertices of the second set, then computing a maximum s-t flow

sage: g = DiGraph()
sage: g.add_edges([('s',i) for i in range(4)])
sage: g.add_edges([(i,4+j) for i in range(4) for j in range(4)])
sage: g.add_edges([(4+i,'t') for i in range(4)])
sage: [cardinal, flow_graph] = g.flow('s','t',integer=True,value_only=False)
sage: flow_graph.delete_vertices(['s','t'])
sage: len(flow_graph.edges())
4

TESTS:

An exception if raised when forcing “FF” with vertex_bound = True:

sage: g = graphs.PetersenGraph()
sage: g.flow(0,1,vertex_bound = True, method = "FF")
Traceback (most recent call last):
...
ValueError: This method does not support both vertex_bound=True and method="FF".

Or if the method is different from the expected values:

sage: g.flow(0,1, method="Divination")
Traceback (most recent call last):
...
ValueError: The method argument has to be equal to either "FF", "LP" or None

The two methods are indeed returning the same results (possibly with some numerical noise, cf. trac ticket #12362):

sage: g = graphs.RandomGNP(20,.3)
sage: for u,v in g.edges(labels=False):
...      g.set_edge_label(u,v,round(random(),5))
sage: flow_ff = g.flow(0,1, method="FF")
sage: flow_lp = g.flow(0,1,method="LP")
sage: abs(flow_ff-flow_lp) < 0.01
True
genus(set_embedding=True, on_embedding=None, minimal=True, maximal=False, circular=False, ordered=True)

Returns the minimal genus of the graph. The genus of a compact surface is the number of handles it has. The genus of a graph is the minimal genus of the surface it can be embedded into.

Note - This function uses Euler’s formula and thus it is necessary to consider only connected graphs.

INPUT:

  • set_embedding (boolean) - whether or not to store an embedding attribute of the computed (minimal) genus of the graph. (Default is True).
  • on_embedding (dict) - a combinatorial embedding to compute the genus of the graph on. Note that this must be a valid embedding for the graph. The dictionary structure is given by: vertex1: [neighbor1, neighbor2, neighbor3], vertex2: [neighbor] where there is a key for each vertex in the graph and a (clockwise) ordered list of each vertex’s neighbors as values. on_embedding takes precedence over a stored _embedding attribute if minimal is set to False. Note that as a shortcut, the user can enter on_embedding=True to compute the genus on the current _embedding attribute. (see eg’s.)
  • minimal (boolean) - whether or not to compute the minimal genus of the graph (i.e., testing all embeddings). If minimal is False, then either maximal must be True or on_embedding must not be None. If on_embedding is not None, it will take priority over minimal. Similarly, if maximal is True, it will take priority over minimal.
  • maximal (boolean) - whether or not to compute the maximal genus of the graph (i.e., testing all embeddings). If maximal is False, then either minimal must be True or on_embedding must not be None. If on_embedding is not None, it will take priority over maximal. However, maximal takes priority over the default minimal.
  • circular (boolean) - whether or not to compute the genus preserving a planar embedding of the boundary. (Default is False). If circular is True, on_embedding is not a valid option.
  • ordered (boolean) - if circular is True, then whether or not the boundary order may be permuted. (Default is True, which means the boundary order is preserved.)

EXAMPLES:

sage: g = graphs.PetersenGraph()
sage: g.genus() # tests for minimal genus by default
1
sage: g.genus(on_embedding=True, maximal=True) # on_embedding overrides minimal and maximal arguments
1
sage: g.genus(maximal=True) # setting maximal to True overrides default minimal=True
3
sage: g.genus(on_embedding=g.get_embedding()) # can also send a valid combinatorial embedding dict
3
sage: (graphs.CubeGraph(3)).genus()
0
sage: K23 = graphs.CompleteBipartiteGraph(2,3)
sage: K23.genus()
0
sage: K33 = graphs.CompleteBipartiteGraph(3,3)
sage: K33.genus()
1

Using the circular argument, we can compute the minimal genus preserving a planar, ordered boundary:

sage: cube = graphs.CubeGraph(2)
sage: cube.set_boundary(['01','10'])
sage: cube.genus()
0
sage: cube.is_circular_planar()
True
sage: cube.genus(circular=True)
0
sage: cube.genus(circular=True, on_embedding=True)
0
sage: cube.genus(circular=True, maximal=True)
Traceback (most recent call last):
...
NotImplementedError: Cannot compute the maximal genus of a genus respecting a boundary.

Note: not everything works for multigraphs, looped graphs or digraphs. But the minimal genus is ultimately computable for every connected graph – but the embedding we obtain for the simple graph can’t be easily converted to an embedding of a non-simple graph. Also, the maximal genus of a multigraph does not trivially correspond to that of its simple graph.

sage: G = DiGraph({ 0 : [0,1,1,1], 1 : [2,2,3,3], 2 : [1,3,3], 3:[0,3]}) sage: G.genus() Traceback (most recent call last): ... NotImplementedError: Can’t work with embeddings of non-simple graphs sage: G.to_simple().genus() 0 sage: G.genus(set_embedding=False) 0 sage: G.genus(maximal=True, set_embedding=False) Traceback (most recent call last): ... NotImplementedError: Can’t compute the maximal genus of a graph with loops or multiple edges

We break graphs with cut vertices into their blocks, which greatly speeds up computation of minimal genus. This is not implemented for maximal genus.

sage: K5 = graphs.CompleteGraph(5) sage: G = K5.copy() sage: s = 4 sage: for i in range(1,100): ... k = K5.relabel(range(s,s+5),inplace=False) ... G.add_edges(k.edges()) ... s += 4 ... sage: G.genus() 100
get_boundary()

Returns the boundary of the (di)graph.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.set_boundary([0,1,2,3,4])
sage: G.get_boundary()
[0, 1, 2, 3, 4]
get_embedding()

Returns the attribute _embedding if it exists.

_embedding is a dictionary organized with vertex labels as keys and a list of each vertex’s neighbors in clockwise order.

Error-checked to insure valid embedding is returned.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.genus()
1
sage: G.get_embedding()
{0: [1, 4, 5], 1: [0, 2, 6], 2: [1, 3, 7], 3: [2, 4, 8], 4: [0, 3, 9], 5: [0, 7, 8], 6: [1, 9, 8], 7: [2, 5, 9], 8: [3, 6, 5], 9: [4, 6, 7]}
get_pos(dim=2)

Returns the position dictionary, a dictionary specifying the coordinates of each vertex.

EXAMPLES: By default, the position of a graph is None:

sage: G = Graph()
sage: G.get_pos()
sage: G.get_pos() is None
True
sage: P = G.plot(save_pos=True)
sage: G.get_pos()
{}

Some of the named graphs come with a pre-specified positioning:

sage: G = graphs.PetersenGraph()
sage: G.get_pos()        
{0: (...e-17, 1.0),
 ...
 9: (0.475..., 0.154...)}
get_vertex(vertex)

Retrieve the object associated with a given vertex.

INPUT:

  • vertex - the given vertex

EXAMPLES:

sage: d = {0 : graphs.DodecahedralGraph(), 1 : graphs.FlowerSnark(), 2 : graphs.MoebiusKantorGraph(), 3 : graphs.PetersenGraph() }
sage: d[2]
Moebius-Kantor Graph: Graph on 16 vertices
sage: T = graphs.TetrahedralGraph()
sage: T.vertices()
[0, 1, 2, 3]
sage: T.set_vertices(d)
sage: T.get_vertex(1)
Flower Snark: Graph on 20 vertices
get_vertices(verts=None)

Return a dictionary of the objects associated to each vertex.

INPUT:

  • verts - iterable container of vertices

EXAMPLES:

sage: d = {0 : graphs.DodecahedralGraph(), 1 : graphs.FlowerSnark(), 2 : graphs.MoebiusKantorGraph(), 3 : graphs.PetersenGraph() }
sage: T = graphs.TetrahedralGraph()
sage: T.set_vertices(d)
sage: T.get_vertices([1,2])
{1: Flower Snark: Graph on 20 vertices,
 2: Moebius-Kantor Graph: Graph on 16 vertices}
girth()

Computes the girth of the graph. For directed graphs, computes the girth of the undirected graph.

The girth is the length of the shortest cycle in the graph. Graphs without cycles have infinite girth.

EXAMPLES:

sage: graphs.TetrahedralGraph().girth()
3
sage: graphs.CubeGraph(3).girth()
4
sage: graphs.PetersenGraph().girth()
5
sage: graphs.HeawoodGraph().girth()
6
sage: graphs.trees(9).next().girth()
+Infinity

See also

TESTS:

Prior to trac ticket #12243, the girth computation assumed vertices were integers (and failed). The example below tests the computation for graphs with vertices that are not integers. In this example the vertices are sets.

sage: G = graphs.OddGraph(3)
sage: type(G.vertices()[0])
<class 'sage.sets.set.Set_object_enumerated_with_category'>
sage: G.girth()
5

Ticket trac ticket #12355:

sage: H=Graph([(0, 1), (0, 3), (0, 4), (0, 5), (1, 2), (1, 3), (1, 4), (1, 6), (2, 5), (3, 4), (5, 6)])
sage: H.girth()
3

Girth < 3 (see trac ticket #12355):

sage: g = graphs.PetersenGraph()
sage: g.allow_multiple_edges(True)
sage: g.allow_loops(True)
sage: g.girth()
5
sage: g.add_edge(0,0)
sage: g.girth()
1
sage: g.delete_edge(0,0)
sage: g.add_edge(0,1)
sage: g.girth()
2
sage: g.delete_edge(0,1)
sage: g.girth()
5
sage: g = DiGraph(g)
sage: g.girth()
2
graphplot(**options)

Returns a GraphPlot object.

EXAMPLES:

Creating a graphplot object uses the same options as graph.plot():

sage: g = Graph({}, loops=True, multiedges=True, sparse=True)
sage: g.add_edges([(0,0,'a'),(0,0,'b'),(0,1,'c'),(0,1,'d'),
...     (0,1,'e'),(0,1,'f'),(0,1,'f'),(2,1,'g'),(2,2,'h')])
sage: g.set_boundary([0,1])
sage: GP = g.graphplot(edge_labels=True, color_by_label=True, edge_style='dashed')
sage: GP.plot()

We can modify the graphplot object. Notice that the changes are cumulative:

sage: GP.set_edges(edge_style='solid')
sage: GP.plot()
sage: GP.set_vertices(talk=True)
sage: GP.plot()
graphviz_string(edge_color=None, vertex_labels=True, edge_labels=False, labels='string', color_by_label=False, edge_colors=None, edge_options=(), **options)

Returns a representation in the dot language.

The dot language is a text based format for graphs. It is used by the software suite graphviz. The specifications of the language are available on the web (see the reference [dotspec]).

INPUT:

  • labels - “string” or “latex” (default: “string”). If labels is string latex command are not interpreted. This option stands for both vertex labels and edge labels.

  • vertex_labels - boolean (default: True) whether to add the labels on vertices.

  • edge_labels - boolean (default: False) whether to add the labels on edges.

  • edge_color - (default: None) specify a default color for the edges.

  • edge_colors - (default: None) a dictionary whose keys are colors and values are list of edges. The list of edges need not to be complete in which case the default color is used.

  • color_by_label - a boolean or dictionary or function (default: False)

    whether to color each edge with a different color according to its label; the colors are chosen along a rainbow, unless they are specified by a function or dictionary mapping labels to colors; this option is incompatible with edge_color and edge_colors.

  • edge_options - a function (or tuple thereof) mapping edges to a dictionary of options for this edge.

EXAMPLES:

sage: G = Graph({0:{1:None,2:None}, 1:{0:None,2:None}, 2:{0:None,1:None,3:'foo'}, 3:{2:'foo'}},sparse=True)
sage: print G.graphviz_string(edge_labels=True)
graph {
  "0" [label="0"];
  "1" [label="1"];
  "2" [label="2"];
  "3" [label="3"];

  "0" -- "1";
  "0" -- "2";
  "1" -- "2";
  "2" -- "3" [label="foo"];
}

A variant, with the labels in latex, for post-processing with dot2tex:

sage: print G.graphviz_string(edge_labels=True,labels = "latex")
graph {
  node [shape="plaintext"];
  "0" [label=" ", texlbl="$0$"];
  "1" [label=" ", texlbl="$1$"];
  "2" [label=" ", texlbl="$2$"];
  "3" [label=" ", texlbl="$3$"];

  "0" -- "1";
  "0" -- "2";
  "1" -- "2";
  "2" -- "3" [label=" ", texlbl="$\text{\texttt{foo}}$"];
}

Same, with a digraph and a color for edges:

sage: G = DiGraph({0:{1:None,2:None}, 1:{2:None}, 2:{3:'foo'}, 3:{}} ,sparse=True)
sage: print G.graphviz_string(edge_color="red")
digraph {
  "0" [label="0"];
  "1" [label="1"];
  "2" [label="2"];
  "3" [label="3"];

  edge [color="red"];
  "0" -> "1";
  "0" -> "2";
  "1" -> "2";
  "2" -> "3";
}

A digraph using latex labels for vertices and edges:

sage: f(x) = -1/x
sage: g(x) = 1/(x+1)
sage: G = DiGraph()
sage: G.add_edges([(i,f(i),f) for i in (1,2,1/2,1/4)])
sage: G.add_edges([(i,g(i),g) for i in (1,2,1/2,1/4)])
sage: print G.graphviz_string(labels="latex",edge_labels=True)
digraph {
  node [shape="plaintext"];
  "2/3" [label=" ", texlbl="$\frac{2}{3}$"];
  "1/3" [label=" ", texlbl="$\frac{1}{3}$"];
  "1/2" [label=" ", texlbl="$\frac{1}{2}$"];
  "1" [label=" ", texlbl="$1$"];
  "1/4" [label=" ", texlbl="$\frac{1}{4}$"];
  "4/5" [label=" ", texlbl="$\frac{4}{5}$"];
  "-4" [label=" ", texlbl="$-4$"];
  "2" [label=" ", texlbl="$2$"];
  "-2" [label=" ", texlbl="$-2$"];
  "-1/2" [label=" ", texlbl="$-\frac{1}{2}$"];
  "-1" [label=" ", texlbl="$-1$"];

  "1/2" -> "-2" [label=" ", texlbl="$x \ {\mapsto}\ -\frac{1}{x}$"];
  "1/2" -> "2/3" [label=" ", texlbl="$x \ {\mapsto}\ \frac{1}{x + 1}$"];
  "1" -> "-1" [label=" ", texlbl="$x \ {\mapsto}\ -\frac{1}{x}$"];
  "1" -> "1/2" [label=" ", texlbl="$x \ {\mapsto}\ \frac{1}{x + 1}$"];
  "1/4" -> "-4" [label=" ", texlbl="$x \ {\mapsto}\ -\frac{1}{x}$"];
  "1/4" -> "4/5" [label=" ", texlbl="$x \ {\mapsto}\ \frac{1}{x + 1}$"];
  "2" -> "-1/2" [label=" ", texlbl="$x \ {\mapsto}\ -\frac{1}{x}$"];
  "2" -> "1/3" [label=" ", texlbl="$x \ {\mapsto}\ \frac{1}{x + 1}$"];
}

sage: print G.graphviz_string(labels="latex",color_by_label=True)
digraph {
  node [shape="plaintext"];
  "2/3" [label=" ", texlbl="$\frac{2}{3}$"];
  "1/3" [label=" ", texlbl="$\frac{1}{3}$"];
  "1/2" [label=" ", texlbl="$\frac{1}{2}$"];
  "1" [label=" ", texlbl="$1$"];
  "1/4" [label=" ", texlbl="$\frac{1}{4}$"];
  "4/5" [label=" ", texlbl="$\frac{4}{5}$"];
  "-4" [label=" ", texlbl="$-4$"];
  "2" [label=" ", texlbl="$2$"];
  "-2" [label=" ", texlbl="$-2$"];
  "-1/2" [label=" ", texlbl="$-\frac{1}{2}$"];
  "-1" [label=" ", texlbl="$-1$"];

  "1/2" -> "-2" [color = "#ff0000"];
  "1/2" -> "2/3" [color = "#00ffff"];
  "1" -> "-1" [color = "#ff0000"];
  "1" -> "1/2" [color = "#00ffff"];
  "1/4" -> "-4" [color = "#ff0000"];
  "1/4" -> "4/5" [color = "#00ffff"];
  "2" -> "-1/2" [color = "#ff0000"];
  "2" -> "1/3" [color = "#00ffff"];
}

sage: print G.graphviz_string(labels="latex",color_by_label={ f: "red", g: "blue" })
digraph {
  node [shape="plaintext"];
  "2/3" [label=" ", texlbl="$\frac{2}{3}$"];
  "1/3" [label=" ", texlbl="$\frac{1}{3}$"];
  "1/2" [label=" ", texlbl="$\frac{1}{2}$"];
  "1" [label=" ", texlbl="$1$"];
  "1/4" [label=" ", texlbl="$\frac{1}{4}$"];
  "4/5" [label=" ", texlbl="$\frac{4}{5}$"];
  "-4" [label=" ", texlbl="$-4$"];
  "2" [label=" ", texlbl="$2$"];
  "-2" [label=" ", texlbl="$-2$"];
  "-1/2" [label=" ", texlbl="$-\frac{1}{2}$"];
  "-1" [label=" ", texlbl="$-1$"];

  "1/2" -> "-2" [color = "red"];
  "1/2" -> "2/3" [color = "blue"];
  "1" -> "-1" [color = "red"];
  "1" -> "1/2" [color = "blue"];
  "1/4" -> "-4" [color = "red"];
  "1/4" -> "4/5" [color = "blue"];
  "2" -> "-1/2" [color = "red"];
  "2" -> "1/3" [color = "blue"];
}

Edge-specific options can also be specified by providing a function (or tuple thereof) which maps each edge to a dictionary of options. Valid options are “color”, “backward” (a boolean), “dot” (a string containing a sequence of options in dot format), “label” (a string), “label_style” (“string” or “latex”), “edge_string” (“–” or “->”). Here we state that the graph should be laid out so that edges starting from 1 are going backward (e.g. going up instead of down):

sage: def edge_options((u,v,label)):
...       return { "backward": u == 1 }
sage: print G.graphviz_string(edge_options = edge_options)
digraph {
  "2/3" [label="2/3"];
  "1/3" [label="1/3"];
  "1/2" [label="1/2"];
  "1" [label="1"];
  "1/4" [label="1/4"];
  "4/5" [label="4/5"];
  "-4" [label="-4"];
  "2" [label="2"];
  "-2" [label="-2"];
  "-1/2" [label="-1/2"];
  "-1" [label="-1"];

  "1/2" -> "-2";
  "1/2" -> "2/3";
  "-1" -> "1" [dir=back];
  "1/2" -> "1" [dir=back];
  "1/4" -> "-4";
  "1/4" -> "4/5";
  "2" -> "-1/2";
  "2" -> "1/3";
}

We now test all options:

sage: def edge_options((u,v,label)):
...       options = { "color": { f: "red", g: "blue" }[label] }
...       if (u,v) == (1/2, -2): options["label"]       = "coucou"; options["label_style"] = "string"
...       if (u,v) == (1/2,2/3): options["dot"]         = "x=1,y=2"
...       if (u,v) == (1,   -1): options["label_style"] = "latex"
...       if (u,v) == (1,  1/2): options["edge_string"] = "<-"
...       if (u,v) == (1/2,  1): options["backward"]    = True
...       return options
sage: print G.graphviz_string(edge_options = edge_options)
digraph {
  "2/3" [label="2/3"];
  "1/3" [label="1/3"];
  "1/2" [label="1/2"];
  "1" [label="1"];
  "1/4" [label="1/4"];
  "4/5" [label="4/5"];
  "-4" [label="-4"];
  "2" [label="2"];
  "-2" [label="-2"];
  "-1/2" [label="-1/2"];
  "-1" [label="-1"];

  "1/2" -> "-2" [label="coucou", color = "red"];
  "1/2" -> "2/3" [x=1,y=2, color = "blue"];
  "1" -> "-1" [label=" ", texlbl="$x \ {\mapsto}\ -\frac{1}{x}$", color = "red"];
  "1" <- "1/2" [color = "blue"];
  "1/4" -> "-4" [color = "red"];
  "1/4" -> "4/5" [color = "blue"];
  "2" -> "-1/2" [color = "red"];
  "2" -> "1/3" [color = "blue"];
}

TESTS:

The following digraph has tuples as vertices:

sage: print digraphs.ButterflyGraph(1).graphviz_string()
digraph {
  "1,1" [label="('1', 1)"];
  "0,0" [label="('0', 0)"];
  "1,0" [label="('1', 0)"];
  "0,1" [label="('0', 1)"];

  "0,0" -> "1,1";
  "0,0" -> "0,1";
  "1,0" -> "1,1";
  "1,0" -> "0,1";
}

The following digraph has vertices with newlines in their string representations:

sage: m1 = matrix(3,3)
sage: m2 = matrix(3,3, 1)
sage: m1.set_immutable()
sage: m2.set_immutable()
sage: g = DiGraph({ m1: [m2] })
sage: print g.graphviz_string()
digraph {
  "000000000" [label="[0 0 0]\n\
  [0 0 0]\n\
  [0 0 0]"];
  "100010001" [label="[1 0 0]\n\
  [0 1 0]\n\
  [0 0 1]"];

  "000000000" -> "100010001";
}

REFERENCES:

[dotspec]http://www.graphviz.org/doc/info/lang.html
graphviz_to_file_named(filename, **options)

Write a representation in the dot in a file.

The dot language is a plaintext format for graph structures. See the documentation of graphviz_string() for available options.

INPUT:

filename - the name of the file to write in

options - options for the graphviz string

EXAMPLES:

sage: G = Graph({0:{1:None,2:None}, 1:{0:None,2:None}, 2:{0:None,1:None,3:'foo'}, 3:{2:'foo'}},sparse=True)
sage: tempfile = os.path.join(SAGE_TMP, 'temp_graphviz')
sage: G.graphviz_to_file_named(tempfile, edge_labels=True)
sage: print open(tempfile).read()
graph {
  "0" [label="0"];
  "1" [label="1"];
  "2" [label="2"];
  "3" [label="3"];

  "0" -- "1";
  "0" -- "2";
  "1" -- "2";
  "2" -- "3" [label="foo"];
}
hamiltonian_cycle(algorithm='tsp')

Returns a Hamiltonian cycle/circuit of the current graph/digraph

A graph (resp. digraph) is said to be Hamiltonian if it contains as a subgraph a cycle (resp. a circuit) going through all the vertices.

Computing a Hamiltonian cycle/circuit being NP-Complete, this algorithm could run for some time depending on the instance.

ALGORITHM:

See Graph.traveling_salesman_problem for ‘tsp’ algorithm and find_hamiltonian from sage.graphs.generic_graph_pyx for ‘backtrack’ algorithm.

INPUT:

  • algorithm - one of ‘tsp’ or ‘backtrack’.

OUTPUT:

If using the ‘tsp’ algorithm, returns a Hamiltonian cycle/circuit if it exists; otherwise, raises a ValueError exception. If using the ‘backtrack’ algorithm, returns a pair (B,P). If B is True then P is a Hamiltonian cycle and if B is False, P is a longest path found by the algorithm. Observe that if B is False, the graph may still be Hamiltonian. The ‘backtrack’ algorithm is only implemented for undirected graphs.

Warning

The ‘backtrack’ algorithm may loop endlessly on graphs with vertices of degree 1.

NOTE:

This function, as is_hamiltonian, computes a Hamiltonian cycle if it exists : the user should NOT test for Hamiltonicity using is_hamiltonian before calling this function, as it would result in computing it twice.

The backtrack algorithm is only implemented for undirected graphs.

EXAMPLES:

The Heawood Graph is known to be Hamiltonian

sage: g = graphs.HeawoodGraph()
sage: g.hamiltonian_cycle()
TSP from Heawood graph: Graph on 14 vertices

The Petersen Graph, though, is not

sage: g = graphs.PetersenGraph()
sage: g.hamiltonian_cycle()
Traceback (most recent call last):
...
ValueError: The given graph is not hamiltonian

Now, using the backtrack algorithm in the Heawood graph

sage: G=graphs.HeawoodGraph()
sage: G.hamiltonian_cycle(algorithm='backtrack')
(True, [11, 10, 1, 2, 3, 4, 9, 8, 7, 6, 5, 0, 13, 12])

And now in the Petersen graph

sage: G=graphs.PetersenGraph()
sage: G.hamiltonian_cycle(algorithm='backtrack')
(False, [6, 8, 5, 0, 1, 2, 7, 9, 4, 3])

Finally, we test the algorithm in a cube graph, which is Hamiltonian

sage: G=graphs.CubeGraph(3)
sage: G.hamiltonian_cycle(algorithm='backtrack')
(True, ['010', '110', '100', '000', '001', '101', '111', '011'])
has_edge(u, v=None, label=None)

Returns True if (u, v) is an edge, False otherwise.

INPUT: The following forms are accepted by NetworkX:

  • G.has_edge( 1, 2 )
  • G.has_edge( (1, 2) )
  • G.has_edge( 1, 2, ‘label’ )
  • G.has_edge( (1, 2, ‘label’) )

EXAMPLES:

sage: graphs.EmptyGraph().has_edge(9,2)
False
sage: DiGraph().has_edge(9,2)
False
sage: G = Graph(sparse=True)
sage: G.add_edge(0,1,"label")
sage: G.has_edge(0,1,"different label")
False
sage: G.has_edge(0,1,"label")
True
has_loops()

Returns whether there are loops in the (di)graph.

EXAMPLES:

sage: G = Graph(loops=True); G
Looped graph on 0 vertices
sage: G.has_loops()
False
sage: G.allows_loops()
True
sage: G.add_edge((0,0))
sage: G.has_loops()
True
sage: G.loops()
[(0, 0, None)]
sage: G.allow_loops(False); G
Graph on 1 vertex
sage: G.has_loops()
False
sage: G.edges()
[]

sage: D = DiGraph(loops=True); D
Looped digraph on 0 vertices
sage: D.has_loops()
False
sage: D.allows_loops()
True
sage: D.add_edge((0,0))
sage: D.has_loops()
True
sage: D.loops()
[(0, 0, None)]
sage: D.allow_loops(False); D
Digraph on 1 vertex
sage: D.has_loops()
False
sage: D.edges()
[]
has_multiple_edges(to_undirected=False)

Returns whether there are multiple edges in the (di)graph.

INPUT:

  • to_undirected – (default: False) If True, runs the test on the undirected version of a DiGraph. Otherwise, treats DiGraph edges (u,v) and (v,u) as unique individual edges.

EXAMPLES:

sage: G = Graph(multiedges=True,sparse=True); G
Multi-graph on 0 vertices
sage: G.has_multiple_edges()
False
sage: G.allows_multiple_edges()
True
sage: G.add_edges([(0,1)]*3)
sage: G.has_multiple_edges()
True
sage: G.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: G.allow_multiple_edges(False); G
Graph on 2 vertices
sage: G.has_multiple_edges()
False
sage: G.edges()
[(0, 1, None)]

sage: D = DiGraph(multiedges=True,sparse=True); D
Multi-digraph on 0 vertices
sage: D.has_multiple_edges()
False
sage: D.allows_multiple_edges()
True
sage: D.add_edges([(0,1)]*3)
sage: D.has_multiple_edges()
True
sage: D.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: D.allow_multiple_edges(False); D
Digraph on 2 vertices
sage: D.has_multiple_edges()
False
sage: D.edges()
[(0, 1, None)]

sage: G = DiGraph({1:{2: 'h'}, 2:{1:'g'}},sparse=True)
sage: G.has_multiple_edges()
False
sage: G.has_multiple_edges(to_undirected=True)
True
sage: G.multiple_edges()
[]
sage: G.multiple_edges(to_undirected=True)
[(1, 2, 'h'), (2, 1, 'g')]
has_vertex(vertex)

Return True if vertex is one of the vertices of this graph.

INPUT:

  • vertex - an integer

OUTPUT:

  • bool - True or False

EXAMPLES:

sage: g = Graph({0:[1,2,3], 2:[4]}); g
Graph on 5 vertices
sage: 2 in g
True
sage: 10 in g
False
sage: graphs.PetersenGraph().has_vertex(99)
False
incidence_matrix(sparse=True)

Returns the incidence matrix of the (di)graph.

Each row is a vertex, and each column is an edge. Note that in the case of graphs, there is a choice of orientation for each edge.

EXAMPLES:

sage: G = graphs.CubeGraph(3)
sage: G.incidence_matrix()
[-1 -1 -1  0  0  0  0  0  0  0  0  0]
[ 0  0  1 -1 -1  0  0  0  0  0  0  0]
[ 0  1  0  0  0 -1 -1  0  0  0  0  0]
[ 0  0  0  0  1  0  1 -1  0  0  0  0]
[ 1  0  0  0  0  0  0  0 -1 -1  0  0]
[ 0  0  0  1  0  0  0  0  0  1 -1  0]
[ 0  0  0  0  0  1  0  0  1  0  0 -1]
[ 0  0  0  0  0  0  0  1  0  0  1  1]

A well known result states that the product of the incidence matrix with its transpose is in fact the Kirchhoff matrix:

sage: G = graphs.PetersenGraph()
sage: G.incidence_matrix()*G.incidence_matrix().transpose() == G.kirchhoff_matrix()
True
sage: D = DiGraph( { 0: [1,2,3], 1: [0,2], 2: [3], 3: [4], 4: [0,5], 5: [1] } )
sage: D.incidence_matrix()
[-1 -1 -1  0  0  0  0  0  1  1]
[ 0  0  1 -1  0  0  0  1 -1  0]
[ 0  1  0  1 -1  0  0  0  0  0]
[ 1  0  0  0  1 -1  0  0  0  0]
[ 0  0  0  0  0  1 -1  0  0 -1]
[ 0  0  0  0  0  0  1 -1  0  0]
interior_paths(start, end)

Returns an exhaustive list of paths (also lists) through only interior vertices from vertex start to vertex end in the (di)graph.

Note - start and end do not necessarily have to be boundary vertices.

INPUT:

  • start - the vertex of the graph to search for paths from
  • end - the vertex of the graph to search for paths to

EXAMPLES:

sage: eg1 = Graph({0:[1,2], 1:[4], 2:[3,4], 4:[5], 5:[6]})
sage: sorted(eg1.all_paths(0,6))
[[0, 1, 4, 5, 6], [0, 2, 4, 5, 6]]
sage: eg2 = copy(eg1)
sage: eg2.set_boundary([0,1,3])
sage: sorted(eg2.interior_paths(0,6))
[[0, 2, 4, 5, 6]]
sage: sorted(eg2.all_paths(0,6))
[[0, 1, 4, 5, 6], [0, 2, 4, 5, 6]]
sage: eg3 = graphs.PetersenGraph()
sage: eg3.set_boundary([0,1,2,3,4])
sage: sorted(eg3.all_paths(1,4))
[[1, 0, 4],
 [1, 0, 5, 7, 2, 3, 4],
 [1, 0, 5, 7, 2, 3, 8, 6, 9, 4],
 [1, 0, 5, 7, 9, 4],
 [1, 0, 5, 7, 9, 6, 8, 3, 4],
 [1, 0, 5, 8, 3, 2, 7, 9, 4],
 [1, 0, 5, 8, 3, 4],
 [1, 0, 5, 8, 6, 9, 4],
 [1, 0, 5, 8, 6, 9, 7, 2, 3, 4],
 [1, 2, 3, 4],
 [1, 2, 3, 8, 5, 0, 4],
 [1, 2, 3, 8, 5, 7, 9, 4],
 [1, 2, 3, 8, 6, 9, 4],
 [1, 2, 3, 8, 6, 9, 7, 5, 0, 4],
 [1, 2, 7, 5, 0, 4],
 [1, 2, 7, 5, 8, 3, 4],
 [1, 2, 7, 5, 8, 6, 9, 4],
 [1, 2, 7, 9, 4],
 [1, 2, 7, 9, 6, 8, 3, 4],
 [1, 2, 7, 9, 6, 8, 5, 0, 4],
 [1, 6, 8, 3, 2, 7, 5, 0, 4],
 [1, 6, 8, 3, 2, 7, 9, 4],
 [1, 6, 8, 3, 4],
 [1, 6, 8, 5, 0, 4],
 [1, 6, 8, 5, 7, 2, 3, 4],
 [1, 6, 8, 5, 7, 9, 4],
 [1, 6, 9, 4],
 [1, 6, 9, 7, 2, 3, 4],
 [1, 6, 9, 7, 2, 3, 8, 5, 0, 4],
 [1, 6, 9, 7, 5, 0, 4],
 [1, 6, 9, 7, 5, 8, 3, 4]]
sage: sorted(eg3.interior_paths(1,4))
[[1, 6, 8, 5, 7, 9, 4], [1, 6, 9, 4]]
sage: dg = DiGraph({0:[1,3,4], 1:[3], 2:[0,3,4],4:[3]}, boundary=[4])
sage: sorted(dg.all_paths(0,3))
[[0, 1, 3], [0, 3], [0, 4, 3]]
sage: sorted(dg.interior_paths(0,3))
[[0, 1, 3], [0, 3]]
sage: ug = dg.to_undirected()
sage: sorted(ug.all_paths(0,3))
[[0, 1, 3], [0, 2, 3], [0, 2, 4, 3], [0, 3], [0, 4, 2, 3], [0, 4, 3]]
sage: sorted(ug.interior_paths(0,3))
[[0, 1, 3], [0, 2, 3], [0, 3]]
is_chordal(certificate=False, algorithm='B')

Tests whether the given graph is chordal.

A Graph G is said to be chordal if it contains no induced hole (a cycle of length at least 4).

Alternatively, chordality can be defined using a Perfect Elimination Order :

A Perfect Elimination Order of a graph G is an ordering v_1,...,v_n of its vertex set such that for all i, the neighbors of v_i whose index is greater that i induce a complete subgraph in G. Hence, the graph G can be totally erased by successively removing vertices whose neighborhood is a clique (also called simplicial vertices) [Fulkerson65].

(It can be seen that if G contains an induced hole, then it can not have a perfect elimination order. Indeed, if we write h_1,...,h_k the k vertices of such a hole, then the first of those vertices to be removed would have two non-adjacent neighbors in the graph.)

A Graph is then chordal if and only if it has a Perfect Elimination Order.

INPUT:

  • certificate (boolean) – Whether to return a certificate.

    • If certificate = False (default), returns True or False accordingly.

    • If certificate = True, returns :

      • (True, peo) when the graph is chordal, where peo is a perfect elimination order of its vertices.
      • (False, Hole) when the graph is not chordal, where Hole (a Graph object) is an induced subgraph of self isomorphic to a hole.
  • algorithm – Two algorithms are available for this method (see next section), which can be selected by setting algorithm = "A" or algorithm = "B" (default). While they will agree on whether the given graph is chordal, they can not be expected to return the same certificates.

ALGORITHM:

This algorithm works through computing a Lex BFS on the graph, then checking whether the order is a Perfect Elimination Order by computing for each vertex v the subgraph induces by its non-deleted neighbors, then testing whether this graph is complete.

This problem can be solved in O(m) [Rose75] ( where m is the number of edges in the graph ) but this implementation is not linear because of the complexity of Lex BFS.

Note

Because of a past bug (#11735, #11961), the first implementation (algorithm A) of this method sometimes returned as certificates subgraphs which were not holes. Since then, this bug has been fixed and the values are now double-checked before being returned, so that the algorithm only returns correct values or raises an exception. In the case where an exception is raised, the user is advised to switch to the other algorithm. And to please report the bug :-)

EXAMPLES:

The lexicographic product of a Path and a Complete Graph is chordal

sage: g = graphs.PathGraph(5).lexicographic_product(graphs.CompleteGraph(3))
sage: g.is_chordal()
True

The same goes with the product of a random lobster ( which is a tree ) and a Complete Graph

sage: g = graphs.RandomLobster(10,.5,.5).lexicographic_product(graphs.CompleteGraph(3))
sage: g.is_chordal()
True

The disjoint union of chordal graphs is still chordal:

sage: (2*g).is_chordal()
True

Let us check the certificate given by Sage is indeed a perfect elimintion order:

sage: (_, peo) = g.is_chordal(certificate = True)
sage: for v in peo:
...       if not g.subgraph(g.neighbors(v)).is_clique():
...            print "This should never happen !"
...       g.delete_vertex(v)
sage: print "Everything is fine !"
Everything is fine !

Of course, the Petersen Graph is not chordal as it has girth 5

sage: g = graphs.PetersenGraph()
sage: g.girth()
5
sage: g.is_chordal()
False

We can even obtain such a cycle as a certificate

sage: (_, hole) = g.is_chordal(certificate = True)
sage: hole
Subgraph of (Petersen graph): Graph on 5 vertices
sage: hole.is_isomorphic(graphs.CycleGraph(5))
True

TESTS:

This shouldn’t fail (trac 10899):

sage: Graph(1).is_chordal()
True
sage: for g in graphs(5):
...     try:
...         forget = g.is_chordal()
...     except StandardError:
...         print("Oh no.")

REFERENCES:

[Rose75]Rose, D.J. and Tarjan, R.E., Algorithmic aspects of vertex elimination, Proceedings of seventh annual ACM symposium on Theory of computing Page 254, ACM 1975
[Fulkerson65]Fulkerson, D.R. and Gross, OA Incidence matrices and interval graphs Pacific J. Math 1965 Vol. 15, number 3, pages 835–855

TESTS:

Trac Ticket #11735:

sage: g = Graph({3:[2,1,4],2:[1],4:[1],5:[2,1,4]})  
sage: _, g1 = g.is_chordal(certificate=True); g1.is_chordal() 
False 
sage: g1.is_isomorphic(graphs.CycleGraph(g1.order())) 
True 
is_circulant(certificate=False)

Tests whether the graph is circulant.

For more information on circulant graphs, see the Wikipedia page on circulant graphs.

INPUT:

  • certificate (boolean) – whether to return a certificate for yes-answers. See OUTPUT section. Set to False by default.

OUTPUT:

When certificate is set to False (default) this method only returns True or False answers. When certificate is set to True, the method either returns (False, None) or (True, lists_of_parameters) each element of lists_of_parameters can be used to define the graph as a circulant graph.

See the documentation of graphs.CirculantGraph() and digraphs.CirculantGraph() for more information, and the examples below.

See also

CirculantGraph() – a constructor for circulant graphs.

EXAMPLES:

The Petersen graph is not a circulant graph:

sage: g = graphs.PetersenGraph()
sage: g.is_circulant()
False

A cycle is obviously a circulant graph, but several sets of parameters can be used to define it:

sage: g = graphs.CycleGraph(5)
sage: g.is_circulant(certificate = True)
(True, [(5, [1, 4]), (5, [2, 3])])

The same goes for directed graphs:

sage: g = digraphs.Circuit(5)
sage: g.is_circulant(certificate = True)
(True, [(5, [1]), (5, [3]), (5, [2]), (5, [4])])

With this information, it is very easy to create (and plot) all possible drawings of a circulant graph:

sage: g = graphs.CirculantGraph(13, [2, 3, 10, 11])
sage: for param in g.is_circulant(certificate = True)[1]:
...      graphs.CirculantGraph(*param)
Circulant graph ([2, 3, 10, 11]): Graph on 13 vertices
Circulant graph ([1, 5, 8, 12]): Graph on 13 vertices
Circulant graph ([4, 6, 7, 9]): Graph on 13 vertices

TESTS:

sage: digraphs.DeBruijn(3,1).is_circulant(certificate = True)
(True, [(3, [0, 1, 2])])
sage: Graph(1).is_circulant(certificate = True)
(True, (1, []))
sage: Graph(0).is_circulant(certificate = True)
(True, (0, []))
sage: Graph([(0,0)]).is_circulant(certificate = True)
(True, (1, [0]))
is_circular_planar(on_embedding=None, kuratowski=False, set_embedding=True, boundary=None, ordered=False, set_pos=False)

Tests whether the graph is circular planar (outerplanar)

A graph is circular planar if it has a planar embedding in which all vertices can be drawn in order on a circle. This method can also be used to check the existence of a planar embedding in which the vertices of a specific set (the boundary) can be drawn on a circle, all other vertices being drawn inside of the circle. An order can be defined on the vertices of the boundary in order to define how they are to appear on the circle.

INPUT:

  • kuratowski (boolean) - if set to True, returns a tuple with

    boolean first entry and the Kuratowski subgraph or minor as the second entry (see OUTPUT below). It is set to False by default.

  • on_embedding (boolean) - the embedding dictionary to test

    planarity on. (i.e.: will return True or False only for the given embedding). It is set to False by default.

  • set_embedding (boolean) - whether or not to set the instance field

    variable that contains a combinatorial embedding (clockwise ordering of neighbors at each vertex). This value will only be set if a circular planar embedding is found. It is stored as a Python dict: v1: [n1,n2,n3] where v1 is a vertex and n1,n2,n3 are its neighbors. It is set to True by default.

  • boundary - a set of vertices that are required to be drawn on the circle, all others being drawn inside of it. It is set to None by default, meaning that all vertices should be drawn on the boundary.

  • ordered (boolean) - whether or not to consider the order of the

    boundary. It is set to False by default, and required boundary to be defined.

  • set_pos - whether or not to set the position dictionary (for

    plotting) to reflect the combinatorial embedding. Note that this value will default to False if set_emb is set to False. Also, the position dictionary will only be updated if a circular planar embedding is found.

OUTPUT:

The method returns True if the graph is circular planar, and False if it is not.

If kuratowski is set to True, then this function will return a tuple, whose first entry is a boolean and whose second entry is the Kuratowski subgraph or minor isolated by the Boyer-Myrvold algorithm. Note that this graph might contain a vertex or edges that were not in the initial graph. These would be elements referred to below as parts of the wheel and the star, which were added to the graph to require that the boundary can be drawn on the boundary of a disc, with all other vertices drawn inside (and no edge crossings). For more information, see [Kirkman].

ALGORITHM:

This is a linear time algorithm to test for circular planarity. It relies on the edge-addition planarity algorithm due to Boyer-Myrvold. We accomplish linear time for circular planarity by modifying the graph before running the general planarity algorithm.

REFERENCE:

[BM04]John M. Boyer and Wendy J. Myrvold, On the Cutting Edge: Simplified O(n) Planarity by Edge Addition. Journal of Graph Algorithms and Applications, Vol. 8, No. 3, pp. 241-273, 2004.
[Kirkman]Kirkman, Emily A. O(n) Circular Planarity Testing. [Online] Available: soon!

EXAMPLES:

sage: g439 = Graph({1:[5,7], 2:[5,6], 3:[6,7], 4:[5,6,7]})
sage: g439.set_boundary([1,2,3,4])
sage: g439.show()
sage: g439.is_circular_planar(boundary = [1,2,3,4])
False
sage: g439.is_circular_planar(kuratowski=True, boundary = [1,2,3,4])
(False, Graph on 8 vertices)
sage: g439.is_circular_planar(kuratowski=True, boundary = [1,2,3])
(True, None)
sage: g439.get_embedding()
{1: [7, 5],
2: [5, 6],
3: [6, 7],
4: [7, 6, 5],
5: [1, 4, 2],
6: [2, 4, 3],
7: [3, 4, 1]}

Order matters:

sage: K23 = graphs.CompleteBipartiteGraph(2,3)
sage: K23.is_circular_planar(boundary = [0,1,2,3])
True
sage: K23.is_circular_planar(ordered=True, boundary = [0,1,2,3])
False

With a different order:

sage: K23.is_circular_planar(set_embedding=True, boundary = [0,2,1,3])
True
is_clique(vertices=None, directed_clique=False)

Tests whether a set of vertices is a clique

A clique is a set of vertices such that there is an edge between any two vertices.

INPUT:

  • vertices - Vertices can be a single vertex or an iterable container of vertices, e.g. a list, set, graph, file or numeric array. If not passed, defaults to the entire graph.
  • directed_clique - (default False) If set to False, only consider the underlying undirected graph. If set to True and the graph is directed, only return True if all possible edges in _both_ directions exist.

EXAMPLES:

sage: g = graphs.CompleteGraph(4)
sage: g.is_clique([1,2,3])
True
sage: g.is_clique()
True
sage: h = graphs.CycleGraph(4)
sage: h.is_clique([1,2])
True
sage: h.is_clique([1,2,3])
False
sage: h.is_clique()
False
sage: i = graphs.CompleteGraph(4).to_directed()
sage: i.delete_edge([0,1])
sage: i.is_clique()
True
sage: i.is_clique(directed_clique=True)
False
is_connected()

Indicates whether the (di)graph is connected. Note that in a graph, path connected is equivalent to connected.

EXAMPLES:

sage: G = Graph( { 0 : [1, 2], 1 : [2], 3 : [4, 5], 4 : [5] } )
sage: G.is_connected()
False
sage: G.add_edge(0,3)
sage: G.is_connected()
True
sage: D = DiGraph( { 0 : [1, 2], 1 : [2], 3 : [4, 5], 4 : [5] } )
sage: D.is_connected()
False
sage: D.add_edge(0,3)
sage: D.is_connected()
True
sage: D = DiGraph({1:[0], 2:[0]})
sage: D.is_connected()
True
is_cut_edge(u, v=None, label=None)

Returns True if the input edge is a cut-edge or a bridge.

A cut edge (or bridge) is an edge that when removed increases the number of connected components. This function works with simple graphs as well as graphs with loops and multiedges. In a digraph, a cut edge is an edge that when removed increases the number of (weakly) connected components.

INPUT: The following forms are accepted

  • G.is_cut_edge( 1, 2 )
  • G.is_cut_edge( (1, 2) )
  • G.is_cut_edge( 1, 2, ‘label’ )
  • G.is_cut_edge( (1, 2, ‘label’) )

OUTPUT:

  • Returns True if (u,v) is a cut edge, False otherwise

EXAMPLES:

sage: G = graphs.CompleteGraph(4)
sage: G.is_cut_edge(0,2)
False

sage: G = graphs.CompleteGraph(4)
sage: G.add_edge((0,5,'silly'))
sage: G.is_cut_edge((0,5,'silly'))
True

sage: G = Graph([[0,1],[0,2],[3,4],[4,5],[3,5]])
sage: G.is_cut_edge((0,1))
True

sage: G = Graph([[0,1],[0,2],[1,1]])
sage: G.allow_loops(True)
sage: G.is_cut_edge((1,1))
False

sage: G = digraphs.Circuit(5)
sage: G.is_cut_edge((0,1))
False

sage: G = graphs.CompleteGraph(6)
sage: G.is_cut_edge((0,7))
Traceback (most recent call last):
...
ValueError: edge not in graph
is_cut_vertex(u, weak=False)

Returns True if the input vertex is a cut-vertex.

A vertex is a cut-vertex if its removal from the (di)graph increases the number of (strongly) connected components. Isolated vertices or leafs are not cut-vertices. This function works with simple graphs as well as graphs with loops and multiple edges.

INPUT:

  • u – a vertex
  • weak – (default: False) boolean set to True if the connectivity of directed graphs is to be taken in the weak sense, that is ignoring edges orientations.

OUTPUT:

Returns True if u is a cut-vertex, and False otherwise.

EXAMPLES:

Giving a LollipopGraph(4,2), that is a complete graph with 4 vertices with a pending edge:

sage: G = graphs.LollipopGraph(4,2)
sage: G.is_cut_vertex(0)
False
sage: G.is_cut_vertex(3)
True

Comparing the weak and strong connectivity of a digraph:

sage: D = digraphs.Circuit(6)
sage: D.is_strongly_connected()
True
sage: D.is_cut_vertex(2)
True
sage: D.is_cut_vertex(2, weak=True)
False

Giving a vertex that is not in the graph:

sage: G = graphs.CompleteGraph(6)
sage: G.is_cut_vertex(7)
Traceback (most recent call last):
...
ValueError: The input vertex is not in the vertex set.
is_drawn_free_of_edge_crossings()

Returns True is the position dictionary for this graph is set and that position dictionary gives a planar embedding.

This simply checks all pairs of edges that don’t share a vertex to make sure that they don’t intersect.

Note

This function require that _pos attribute is set. (Returns False otherwise.)

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: D.set_planar_positions()
sage: D.is_drawn_free_of_edge_crossings()
True
is_equitable(partition, quotient_matrix=False)

Checks whether the given partition is equitable with respect to self.

A partition is equitable with respect to a graph if for every pair of cells C1, C2 of the partition, the number of edges from a vertex of C1 to C2 is the same, over all vertices in C1.

INPUT:

  • partition - a list of lists
  • quotient_matrix - (default False) if True, and the partition is equitable, returns a matrix over the integers whose rows and columns represent cells of the partition, and whose i,j entry is the number of vertices in cell j adjacent to each vertex in cell i (since the partition is equitable, this is well defined)

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.is_equitable([[0,4],[1,3,5,9],[2,6,8],[7]])
False
sage: G.is_equitable([[0,4],[1,3,5,9],[2,6,8,7]])
True
sage: G.is_equitable([[0,4],[1,3,5,9],[2,6,8,7]], quotient_matrix=True)
[1 2 0]
[1 0 2]
[0 2 1]
sage: ss = (graphs.WheelGraph(6)).line_graph(labels=False)
sage: prt = [[(0, 1)], [(0, 2), (0, 3), (0, 4), (1, 2), (1, 4)], [(2, 3), (3, 4)]]
sage: ss.is_equitable(prt)
Traceback (most recent call last):
...
TypeError: Partition ([[(0, 1)], [(0, 2), (0, 3), (0, 4), (1, 2), (1, 4)], [(2, 3), (3, 4)]]) is not valid for this graph: vertices are incorrect.
sage: ss = (graphs.WheelGraph(5)).line_graph(labels=False)
sage: ss.is_equitable(prt)
False
is_eulerian(path=False)

Return true if the graph has a (closed) tour that visits each edge exactly once.

INPUT:

  • path – by default this function finds if the graph contains a closed tour visiting each edge once, i.e. an eulerian cycle. If you want to test the existence of an eulerian path, set this argument to True. Graphs with this property are sometimes called semi-eulerian.

OUTPUT:

True or False for the closed tour case. For an open tour search (path``=``True) the function returns False if the graph is not semi-eulerian, or a tuple (u, v) in the other case. This tuple defines the edge that would make the graph eulerian, i.e. close an existing open tour. This edge may or may not be already present in the graph.

EXAMPLES:

sage: graphs.CompleteGraph(4).is_eulerian()
False
sage: graphs.CycleGraph(4).is_eulerian()
True
sage: g = DiGraph({0:[1,2], 1:[2]}); g.is_eulerian()
False
sage: g = DiGraph({0:[2], 1:[3], 2:[0,1], 3:[2]}); g.is_eulerian()
True
sage: g = DiGraph({0:[1], 1:[2], 2:[0], 3:[]}); g.is_eulerian()
True
sage: g = Graph([(1,2), (2,3), (3,1), (4,5), (5,6), (6,4)]); g.is_eulerian()
False
sage: g = DiGraph({0: [1]}); g.is_eulerian(path=True)
(1, 0)
sage: graphs.CycleGraph(4).is_eulerian(path=True)
False
sage: g = DiGraph({0: [1], 1: [2,3], 2: [4]}); g.is_eulerian(path=True)
False
sage: g = Graph({0:[1,2,3], 1:[2,3], 2:[3,4], 3:[4]}, multiedges=True)
sage: g.is_eulerian()
False
sage: e = g.is_eulerian(path=True); e
(0, 1)
sage: g.add_edge(e)
sage: g.is_eulerian(path=False)
True
sage: g.is_eulerian(path=True)
False

TESTS:

sage: g = Graph({0:[], 1:[], 2:[], 3:[]}); g.is_eulerian()
True
is_gallai_tree()

Returns whether the current graph is a Gallai tree.

A graph is a Gallai tree if and only if it is connected and its 2-connected components are all isomorphic to complete graphs or odd cycles.

A connected graph is not degree-choosable if and only if it is a Gallai tree [erdos1978choos].

REFERENCES:

[erdos1978choos]Erdos, P. and Rubin, A.L. and Taylor, H. Proc. West Coast Conf. on Combinatorics Graph Theory and Computing, Congressus Numerantium vol 26, pages 125–157, 1979

EXAMPLES:

A complete graph is, or course, a Gallai Tree:

sage: g = graphs.CompleteGraph(15)
sage: g.is_gallai_tree()
True

The Petersen Graph is not:

sage: g = graphs.PetersenGraph()
sage: g.is_gallai_tree()
False

A Graph built from vertex-disjoint complete graphs linked by one edge to a special vertex -1 is a ‘’star-shaped’’ Gallai tree

sage: g = 8 * graphs.CompleteGraph(6)
sage: g.add_edges([(-1,c[0]) for c in g.connected_components()])
sage: g.is_gallai_tree()
True
is_hamiltonian()

Tests whether the current graph is Hamiltonian.

A graph (resp. digraph) is said to be Hamiltonian if it contains as a subgraph a cycle (resp. a circuit) going through all the vertices.

Testing for Hamiltonicity being NP-Complete, this algorithm could run for some time depending on the instance.

ALGORITHM:

See Graph.traveling_salesman_problem.

OUTPUT:

Returns True if a Hamiltonian cycle/circuit exists, and False otherwise.

NOTE:

This function, as hamiltonian_cycle and traveling_salesman_problem, computes a Hamiltonian cycle if it exists : the user should NOT test for Hamiltonicity using is_hamiltonian before calling hamiltonian_cycle or traveling_salesman_problem as it would result in computing it twice.

EXAMPLES:

The Heawood Graph is known to be Hamiltonian

sage: g = graphs.HeawoodGraph()
sage: g.is_hamiltonian()
True

The Petergraph, though, is not

sage: g = graphs.PetersenGraph()
sage: g.is_hamiltonian()
False

TESTS:

When no solver is installed, a OptionalPackageNotFoundError exception is raised:

sage: from sage.misc.exceptions import OptionalPackageNotFoundError
sage: try:
...       g = graphs.ChvatalGraph()
...       if not g.is_hamiltonian():
...          print "There is something wrong here !"
... except OptionalPackageNotFoundError:
...       pass
is_independent_set(vertices=None)

Returns True if the set vertices is an independent set, False if not. An independent set is a set of vertices such that there is no edge between any two vertices.

INPUT:

  • vertices - Vertices can be a single vertex or an iterable container of vertices, e.g. a list, set, graph, file or numeric array. If not passed, defaults to the entire graph.

EXAMPLES:

sage: graphs.CycleGraph(4).is_independent_set([1,3])
True
sage: graphs.CycleGraph(4).is_independent_set([1,2,3])
False
is_interval(certificate=False)

Check whether self is an interval graph

INPUT:

  • certificate (boolean) – The function returns True or False according to the graph, when certificate = False (default). When certificate = True and the graph is an interval graph, a dictionary whose keys are the vertices and values are pairs of integers are returned instead of True. They correspond to an embedding of the interval graph, each vertex being represented by an interval going from the first of the two values to the second.

ALGORITHM:

Through the use of PQ-Trees

AUTHOR:

Nathann Cohen (implementation)

EXAMPLES:

A Petersen Graph is not chordal, nor car it be an interval graph

sage: g = graphs.PetersenGraph()
sage: g.is_interval()
False

Though we can build intervals from the corresponding random generator:

sage: g = graphs.RandomInterval(20)
sage: g.is_interval()
True

This method can also return, given an interval graph, a possible embedding (we can actually compute all of them through the PQ-Tree structures):

sage: g = Graph(':S__@_@A_@AB_@AC_@ACD_@ACDE_ACDEF_ACDEFG_ACDEGH_ACDEGHI_ACDEGHIJ_ACDEGIJK_ACDEGIJKL_ACDEGIJKLMaCEGIJKNaCEGIJKNaCGIJKNPaCIP')
sage: d = g.is_interval(certificate = True)
sage: print d                                    # not tested
{0: (0, 20), 1: (1, 9), 2: (2, 36), 3: (3, 5), 4: (4, 38), 5: (6, 21), 6: (7, 27), 7: (8, 12), 8: (10, 29), 9: (11, 16), 10: (13, 39), 11: (14, 31), 12: (15, 32), 13: (17, 23), 14: (18, 22), 15: (19, 33), 16: (24, 25), 17: (26, 35), 18: (28, 30), 19: (34, 37)}

From this embedding, we can clearly build an interval graph isomorphic to the previous one:

sage: g2 = graphs.IntervalGraph(d.values())
sage: g2.is_isomorphic(g)
True

See also

is_isomorphic(other, certify=False, verbosity=0, edge_labels=False)

Tests for isomorphism between self and other.

INPUT:

  • certify - if True, then output is (a,b), where a is a boolean and b is either a map or None.
  • edge_labels - default False, otherwise allows only permutations respecting edge labels.

EXAMPLES: Graphs:

sage: from sage.groups.perm_gps.permgroup_named import SymmetricGroup
sage: D = graphs.DodecahedralGraph()
sage: E = copy(D)
sage: gamma = SymmetricGroup(20).random_element()
sage: E.relabel(gamma)
sage: D.is_isomorphic(E)
True
sage: D = graphs.DodecahedralGraph()
sage: S = SymmetricGroup(20)
sage: gamma = S.random_element()
sage: E = copy(D)
sage: E.relabel(gamma)
sage: a,b = D.is_isomorphic(E, certify=True); a
True
sage: from sage.plot.graphics import GraphicsArray
sage: from sage.graphs.generic_graph_pyx import spring_layout_fast
sage: position_D = spring_layout_fast(D)
sage: position_E = {}
sage: for vert in position_D:
...    position_E[b[vert]] = position_D[vert]
sage: GraphicsArray([D.plot(pos=position_D), E.plot(pos=position_E)]).show() # long time
sage: g=graphs.HeawoodGraph()
sage: g.is_isomorphic(g)
True

Multigraphs:

sage: G = Graph(multiedges=True,sparse=True)
sage: G.add_edge((0,1,1))
sage: G.add_edge((0,1,2))
sage: G.add_edge((0,1,3))
sage: G.add_edge((0,1,4))
sage: H = Graph(multiedges=True,sparse=True)
sage: H.add_edge((3,4))
sage: H.add_edge((3,4))
sage: H.add_edge((3,4))
sage: H.add_edge((3,4))
sage: G.is_isomorphic(H)
True

Digraphs:

sage: A = DiGraph( { 0 : [1,2] } )
sage: B = DiGraph( { 1 : [0,2] } )
sage: A.is_isomorphic(B, certify=True)
(True, {0: 1, 1: 0, 2: 2})

Edge labeled graphs:

sage: G = Graph(sparse=True)
sage: G.add_edges( [(0,1,'a'),(1,2,'b'),(2,3,'c'),(3,4,'b'),(4,0,'a')] )
sage: H = G.relabel([1,2,3,4,0], inplace=False)
sage: G.is_isomorphic(H, edge_labels=True)
True

Edge labeled digraphs:

sage: G = DiGraph()
sage: G.add_edges( [(0,1,'a'),(1,2,'b'),(2,3,'c'),(3,4,'b'),(4,0,'a')] )
sage: H = G.relabel([1,2,3,4,0], inplace=False)
sage: G.is_isomorphic(H, edge_labels=True)
True
sage: G.is_isomorphic(H, edge_labels=True, certify=True)
(True, {0: 1, 1: 2, 2: 3, 3: 4, 4: 0})

TESTS:

sage: g1 = '~?A[~~{ACbCwV_~__OOcCW_fAA{CF{CCAAAC__bCCCwOOV___~____OOOOcCCCW___fAAAA'+            ...   '{CCCF{CCCCAAAAAC____bCCCCCwOOOOV_____~_O@ACG_@ACGOo@ACG?{?`A?GV_GO@AC}@?_OGC'+            ...   'C?_OI@?K?I@?_OM?_OGD?F_A@OGC@{A@?_OG?O@?gCA?@_GCA@O?B_@OGCA?BoA@?gC?@{A?GO`?'+            ...   '??_GO@AC??E?O`?CG??[?O`A?G??{?GO`A???|A?_GOC`AC@_OCGACEAGS?HA?_SA`aO@G?cOC_N'+            ...   'G_C@AOP?GnO@_GACOE?g?`OGACCOGaGOc?HA?`GORCG_AO@B?K@[`A?OCI@A@By?_K@?SCABA?H?'+            ...   'SA?a@GC`CH?Q?C_c?cGRC@G_AOCOa@Ax?QC?_GOo_CNg@A?oC@CaCGO@CGA_O`?GSGPAGOC_@OO_'+            ...   'aCHaG?cO@CB?_`Ax?GQC?_cAOCG^OGAC@_D?IGO`?D?O_I?HAOO`AGOHA?cC?oAO`AW_Q?HCACAC'+            ...   'GO`[_OCHA?_cCACG^O_@CAGO`A?GCOGc@?I?OQOC?IGC_o@CAGCCE?A@DBG_OA@C_CP?OG_VA_CO'+            ...   'G@D?_OA_DFgA@CO?aH?Ga@?a?_I?S@A@@Oa@?@P@GCO_AACO_a_?`K_GCQ@?cAOG_OGAwQ@?K?cC'+            ...   'GH?I?ABy@C?G_S@@GCA@C`?OI?_D?OP@G?IGGP@O_AGCP?aG?GCPAX?cA?OGSGCGCAGCJ`?oAGCC'+            ...   'HAA?A_CG^O@CAG_GCSCAGCCGOCG@OA_`?`?g_OACG_`CAGOAO_H?a_?`AXA?OGcAAOP?a@?CGVAC'+            ...   'OG@_AGG`OA_?O`|?Ga?COKAAGCA@O`A?a?S@?HCG`?_?gO`AGGaC?PCAOGI?A@GO`K_CQ@?GO_`O'+            ...   'GCAACGVAG@_COOCQ?g?I?O`ByC?G_P?O`A?H@G?_P?`OAGC?gD?_C@_GCAGDG_OA@CCPC?AOQ??g'+            ...   '_R@_AGCO____OCC_@OAbaOC?g@C_H?AOOC@?a`y?PC?G`@OOH??cOG_OOAG@_COAP?WA?_KAGC@C'+            ...   '_CQ@?HAACH??c@P?_AWGaC?P?gA_C_GAD?I?Awa?S@?K?`C_GAOGCS?@|?COGaA@CAAOQ?AGCAGO'+            ...   'ACOG@_G_aC@_G@CA@@AHA?OGc?WAAH@G?P?_?cH_`CAGOGACc@@GA?S?CGVCG@OA_CICAOOC?PO?'+            ...   'OG^OG_@CAC_cC?AOP?_OICG@?oAGCO_GO_GB@?_OG`AH?cA?OH?`P??cC_O?SCGR@O_AGCAI?Q?_'+            ...   'GGS?D?O`[OI?_D@@CCA?cCA_?_O`By?_PC?IGAGOQ?@A@?aO`A?Q@?K?__`_E?_GCA@CGO`C_GCQ'+            ...   '@A?gAOQ?@C?DCACGR@GCO_AGPA@@GAA?A_CO`Aw_I?S@?SCB@?OC_?_P@ACNgOC@A?aCGOCAGCA@'+            ...   'CA?H@GG_C@AOGa?OOG_O?g_OA?oDC_AO@GOCc?@P?_A@D??cC``O?cGAOGD?@OA_CAGCA?_cwKA?'+            ...   '`?OWGG?_PO?I?S?H@?^OGAC@_Aa@CAGC?a@?_Q?@H?_OCHA?OQA_P?_G_O?WA?_IG_Q?HC@A@ADC'+            ...   'A?AI?AC_?QAWOHA?cAGG_I?S?G_OG@GA?`[D?O_IA?`GGCS?OA_?c@?Q?^OAC@_G_Ca@CA@?OGCO'+            ...   'H@G@A@?GQC?_Q@GP?_OG?IGGB?OCGaG?cO@A__QGC?E?A@CH@G?GRAGOC_@GGOW@O?O_OGa?_c?G'+            ...   'V@CGA_OOaC?a_?a?A_CcC@?CNgA?oC@GGE@?_OH?a@?_?QA`A@?QC?_KGGO_OGCAa@?A?_KCGPC@'+            ...   'G_AOAGPGC?D@?a_A?@GGO`KH?Q?C_QGAA_?gOG_OA?_GG`AwH?SA?`?cAI?A@D?I?@?QA?`By?K@'+            ...   '?O`GGACA@CGCA@CC_?WO`?`A?OCH?`OCA@COG?I?oC@ACGPCG_AO@_aAA?Aa?g?GD@G?CO`AWOc?'+            ...   'HA?OcG_?g@OGCAAAOC@ACJ_`OGACAGCS?CAGI?A`@?OCACG^'
sage: g2 = '~?A[??osR?WARSETCJ_QWASehOXQg`QwChK?qSeFQ_sTIaWIV?XIR?KAC?B?`?COCG?o?O_'+            ...   '@_?`??B?`?o@_O_WCOCHC@_?`W?E?AD_O?WCCeO?WCSEGAGAIaA@_?aw?OK?ER?`?@_HQXA?B@Q_'+            ...   'pA?a@Qg_`?o?h[?GOK@IR?@A?BEQcoCG?K\IB?GOCWiTC?GOKWIV??CGEKdH_H_?CB?`?DC??_WC'+            ...   'G?SO?AP?O_?g_?D_?`?C__?D_?`?CCo??@_O_XDC???WCGEGg_??a?`G_aa??E?AD_@cC??K?CJ?'+            ...   '@@K?O?WCCe?aa?G?KAIB?Gg_A?a?ag_@DC?OK?CV??EOO@?o?XK??GH`A?B?Qco?Gg`A?B@Q_o?C'+            ...   'SO`?P?hSO?@DCGOK?IV???K_`A@_HQWC??_cCG?KXIRG?@D?GO?WySEG?@D?GOCWiTCC??a_CGEK'+            ...   'DJ_@??K_@A@bHQWAW?@@K??_WCG?g_?CSO?A@_O_@P??Gg_?Ca?`?@P??Gg_?D_?`?C__?EOO?Ao'+            ...   '?O_AAW?@@K???WCGEPP??Gg_??B?`?pDC??aa??AGACaAIG?@DC??K?CJ?BGG?@cC??K?CJ?@@K?'+            ...   '?_e?G?KAAR?PP??Gg_A?B?a_oAIG?@DC?OCOCTC?Gg_?CSO@?o?P[??X@??K__A@_?qW??OR??GH'+            ...   '`A?B?Qco?Gg_?CSO`?@_hOW?AIG?@DCGOCOITC??PP??Gg`A@_@Qw??@cC??qACGE?dH_O?AAW?@'+            ...   '@GGO?WqSeO?AIG?@D?GO?WySEG?@DC??a_CGAKTIaA??PP??Gg@A@b@Qw?O?BGG?@c?GOKXIR?KA'+            ...   'C?H_?CCo?A@_O_?WCG@P??Gg_?CB?`?COCG@P??Gg_?Ca?`?E?AC?g_?CSO?Ao?O_@_?`@GG?@cC'+            ...   '??k?CG??WCGOR??GH_??B?`?o@_O`DC??aa???KACB?a?`AIG?@DC??COCHC@_?`AIG?@DC??K?C'+            ...   'J??o?O`cC??qA??E?AD_O?WC?OR??GH_A?B?_cq?B?_AIG?@DC?O?WCSEGAGA?Gg_?CSO@?P?PSO'+            ...   'OK?C?PP??Gg_A@_?aw?OK?C?X@??K__A@_?qWCG?K??GH_?CCo`?@_HQXA?B??AIG?@DCGO?WISE'+            ...   'GOCO??PP??Gg`A?a@Qg_`?o??@DC??aaCGE?DJ_@A@_??BGG?@cCGOK@IR?@A?BO?AAW?@@GGO?W'+            ...   'qSe?`?@g?@DC??a_CG?K\IB?GOCQ??PP??Gg@A?bDQg_@A@_O?AIG?@D?GOKWIV??CGE@??K__?E'+            ...   'O?`?pchK?_SA_OI@OGD?gCA_SA@OI?c@H?Q?c_H?QOC_HGAOCc?QOC_HGAOCc@GAQ?c@H?QD?gCA'+            ...   '_SA@OI@?gD?_SA_OKA_SA@OI@?gD?_SA_OI@OHI?c_H?QOC_HGAOCc@GAQ?eC_H?QOC_HGAOCc@G'+            ...   'AQ?c@XD?_SA_OI@OGD?gCA_SA@PKGO`A@ACGSGO`?`ACICGO_?ACGOcGO`?O`AC`ACHACGO???^?'+            ...   '????}Bw????Fo^???????Fo?}?????Bw?^?Bw?????GO`AO`AC`ACGACGOcGO`??aCGO_O`ADACG'+            ...   'OGO`A@ACGOA???@{?N_@{?????Fo?}????OFo????N_}????@{????Bw?OACGOgO`A@ACGSGO`?`'+            ...   'ACG?OaCGO_GO`AO`AC`ACGACGO_@G???Fo^?????}Bw????Fo??AC@{?????Fo?}?Fo?????^??A'+            ...   'OGO`AO`AC@ACGQCGO_GO`A?HAACGOgO`A@ACGOGO`A`ACG?GQ??^?Bw?????N_@{?????Fo?QC??'+            ...   'Fo^?????}????@{Fo???CHACGO_O`ACACGOgO`A@ACGO@AOcGO`?O`AC`ACGACGOcGO`?@GQFo??'+            ...   '??N_????^@{????Bw??`GRw?????N_@{?????Fo?}???HAO_OI@OGD?gCA_SA@OI@?gDK_??C@GA'+            ...   'Q?c@H?Q?c_H?QOC_HEW????????????????????????~~~~~'
sage: G1 = Graph(g1)
sage: G2 = Graph(g2)
sage: G1.is_isomorphic(G2)
True

Ensure that isomorphic looped graphs with non-range vertex labels report correctly (trac ticket #10814, fixed by trac ticket #8395):

sage: G1 = Graph([(0,1), (1,1)])
sage: G2 = Graph([(0,2), (2,2)])
sage: G1.is_isomorphic(G2)
True
sage: G = Graph(multiedges = True, loops = True)
sage: H = Graph(multiedges = True, loops = True)
sage: G.add_edges([(0,1,0),(1,0,1),(1,1,2),(0,0,3)])
sage: H.add_edges([(0,1,3),(1,0,2),(1,1,1),(0,0,0)])
sage: G.is_isomorphic(H, certify=True)
(True, {0: 0, 1: 1})
sage: set_random_seed(0)
sage: D = digraphs.RandomDirectedGNP(6, .2)
sage: D.is_isomorphic(D, certify = True)
(True, {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5})
sage: D.is_isomorphic(D,edge_labels=True, certify = True)
(True, {0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5})

Ensure that trac trac ticket #11620 is fixed:

sage: G1 = DiGraph([(0, 0, 'c'), (0, 4, 'b'), (0, 5, 'c'),
...   (0, 5, 't'), (1, 1, 'c'), (1, 3,'c'), (1, 3, 't'), (1, 5, 'b'),
...   (2, 2, 'c'), (2, 3, 'b'), (2, 4, 'c'),(2, 4, 't'), (3, 1, 't'),
...   (3, 2, 'b'), (3, 2, 'c'), (3, 4, 'c'), (4, 0,'b'), (4, 0, 'c'),
...   (4, 2, 't'), (4, 5, 'c'), (5, 0, 't'), (5, 1, 'b'), (5, 1, 'c'),
...   (5, 3, 'c')], loops=True, multiedges=True)
sage: G2 = G1.relabel({0:4, 1:5, 2:3, 3:2, 4:1,5:0}, inplace=False)
sage: G1.canonical_label(edge_labels=True) == G2.canonical_label(edge_labels=True)
True
sage: G1.is_isomorphic(G2,edge_labels=True)
True

Ensure that trac ticket #13114 is fixed

sage: g = Graph([(0, 0, 0), (0, 2, 0), (1, 1, 0), (1, 2, 0), (1, 2, 1), (2, 2, 0)])
sage: gg = Graph([(0, 0, 0), (0, 1, 0), (1, 1, 0), (1, 2, 0), (2, 2, 0), (2, 2, 1)])
sage: g.is_isomorphic(gg)
False

Ensure that trac:14777 is fixed

sage: g = Graph()
sage: h = Graph()
sage: g.is_isomorphic(h)
True
is_planar(on_embedding=None, kuratowski=False, set_embedding=False, set_pos=False)

Returns True if the graph is planar, and False otherwise. This wraps the reference implementation provided by John Boyer of the linear time planarity algorithm by edge addition due to Boyer Myrvold. (See reference code in graphs.planarity).

Note - the argument on_embedding takes precedence over set_embedding. This means that only the on_embedding combinatorial embedding will be tested for planarity and no _embedding attribute will be set as a result of this function call, unless on_embedding is None.

REFERENCE:

  • [1] John M. Boyer and Wendy J. Myrvold, On the Cutting Edge: Simplified O(n) Planarity by Edge Addition. Journal of Graph Algorithms and Applications, Vol. 8, No. 3, pp. 241-273, 2004.

INPUT:

  • kuratowski - returns a tuple with boolean as first entry. If the graph is nonplanar, will return the Kuratowski subgraph or minor as the second tuple entry. If the graph is planar, returns None as the second entry.
  • on_embedding - the embedding dictionary to test planarity on. (i.e.: will return True or False only for the given embedding.)
  • set_embedding - whether or not to set the instance field variable that contains a combinatorial embedding (clockwise ordering of neighbors at each vertex). This value will only be set if a planar embedding is found. It is stored as a Python dict: v1: [n1,n2,n3] where v1 is a vertex and n1,n2,n3 are its neighbors.
  • set_pos - whether or not to set the position dictionary (for plotting) to reflect the combinatorial embedding. Note that this value will default to False if set_emb is set to False. Also, the position dictionary will only be updated if a planar embedding is found.

EXAMPLES:

sage: g = graphs.CubeGraph(4)
sage: g.is_planar()
False
sage: g = graphs.CircularLadderGraph(4)
sage: g.is_planar(set_embedding=True)
True
sage: g.get_embedding()
{0: [1, 4, 3],
 1: [2, 5, 0],
 2: [3, 6, 1],
 3: [0, 7, 2],
 4: [0, 5, 7],
 5: [1, 6, 4],
 6: [2, 7, 5],
 7: [4, 6, 3]}
sage: g = graphs.PetersenGraph()
sage: (g.is_planar(kuratowski=True))[1].adjacency_matrix()
[0 1 0 0 0 1 0 0 0]
[1 0 1 0 0 0 1 0 0]
[0 1 0 1 0 0 0 1 0]
[0 0 1 0 0 0 0 0 1]
[0 0 0 0 0 0 1 1 0]
[1 0 0 0 0 0 0 1 1]
[0 1 0 0 1 0 0 0 1]
[0 0 1 0 1 1 0 0 0]
[0 0 0 1 0 1 1 0 0]
sage: k43 = graphs.CompleteBipartiteGraph(4,3)
sage: result = k43.is_planar(kuratowski=True); result
(False, Graph on 6 vertices)
sage: result[1].is_isomorphic(graphs.CompleteBipartiteGraph(3,3))
True

Multi-edged and looped graphs are partially supported:

sage: G = Graph({0:[1,1]}, multiedges=True)
sage: G.is_planar()
True
sage: G.is_planar(on_embedding={})
Traceback (most recent call last):
...
NotImplementedError: Cannot compute with embeddings of multiple-edged or looped graphs.
sage: G.is_planar(set_pos=True)
Traceback (most recent call last):
...
NotImplementedError: Cannot compute with embeddings of multiple-edged or looped graphs.
sage: G.is_planar(set_embedding=True)
Traceback (most recent call last):
...
NotImplementedError: Cannot compute with embeddings of multiple-edged or looped graphs.
sage: G.is_planar(kuratowski=True)
(True, None)
sage: G = graphs.CompleteGraph(5)
sage: G = Graph(G, multiedges=True)
sage: G.add_edge(0,1)
sage: G.is_planar()
False
sage: b,k = G.is_planar(kuratowski=True)
sage: b
False
sage: k.vertices()
[0, 1, 2, 3, 4]
is_regular(k=None)

Return True if this graph is (k-)regular.

INPUT:

  • k (default: None) - the degree of regularity to check for

EXAMPLES:

sage: G = graphs.HoffmanSingletonGraph()
sage: G.is_regular()
True
sage: G.is_regular(9)
False

So the Hoffman-Singleton graph is regular, but not 9-regular. In fact, we can now find the degree easily as follows:

sage: G.degree_iterator().next()
7

The house graph is not regular:

sage: graphs.HouseGraph().is_regular()
False

A graph without vertices is k-regular for every k:

sage: Graph().is_regular()
True
is_subgraph(other, induced=True)

Tests whether self is a subgraph of other.

Warning

Please note that this method does not check whether self contains a subgraph isomorphic to other, but only if it directly contains it as a subgraph !

By default induced is True for backwards compatibility.

INPUT:

  • induced - boolean (default: True) If set to True tests whether self is an induced subgraph of other that is if the vertices of self are also vertices of other, and the edges of self are equal to the edges of other between the vertices contained in self`. If set to ``False tests whether self is a subgraph of other that is if all vertices of self are also in other and all edges of self are also in other.

OUTPUT:

boolean – True iff self is a (possibly induced) subgraph of other.

See also

If you are interested in the (possibly induced) subgraphs isomorphic to self in other, you are looking for the following methods:

EXAMPLES:

sage: P = graphs.PetersenGraph() sage: G = P.subgraph(range(6)) sage: G.is_subgraph(P) True

sage: H=graphs.CycleGraph(5) sage: G=graphs.PathGraph(5) sage: G.is_subgraph(H) False sage: G.is_subgraph(H, induced=False) True sage: H.is_subgraph(G, induced=False) False

TESTS:

Raise an error when self and other are of different types:

sage: Graph([(0,1)]).is_subgraph( DiGraph([(0,1)]) )
Traceback (most recent call last):
...
ValueError: The input parameter must be a Graph.
sage: DiGraph([(0,1)]).is_subgraph( Graph([(0,1)]) )
Traceback (most recent call last):
...
ValueError: The input parameter must be a DiGraph.
is_transitively_reduced()

Tests whether the digraph is transitively reduced.

A digraph is transitively reduced if it is equal to its transitive reduction.

EXAMPLES:

sage: d = DiGraph({0:[1],1:[2],2:[3]})
sage: d.is_transitively_reduced()
True

sage: d = DiGraph({0:[1,2],1:[2]})
sage: d.is_transitively_reduced()
False

sage: d = DiGraph({0:[1,2],1:[2],2:[]})
sage: d.is_transitively_reduced()
False
is_vertex_transitive(partition=None, verbosity=0, edge_labels=False, order=False, return_group=True, orbits=False)

Returns whether the automorphism group of self is transitive within the partition provided, by default the unit partition of the vertices of self (thus by default tests for vertex transitivity in the usual sense).

EXAMPLES:

sage: G = Graph({0:[1],1:[2]})
sage: G.is_vertex_transitive()
False
sage: P = graphs.PetersenGraph()
sage: P.is_vertex_transitive()
True
sage: D = graphs.DodecahedralGraph()
sage: D.is_vertex_transitive()
True
sage: R = graphs.RandomGNP(2000, .01)
sage: R.is_vertex_transitive()
False
kirchhoff_matrix(weighted=None, indegree=True, normalized=False, **kwds)

Returns the Kirchhoff matrix (a.k.a. the Laplacian) of the graph.

The Kirchhoff matrix is defined to be D - M, where D is the diagonal degree matrix (each diagonal entry is the degree of the corresponding vertex), and M is the adjacency matrix. If normalized is True, then the returned matrix is D^{-1/2}(D-M)D^{-1/2}.

( In the special case of DiGraphs, D is defined as the diagonal in-degree matrix or diagonal out-degree matrix according to the value of indegree)

INPUT:

  • weighted – Binary variable :
    • If True, the weighted adjacency matrix is used for M, and the diagonal matrix D takes into account the weight of edges (replace in the definition “degree” by “sum of the incident edges” ).
    • Else, each edge is assumed to have weight 1.

    Default is to take weights into consideration if and only if the graph is weighted.

  • indegree – Binary variable :
    • If True, each diagonal entry of D is equal to the in-degree of the corresponding vertex.

    • Else, each diagonal entry of D is equal to the out-degree of the corresponding vertex.

      By default, indegree is set to True

    ( This variable only matters when the graph is a digraph )

  • normalized – Binary variable :

    • If True, the returned matrix is D^{-1/2}(D-M)D^{-1/2}, a normalized version of the Laplacian matrix. (More accurately, the normalizing matrix used is equal to D^{-1/2} only for non-isolated vertices. If vertex i is isolated, then diagonal entry i in the matrix is 1, rather than a division by zero.)
    • Else, the matrix D-M is returned

Note that any additional keywords will be passed on to either the adjacency_matrix or weighted_adjacency_matrix method.

AUTHORS:

  • Tom Boothby
  • Jason Grout

EXAMPLES:

sage: G = Graph(sparse=True)
sage: G.add_edges([(0,1,1),(1,2,2),(0,2,3),(0,3,4)])
sage: M = G.kirchhoff_matrix(weighted=True); M
[ 8 -1 -3 -4]
[-1  3 -2  0]
[-3 -2  5  0]
[-4  0  0  4]
sage: M = G.kirchhoff_matrix(); M
[ 3 -1 -1 -1]
[-1  2 -1  0]
[-1 -1  2  0]
[-1  0  0  1]
sage: G.set_boundary([2,3])
sage: M = G.kirchhoff_matrix(weighted=True, boundary_first=True); M
[ 5  0 -3 -2]
[ 0  4 -4  0]
[-3 -4  8 -1]
[-2  0 -1  3]
sage: M = G.kirchhoff_matrix(boundary_first=True); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(boundary_first=True); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(boundary_first=True, sparse=False); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(normalized=True); M
[                   1 -1/6*sqrt(3)*sqrt(2) -1/6*sqrt(3)*sqrt(2)         -1/3*sqrt(3)]
[-1/6*sqrt(3)*sqrt(2)                    1                 -1/2                    0]
[-1/6*sqrt(3)*sqrt(2)                 -1/2                    1                    0]
[        -1/3*sqrt(3)                    0                    0                    1]

sage: Graph({0:[],1:[2]}).laplacian_matrix(normalized=True)
[ 0  0  0]
[ 0  1 -1]
[ 0 -1  1]

A weighted directed graph with loops, changing the variable indegree

sage: G = DiGraph({1:{1:2,2:3}, 2:{1:4}}, weighted=True,sparse=True)
sage: G.laplacian_matrix()
[ 4 -3]
[-4  3]
sage: G = DiGraph({1:{1:2,2:3}, 2:{1:4}}, weighted=True,sparse=True)
sage: G.laplacian_matrix(indegree=False)
[ 3 -3]
[-4  4]
kronecker_product(other)

Returns the tensor product of self and other.

The tensor product of G and H is the graph L with vertex set V(L) equal to the Cartesian product of the vertices V(G) and V(H), and ((u,v), (w,x)) is an edge iff - (u, w) is an edge of self, and - (v, x) is an edge of other.

The tensor product is also known as the categorical product and the kronecker product (refering to the kronecker matrix product). See Wikipedia article on the Kronecker product.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: C = graphs.CycleGraph(5)
sage: T = C.tensor_product(Z); T
Graph on 10 vertices
sage: T.size()
10
sage: T.plot() # long time
sage: D = graphs.DodecahedralGraph()
sage: P = graphs.PetersenGraph()
sage: T = D.tensor_product(P); T
Graph on 200 vertices
sage: T.size()
900
sage: T.plot() # long time

TESTS:

Tensor product of graphs:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.tensor_product(H)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((0, 'b'), (1, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a'))]
sage: T.is_isomorphic( H.tensor_product(G) )
True

Tensor product of digraphs:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.tensor_product(J)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((1, 'a'), (2, 'b'))]
sage: T.is_isomorphic( J.tensor_product(I) )
True

The tensor product of two DeBruijn digraphs of same diameter is a DeBruijn digraph:

sage: B1 = digraphs.DeBruijn(2, 3)
sage: B2 = digraphs.DeBruijn(3, 3)
sage: T = B1.tensor_product( B2 )
sage: T.is_isomorphic( digraphs.DeBruijn( 2*3, 3) )
True
laplacian_matrix(weighted=None, indegree=True, normalized=False, **kwds)

Returns the Kirchhoff matrix (a.k.a. the Laplacian) of the graph.

The Kirchhoff matrix is defined to be D - M, where D is the diagonal degree matrix (each diagonal entry is the degree of the corresponding vertex), and M is the adjacency matrix. If normalized is True, then the returned matrix is D^{-1/2}(D-M)D^{-1/2}.

( In the special case of DiGraphs, D is defined as the diagonal in-degree matrix or diagonal out-degree matrix according to the value of indegree)

INPUT:

  • weighted – Binary variable :
    • If True, the weighted adjacency matrix is used for M, and the diagonal matrix D takes into account the weight of edges (replace in the definition “degree” by “sum of the incident edges” ).
    • Else, each edge is assumed to have weight 1.

    Default is to take weights into consideration if and only if the graph is weighted.

  • indegree – Binary variable :
    • If True, each diagonal entry of D is equal to the in-degree of the corresponding vertex.

    • Else, each diagonal entry of D is equal to the out-degree of the corresponding vertex.

      By default, indegree is set to True

    ( This variable only matters when the graph is a digraph )

  • normalized – Binary variable :

    • If True, the returned matrix is D^{-1/2}(D-M)D^{-1/2}, a normalized version of the Laplacian matrix. (More accurately, the normalizing matrix used is equal to D^{-1/2} only for non-isolated vertices. If vertex i is isolated, then diagonal entry i in the matrix is 1, rather than a division by zero.)
    • Else, the matrix D-M is returned

Note that any additional keywords will be passed on to either the adjacency_matrix or weighted_adjacency_matrix method.

AUTHORS:

  • Tom Boothby
  • Jason Grout

EXAMPLES:

sage: G = Graph(sparse=True)
sage: G.add_edges([(0,1,1),(1,2,2),(0,2,3),(0,3,4)])
sage: M = G.kirchhoff_matrix(weighted=True); M
[ 8 -1 -3 -4]
[-1  3 -2  0]
[-3 -2  5  0]
[-4  0  0  4]
sage: M = G.kirchhoff_matrix(); M
[ 3 -1 -1 -1]
[-1  2 -1  0]
[-1 -1  2  0]
[-1  0  0  1]
sage: G.set_boundary([2,3])
sage: M = G.kirchhoff_matrix(weighted=True, boundary_first=True); M
[ 5  0 -3 -2]
[ 0  4 -4  0]
[-3 -4  8 -1]
[-2  0 -1  3]
sage: M = G.kirchhoff_matrix(boundary_first=True); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(boundary_first=True); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(boundary_first=True, sparse=False); M
[ 2  0 -1 -1]
[ 0  1 -1  0]
[-1 -1  3 -1]
[-1  0 -1  2]
sage: M = G.laplacian_matrix(normalized=True); M
[                   1 -1/6*sqrt(3)*sqrt(2) -1/6*sqrt(3)*sqrt(2)         -1/3*sqrt(3)]
[-1/6*sqrt(3)*sqrt(2)                    1                 -1/2                    0]
[-1/6*sqrt(3)*sqrt(2)                 -1/2                    1                    0]
[        -1/3*sqrt(3)                    0                    0                    1]

sage: Graph({0:[],1:[2]}).laplacian_matrix(normalized=True)
[ 0  0  0]
[ 0  1 -1]
[ 0 -1  1]

A weighted directed graph with loops, changing the variable indegree

sage: G = DiGraph({1:{1:2,2:3}, 2:{1:4}}, weighted=True,sparse=True)
sage: G.laplacian_matrix()
[ 4 -3]
[-4  3]
sage: G = DiGraph({1:{1:2,2:3}, 2:{1:4}}, weighted=True,sparse=True)
sage: G.laplacian_matrix(indegree=False)
[ 3 -3]
[-4  4]
latex_options()

Returns an instance of GraphLatex for the graph.

Changes to this object will affect the LaTeX version of the graph. For a full explanation of how to use LaTeX to render graphs, see the introduction to the graph_latex module.

EXAMPLES:

sage: g = graphs.PetersenGraph()
sage: opts = g.latex_options()
sage: opts
LaTeX options for Petersen graph: {}
sage: opts.set_option('tkz_style', 'Classic')
sage: opts
LaTeX options for Petersen graph: {'tkz_style': 'Classic'}
layout(layout=None, pos=None, dim=2, save_pos=False, **options)

Returns a layout for the vertices of this graph.

INPUT:

  • layout – one of “acyclic”, “circular”, “ranked”, “graphviz”, “planar”, “spring”, or “tree”
  • pos – a dictionary of positions or None (the default)
  • save_pos – a boolean
  • layout options – (see below)

If layout=algorithm is specified, this algorithm is used to compute the positions.

Otherwise, if pos is specified, use the given positions.

Otherwise, try to fetch previously computed and saved positions.

Otherwise use the default layout (usually the spring layout)

If save_pos = True, the layout is saved for later use.

EXAMPLES:

sage: g = digraphs.ButterflyGraph(1)
sage: g.layout()
{('1', 1): [2.50..., -0.545...],
 ('0', 0): [2.22..., 0.832...],
 ('1', 0): [1.12..., -0.830...],
 ('0', 1): [0.833..., 0.543...]}

sage: 1+1
2
sage: x = g.layout(layout = "acyclic_dummy", save_pos = True)
sage: x =  {('1', 1): [41, 18], ('0', 0): [41, 90], ('1', 0): [140, 90], ('0', 1): [141, 18]}

{('1', 1): [41, 18], ('0', 0): [41, 90], ('1', 0): [140, 90], ('0', 1): [141, 18]}


sage: g.layout(dim = 3)
{('1', 1): [1.07..., -0.260..., 0.927...],
 ('0', 0): [2.02..., 0.528..., 0.343...],
 ('1', 0): [0.674..., -0.528..., -0.343...],
 ('0', 1): [1.61..., 0.260..., -0.927...]}

Here is the list of all the available layout options:

sage: from sage.graphs.graph_plot import layout_options
sage: for key, value in list(sorted(layout_options.iteritems())):
...      print "option", key, ":", value
option by_component : Whether to do the spring layout by connected component -- a boolean.
option dim : The dimension of the layout -- 2 or 3.
option heights : A dictionary mapping heights to the list of vertices at this height.
option iterations : The number of times to execute the spring layout algorithm.
option layout : A layout algorithm -- one of : "acyclic", "circular" (plots the graph with vertices evenly distributed on a circle), "ranked", "graphviz", "planar", "spring" (traditional spring layout, using the graph's current positions as initial positions), or "tree" (the tree will be plotted in levels, depending on minimum distance for the root).
option prog : Which graphviz layout program to use -- one of "circo", "dot", "fdp", "neato", or "twopi".
option save_pos : Whether or not to save the computed position for the graph.
option spring : Use spring layout to finalize the current layout.
option tree_orientation : The direction of tree branches -- 'up', 'down', 'left' or 'right'.
option tree_root : A vertex designation for drawing trees. A vertex of the tree to be used as the root for the ``layout='tree'`` option. If no root is specified, then one is chosen close to the center of the tree. Ignored unless ``layout='tree'``

Some of them only apply to certain layout algorithms. For details, see layout_acyclic(), layout_planar(), layout_circular(), layout_spring(), ...

..warning: unknown optional arguments are silently ignored

..warning: graphviz and dot2tex are currently required to obtain a nice ‘acyclic’ layout. See layout_graphviz() for installation instructions.

A subclass may implement another layout algorithm blah, by implementing a method .layout_blah. It may override the default layout by overriding layout_default(), and similarly override the predefined layouts.

TODO: use this feature for all the predefined graphs classes (like for the Petersen graph, ...), rather than systematically building the layout at construction time.

layout_circular(dim=2, **options)

Computes a circular layout for this graph

OUTPUT: a dictionary mapping vertices to positions

EXAMPLES:

sage: G = graphs.CirculantGraph(7,[1,3])
sage: G.layout_circular()
{0: [6.12...e-17, 1.0],
 1: [-0.78...,  0.62...],
 2: [-0.97..., -0.22...],
 3: [-0.43..., -0.90...],
 4: [0.43...,  -0.90...],
 5: [0.97...,  -0.22...],
 6: [0.78...,   0.62...]}
sage: G.plot(layout = "circular")
layout_default(by_component=True, **options)

Computes a spring layout for this graph

INPUT:

  • iterations – a positive integer
  • dim – 2 or 3 (default: 2)

OUTPUT: a dictionary mapping vertices to positions

Returns a layout computed by randomly arranging the vertices along the given heights

EXAMPLES:

sage: g = graphs.LadderGraph(3) #TODO!!!!
sage: g.layout_spring()
{0: [1.28..., -0.943...],
 1: [1.57..., -0.101...],
 2: [1.83..., 0.747...],
 3: [0.531..., -0.757...],
 4: [0.795..., 0.108...],
 5: [1.08..., 0.946...]}
sage: g = graphs.LadderGraph(7)
sage: g.plot(layout = "spring")
layout_extend_randomly(pos, dim=2)

Extends randomly a partial layout

INPUT:

  • pos: a dictionary mapping vertices to positions

OUTPUT: a dictionary mapping vertices to positions

The vertices not referenced in pos are assigned random positions within the box delimited by the other vertices.

EXAMPLES:

sage: H = digraphs.ButterflyGraph(1)
sage: H.layout_extend_randomly({('0',0): (0,0), ('1',1): (1,1)})
{('1', 1): (1, 1),
 ('0', 0): (0, 0),
 ('1', 0): [0.111..., 0.514...],
 ('0', 1): [0.0446..., 0.332...]}
layout_graphviz(dim=2, prog='dot', **options)

Calls graphviz to compute a layout of the vertices of this graph.

INPUT:

  • prog – one of “dot”, “neato”, “twopi”, “circo”, or “fdp”

EXAMPLES:

sage: g = digraphs.ButterflyGraph(2)
sage: g.layout_graphviz() # optional - dot2tex, graphviz
{('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...],
 ('...', ...): [...,...]}
sage: g.plot(layout = "graphviz") # optional - dot2tex, graphviz

Note: the actual coordinates are not deterministic

By default, an acyclic layout is computed using graphviz‘s dot layout program. One may specify an alternative layout program:

sage: g.plot(layout = "graphviz", prog = "dot")   # optional - dot2tex, graphviz
sage: g.plot(layout = "graphviz", prog = "neato") # optional - dot2tex, graphviz
sage: g.plot(layout = "graphviz", prog = "twopi") # optional - dot2tex, graphviz
sage: g.plot(layout = "graphviz", prog = "fdp")   # optional - dot2tex, graphviz
sage: g = graphs.BalancedTree(5,2)
sage: g.plot(layout = "graphviz", prog = "circo") # optional - dot2tex, graphviz

TODO: put here some cool examples showcasing graphviz features.

This requires graphviz and the dot2tex spkg. Here are some installation tips:

TODO: use the graphviz functionality of Networkx 1.0 once it will be merged into Sage.

layout_planar(set_embedding=False, on_embedding=None, external_face=None, test=False, circular=False, **options)

Uses Schnyder’s algorithm to compute a planar layout for self, raising an error if self is not planar.

INPUT:

  • set_embedding - if True, sets the combinatorial embedding used (see self.get_embedding())
  • on_embedding - dict: provide a combinatorial embedding
  • external_face - ignored
  • test - if True, perform sanity tests along the way
  • circular - ignored

EXAMPLES:

sage: g = graphs.PathGraph(10)
sage: g.set_planar_positions(test=True)
True
sage: g = graphs.BalancedTree(3,4)
sage: g.set_planar_positions(test=True)
True
sage: g = graphs.CycleGraph(7)
sage: g.set_planar_positions(test=True)
True
sage: g = graphs.CompleteGraph(5)
sage: g.set_planar_positions(test=True,set_embedding=True)
Traceback (most recent call last):
...
ValueError: Complete graph is not a planar graph
layout_ranked(heights=None, dim=2, spring=False, **options)

Computes a ranked layout for this graph

INPUT:

  • heights – a dictionary mapping heights to the list of vertices at this height

OUTPUT: a dictionary mapping vertices to positions

Returns a layout computed by randomly arranging the vertices along the given heights

EXAMPLES:

sage: g = graphs.LadderGraph(3)
sage: g.layout_ranked(heights = dict( (i,[i, i+3]) for i in range(3) ))
{0: [0.668..., 0],
 1: [0.667..., 1],
 2: [0.677..., 2],
 3: [1.34..., 0],
 4: [1.33..., 1],
 5: [1.33..., 2]}
sage: g = graphs.LadderGraph(7)
sage: g.plot(layout = "ranked", heights = dict( (i,[i, i+7]) for i in range(7) ))
layout_spring(by_component=True, **options)

Computes a spring layout for this graph

INPUT:

  • iterations – a positive integer
  • dim – 2 or 3 (default: 2)

OUTPUT: a dictionary mapping vertices to positions

Returns a layout computed by randomly arranging the vertices along the given heights

EXAMPLES:

sage: g = graphs.LadderGraph(3) #TODO!!!!
sage: g.layout_spring()
{0: [1.28..., -0.943...],
 1: [1.57..., -0.101...],
 2: [1.83..., 0.747...],
 3: [0.531..., -0.757...],
 4: [0.795..., 0.108...],
 5: [1.08..., 0.946...]}
sage: g = graphs.LadderGraph(7)
sage: g.plot(layout = "spring")
layout_tree(tree_orientation='down', tree_root=None, dim=2, **options)

Computes an ordered tree layout for this graph, which should be a tree (no non-oriented cycles).

INPUT:

  • tree_root – the root vertex. By default None. In this case, a vertex is chosen close to the center of the tree.
  • tree_orientation – the direction in which the tree is growing, can be ‘up’, ‘down’, ‘left’ or ‘right’ (default is ‘down’)

OUTPUT: a dictionary mapping vertices to positions

EXAMPLES:

sage: T = graphs.RandomLobster(25, 0.3, 0.3)
sage: T.show(layout='tree', tree_orientation='up')

sage: G = graphs.HoffmanSingletonGraph()
sage: T = Graph()
sage: T.add_edges(G.min_spanning_tree(starting_vertex=0))
sage: T.show(layout='tree', tree_root=0)

sage: G = graphs.BalancedTree(2, 2)
sage: G.layout_tree(tree_root = 0)
{0: (1.5, 0),
 1: (2.5, -1),
 2: (0.5, -1),
 3: (3.0, -2),
 4: (2.0, -2),
 5: (1.0, -2),
 6: (0.0, -2)}

sage: G = graphs.BalancedTree(2,4)
sage: G.plot(layout="tree", tree_root = 0, tree_orientation = "up")

sage: G = graphs.RandomTree(80)
sage: G.plot(layout="tree", tree_orientation = "right")

TESTS:

sage: G = graphs.CycleGraph(3)
sage: G.plot(layout='tree')
Traceback (most recent call last):
...
RuntimeError: Cannot use tree layout on this graph: self.is_tree() returns False.
lex_BFS(reverse=False, tree=False, initial_vertex=None)

Performs a Lex BFS on the graph.

A Lex BFS ( or Lexicographic Breadth-First Search ) is a Breadth First Search used for the recognition of Chordal Graphs. For more information, see the Wikipedia article on Lex-BFS.

INPUT:

  • reverse (boolean) – whether to return the vertices in discovery order, or the reverse.

    False by default.

  • tree (boolean) – whether to return the discovery directed tree (each vertex being linked to the one that saw it for the first time)

    False by default.

  • initial_vertex – the first vertex to consider.

    None by default.

ALGORITHM:

This algorithm maintains for each vertex left in the graph a code corresponding to the vertices already removed. The vertex of maximal code ( according to the lexicographic order ) is then removed, and the codes are updated.

This algorithm runs in time O(n^2) ( where n is the number of vertices in the graph ), which is not optimal. An optimal algorithm would run in time O(m) ( where m is the number of edges in the graph ), and require the use of a doubly-linked list which are not available in python and can not really be written efficiently. This could be done in Cython, though.

EXAMPLE:

A Lex BFS is obviously an ordering of the vertices:

sage: g = graphs.PetersenGraph()
sage: len(g.lex_BFS()) == g.order()
True

For a Chordal Graph, a reversed Lex BFS is a Perfect Elimination Order

sage: g = graphs.PathGraph(3).lexicographic_product(graphs.CompleteGraph(2))
sage: g.lex_BFS(reverse=True)
[(2, 1), (2, 0), (1, 1), (1, 0), (0, 1), (0, 0)]

And the vertices at the end of the tree of discovery are, for chordal graphs, simplicial vertices (their neighborhood is a complete graph):

sage: g = graphs.ClawGraph().lexicographic_product(graphs.CompleteGraph(2))
sage: v = g.lex_BFS()[-1]
sage: peo, tree = g.lex_BFS(initial_vertex = v,  tree=True)
sage: leaves = [v for v in tree if tree.in_degree(v) ==0]
sage: all([g.subgraph(g.neighbors(v)).is_clique() for v in leaves])
True

TESTS:

There were some problems with the following call in the past (trac 10899) – now it should be fine:

sage: Graph(1).lex_BFS(tree=True)
([0], Digraph on 1 vertex)
lexicographic_product(other)

Returns the lexicographic product of self and other.

The lexicographic product of G and H is the graph L with vertex set V(L)=V(G)\times V(H), and ((u,v), (w,x)) is an edge iff :

  • (u, w) is an edge of G, or
  • u = w and (v, x) is an edge of H.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: C = graphs.CycleGraph(5)
sage: L = C.lexicographic_product(Z); L
Graph on 10 vertices
sage: L.plot() # long time
sage: D = graphs.DodecahedralGraph()
sage: P = graphs.PetersenGraph()
sage: L = D.lexicographic_product(P); L
Graph on 200 vertices
sage: L.plot() # long time

TESTS:

Lexicographic product of graphs:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.lexicographic_product(H)
sage: T.edges(labels=None)
[((0, 'a'), (0, 'b')), ((0, 'a'), (1, 'a')), ((0, 'a'), (1, 'b')), ((0, 'b'), (1, 'a')), ((0, 'b'), (1, 'b')), ((1, 'a'), (1, 'b')), ((1, 'a'), (2, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a')), ((1, 'b'), (2, 'b')), ((2, 'a'), (2, 'b'))]
sage: T.is_isomorphic( H.lexicographic_product(G) )
False

Lexicographic product of digraphs:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.lexicographic_product(J)
sage: T.edges(labels=None)
[((0, 'a'), (0, 'b')), ((0, 'a'), (1, 'a')), ((0, 'a'), (1, 'b')), ((0, 'b'), (1, 'a')), ((0, 'b'), (1, 'b')), ((1, 'a'), (1, 'b')), ((1, 'a'), (2, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a')), ((1, 'b'), (2, 'b')), ((2, 'a'), (2, 'b'))]
sage: T.is_isomorphic( J.lexicographic_product(I) )
False
line_graph(labels=True)

Returns the line graph of the (di)graph.

INPUT:

  • labels (boolean) – whether edge labels should be taken in consideration. If labels=True, the vertices of the line graph will be triples (u,v,label), and pairs of vertices otherwise.

    This is set to True by default.

The line graph of an undirected graph G is an undirected graph H such that the vertices of H are the edges of G and two vertices e and f of H are adjacent if e and f share a common vertex in G. In other words, an edge in H represents a path of length 2 in G.

The line graph of a directed graph G is a directed graph H such that the vertices of H are the edges of G and two vertices e and f of H are adjacent if e and f share a common vertex in G and the terminal vertex of e is the initial vertex of f. In other words, an edge in H represents a (directed) path of length 2 in G.

Note

As a Graph object only accepts hashable objects as vertices (and as the vertices of the line graph are the edges of the graph), this code will fail if edge labels are not hashable. You can also set the argument labels=False to ignore labels.

See also

EXAMPLES:

sage: g = graphs.CompleteGraph(4)
sage: h = g.line_graph()
sage: h.vertices()
[(0, 1, None),
(0, 2, None),
(0, 3, None),
(1, 2, None),
(1, 3, None),
(2, 3, None)]
sage: h.am()
[0 1 1 1 1 0]
[1 0 1 1 0 1]
[1 1 0 0 1 1]
[1 1 0 0 1 1]
[1 0 1 1 0 1]
[0 1 1 1 1 0]
sage: h2 = g.line_graph(labels=False)
sage: h2.vertices()
[(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]
sage: h2.am() == h.am()
True
sage: g = DiGraph([[1..4],lambda i,j: i<j])
sage: h = g.line_graph()
sage: h.vertices()
[(1, 2, None),
(1, 3, None),
(1, 4, None),
(2, 3, None),
(2, 4, None),
(3, 4, None)]
sage: h.edges()
[((1, 2, None), (2, 3, None), None),
 ((1, 2, None), (2, 4, None), None),
 ((1, 3, None), (3, 4, None), None),
 ((2, 3, None), (3, 4, None), None)]

Tests:

trac ticket #13787:

sage: g = graphs.KneserGraph(7,1)
sage: C = graphs.CompleteGraph(7)
sage: C.is_isomorphic(g)
True
sage: C.line_graph().is_isomorphic(g.line_graph())
True
longest_path(s=None, t=None, use_edge_labels=False, algorithm='MILP', solver=None, verbose=0)

Returns a longest path of self.

INPUT:

  • s (vertex) – forces the source of the path (the method then returns the longest path starting at s). The argument is set to None by default, which means that no constraint is set upon the first vertex in the path.

  • t (vertex) – forces the destination of the path (the method then returns the longest path ending at t). The argument is set to None by default, which means that no constraint is set upon the last vertex in the path.

  • use_edge_labels (boolean) – whether the labels on the edges are to be considered as weights (a label set to None or {} being considered as a weight of 1). Set to False by default.

  • algorithm – one of "MILP" (default) or "backtrack". Two remarks on this respect:

    • While the MILP formulation returns an exact answer, the backtrack algorithm is a randomized heuristic.
    • As the backtrack algorithm does not support edge weighting, setting use_edge_labels=True will force the use of the MILP algorithm.
  • solver – (default: None) Specify the Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.

  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

Note

The length of a path is assumed to be the number of its edges, or the sum of their labels.

OUTPUT:

A subgraph of self corresponding to a (directed if self is directed) longest path. If use_edge_labels == True, a pair weight, path is returned.

ALGORITHM:

Mixed Integer Linear Programming. (This problem is known to be NP-Hard).

EXAMPLES:

Petersen’s graph being hypohamiltonian, it has a longest path of length n-2:

sage: g = graphs.PetersenGraph()
sage: lp = g.longest_path()
sage: lp.order() >= g.order() - 2
True

The heuristic totally agrees:

sage: g = graphs.PetersenGraph()
sage: g.longest_path(algorithm="backtrack").edges()
[(0, 1, None), (1, 2, None), (2, 3, None), (3, 4, None), (4, 9, None), (5, 7, None), (5, 8, None), (6, 8, None), (6, 9, None)]

Let us compute longest paths on random graphs with random weights. Each time, we ensure the resulting graph is indeed a path:

sage: for i in range(20):
...       g = graphs.RandomGNP(15, 0.3)
...       for u, v in g.edges(labels=False):
...           g.set_edge_label(u, v, random())
...       lp = g.longest_path()
...       if (not lp.is_forest() or
...           not max(lp.degree()) <= 2 or
...           not lp.is_connected()):
...           print("Error!")
...           break

TESTS:

The argument algorithm must be either 'backtrack' or 'MILP':

sage: graphs.PetersenGraph().longest_path(algorithm="abc")
Traceback (most recent call last):
...
ValueError: algorithm must be either 'backtrack' or 'MILP'

Disconnected graphs not weighted:

sage: g1 = graphs.PetersenGraph()
sage: g2 = 2 * g1
sage: lp1 = g1.longest_path()
sage: lp2 = g2.longest_path()
sage: len(lp1) == len(lp2)
True

Disconnected graphs weighted:

sage: g1 = graphs.PetersenGraph()
sage: for u,v in g.edges(labels=False):
...       g.set_edge_label(u, v, random())
sage: g2 = 2 * g1
sage: lp1 = g1.longest_path(use_edge_labels=True)
sage: lp2 = g2.longest_path(use_edge_labels=True)
sage: lp1[0] == lp2[0]
True

Empty graphs:

sage: Graph().longest_path()
Graph on 0 vertices
sage: Graph().longest_path(use_edge_labels=True)
[0, Graph on 0 vertices]
sage: graphs.EmptyGraph().longest_path()
Graph on 0 vertices
sage: graphs.EmptyGraph().longest_path(use_edge_labels=True)
[0, Graph on 0 vertices]

Trivial graphs:

sage: G = Graph()
sage: G.add_vertex(0)
sage: G.longest_path()
Graph on 0 vertices
sage: G.longest_path(use_edge_labels=True)
[0, Graph on 0 vertices]
sage: graphs.CompleteGraph(1).longest_path()
Graph on 0 vertices
sage: graphs.CompleteGraph(1).longest_path(use_edge_labels=True)
[0, Graph on 0 vertices]

Random test for digraphs:

sage: for i in range(20):
...       g = digraphs.RandomDirectedGNP(15, 0.3)
...       for u, v in g.edges(labels=False):
...           g.set_edge_label(u, v, random())
...       lp = g.longest_path()
...       if (not lp.is_directed_acyclic() or
...           not max(lp.out_degree()) <= 1 or
...           not max(lp.in_degree()) <= 1 or
...           not lp.is_connected()):
...           print("Error!")
...           print g.edges()
...           break

trac ticket #13019:

sage: g = graphs.CompleteGraph(5).to_directed()
sage: g.longest_path(s=1,t=2)
Subgraph of (Complete graph): Digraph on 5 vertices

trac ticket #14412:

sage: l = [(0, 1), (0, 3), (2, 0)]
sage: G = DiGraph(l)
sage: G.longest_path().edges()
[(0, 1, None), (2, 0, None)]
loop_edges()

Returns a list of all loops in the graph.

EXAMPLES:

sage: G = Graph(4, loops=True)
sage: G.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: G.loop_edges()
[(0, 0, None), (1, 1, None), (2, 2, None), (3, 3, None)]
sage: D = DiGraph(4, loops=True)
sage: D.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: D.loop_edges()
[(0, 0, None), (1, 1, None), (2, 2, None), (3, 3, None)]
sage: G = Graph(4, loops=True, multiedges=True, sparse=True)
sage: G.add_edges([(i,i) for i in range(4)])
sage: G.loop_edges()
[(0, 0, None), (1, 1, None), (2, 2, None), (3, 3, None)]
loop_vertices()

Returns a list of vertices with loops.

EXAMPLES:

sage: G = Graph({0 : [0], 1: [1,2,3], 2: [3]}, loops=True)
sage: G.loop_vertices()
[0, 1]
loops(labels=True)

Returns any loops in the (di)graph.

INPUT:

  • new – deprecated
  • labels – whether returned edges have labels ((u,v,l)) or not ((u,v)).

EXAMPLES:

sage: G = Graph(loops=True); G
Looped graph on 0 vertices
sage: G.has_loops()
False
sage: G.allows_loops()
True
sage: G.add_edge((0,0))
sage: G.has_loops()
True
sage: G.loops()
[(0, 0, None)]
sage: G.allow_loops(False); G
Graph on 1 vertex
sage: G.has_loops()
False
sage: G.edges()
[]

sage: D = DiGraph(loops=True); D
Looped digraph on 0 vertices
sage: D.has_loops()
False
sage: D.allows_loops()
True
sage: D.add_edge((0,0))
sage: D.has_loops()
True
sage: D.loops()
[(0, 0, None)]
sage: D.allow_loops(False); D
Digraph on 1 vertex
sage: D.has_loops()
False
sage: D.edges()
[]

sage: G = graphs.PetersenGraph()
sage: G.loops()
[]
max_cut(value_only=True, use_edge_labels=False, vertices=False, solver=None, verbose=0)

Returns a maximum edge cut of the graph. For more information, see the Wikipedia article on cuts.

INPUT:

  • value_only – boolean (default: True)
    • When set to True (default), only the value is returned.
    • When set to False, both the value and a maximum edge cut are returned.
  • use_edge_labels – boolean (default: False)
    • When set to True, computes a maximum weighted cut where each edge has a weight defined by its label. (If an edge has no label, 1 is assumed.)
    • When set to False, each edge has weight 1.
  • vertices – boolean (default: False)
    • When set to True, also returns the two sets of vertices that are disconnected by the cut. This implies value_only=False.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

EXAMPLE:

Quite obviously, the max cut of a bipartite graph is the number of edges, and the two sets of vertices are the the two sides

sage: g = graphs.CompleteBipartiteGraph(5,6)
sage: [ value, edges, [ setA, setB ]] = g.max_cut(vertices=True)
sage: value == 5*6
True
sage: bsetA, bsetB  = map(list,g.bipartite_sets())
sage: (bsetA == setA and bsetB == setB ) or ((bsetA == setB and bsetB == setA ))
True

The max cut of a Petersen graph:

sage: g=graphs.PetersenGraph()
sage: g.max_cut()
12
merge_vertices(vertices)

Merge vertices.

This function replaces a set S of vertices by a single vertex v_{new}, such that the edge uv_{new} exists if and only if \exists v'\in S: (u,v')\in G.

The new vertex is named after the first vertex in the list given in argument. If this first name is None, a new vertex is created.

In the case of multigraphs, the multiplicity is preserved.

INPUT:

  • vertices – the set of vertices to be merged

EXAMPLE:

sage: g=graphs.CycleGraph(3)
sage: g.merge_vertices([0,1])
sage: g.edges()
[(0, 2, None)]

sage: # With a Multigraph :
sage: g=graphs.CycleGraph(3)
sage: g.allow_multiple_edges(True)
sage: g.merge_vertices([0,1])
sage: g.edges(labels=False)
[(0, 2), (0, 2)]

sage: P=graphs.PetersenGraph()
sage: P.merge_vertices([5,7])
sage: P.vertices()
[0, 1, 2, 3, 4, 5, 6, 8, 9]

sage: g=graphs.CycleGraph(5)
sage: g.vertices()
[0, 1, 2, 3, 4]
sage: g.merge_vertices([None, 1, 3])
sage: g.edges(labels=False)
[(0, 4), (0, 5), (2, 5), (4, 5)]
min_spanning_tree(weight_function=<function <lambda> at 0x46781b8>, algorithm='Kruskal', starting_vertex=None, check=False)

Returns the edges of a minimum spanning tree.

INPUT:

  • weight_function – A function that takes an edge and returns a numeric weight. Defaults to assigning each edge a weight of 1.
  • algorithm – The algorithm to use in computing a minimum spanning tree of G. The default is to use Kruskal’s algorithm. The following algorithms are supported:
    • "Kruskal" – Kruskal’s algorithm.
    • "Prim_fringe" – a variant of Prim’s algorithm. "Prim_fringe" ignores the labels on the edges.
    • "Prim_edge" – a variant of Prim’s algorithm.
    • NetworkX – Uses NetworkX’s minimum spanning tree implementation.
  • starting_vertex – The vertex from which to begin the search for a minimum spanning tree.
  • check – Boolean; default: False. Whether to first perform sanity checks on the input graph G. If appropriate, check is passed on to any minimum spanning tree functions that are invoked from the current method. See the documentation of the corresponding functions for details on what sort of sanity checks will be performed.

OUTPUT:

The edges of a minimum spanning tree of G, if one exists, otherwise returns the empty list.

EXAMPLES:

Kruskal’s algorithm:

sage: g = graphs.CompleteGraph(5)
sage: len(g.min_spanning_tree())
4
sage: weight = lambda e: 1 / ((e[0] + 1) * (e[1] + 1))
sage: g.min_spanning_tree(weight_function=weight)
[(3, 4, None), (2, 4, None), (1, 4, None), (0, 4, None)]
sage: g = graphs.PetersenGraph()
sage: g.allow_multiple_edges(True)
sage: g.weighted(True)
sage: g.add_edges(g.edges())
sage: g.min_spanning_tree()
[(0, 1, None), (0, 4, None), (0, 5, None), (1, 2, None), (1, 6, None), (2, 3, None), (2, 7, None), (3, 8, None), (4, 9, None)]

Prim’s algorithm:

sage: g = graphs.CompleteGraph(5)
sage: g.min_spanning_tree(algorithm='Prim_edge', starting_vertex=2, weight_function=weight)
[(2, 4, None), (3, 4, None), (1, 4, None), (0, 4, None)]
sage: g.min_spanning_tree(algorithm='Prim_fringe', starting_vertex=2, weight_function=weight)
[(2, 4), (4, 3), (4, 1), (4, 0)]
multicommodity_flow(terminals, integer=True, use_edge_labels=False, vertex_bound=False, solver=None, verbose=0)

Solves a multicommodity flow problem.

In the multicommodity flow problem, we are given a set of pairs (s_i, t_i), called terminals meaning that s_i is willing some flow to t_i.

Even though it is a natural generalisation of the flow problem this version of it is NP-Complete to solve when the flows are required to be integer.

For more information, see the Wikipedia page on multicommodity flows.

INPUT:

  • terminals – a list of pairs (s_i, t_i) or triples (s_i, t_i, w_i) representing a flow from s_i to t_i of intensity w_i. When the pairs are of size 2, a intensity of 1 is assumed.
  • integer (boolean) – whether to require an integer multicommodity flow
  • use_edge_labels (boolean) – whether to consider the label of edges as numerical values representing a capacity. If set to False, a capacity of 1 is assumed
  • vertex_bound (boolean) – whether to require that a vertex can stand at most 1 commodity of flow through it of intensity 1. Terminals can obviously still send or receive several units of flow even though vertex_bound is set to True, as this parameter is meant to represent topological properties.
  • solver – Specify a Linear Program solver to be used. If set to None, the default one is used. function of MixedIntegerLinearProgram. See the documentation of MixedIntegerLinearProgram.solve for more informations.
  • verbose (integer) – sets the level of verbosity. Set to 0 by default (quiet).

ALGORITHM:

(Mixed Integer) Linear Program, depending on the value of integer.

EXAMPLE:

An easy way to obtain a satisfiable multiflow is to compute a matching in a graph, and to consider the paired vertices as terminals

sage: g = graphs.PetersenGraph() 
sage: matching = [(u,v) for u,v,_ in g.matching()]
sage: h = g.multicommodity_flow(matching)
sage: len(h)
5 

We could also have considered g as symmetric and computed the multiflow in this version instead. In this case, however edges can be used in both directions at the same time:

sage: h = DiGraph(g).multicommodity_flow(matching)
sage: len(h)
5 

An exception is raised when the problem has no solution

sage: h = g.multicommodity_flow([(u,v,3) for u,v in matching])
... 
ValueError: The multiflow problem has no solution 
multiple_edges(to_undirected=False, labels=True)

Returns any multiple edges in the (di)graph.

EXAMPLES:

sage: G = Graph(multiedges=True,sparse=True); G
Multi-graph on 0 vertices
sage: G.has_multiple_edges()
False
sage: G.allows_multiple_edges()
True
sage: G.add_edges([(0,1)]*3)
sage: G.has_multiple_edges()
True
sage: G.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: G.allow_multiple_edges(False); G
Graph on 2 vertices
sage: G.has_multiple_edges()
False
sage: G.edges()
[(0, 1, None)]

sage: D = DiGraph(multiedges=True,sparse=True); D
Multi-digraph on 0 vertices
sage: D.has_multiple_edges()
False
sage: D.allows_multiple_edges()
True
sage: D.add_edges([(0,1)]*3)
sage: D.has_multiple_edges()
True
sage: D.multiple_edges()
[(0, 1, None), (0, 1, None), (0, 1, None)]
sage: D.allow_multiple_edges(False); D
Digraph on 2 vertices
sage: D.has_multiple_edges()
False
sage: D.edges()
[(0, 1, None)]

sage: G = DiGraph({1:{2: 'h'}, 2:{1:'g'}},sparse=True)
sage: G.has_multiple_edges()
False
sage: G.has_multiple_edges(to_undirected=True)
True
sage: G.multiple_edges()
[]
sage: G.multiple_edges(to_undirected=True)
[(1, 2, 'h'), (2, 1, 'g')]
multiway_cut(vertices, value_only=False, use_edge_labels=False, solver=None, verbose=0)

Returns a minimum edge multiway cut corresponding to the given set of vertices ( cf. http://www.d.kth.se/~viggo/wwwcompendium/node92.html ) represented by a list of edges.

A multiway cut for a vertex set S in a graph or a digraph G is a set C of edges such that any two vertices u,v in S are disconnected when removing the edges from C from G.

Such a cut is said to be minimum when its cardinality (or weight) is minimum.

INPUT:

  • vertices (iterable)– the set of vertices

  • value_only (boolean)

    • When set to True, only the value of a minimum multiway cut is returned.
    • When set to False (default), the list of edges is returned
  • use_edge_labels (boolean)
    • When set to True, computes a weighted minimum cut where each edge has a weight defined by its label. ( if an edge has no label, 1 is assumed )
    • when set to False (default), each edge has weight 1.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.

  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

EXAMPLES:

Of course, a multiway cut between two vertices correspond to a minimum edge cut

sage: g = graphs.PetersenGraph()
sage: g.edge_cut(0,3) == g.multiway_cut([0,3], value_only = True)
True

As Petersen’s graph is 3-regular, a minimum multiway cut between three vertices contains at most 2\times 3 edges (which could correspond to the neighborhood of 2 vertices):

sage: g.multiway_cut([0,3,9], value_only = True) == 2*3
True

In this case, though, the vertices are an independent set. If we pick instead vertices 0,9, and 7, we can save 4 edges in the multiway cut

sage: g.multiway_cut([0,7,9], value_only = True) == 2*3 - 1
True

This example, though, does not work in the directed case anymore, as it is not possible in Petersen’s graph to mutualise edges

sage: g = DiGraph(g)
sage: g.multiway_cut([0,7,9], value_only = True) == 3*3
True

Of course, a multiway cut between the whole vertex set contains all the edges of the graph:

sage: C = g.multiway_cut(g.vertices())
sage: set(C) == set(g.edges())
True
name(new=None)

Returns or sets the graph’s name.

INPUT:

  • new - if not None, then this becomes the new name of the (di)graph. (if new == ‘’, removes any name)

EXAMPLES:

sage: d = {0: [1,4,5], 1: [2,6], 2: [3,7], 3: [4,8], 4: [9], 5: [7, 8], 6: [8,9], 7: [9]}
sage: G = Graph(d); G
Graph on 10 vertices
sage: G.name("Petersen Graph"); G
Petersen Graph: Graph on 10 vertices
sage: G.name(new=""); G
Graph on 10 vertices
sage: G.name()
''
neighbor_iterator(vertex)

Return an iterator over neighbors of vertex.

EXAMPLES:

sage: G = graphs.CubeGraph(3)
sage: for i in G.neighbor_iterator('010'):
...    print i
011
000
110
sage: D = G.to_directed()
sage: for i in D.neighbor_iterator('010'):
...    print i
011
000
110
sage: D = DiGraph({0:[1,2], 3:[0]})
sage: list(D.neighbor_iterator(0))
[1, 2, 3]
neighbors(vertex)

Return a list of neighbors (in and out if directed) of vertex.

G[vertex] also works.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: sorted(P.neighbors(3))
[2, 4, 8]
sage: sorted(P[4])
[0, 3, 9]
networkx_graph(copy=True)

Creates a new NetworkX graph from the Sage graph.

INPUT:

  • copy - if False, and the underlying implementation is a NetworkX graph, then the actual object itself is returned.

EXAMPLES:

sage: G = graphs.TetrahedralGraph()
sage: N = G.networkx_graph()
sage: type(N)
<class 'networkx.classes.graph.Graph'>
sage: G = graphs.TetrahedralGraph()
sage: G = Graph(G, implementation='networkx')
sage: N = G.networkx_graph()
sage: G._backend._nxg is N
False
sage: G = Graph(graphs.TetrahedralGraph(), implementation='networkx')
sage: N = G.networkx_graph(copy=False)
sage: G._backend._nxg is N
True
num_edges()

Returns the number of edges.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.size()
15
num_verts()

Returns the number of vertices. Note that len(G) returns the number of vertices in G also.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.order()
10
sage: G = graphs.TetrahedralGraph()
sage: len(G)
4
number_of_loops()

Returns the number of edges that are loops.

EXAMPLES:

sage: G = Graph(4, loops=True)
sage: G.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: G.edges(labels=False)
[(0, 0), (1, 1), (2, 2), (2, 3), (3, 3)]
sage: G.number_of_loops()
4
sage: D = DiGraph(4, loops=True)
sage: D.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: D.edges(labels=False)
[(0, 0), (1, 1), (2, 2), (2, 3), (3, 3)]
sage: D.number_of_loops()
4
order()

Returns the number of vertices. Note that len(G) returns the number of vertices in G also.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.order()
10
sage: G = graphs.TetrahedralGraph()
sage: len(G)
4
periphery()

Returns the set of vertices in the periphery, i.e. whose eccentricity is equal to the diameter of the (di)graph.

In other words, the periphery is the set of vertices achieving the maximum eccentricity.

EXAMPLES:

sage: G = graphs.DiamondGraph()
sage: G.periphery()
[0, 3]
sage: P = graphs.PetersenGraph()
sage: P.subgraph(P.periphery()) == P
True
sage: S = graphs.StarGraph(19)
sage: S.periphery()
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
sage: G = Graph()
sage: G.periphery()
[]
sage: G.add_vertex()
0
sage: G.periphery()
[0]
plot(**options)

Returns a graphics object representing the (di)graph.

INPUT:

  • pos - an optional positioning dictionary

  • layout - what kind of layout to use, takes precedence over pos

    • ‘circular’ – plots the graph with vertices evenly distributed on a circle
    • ‘spring’ - uses the traditional spring layout, using the graph’s current positions as initial positions
    • ‘tree’ - the (di)graph must be a tree. One can specify the root of the tree using the keyword tree_root, otherwise a root will be selected at random. Then the tree will be plotted in levels, depending on minimum distance for the root.
  • vertex_labels - whether to print vertex labels

  • edge_labels - whether to print edge labels. By default, False, but if True, the result of str(l) is printed on the edge for each label l. Labels equal to None are not printed (to set edge labels, see set_edge_label).

  • vertex_size - size of vertices displayed

  • vertex_shape - the shape to draw the vertices (Not available for multiedge digraphs.)

  • graph_border - whether to include a box around the graph

  • vertex_colors - optional dictionary to specify vertex colors: each key is a color recognizable by matplotlib, and each corresponding entry is a list of vertices. If a vertex is not listed, it looks invisible on the resulting plot (it doesn’t get drawn).

  • edge_colors - a dictionary specifying edge colors: each key is a color recognized by matplotlib, and each entry is a list of edges.

  • partition - a partition of the vertex set. if specified, plot will show each cell in a different color. vertex_colors takes precedence.

  • talk - if true, prints large vertices with white backgrounds so that labels are legible on slides

  • iterations - how many iterations of the spring layout algorithm to go through, if applicable

  • color_by_label - a boolean or dictionary or function (default: False)

    whether to color each edge with a different color according to its label; the colors are chosen along a rainbow, unless they are specified by a function or dictionary mapping labels to colors; this option is incompatible with edge_color and edge_colors.

  • heights - if specified, this is a dictionary from a set of floating point heights to a set of vertices

  • edge_style - keyword arguments passed into the edge-drawing routine. This currently only works for directed graphs, since we pass off the undirected graph to networkx

  • tree_root - a vertex of the tree to be used as the root for the layout=”tree” option. If no root is specified, then one is chosen at random. Ignored unless layout=’tree’.

  • tree_orientation - “up” or “down” (default is “down”). If “up” (resp., “down”), then the root of the tree will appear on the bottom (resp., top) and the tree will grow upwards (resp. downwards). Ignored unless layout=’tree’.

  • save_pos - save position computed during plotting

Note

  • See the documentation of the sage.graphs.graph_plot module for information and examples of how to define parameters that will be applied to all graph plots.
  • Default parameters for this method and a specific graph can also be set through the options mechanism. For more information on this different way to set default parameters, see the help of the options decorator.
  • See also the sage.graphs.graph_latex module for ways to use LaTeX to produce an image of a graph.

EXAMPLES:

sage: from sage.graphs.graph_plot import graphplot_options
sage: list(sorted(graphplot_options.iteritems()))
[...]

sage: from math import sin, cos, pi
sage: P = graphs.PetersenGraph()
sage: d = {'#FF0000':[0,5], '#FF9900':[1,6], '#FFFF00':[2,7], '#00FF00':[3,8], '#0000FF':[4,9]}
sage: pos_dict = {}
sage: for i in range(5):
...    x = float(cos(pi/2 + ((2*pi)/5)*i))
...    y = float(sin(pi/2 + ((2*pi)/5)*i))
...    pos_dict[i] = [x,y]
...
sage: for i in range(10)[5:]:
...    x = float(0.5*cos(pi/2 + ((2*pi)/5)*i))
...    y = float(0.5*sin(pi/2 + ((2*pi)/5)*i))
...    pos_dict[i] = [x,y]
...
sage: pl = P.plot(pos=pos_dict, vertex_colors=d)
sage: pl.show()
sage: C = graphs.CubeGraph(8)
sage: P = C.plot(vertex_labels=False, vertex_size=0, graph_border=True)
sage: P.show()
sage: G = graphs.HeawoodGraph()
sage: for u,v,l in G.edges():
...    G.set_edge_label(u,v,'(' + str(u) + ',' + str(v) + ')')
sage: G.plot(edge_labels=True).show()
sage: D = DiGraph( { 0: [1, 10, 19], 1: [8, 2], 2: [3, 6], 3: [19, 4], 4: [17, 5], 5: [6, 15], 6: [7], 7: [8, 14], 8: [9], 9: [10, 13], 10: [11], 11: [12, 18], 12: [16, 13], 13: [14], 14: [15], 15: [16], 16: [17], 17: [18], 18: [19], 19: []} , sparse=True)
sage: for u,v,l in D.edges():
...    D.set_edge_label(u,v,'(' + str(u) + ',' + str(v) + ')')
sage: D.plot(edge_labels=True, layout='circular').show()
sage: from sage.plot.colors import rainbow
sage: C = graphs.CubeGraph(5)
sage: R = rainbow(5)
sage: edge_colors = {}
sage: for i in range(5):
...    edge_colors[R[i]] = []
sage: for u,v,l in C.edges():
...    for i in range(5):
...        if u[i] != v[i]:
...            edge_colors[R[i]].append((u,v,l))
sage: C.plot(vertex_labels=False, vertex_size=0, edge_colors=edge_colors).show()
sage: D = graphs.DodecahedralGraph()
sage: Pi = [[6,5,15,14,7],[16,13,8,2,4],[12,17,9,3,1],[0,19,18,10,11]]
sage: D.show(partition=Pi)
sage: G = graphs.PetersenGraph()
sage: G.allow_loops(True)
sage: G.add_edge(0,0)
sage: G.show()
sage: D = DiGraph({0:[0,1], 1:[2], 2:[3]}, loops=True)
sage: D.show()
sage: D.show(edge_colors={(0,1,0):[(0,1,None),(1,2,None)],(0,0,0):[(2,3,None)]}) 
sage: pos = {0:[0.0, 1.5], 1:[-0.8, 0.3], 2:[-0.6, -0.8], 3:[0.6, -0.8], 4:[0.8, 0.3]}
sage: g = Graph({0:[1], 1:[2], 2:[3], 3:[4], 4:[0]})
sage: g.plot(pos=pos, layout='spring', iterations=0)
sage: G = Graph()
sage: P = G.plot()
sage: P.axes()
False
sage: G = DiGraph()
sage: P = G.plot()
sage: P.axes()
False
sage: G = graphs.PetersenGraph()
sage: G.get_pos()
{0: (6.12..., 1.0...),
 1: (-0.95..., 0.30...),
 2: (-0.58..., -0.80...),
 3: (0.58..., -0.80...),
 4: (0.95..., 0.30...),
 5: (1.53..., 0.5...),
 6: (-0.47..., 0.15...),
 7: (-0.29..., -0.40...),
 8: (0.29..., -0.40...),
 9: (0.47..., 0.15...)}
sage: P = G.plot(save_pos=True, layout='spring')

The following illustrates the format of a position dictionary.

sage: G.get_pos() # currently random across platforms, see #9593
{0: [1.17..., -0.855...], 
 1: [1.81..., -0.0990...], 
 2: [1.35..., 0.184...], 
 3: [1.51..., 0.644...], 
 4: [2.00..., -0.507...], 
 5: [0.597..., -0.236...], 
 6: [2.04..., 0.687...], 
 7: [1.46..., -0.473...], 
 8: [0.902..., 0.773...], 
 9: [2.48..., -0.119...]}
sage: T = list(graphs.trees(7))
sage: t = T[3]
sage: t.plot(heights={0:[0], 1:[4,5,1], 2:[2], 3:[3,6]})
sage: T = list(graphs.trees(7))
sage: t = T[3]
sage: t.plot(heights={0:[0], 1:[4,5,1], 2:[2], 3:[3,6]})
sage: t.set_edge_label(0,1,-7)
sage: t.set_edge_label(0,5,3)
sage: t.set_edge_label(0,5,99)
sage: t.set_edge_label(1,2,1000)
sage: t.set_edge_label(3,2,'spam')
sage: t.set_edge_label(2,6,3/2)
sage: t.set_edge_label(0,4,66)
sage: t.plot(heights={0:[0], 1:[4,5,1], 2:[2], 3:[3,6]}, edge_labels=True)
sage: T = list(graphs.trees(7))
sage: t = T[3]
sage: t.plot(layout='tree')
sage: t = DiGraph('JCC???@A??GO??CO??GO??')
sage: t.plot(layout='tree', tree_root=0, tree_orientation="up")
sage: D = DiGraph({0:[1,2,3], 2:[1,4], 3:[0]})
sage: D.plot()

sage: D = DiGraph(multiedges=True,sparse=True)
sage: for i in range(5):
...     D.add_edge((i,i+1,'a'))
...     D.add_edge((i,i-1,'b'))
sage: D.plot(edge_labels=True,edge_colors=D._color_by_label())
sage: D.plot(edge_labels=True, color_by_label={'a':'blue', 'b':'red'}, edge_style='dashed')

sage: g = Graph({}, loops=True, multiedges=True,sparse=True)
sage: g.add_edges([(0,0,'a'),(0,0,'b'),(0,1,'c'),(0,1,'d'),
...     (0,1,'e'),(0,1,'f'),(0,1,'f'),(2,1,'g'),(2,2,'h')])
sage: g.plot(edge_labels=True, color_by_label=True, edge_style='dashed')
sage: S = SupersingularModule(389)
sage: H = S.hecke_matrix(2)
sage: D = DiGraph(H,sparse=True)
sage: P = D.plot()
sage: G=Graph({'a':['a','b','b','b','e'],'b':['c','d','e'],'c':['c','d','d','d'],'d':['e']},sparse=True)
sage: G.show(pos={'a':[0,1],'b':[1,1],'c':[2,0],'d':[1,0],'e':[0,0]})

TESTS:

sage: G = DiGraph({0:{1:'a', 2:'a'}, 1:{0:'b'}, 2:{0:'c'}})
sage: p = G.plot(edge_labels=True, color_by_label={'a':'yellow', 'b':'purple'}); p
sage: sorted([x.options()['rgbcolor'] for x in p if isinstance(x, sage.plot.arrow.CurveArrow)])
['black', 'purple', 'yellow', 'yellow']
plot3d(bgcolor=(1, 1, 1), vertex_colors=None, vertex_size=0.06, vertex_labels=False, edge_colors=None, edge_size=0.02, edge_size2=0.0325, pos3d=None, color_by_label=False, engine='jmol', **kwds)

Plot a graph in three dimensions.

See also the sage.graphs.graph_latex module for ways to use LaTeX to produce an image of a graph.

INPUT:

  • bgcolor - rgb tuple (default: (1,1,1))

  • vertex_size - float (default: 0.06)

  • vertex_labels – a boolean (default: False) whether to display vertices using text labels instead of spheres

  • vertex_colors - optional dictionary to specify vertex colors: each key is a color recognizable by tachyon (rgb tuple (default: (1,0,0))), and each corresponding entry is a list of vertices. If a vertex is not listed, it looks invisible on the resulting plot (it doesn’t get drawn).

  • edge_colors - a dictionary specifying edge colors: each key is a color recognized by tachyon ( default: (0,0,0) ), and each entry is a list of edges.

  • color_by_label - a boolean or dictionary or function (default: False)

    whether to color each edge with a different color according to its label; the colors are chosen along a rainbow, unless they are specified by a function or dictionary mapping labels to colors; this option is incompatible with edge_color and edge_colors.

  • edge_size - float (default: 0.02)

  • edge_size2 - float (default: 0.0325), used for Tachyon sleeves

  • pos3d - a position dictionary for the vertices

  • layout, iterations, ... - layout options; see layout()

  • engine - which renderer to use. Options:

    • 'jmol' - default
    • 'tachyon'
  • xres - resolution

  • yres - resolution

  • **kwds - passed on to the rendering engine

EXAMPLES:

sage: G = graphs.CubeGraph(5)
sage: G.plot3d(iterations=500, edge_size=None, vertex_size=0.04) # long time

We plot a fairly complicated Cayley graph:

sage: A5 = AlternatingGroup(5); A5
Alternating group of order 5!/2 as a permutation group
sage: G = A5.cayley_graph()
sage: G.plot3d(vertex_size=0.03, edge_size=0.01, vertex_colors={(1,1,1):G.vertices()}, bgcolor=(0,0,0), color_by_label=True, iterations=200) # long time

Some Tachyon examples:

sage: D = graphs.DodecahedralGraph()
sage: P3D = D.plot3d(engine='tachyon')
sage: P3D.show() # long time
sage: G = graphs.PetersenGraph()
sage: G.plot3d(engine='tachyon', vertex_colors={(0,0,1):G.vertices()}).show() # long time
sage: C = graphs.CubeGraph(4)
sage: C.plot3d(engine='tachyon', edge_colors={(0,1,0):C.edges()}, vertex_colors={(1,1,1):C.vertices()}, bgcolor=(0,0,0)).show() # long time
sage: K = graphs.CompleteGraph(3)
sage: K.plot3d(engine='tachyon', edge_colors={(1,0,0):[(0,1,None)], (0,1,0):[(0,2,None)], (0,0,1):[(1,2,None)]}).show() # long time

A directed version of the dodecahedron

sage: D = DiGraph( { 0: [1, 10, 19], 1: [8, 2], 2: [3, 6], 3: [19, 4], 4: [17, 5], 5: [6, 15], 6: [7], 7: [8, 14], 8: [9], 9: [10, 13], 10: [11], 11: [12, 18], 12: [16, 13], 13: [14], 14: [15], 15: [16], 16: [17], 17: [18], 18: [19], 19: []} )
sage: D.plot3d().show() # long time
sage: P = graphs.PetersenGraph().to_directed()
sage: from sage.plot.colors import rainbow
sage: edges = P.edges()
sage: R = rainbow(len(edges), 'rgbtuple')
sage: edge_colors = {}
sage: for i in range(len(edges)):
...       edge_colors[R[i]] = [edges[i]]
sage: P.plot3d(engine='tachyon', edge_colors=edge_colors).show() # long time
sage: G=Graph({'a':['a','b','b','b','e'],'b':['c','d','e'],'c':['c','d','d','d'],'d':['e']},sparse=True)
sage: G.show3d()
Traceback (most recent call last):
...
NotImplementedError: 3D plotting of multiple edges or loops not implemented.

TESTS:

sage: G = DiGraph({0:{1:'a', 2:'a'}, 1:{0:'b'}, 2:{0:'c'}})
sage: p = G.plot3d(edge_labels=True, color_by_label={'a':'yellow', 'b':'cyan'})
sage: s = p.x3d_str()

This 3D plot contains four yellow objects (two cylinders and two cones), two black objects and 2 cyan objects:

sage: s.count("Material diffuseColor='1.0 1.0 0.0'")
4
sage: s.count("Material diffuseColor='0.0 0.0 0.0'")
2
sage: s.count("Material diffuseColor='0.0 1.0 1.0'")
2

See also

radius()

Returns the radius of the (di)graph.

The radius is defined to be the minimum eccentricity of any vertex, where the eccentricity is the maximum distance to any other vertex.

EXAMPLES: The more symmetric a graph is, the smaller (diameter - radius) is.

sage: G = graphs.BarbellGraph(9, 3)
sage: G.radius()
3
sage: G.diameter()
6
sage: G = graphs.OctahedralGraph()
sage: G.radius()
2
sage: G.diameter()
2

TEST:

sage: g = Graph()
sage: g.radius()
Traceback (most recent call last):
...
ValueError: This method has no meaning on empty graphs.
random_edge(**kwds)

Returns a random edge of self.

INPUT:

  • **kwds - arguments to be passed down to the edge_iterator method.

EXAMPLE:

The returned value is an edge of self:

sage: g = graphs.PetersenGraph()
sage: u,v = g.random_edge(labels=False)
sage: g.has_edge(u,v)
True

As the edges() method would, this function returns by default a triple (u,v,l) of values, in which l is the label of edge (u,v):

sage: g.random_edge()
(3, 4, None)
random_subgraph(p, inplace=False)

Return a random subgraph that contains each vertex with prob. p.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.random_subgraph(.25)
Subgraph of (Petersen graph): Graph on 4 vertices
random_vertex(**kwds)

Returns a random vertex of self.

INPUT:

  • **kwds - arguments to be passed down to the vertex_iterator method.

EXAMPLE:

The returned value is a vertex of self:

sage: g = graphs.PetersenGraph()
sage: v = g.random_vertex()
sage: v in g
True
relabel(perm=None, inplace=True, return_map=False, check_input=True, complete_partial_function=True)

Relabels the vertices of self

INPUT:

  • perm – a function, dictionary, list, permutation, or None (default: None)
  • inplace – a boolean (default: True)
  • return_map – a boolean (default: False)
  • check_input (boolean) – whether to test input for correctness. This can potentially be very time-consuming !.
  • complete_partial_function (boolean) – whether to automatically complete the permutation if some elements of the graph are not associated with any new name. In this case, those elements are not relabeled This can potentially be very time-consuming !.

If perm is a function f, then each vertex v is relabeled to f(v).

If perm is a dictionary d, then each vertex v (which should be a key of d) is relabeled to d[v]. Similarly, if perm is a list or tuple l of length n, then each vertex (which should be in \{0,1,...,n-1\}) is relabeled to l[v].

If perm is a permutation, then each vertex v is relabeled to perm(v). Caveat: this assumes that the vertices are labelled \{0,1,...,n-1\}; since permutations act by default on the set \{1,2,...,n\}, this is achieved by identifying n and 0.

If perm is None, the graph is relabeled to be on the vertices \{0,1,...,n-1\}.

Note

at this point, only injective relabeling are supported.

If inplace is True, the graph is modified in place and None is returned. Otherwise a relabeled copy of the graph is returned.

If return_map is True a dictionary representing the relabelling map is returned (incompatible with inplace==False).

EXAMPLES:

sage: G = graphs.PathGraph(3)
sage: G.am()
[0 1 0]
[1 0 1]
[0 1 0]

Relabeling using a dictionary. Note that the dictionary does not define the new label of vertex 0:

sage: G.relabel({1:2,2:1}, inplace=False).am()
[0 0 1]
[0 0 1]
[1 1 0]

This is because the method automatically “extends” the relabeling to the missing vertices (whose label will not change). Checking that all vertices have an image can require some time, and this feature can be disabled (at your own risk):

sage: G.relabel({1:2,2:1}, inplace=False, complete_partial_function = False).am()
Traceback (most recent call last):
...
KeyError: 0

Relabeling using a list:

sage: G.relabel([0,2,1], inplace=False).am()
[0 0 1]
[0 0 1]
[1 1 0]

Relabeling using a tuple:

sage: G.relabel((0,2,1), inplace=False).am()
[0 0 1]
[0 0 1]
[1 1 0]

Relabeling using a Sage permutation:

sage: G = graphs.PathGraph(3)
sage: from sage.groups.perm_gps.permgroup_named import SymmetricGroup
sage: S = SymmetricGroup(3)
sage: gamma = S('(1,2)')
sage: G.relabel(gamma, inplace=False).am()
[0 0 1]
[0 0 1]
[1 1 0]

Relabeling using an injective function:

sage: G.edges()
[(0, 1, None), (1, 2, None)]
sage: H = G.relabel(lambda i: i+10, inplace=False)
sage: H.vertices()
[10, 11, 12]
sage: H.edges()
[(10, 11, None), (11, 12, None)]

Relabeling using a non injective function has no meaning:

sage: G.edges()
[(0, 1, None), (1, 2, None)]
sage: G.relabel(lambda i: 0, inplace=False)
Traceback (most recent call last):
...
NotImplementedError: Non injective relabeling

But this test can be disabled, which leads to ... problems:

sage: G.edges()
[(0, 1, None), (1, 2, None)]
sage: G.relabel(lambda i: 0, check_input = False)
sage: G.edges()
[(0, 0, None)]

Relabeling to simpler labels:

sage: G = graphs.CubeGraph(3)
sage: G.vertices()
['000', '001', '010', '011', '100', '101', '110', '111']
sage: G.relabel()
sage: G.vertices()
[0, 1, 2, 3, 4, 5, 6, 7]

Recovering the relabeling with return_map:

sage: G = graphs.CubeGraph(3)
sage: expecting = {'000': 0, '001': 1, '010': 2, '011': 3, '100': 4, '101': 5, '110': 6, '111': 7}
sage: G.relabel(return_map=True) == expecting
True
sage: G = graphs.PathGraph(3)
sage: G.relabel(lambda i: i+10, return_map=True)
{0: 10, 1: 11, 2: 12}

TESTS:

sage: P = Graph(graphs.PetersenGraph())
sage: P.delete_edge([0,1])
sage: P.add_edge((4,5))
sage: P.add_edge((2,6))
sage: P.delete_vertices([0,1])
sage: P.relabel()

The attributes are properly updated too

sage: G = graphs.PathGraph(5)
sage: G.set_vertices({0: 'before', 1: 'delete', 2: 'after'})
sage: G.set_boundary([1,2,3])
sage: G.delete_vertex(1)
sage: G.relabel()
sage: G.get_vertices()
{0: 'before', 1: 'after', 2: None, 3: None}
sage: G.get_boundary()
[1, 2]
sage: G.get_pos()
{0: (0, 0), 1: (2, 0), 2: (3, 0), 3: (4, 0)}

Check that #12477 is fixed:

sage: g = Graph({1:[2,3]})
sage: rel = {1:'a', 2:'b'}
sage: g.relabel(rel)
sage: g.vertices()
[3, 'a', 'b']
sage: rel
{1: 'a', 2: 'b'}
remove_loops(vertices=None)

Removes loops on vertices in vertices. If vertices is None, removes all loops.

EXAMPLE

sage: G = Graph(4, loops=True)
sage: G.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: G.edges(labels=False)
[(0, 0), (1, 1), (2, 2), (2, 3), (3, 3)]
sage: G.remove_loops()
sage: G.edges(labels=False)
[(2, 3)]
sage: G.allows_loops()
True
sage: G.has_loops()
False

sage: D = DiGraph(4, loops=True)
sage: D.add_edges( [ (0,0), (1,1), (2,2), (3,3), (2,3) ] )
sage: D.edges(labels=False)
[(0, 0), (1, 1), (2, 2), (2, 3), (3, 3)]
sage: D.remove_loops()
sage: D.edges(labels=False)
[(2, 3)]
sage: D.allows_loops()
True
sage: D.has_loops()
False
remove_multiple_edges()

Removes all multiple edges, retaining one edge for each.

EXAMPLES:

sage: G = Graph(multiedges=True, sparse=True)
sage: G.add_edges( [ (0,1), (0,1), (0,1), (0,1), (1,2) ] )
sage: G.edges(labels=False)
[(0, 1), (0, 1), (0, 1), (0, 1), (1, 2)]
sage: G.remove_multiple_edges()
sage: G.edges(labels=False)
[(0, 1), (1, 2)]
sage: D = DiGraph(multiedges=True, sparse=True)
sage: D.add_edges( [ (0,1,1), (0,1,2), (0,1,3), (0,1,4), (1,2) ] )
sage: D.edges(labels=False)
[(0, 1), (0, 1), (0, 1), (0, 1), (1, 2)]
sage: D.remove_multiple_edges()
sage: D.edges(labels=False)
[(0, 1), (1, 2)]
set_boundary(boundary)

Sets the boundary of the (di)graph.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.set_boundary([0,1,2,3,4])
sage: G.get_boundary()
[0, 1, 2, 3, 4]
sage: G.set_boundary((1..4))
sage: G.get_boundary()
[1, 2, 3, 4]
set_edge_label(u, v, l)

Set the edge label of a given edge.

Note

There can be only one edge from u to v for this to make sense. Otherwise, an error is raised.

INPUT:

  • u, v - the vertices (and direction if digraph) of the edge
  • l - the new label

EXAMPLES:

sage: SD = DiGraph( { 1:[18,2], 2:[5,3], 3:[4,6], 4:[7,2], 5:[4], 6:[13,12], 7:[18,8,10], 8:[6,9,10], 9:[6], 10:[11,13], 11:[12], 12:[13], 13:[17,14], 14:[16,15], 15:[2], 16:[13], 17:[15,13], 18:[13] }, sparse=True)
sage: SD.set_edge_label(1, 18, 'discrete')
sage: SD.set_edge_label(4, 7, 'discrete')
sage: SD.set_edge_label(2, 5, 'h = 0')
sage: SD.set_edge_label(7, 18, 'h = 0')
sage: SD.set_edge_label(7, 10, 'aut')
sage: SD.set_edge_label(8, 10, 'aut')
sage: SD.set_edge_label(8, 9, 'label')
sage: SD.set_edge_label(8, 6, 'no label')
sage: SD.set_edge_label(13, 17, 'k > h')
sage: SD.set_edge_label(13, 14, 'k = h')
sage: SD.set_edge_label(17, 15, 'v_k finite')
sage: SD.set_edge_label(14, 15, 'v_k m.c.r.')
sage: posn = {1:[ 3,-3],  2:[0,2],  3:[0, 13],  4:[3,9],  5:[3,3],  6:[16, 13], 7:[6,1],  8:[6,6],  9:[6,11], 10:[9,1], 11:[10,6], 12:[13,6], 13:[16,2], 14:[10,-6], 15:[0,-10], 16:[14,-6], 17:[16,-10], 18:[6,-4]}
sage: SD.plot(pos=posn, vertex_size=400, vertex_colors={'#FFFFFF':range(1,19)}, edge_labels=True).show() # long time
sage: G = graphs.HeawoodGraph()
sage: for u,v,l in G.edges():
...    G.set_edge_label(u,v,'(' + str(u) + ',' + str(v) + ')')
sage: G.edges()
    [(0, 1, '(0,1)'),
     (0, 5, '(0,5)'),
     (0, 13, '(0,13)'),
     ...
     (11, 12, '(11,12)'),
     (12, 13, '(12,13)')]
sage: g = Graph({0: [0,1,1,2]}, loops=True, multiedges=True, sparse=True)
sage: g.set_edge_label(0,0,'test')
sage: g.edges()
[(0, 0, 'test'), (0, 1, None), (0, 1, None), (0, 2, None)]
sage: g.add_edge(0,0,'test2')
sage: g.set_edge_label(0,0,'test3')
Traceback (most recent call last):
...
RuntimeError: Cannot set edge label, since there are multiple edges from 0 to 0.
sage: dg = DiGraph({0 : [1], 1 : [0]}, sparse=True)
sage: dg.set_edge_label(0,1,5)
sage: dg.set_edge_label(1,0,9)
sage: dg.outgoing_edges(1)
[(1, 0, 9)]
sage: dg.incoming_edges(1)
[(0, 1, 5)]
sage: dg.outgoing_edges(0)
[(0, 1, 5)]
sage: dg.incoming_edges(0)
[(1, 0, 9)]
sage: G = Graph({0:{1:1}}, sparse=True)
sage: G.num_edges()
1
sage: G.set_edge_label(0,1,1)
sage: G.num_edges()
1
set_embedding(embedding)

Sets a combinatorial embedding dictionary to _embedding attribute.

Dictionary is organized with vertex labels as keys and a list of each vertex’s neighbors in clockwise order.

Dictionary is error-checked for validity.

INPUT:

  • embedding - a dictionary

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.set_embedding({0: [1, 5, 4], 1: [0, 2, 6], 2: [1, 3, 7], 3: [8, 2, 4], 4: [0, 9, 3], 5: [0, 8, 7], 6: [8, 1, 9], 7: [9, 2, 5], 8: [3, 5, 6], 9: [4, 6, 7]})
sage: G.set_embedding({'s': [1, 5, 4], 1: [0, 2, 6], 2: [1, 3, 7], 3: [8, 2, 4], 4: [0, 9, 3], 5: [0, 8, 7], 6: [8, 1, 9], 7: [9, 2, 5], 8: [3, 5, 6], 9: [4, 6, 7]})
Traceback (most recent call last):
...
ValueError: embedding is not valid for Petersen graph
set_latex_options(**kwds)

Sets multiple options for rendering a graph with LaTeX.

INPUTS:

  • kwds - any number of option/value pairs to set many graph latex options at once (a variable number, in any order). Existing values are overwritten, new values are added. Existing values can be cleared by setting the value to None. Possible options are documented at sage.graphs.graph_latex.GraphLatex.set_option().

This method is a convenience for setting the options of a graph directly on an instance of the graph. For a full explanation of how to use LaTeX to render graphs, see the introduction to the graph_latex module.

EXAMPLES:

sage: g = graphs.PetersenGraph()
sage: g.set_latex_options(tkz_style = 'Welsh')
sage: opts = g.latex_options()
sage: opts.get_option('tkz_style')
'Welsh'
set_planar_positions(test=False, **layout_options)

Compute a planar layout for self using Schnyder’s algorithm, and save it as default layout.

EXAMPLES:

sage: g = graphs.CycleGraph(7)
sage: g.set_planar_positions(test=True)
True

This method is deprecated since Sage-4.4.1.alpha2. Please use instead:

sage: g.layout(layout = “planar”, save_pos = True) {0: [1, 1], 1: [2, 2], 2: [3, 2], 3: [1, 4], 4: [5, 1], 5: [0, 5], 6: [1, 0]}
set_pos(pos, dim=2)

Sets the position dictionary, a dictionary specifying the coordinates of each vertex.

EXAMPLES: Note that set_pos will allow you to do ridiculous things, which will not blow up until plotting:

sage: G = graphs.PetersenGraph()
sage: G.get_pos()
{0: (..., ...),
 ...
 9: (..., ...)}
sage: G.set_pos('spam')
sage: P = G.plot()
Traceback (most recent call last):
...
TypeError: string indices must be integers, not str
set_vertex(vertex, object)

Associate an arbitrary object with a vertex.

INPUT:

  • vertex - which vertex
  • object - object to associate to vertex

EXAMPLES:

sage: T = graphs.TetrahedralGraph()
sage: T.vertices()
[0, 1, 2, 3]
sage: T.set_vertex(1, graphs.FlowerSnark())
sage: T.get_vertex(1)
Flower Snark: Graph on 20 vertices
set_vertices(vertex_dict)

Associate arbitrary objects with each vertex, via an association dictionary.

INPUT:

  • vertex_dict - the association dictionary

EXAMPLES:

sage: d = {0 : graphs.DodecahedralGraph(), 1 : graphs.FlowerSnark(), 2 : graphs.MoebiusKantorGraph(), 3 : graphs.PetersenGraph() }
sage: d[2]
Moebius-Kantor Graph: Graph on 16 vertices
sage: T = graphs.TetrahedralGraph()
sage: T.vertices()
[0, 1, 2, 3]
sage: T.set_vertices(d)
sage: T.get_vertex(1)
Flower Snark: Graph on 20 vertices
shortest_path(u, v, by_weight=False, bidirectional=True)

Returns a list of vertices representing some shortest path from u to v: if there is no path from u to v, the list is empty.

INPUT:

  • by_weight - if False, uses a breadth first search. If True, takes edge weightings into account, using Dijkstra’s algorithm.
  • bidirectional - if True, the algorithm will expand vertices from u and v at the same time, making two spheres of half the usual radius. This generally doubles the speed (consider the total volume in each case).

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: D.shortest_path(4, 9)
[4, 17, 16, 12, 13, 9]
sage: D.shortest_path(5, 5)
[5]
sage: D.delete_edges(D.edges_incident(13))
sage: D.shortest_path(13, 4)
[]
sage: G = Graph( { 0: [1], 1: [2], 2: [3], 3: [4], 4: [0] })
sage: G.plot(edge_labels=True).show() # long time
sage: G.shortest_path(0, 3)
[0, 4, 3]
sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse = True)
sage: G.shortest_path(0, 3, by_weight=True)
[0, 1, 2, 3]
shortest_path_all_pairs(by_weight=False, default_weight=1, algorithm='auto')

Computes a shortest path between each pair of vertices.

INPUT:

  • by_weight - Whether to use the labels defined over the edges as

    weights. If False (default), the distance between u and v is the minimum number of edges of a path from u to v.

  • default_weight - (defaults to 1) The default weight to assign

    edges that don’t have a weight (i.e., a label).

    Implies by_weight == True.

  • algorithm – four options :

    • "BFS" – the computation is done through a BFS centered on each vertex successively. Only implemented when default_weight = 1 and by_weight = False.

    • "Floyd-Warshall-Cython" – through the Cython implementation of the Floyd-Warshall algorithm.

    • "Floyd-Warshall-Python" – through the Python implementation of the Floyd-Warshall algorithm.

    • "auto" – use the fastest algorithm depending on the input ("BFS" if possible, and "Floyd-Warshall-Python" otherwise)

      This is the default value.

OUTPUT:

A tuple (dist, pred). They are both dicts of dicts. The first indicates the length dist[u][v] of the shortest weighted path from u to v. The second is a compact representation of all the paths- it indicates the predecessor pred[u][v] of v in the shortest path from u to v.

Note

Three different implementations are actually available through this method :

  • BFS (Cython)
  • Floyd-Warshall (Cython)
  • Floyd-Warshall (Python)

The BFS algorithm is the fastest of the three, then comes the Cython implementation of Floyd-Warshall, and last the Python implementation. The first two implementations, however, only compute distances based on the topological distance (each edge is of weight 1, or equivalently the length of a path is its number of edges). Besides, they do not deal with graphs larger than 65536 vertices (which already represents 16GB of ram).

Note

There is a Cython version of this method that is usually much faster for large graphs, as most of the time is actually spent building the final double dictionary. Everything on the subject is to be found in the distances_all_pairs module.

EXAMPLES:

sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse=True )
sage: G.plot(edge_labels=True).show() # long time
sage: dist, pred = G.shortest_path_all_pairs(by_weight = True)
sage: dist
{0: {0: 0, 1: 1, 2: 2, 3: 3, 4: 2}, 1: {0: 1, 1: 0, 2: 1, 3: 2, 4: 3}, 2: {0: 2, 1: 1, 2: 0, 3: 1, 4: 3}, 3: {0: 3, 1: 2, 2: 1, 3: 0, 4: 2}, 4: {0: 2, 1: 3, 2: 3, 3: 2, 4: 0}}
sage: pred
{0: {0: None, 1: 0, 2: 1, 3: 2, 4: 0}, 1: {0: 1, 1: None, 2: 1, 3: 2, 4: 0}, 2: {0: 1, 1: 2, 2: None, 3: 2, 4: 3}, 3: {0: 1, 1: 2, 2: 3, 3: None, 4: 3}, 4: {0: 4, 1: 0, 2: 3, 3: 4, 4: None}}
sage: pred[0]
{0: None, 1: 0, 2: 1, 3: 2, 4: 0}

So for example the shortest weighted path from 0 to 3 is obtained as follows. The predecessor of 3 is pred[0][3] == 2, the predecessor of 2 is pred[0][2] == 1, and the predecessor of 1 is pred[0][1] == 0.

sage: G = Graph( { 0: {1:None}, 1: {2:None}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse=True )
sage: G.shortest_path_all_pairs()
({0: {0: 0, 1: 1, 2: 2, 3: 2, 4: 1},
1: {0: 1, 1: 0, 2: 1, 3: 2, 4: 2},
2: {0: 2, 1: 1, 2: 0, 3: 1, 4: 2},
3: {0: 2, 1: 2, 2: 1, 3: 0, 4: 1},
4: {0: 1, 1: 2, 2: 2, 3: 1, 4: 0}},
{0: {0: None, 1: 0, 2: 1, 3: 4, 4: 0},
1: {0: 1, 1: None, 2: 1, 3: 2, 4: 0},
2: {0: 1, 1: 2, 2: None, 3: 2, 4: 3},
3: {0: 4, 1: 2, 2: 3, 3: None, 4: 3},
4: {0: 4, 1: 0, 2: 3, 3: 4, 4: None}})
sage: G.shortest_path_all_pairs(by_weight = True)
({0: {0: 0, 1: 1, 2: 2, 3: 3, 4: 2},
1: {0: 1, 1: 0, 2: 1, 3: 2, 4: 3},
2: {0: 2, 1: 1, 2: 0, 3: 1, 4: 3},
3: {0: 3, 1: 2, 2: 1, 3: 0, 4: 2},
4: {0: 2, 1: 3, 2: 3, 3: 2, 4: 0}},
{0: {0: None, 1: 0, 2: 1, 3: 2, 4: 0},
1: {0: 1, 1: None, 2: 1, 3: 2, 4: 0},
2: {0: 1, 1: 2, 2: None, 3: 2, 4: 3},
3: {0: 1, 1: 2, 2: 3, 3: None, 4: 3},
4: {0: 4, 1: 0, 2: 3, 3: 4, 4: None}})
sage: G.shortest_path_all_pairs(default_weight=200)
({0: {0: 0, 1: 200, 2: 5, 3: 4, 4: 2},
1: {0: 200, 1: 0, 2: 200, 3: 201, 4: 202},
2: {0: 5, 1: 200, 2: 0, 3: 1, 4: 3},
3: {0: 4, 1: 201, 2: 1, 3: 0, 4: 2},
4: {0: 2, 1: 202, 2: 3, 3: 2, 4: 0}},
{0: {0: None, 1: 0, 2: 3, 3: 4, 4: 0},
1: {0: 1, 1: None, 2: 1, 3: 2, 4: 0},
2: {0: 4, 1: 2, 2: None, 3: 2, 4: 3},
3: {0: 4, 1: 2, 2: 3, 3: None, 4: 3},
4: {0: 4, 1: 0, 2: 3, 3: 4, 4: None}})

Checking the distances are equal regardless of the algorithm used:

sage: g = graphs.Grid2dGraph(5,5)
sage: d1, _ = g.shortest_path_all_pairs(algorithm="BFS")
sage: d2, _ = g.shortest_path_all_pairs(algorithm="Floyd-Warshall-Cython")
sage: d3, _ = g.shortest_path_all_pairs(algorithm="Floyd-Warshall-Python")
sage: d1 == d2 == d3
True

Checking a random path is valid

sage: dist, path = g.shortest_path_all_pairs(algorithm="BFS")
sage: u,v = g.random_vertex(), g.random_vertex()
sage: p = [v]
sage: while p[0] != None:
...     p.insert(0,path[u][p[0]])
sage: len(p) == dist[u][v] + 2
True

TESTS:

Wrong name for algorithm:

sage: g.shortest_path_all_pairs(algorithm="Bob")
Traceback (most recent call last):
...
ValueError: The algorithm keyword can only be set to "auto", "BFS", "Floyd-Warshall-Python" or "Floyd-Warshall-Cython"
shortest_path_length(u, v, by_weight=False, bidirectional=True, weight_sum=None)

Returns the minimal length of paths from u to v.

If there is no path from u to v, returns Infinity.

INPUT:

  • by_weight - if False, uses a breadth first search. If True, takes edge weightings into account, using Dijkstra’s algorithm.
  • bidirectional - if True, the algorithm will expand vertices from u and v at the same time, making two spheres of half the usual radius. This generally doubles the speed (consider the total volume in each case).
  • weight_sum - if False, returns the number of edges in the path. If True, returns the sum of the weights of these edges. Default behavior is to have the same value as by_weight.

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: D.shortest_path_length(4, 9)
5
sage: D.shortest_path_length(5, 5)
0
sage: D.delete_edges(D.edges_incident(13))
sage: D.shortest_path_length(13, 4)
+Infinity
sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse = True)
sage: G.plot(edge_labels=True).show() # long time
sage: G.shortest_path_length(0, 3)
2
sage: G.shortest_path_length(0, 3, by_weight=True)
3
shortest_path_lengths(u, by_weight=False, weight_sums=None)

Returns a dictionary of shortest path lengths keyed by targets that are connected by a path from u.

INPUT:

  • by_weight - if False, uses a breadth first search. If True, takes edge weightings into account, using Dijkstra’s algorithm.

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: D.shortest_path_lengths(0)
{0: 0, 1: 1, 2: 2, 3: 2, 4: 3, 5: 4, 6: 3, 7: 3, 8: 2, 9: 2, 10: 1, 11: 2, 12: 3, 13: 3, 14: 4, 15: 5, 16: 4, 17: 3, 18: 2, 19: 1}
sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse=True )
sage: G.plot(edge_labels=True).show() # long time
sage: G.shortest_path_lengths(0, by_weight=True)
{0: 0, 1: 1, 2: 2, 3: 3, 4: 2}
shortest_paths(u, by_weight=False, cutoff=None)

Returns a dictionary associating to each vertex v a shortest path from u to v, if it exists.

INPUT:

  • by_weight - if False, uses a breadth first search. If True, uses Dijkstra’s algorithm to find the shortest paths by weight.

  • cutoff - integer depth to stop search.

    (ignored if by_weight == True)

EXAMPLES:

sage: D = graphs.DodecahedralGraph()
sage: D.shortest_paths(0)
{0: [0], 1: [0, 1], 2: [0, 1, 2], 3: [0, 19, 3], 4: [0, 19, 3, 4], 5: [0, 1, 2, 6, 5], 6: [0, 1, 2, 6], 7: [0, 1, 8, 7], 8: [0, 1, 8], 9: [0, 10, 9], 10: [0, 10], 11: [0, 10, 11], 12: [0, 10, 11, 12], 13: [0, 10, 9, 13], 14: [0, 1, 8, 7, 14], 15: [0, 19, 18, 17, 16, 15], 16: [0, 19, 18, 17, 16], 17: [0, 19, 18, 17], 18: [0, 19, 18], 19: [0, 19]}

All these paths are obviously induced graphs:

sage: all([D.subgraph(p).is_isomorphic(graphs.PathGraph(len(p)) )for p in D.shortest_paths(0).values()])
True
sage: D.shortest_paths(0, cutoff=2)
{0: [0], 1: [0, 1], 2: [0, 1, 2], 3: [0, 19, 3], 8: [0, 1, 8], 9: [0, 10, 9], 10: [0, 10], 11: [0, 10, 11], 18: [0, 19, 18], 19: [0, 19]}
sage: G = Graph( { 0: {1: 1}, 1: {2: 1}, 2: {3: 1}, 3: {4: 2}, 4: {0: 2} }, sparse=True)
sage: G.plot(edge_labels=True).show() # long time
sage: G.shortest_paths(0, by_weight=True)
{0: [0], 1: [0, 1], 2: [0, 1, 2], 3: [0, 1, 2, 3], 4: [0, 4]}
show(**kwds)

Shows the (di)graph.

INPUT:

This method accepts any option understood by plot() (graph-specific) or by sage.plot.graphics.Graphics.show().

Note

See the documentation of the sage.graphs.graph_plot module for information on default arguments of this method.

EXAMPLES:

sage: C = graphs.CubeGraph(8)
sage: P = C.plot(vertex_labels=False, vertex_size=0, graph_border=True)
sage: P.show()  # long time (3s on sage.math, 2011)
show3d(bgcolor=(1, 1, 1), vertex_colors=None, vertex_size=0.06, edge_colors=None, edge_size=0.02, edge_size2=0.0325, pos3d=None, color_by_label=False, engine='jmol', **kwds)

Plots the graph using Tachyon, and shows the resulting plot.

INPUT:

  • bgcolor - rgb tuple (default: (1,1,1))
  • vertex_size - float (default: 0.06)
  • vertex_colors - optional dictionary to specify vertex colors: each key is a color recognizable by tachyon (rgb tuple (default: (1,0,0))), and each corresponding entry is a list of vertices. If a vertex is not listed, it looks invisible on the resulting plot (it doesn’t get drawn).
  • edge_colors - a dictionary specifying edge colors: each key is a color recognized by tachyon ( default: (0,0,0) ), and each entry is a list of edges.
  • edge_size - float (default: 0.02)
  • edge_size2 - float (default: 0.0325), used for Tachyon sleeves
  • pos3d - a position dictionary for the vertices
  • iterations - how many iterations of the spring layout algorithm to go through, if applicable
  • engine - which renderer to use. Options:
  • 'jmol' - default ‘tachyon’
  • xres - resolution
  • yres - resolution
  • **kwds - passed on to the Tachyon command

EXAMPLES:

sage: G = graphs.CubeGraph(5)
sage: G.show3d(iterations=500, edge_size=None, vertex_size=0.04) # long time

We plot a fairly complicated Cayley graph:

sage: A5 = AlternatingGroup(5); A5
Alternating group of order 5!/2 as a permutation group
sage: G = A5.cayley_graph()
sage: G.show3d(vertex_size=0.03, edge_size=0.01, edge_size2=0.02, vertex_colors={(1,1,1):G.vertices()}, bgcolor=(0,0,0), color_by_label=True, iterations=200) # long time

Some Tachyon examples:

sage: D = graphs.DodecahedralGraph()
sage: D.show3d(engine='tachyon') # long time
sage: G = graphs.PetersenGraph()
sage: G.show3d(engine='tachyon', vertex_colors={(0,0,1):G.vertices()}) # long time
sage: C = graphs.CubeGraph(4)
sage: C.show3d(engine='tachyon', edge_colors={(0,1,0):C.edges()}, vertex_colors={(1,1,1):C.vertices()}, bgcolor=(0,0,0)) # long time
sage: K = graphs.CompleteGraph(3)
sage: K.show3d(engine='tachyon', edge_colors={(1,0,0):[(0,1,None)], (0,1,0):[(0,2,None)], (0,0,1):[(1,2,None)]}) # long time
size()

Returns the number of edges.

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.size()
15
spanning_trees_count(root_vertex=None)

Returns the number of spanning trees in a graph.

In the case of a digraph, counts the number of spanning out-trees rooted in root_vertex. Default is to set first vertex as root.

This computation uses Kirchhoff’s Matrix Tree Theorem [1] to calculate the number of spanning trees. For complete graphs on n vertices the result can also be reached using Cayley’s formula: the number of spanning trees are n^(n-2).

For digraphs, the augmented Kirchhoff Matrix as defined in [2] is used for calculations. Here the result is the number of out-trees rooted at a specific vertex.

INPUT:

  • root_vertex – integer (default: the first vertex) This is the vertex that will be used as root for all spanning out-trees if the graph is a directed graph. This argument is ignored if the graph is not a digraph.

REFERENCES:

AUTHORS:

  • Anders Jonsson (2009-10-10)

EXAMPLES:

sage: G = graphs.PetersenGraph()
sage: G.spanning_trees_count()
2000
sage: n = 11
sage: G = graphs.CompleteGraph(n)
sage: ST = G.spanning_trees_count()
sage: ST == n^(n-2)
True
sage: M=matrix(3,3,[0,1,0,0,0,1,1,1,0])
sage: D=DiGraph(M)
sage: D.spanning_trees_count()
1
sage: D.spanning_trees_count(0)
1
sage: D.spanning_trees_count(2)
2
spectrum(laplacian=False)

Returns a list of the eigenvalues of the adjacency matrix.

INPUT:

OUTPUT:

A list of the eigenvalues, including multiplicities, sorted with the largest eigenvalue first.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.spectrum()
[3, 1, 1, 1, 1, 1, -2, -2, -2, -2]
sage: P.spectrum(laplacian=True)
[5, 5, 5, 5, 2, 2, 2, 2, 2, 0]
sage: D = P.to_directed()
sage: D.delete_edge(7,9)
sage: D.spectrum()
[2.9032119259..., 1, 1, 1, 1, 0.8060634335..., -1.7092753594..., -2, -2, -2]
sage: C = graphs.CycleGraph(8)
sage: C.spectrum()
[2, 1.4142135623..., 1.4142135623..., 0, 0, -1.4142135623..., -1.4142135623..., -2]

A digraph may have complex eigenvalues. Previously, the complex parts of graph eigenvalues were being dropped. For a 3-cycle, we have:

sage: T = DiGraph({0:[1], 1:[2], 2:[0]})
sage: T.spectrum()
[1, -0.5000000000... + 0.8660254037...*I, -0.5000000000... - 0.8660254037...*I]

TESTS:

The Laplacian matrix of a graph is the negative of the adjacency matrix with the degree of each vertex on the diagonal. So for a regular graph, if \delta is an eigenvalue of a regular graph of degree r, then r-\delta will be an eigenvalue of the Laplacian. The Hoffman-Singleton graph is regular of degree 7, so the following will test both the Laplacian construction and the computation of eigenvalues.

sage: H = graphs.HoffmanSingletonGraph()
sage: evals = H.spectrum()
sage: lap = map(lambda x : 7 - x, evals)
sage: lap.sort(reverse=True)
sage: lap == H.spectrum(laplacian=True)
True
steiner_tree(vertices, weighted=False, solver=None, verbose=0)

Returns a tree of minimum weight connecting the given set of vertices.

Definition :

Computing a minimum spanning tree in a graph can be done in n
\log(n) time (and in linear time if all weights are equal) where n = V + E. On the other hand, if one is given a large (possibly weighted) graph and a subset of its vertices, it is NP-Hard to find a tree of minimum weight connecting the given set of vertices, which is then called a Steiner Tree.

Wikipedia article on Steiner Trees.

INPUT:

  • vertices – the vertices to be connected by the Steiner Tree.
  • weighted (boolean) – Whether to consider the graph as weighted, and use each edge’s label as a weight, considering None as a weight of 1. If weighted=False (default) all edges are considered to have a weight of 1.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

Note

  • This problem being defined on undirected graphs, the orientation is not considered if the current graph is actually a digraph.
  • The graph is assumed not to have multiple edges.

ALGORITHM:

Solved through Linear Programming.

COMPLEXITY:

NP-Hard.

Note that this algorithm first checks whether the given set of vertices induces a connected graph, returning one of its spanning trees if weighted is set to False, and thus answering very quickly in some cases

EXAMPLES:

The Steiner Tree of the first 5 vertices in a random graph is, of course, always a tree

sage: g = graphs.RandomGNP(30,.5)
sage: st = g.steiner_tree(g.vertices()[:5])
sage: st.is_tree()
True

And all the 5 vertices are contained in this tree

sage: all([v in st for v in g.vertices()[:5] ])
True

An exception is raised when the problem is impossible, i.e. if the given vertices are not all included in the same connected component

sage: g = 2 * graphs.PetersenGraph()
sage: st = g.steiner_tree([5,15])
Traceback (most recent call last):
...
ValueError: The given vertices do not all belong to the same connected component. This problem has no solution !
strong_product(other)

Returns the strong product of self and other.

The strong product of G and H is the graph L with vertex set V(L)=V(G)\times V(H), and ((u,v), (w,x)) is an edge of L iff either :

  • (u, w) is an edge of G and v = x, or
  • (v, x) is an edge of H and u = w, or
  • (u, w) is an edge of G and (v, x) is an edge of H.

In other words, the edges of the strong product is the union of the edges of the tensor and Cartesian products.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: C = graphs.CycleGraph(5)
sage: S = C.strong_product(Z); S
Graph on 10 vertices
sage: S.plot() # long time
sage: D = graphs.DodecahedralGraph()
sage: P = graphs.PetersenGraph()
sage: S = D.strong_product(P); S
Graph on 200 vertices
sage: S.plot() # long time

TESTS:

Strong product of graphs is commutative:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.strong_product(H)
sage: T.is_isomorphic( H.strong_product(G) )
True

Strong product of digraphs is commutative:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.strong_product(J)
sage: T.is_isomorphic( J.strong_product(I) )
True

Counting the edges (see trac ticket #13699):

sage: g = graphs.RandomGNP(5,.5)
sage: gn,gm = g.order(), g.size()
sage: h = graphs.RandomGNP(5,.5)
sage: hn,hm = h.order(), h.size()
sage: product_size = g.strong_product(h).size()
sage: expected = gm*hn + hm*gn + 2*gm*hm
sage: if product_size != expected:
...       print "Something is really wrong here...", product_size, "!=", expected
subdivide_edge(*args)

Subdivides an edge k times.

INPUT:

The following forms are all accepted to subdivide 8 times the edge between vertices 1 and 2 labeled with "my_label".

  • G.subdivide_edge( 1, 2, 8 )
  • G.subdivide_edge( (1, 2), 8 )
  • G.subdivide_edge( (1, 2, "my_label"), 8 )

Note

  • If the given edge is labelled with l, all the edges created by the subdivision will have the same label.
  • If no label is given, the label used will be the one returned by the method edge_label() on the pair u,v

EXAMPLE:

Subdividing 5 times an edge in a path of length 3 makes it a path of length 8:

sage: g = graphs.PathGraph(3)
sage: edge = g.edges()[0]
sage: g.subdivide_edge(edge, 5)
sage: g.is_isomorphic(graphs.PathGraph(8))
True

Subdividing a labelled edge in two ways

sage: g = Graph()
sage: g.add_edge(0,1,"label1")
sage: g.add_edge(1,2,"label2")
sage: print sorted(g.edges())
[(0, 1, 'label1'), (1, 2, 'label2')]

Specifying the label:

sage: g.subdivide_edge(0,1,"label1", 3)
sage: print sorted(g.edges())
[(0, 3, 'label1'), (1, 2, 'label2'), (1, 5, 'label1'), (3, 4, 'label1'), (4, 5, 'label1')]

The lazy way:

sage: g.subdivide_edge(1,2,"label2", 5)
sage: print sorted(g.edges())
[(0, 3, 'label1'), (1, 5, 'label1'), (1, 6, 'label2'), (2, 10, 'label2'), (3, 4, 'label1'), (4, 5, 'label1'), (6, 7, 'label2'), (7, 8, 'label2'), (8, 9, 'label2'), (9, 10, 'label2')]

If too many arguments are given, an exception is raised

sage: g.subdivide_edge(0,1,1,1,1,1,1,1,1,1,1)
Traceback (most recent call last):
...
ValueError: This method takes at most 4 arguments !

The same goes when the given edge does not exist:

sage: g.subdivide_edge(0,1,"fake_label",5)
Traceback (most recent call last):
...
ValueError: The given edge does not exist.

See also

subdivide_edges(edges, k)

Subdivides k times edges from an iterable container.

For more information on the behaviour of this method, please refer to the documentation of subdivide_edge().

INPUT:

  • edges – a list of edges
  • k (integer) – common length of the subdivisions

Note

If a given edge is labelled with l, all the edges created by its subdivision will have the same label.

EXAMPLE:

If we are given the disjoint union of several paths:

sage: paths = [2,5,9]
sage: paths = map(graphs.PathGraph, paths)
sage: g = Graph()
sage: for P in paths:
...     g = g + P

... subdividing edges in each of them will only change their lengths:

sage: edges = [P.edges()[0] for P in g.connected_components_subgraphs()]
sage: k = 6
sage: g.subdivide_edges(edges, k)

Let us check this by creating the graph we expect to have built through subdivision:

sage: paths2 = [2+k, 5+k, 9+k]
sage: paths2 = map(graphs.PathGraph, paths2)
sage: g2 = Graph()
sage: for P in paths2:
...     g2 = g2 + P
sage: g.is_isomorphic(g2)
True

See also

subgraph(vertices=None, edges=None, inplace=False, vertex_property=None, edge_property=None, algorithm=None)

Returns the subgraph containing the given vertices and edges.

If either vertices or edges are not specified, they are assumed to be all vertices or edges. If edges are not specified, returns the subgraph induced by the vertices.

INPUT:

  • inplace - Using inplace is True will simply delete the extra vertices and edges from the current graph. This will modify the graph.
  • vertices - Vertices can be a single vertex or an iterable container of vertices, e.g. a list, set, graph, file or numeric array. If not passed, defaults to the entire graph.
  • edges - As with vertices, edges can be a single edge or an iterable container of edges (e.g., a list, set, file, numeric array, etc.). If not edges are not specified, then all edges are assumed and the returned graph is an induced subgraph. In the case of multiple edges, specifying an edge as (u,v) means to keep all edges (u,v), regardless of the label.
  • vertex_property - If specified, this is expected to be a function on vertices, which is intersected with the vertices specified, if any are.
  • edge_property - If specified, this is expected to be a function on edges, which is intersected with the edges specified, if any are.
  • algorithm - If algorithm=delete or inplace=True, then the graph is constructed by deleting edges and vertices. If add, then the graph is constructed by building a new graph from the appropriate vertices and edges. If not specified, then the algorithm is chosen based on the number of vertices in the subgraph.

EXAMPLES:

sage: G = graphs.CompleteGraph(9)
sage: H = G.subgraph([0,1,2]); H
Subgraph of (Complete graph): Graph on 3 vertices
sage: G
Complete graph: Graph on 9 vertices
sage: J = G.subgraph(edges=[(0,1)])
sage: J.edges(labels=False)
[(0, 1)]
sage: J.vertices()==G.vertices()
True
sage: G.subgraph([0,1,2], inplace=True); G
Subgraph of (Complete graph): Graph on 3 vertices
sage: G.subgraph()==G
True
sage: D = graphs.CompleteGraph(9).to_directed()
sage: H = D.subgraph([0,1,2]); H
Subgraph of (Complete graph): Digraph on 3 vertices
sage: H = D.subgraph(edges=[(0,1), (0,2)])
sage: H.edges(labels=False)
[(0, 1), (0, 2)]
sage: H.vertices()==D.vertices()
True
sage: D
Complete graph: Digraph on 9 vertices
sage: D.subgraph([0,1,2], inplace=True); D
Subgraph of (Complete graph): Digraph on 3 vertices
sage: D.subgraph()==D
True

A more complicated example involving multiple edges and labels.

sage: G = Graph(multiedges=True, sparse=True)
sage: G.add_edges([(0,1,'a'), (0,1,'b'), (1,0,'c'), (0,2,'d'), (0,2,'e'), (2,0,'f'), (1,2,'g')])
sage: G.subgraph(edges=[(0,1), (0,2,'d'), (0,2,'not in graph')]).edges()
[(0, 1, 'a'), (0, 1, 'b'), (0, 1, 'c'), (0, 2, 'd')]
sage: J = G.subgraph(vertices=[0,1], edges=[(0,1,'a'), (0,2,'c')])
sage: J.edges()
[(0, 1, 'a')]
sage: J.vertices()
[0, 1]
sage: G.subgraph(vertices=G.vertices())==G
True
sage: D = DiGraph(multiedges=True, sparse=True)
sage: D.add_edges([(0,1,'a'), (0,1,'b'), (1,0,'c'), (0,2,'d'), (0,2,'e'), (2,0,'f'), (1,2,'g')])
sage: D.subgraph(edges=[(0,1), (0,2,'d'), (0,2,'not in graph')]).edges()
[(0, 1, 'a'), (0, 1, 'b'), (0, 2, 'd')]
sage: H = D.subgraph(vertices=[0,1], edges=[(0,1,'a'), (0,2,'c')])
sage: H.edges()
[(0, 1, 'a')]
sage: H.vertices()
[0, 1]

Using the property arguments:

sage: P = graphs.PetersenGraph()
sage: S = P.subgraph(vertex_property = lambda v : v%2 == 0)
sage: S.vertices()
[0, 2, 4, 6, 8]
sage: C = graphs.CubeGraph(2)
sage: S = C.subgraph(edge_property=(lambda e: e[0][0] == e[1][0]))
sage: C.edges()
[('00', '01', None), ('00', '10', None), ('01', '11', None), ('10', '11', None)]
sage: S.edges()
[('00', '01', None), ('10', '11', None)]

The algorithm is not specified, then a reasonable choice is made for speed.

sage: g=graphs.PathGraph(1000)
sage: g.subgraph(range(10)) # uses the 'add' algorithm
Subgraph of (Path Graph): Graph on 10 vertices

TESTS: The appropriate properties are preserved.

sage: g = graphs.PathGraph(10)
sage: g.is_planar(set_embedding=True)
True
sage: g.set_vertices(dict((v, 'v%d'%v) for v in g.vertices()))
sage: h = g.subgraph([3..5])
sage: h.get_pos().keys()
[3, 4, 5]
sage: h.get_vertices()
{3: 'v3', 4: 'v4', 5: 'v5'}

Returns a copy of G in self.

INPUT:

  • G – the graph whose copy we are looking for in self.
  • induced – boolean (default: False). Whether or not to search for an induced copy of G in self.

OUTPUT:

  • If induced=False, return a copy of G in this graph. Otherwise, return an induced copy of G in self. If G is the empty graph, return the empty graph since it is a subgraph of every graph. Now suppose G is not the empty graph. If there is no copy (induced or otherwise) of G in self, we return None.

Note

This method also works on digraphs.

See also

ALGORITHM:

Brute-force search.

EXAMPLES:

The Petersen graph contains the path graph P_5:

sage: g = graphs.PetersenGraph()
sage: h1 = g.subgraph_search(graphs.PathGraph(5)); h1
Subgraph of (Petersen graph): Graph on 5 vertices
sage: h1.vertices(); h1.edges(labels=False)
[0, 1, 2, 3, 4]
[(0, 1), (1, 2), (2, 3), (3, 4)]
sage: I1 = g.subgraph_search(graphs.PathGraph(5), induced=True); I1
Subgraph of (Petersen graph): Graph on 5 vertices
sage: I1.vertices(); I1.edges(labels=False)
[0, 1, 2, 3, 8]
[(0, 1), (1, 2), (2, 3), (3, 8)]

It also contains the claw K_{1,3}:

sage: h2 = g.subgraph_search(graphs.ClawGraph()); h2
Subgraph of (Petersen graph): Graph on 4 vertices
sage: h2.vertices(); h2.edges(labels=False)
[0, 1, 4, 5]
[(0, 1), (0, 4), (0, 5)]
sage: I2 = g.subgraph_search(graphs.ClawGraph(), induced=True); I2
Subgraph of (Petersen graph): Graph on 4 vertices
sage: I2.vertices(); I2.edges(labels=False)
[0, 1, 4, 5]
[(0, 1), (0, 4), (0, 5)]

Of course the induced copies are isomorphic to the graphs we were looking for:

sage: I1.is_isomorphic(graphs.PathGraph(5))
True
sage: I2.is_isomorphic(graphs.ClawGraph())
True

However, the Petersen graph does not contain a subgraph isomorphic to K_3:

sage: g.subgraph_search(graphs.CompleteGraph(3)) is None
True

Nor does it contain a nonempty induced subgraph isomorphic to P_6:

sage: g.subgraph_search(graphs.PathGraph(6), induced=True) is None
True

The empty graph is a subgraph of every graph:

sage: g.subgraph_search(graphs.EmptyGraph())
Graph on 0 vertices
sage: g.subgraph_search(graphs.EmptyGraph(), induced=True)
Graph on 0 vertices

The subgraph may just have edges missing:

sage: k3=graphs.CompleteGraph(3); p3=graphs.PathGraph(3)
sage: k3.relabel(list('abc'))
sage: s=k3.subgraph_search(p3)
sage: s.edges(labels=False)
[('a', 'b'), ('b', 'c')]

Of course, P_3 is not an induced subgraph of K_3, though:

sage: k3=graphs.CompleteGraph(3); p3=graphs.PathGraph(3)
sage: k3.relabel(list('abc'))
sage: k3.subgraph_search(p3, induced=True) is None
True

TESTS:

Inside of a small graph (trac ticket #13906):

sage: Graph(5).subgraph_search(Graph(1))
Graph on 1 vertex
subgraph_search_count(G, induced=False)

Returns the number of labelled occurences of G in self.

INPUT:

  • G – the graph whose copies we are looking for in self.
  • induced – boolean (default: False). Whether or not to count induced copies of G in self.

ALGORITHM:

Brute-force search.

Note

This method also works on digraphs.

See also

EXAMPLES:

Counting the number of paths P_5 in a PetersenGraph:

sage: g = graphs.PetersenGraph()
sage: g.subgraph_search_count(graphs.PathGraph(5))
240

Requiring these subgraphs be induced:

sage: g.subgraph_search_count(graphs.PathGraph(5), induced = True)
120

If we define the graph T_k (the transitive tournament on k vertices) as the graph on \{0, ..., k-1\} such that ij \in
T_k iif i<j, how many directed triangles can be found in T_5 ? The answer is of course 0

sage: T5 = DiGraph()
sage: T5.add_edges([(i,j) for i in xrange(5) for j in xrange(i+1, 5)])
sage: T5.subgraph_search_count(digraphs.Circuit(3))
0

If we count instead the number of T_3 in T_5, we expect the answer to be {5 \choose 3}:

sage: T3 = T5.subgraph([0,1,2])
sage: T5.subgraph_search_count(T3)
10
sage: binomial(5,3)
10

The empty graph is a subgraph of every graph:

sage: g.subgraph_search_count(graphs.EmptyGraph())
1

TESTS:

Inside of a small graph (trac ticket #13906):

sage: Graph(5).subgraph_search_count(Graph(1))
5
subgraph_search_iterator(G, induced=False)

Returns an iterator over the labelled copies of G in self.

INPUT:

  • G – the graph whose copies we are looking for in self.
  • induced – boolean (default: False). Whether or not to iterate over the induced copies of G in self.

ALGORITHM:

Brute-force search.

OUTPUT:

Iterator over the labelled copies of G in self, as lists. For each value (v_1, v_2, ..., v_k) returned, the first vertex of G is associated with v_1, the second with v_2, etc ...

Note

This method also works on digraphs.

See also

EXAMPLE:

Iterating through all the labelled P_3 of P_5:

sage: g = graphs.PathGraph(5)
sage: for p in g.subgraph_search_iterator(graphs.PathGraph(3)):
...      print p
[0, 1, 2]
[1, 2, 3]
[2, 1, 0]
[2, 3, 4]
[3, 2, 1]
[4, 3, 2]

TESTS:

Inside of a small graph (trac ticket #13906):

sage: list(Graph(5).subgraph_search_iterator(Graph(1)))
[Graph on 1 vertex, Graph on 1 vertex, Graph on 1 vertex, Graph on 1 vertex, Graph on 1 vertex]
szeged_index()

Returns the Szeged index of the graph.

For any uv\in E(G), let N_u(uv) = \{w\in G:d(u,w)<d(v,w)\}, n_u(uv)=|N_u(uv)|

The Szeged index of a graph is then defined as [1]: \sum_{uv \in E(G)}n_u(uv)\times n_v(uv)

EXAMPLE:

True for any connected graph [1]:

sage: g=graphs.PetersenGraph()
sage: g.wiener_index()<= g.szeged_index()
True

True for all trees [1]:

sage: g=Graph()
sage: g.add_edges(graphs.CubeGraph(5).min_spanning_tree())
sage: g.wiener_index() == g.szeged_index()
True

REFERENCE:

[1] Klavzar S., Rajapakse A., Gutman I. (1996). The Szeged and the Wiener index of graphs. Applied Mathematics Letters, 9 (5), pp. 45-49.

tensor_product(other)

Returns the tensor product of self and other.

The tensor product of G and H is the graph L with vertex set V(L) equal to the Cartesian product of the vertices V(G) and V(H), and ((u,v), (w,x)) is an edge iff - (u, w) is an edge of self, and - (v, x) is an edge of other.

The tensor product is also known as the categorical product and the kronecker product (refering to the kronecker matrix product). See Wikipedia article on the Kronecker product.

EXAMPLES:

sage: Z = graphs.CompleteGraph(2)
sage: C = graphs.CycleGraph(5)
sage: T = C.tensor_product(Z); T
Graph on 10 vertices
sage: T.size()
10
sage: T.plot() # long time
sage: D = graphs.DodecahedralGraph()
sage: P = graphs.PetersenGraph()
sage: T = D.tensor_product(P); T
Graph on 200 vertices
sage: T.size()
900
sage: T.plot() # long time

TESTS:

Tensor product of graphs:

sage: G = Graph([(0,1), (1,2)])
sage: H = Graph([('a','b')])
sage: T = G.tensor_product(H)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((0, 'b'), (1, 'a')), ((1, 'a'), (2, 'b')), ((1, 'b'), (2, 'a'))]
sage: T.is_isomorphic( H.tensor_product(G) )
True

Tensor product of digraphs:

sage: I = DiGraph([(0,1), (1,2)])
sage: J = DiGraph([('a','b')])
sage: T = I.tensor_product(J)
sage: T.edges(labels=None)
[((0, 'a'), (1, 'b')), ((1, 'a'), (2, 'b'))]
sage: T.is_isomorphic( J.tensor_product(I) )
True

The tensor product of two DeBruijn digraphs of same diameter is a DeBruijn digraph:

sage: B1 = digraphs.DeBruijn(2, 3)
sage: B2 = digraphs.DeBruijn(3, 3)
sage: T = B1.tensor_product( B2 )
sage: T.is_isomorphic( digraphs.DeBruijn( 2*3, 3) )
True
to_dictionary(edge_labels=False, multiple_edges=False)

Returns the graph as a dictionary.

INPUT:

  • edge_labels (boolean) – whether to include edge labels in the output.
  • multiple_edges (boolean) – whether to include multiple edges in the output.

OUTPUT:

The output depends on the input:

  • If edge_labels == False and multiple_edges == False, the output is a dictionary associating to each vertex the list of its neighbors.
  • If edge_labels == False and multiple_edges == True, the output is a dictionary the same as previously with one difference : the neighbors are listed with multiplicity.
  • If edge_labels == True and multiple_edges == False, the output is a dictionary associating to each vertex u [a dictionary associating to each vertex v incident to u the label of edge (u,v)].
  • If edge_labels == True and multiple_edges == True, the output is a dictionary associating to each vertex u [a dictionary associating to each vertex v incident to u [the list of labels of all edges between u and v]].

Note

When used on directed graphs, the explanations above can be understood by replacing the word “neigbours” by “out-neighbors”

EXAMPLES:

sage: g = graphs.PetersenGraph().to_dictionary()
sage: [(key, sorted(g[key])) for key in g]
[(0, [1, 4, 5]),
 (1, [0, 2, 6]),
 (2, [1, 3, 7]),
 (3, [2, 4, 8]),
 (4, [0, 3, 9]),
 (5, [0, 7, 8]),
 (6, [1, 8, 9]),
 (7, [2, 5, 9]),
 (8, [3, 5, 6]),
 (9, [4, 6, 7])]
sage: graphs.PetersenGraph().to_dictionary(multiple_edges=True)
{0: [1, 4, 5], 1: [0, 2, 6],
 2: [1, 3, 7], 3: [2, 4, 8],
 4: [0, 3, 9], 5: [0, 7, 8],
 6: [1, 8, 9], 7: [2, 5, 9],
 8: [3, 5, 6], 9: [4, 6, 7]}
sage: graphs.PetersenGraph().to_dictionary(edge_labels=True)
{0: {1: None, 4: None, 5: None},
 1: {0: None, 2: None, 6: None},
 2: {1: None, 3: None, 7: None},
 3: {8: None, 2: None, 4: None},
 4: {0: None, 9: None, 3: None},
 5: {0: None, 8: None, 7: None},
 6: {8: None, 1: None, 9: None},
 7: {9: None, 2: None, 5: None},
 8: {3: None, 5: None, 6: None},
 9: {4: None, 6: None, 7: None}}
sage: graphs.PetersenGraph().to_dictionary(edge_labels=True,multiple_edges=True)
{0: {1: [None], 4: [None], 5: [None]},
 1: {0: [None], 2: [None], 6: [None]},
 2: {1: [None], 3: [None], 7: [None]},
 3: {8: [None], 2: [None], 4: [None]},
 4: {0: [None], 9: [None], 3: [None]},
 5: {0: [None], 8: [None], 7: [None]},
 6: {8: [None], 1: [None], 9: [None]},
 7: {9: [None], 2: [None], 5: [None]},
 8: {3: [None], 5: [None], 6: [None]},
 9: {4: [None], 6: [None], 7: [None]}}
to_simple()

Returns a simple version of itself (i.e., undirected and loops and multiple edges are removed).

EXAMPLES:

sage: G = DiGraph(loops=True,multiedges=True,sparse=True)
sage: G.add_edges( [ (0,0), (1,1), (2,2), (2,3,1), (2,3,2), (3,2) ] )
sage: G.edges(labels=False)
[(0, 0), (1, 1), (2, 2), (2, 3), (2, 3), (3, 2)]
sage: H=G.to_simple()
sage: H.edges(labels=False)
[(2, 3)]
sage: H.is_directed()
False
sage: H.allows_loops()
False
sage: H.allows_multiple_edges()
False
trace_faces(comb_emb)

A helper function for finding the genus of a graph. Given a graph and a combinatorial embedding (rot_sys), this function will compute the faces (returned as a list of lists of edges (tuples) of the particular embedding.

Note - rot_sys is an ordered list based on the hash order of the vertices of graph. To avoid confusion, it might be best to set the rot_sys based on a ‘nice_copy’ of the graph.

INPUT:

  • comb_emb - a combinatorial embedding dictionary. Format: v1:[v2,v3], v2:[v1], v3:[v1] (clockwise ordering of neighbors at each vertex.)

EXAMPLES:

sage: T = graphs.TetrahedralGraph()
sage: T.trace_faces({0: [1, 3, 2], 1: [0, 2, 3], 2: [0, 3, 1], 3: [0, 1, 2]})
[[(0, 1), (1, 2), (2, 0)],
 [(3, 2), (2, 1), (1, 3)],
 [(2, 3), (3, 0), (0, 2)],
 [(0, 3), (3, 1), (1, 0)]]
transitive_closure()

Computes the transitive closure of a graph and returns it. The original graph is not modified.

The transitive closure of a graph G has an edge (x,y) if and only if there is a path between x and y in G.

The transitive closure of any strongly connected component of a graph is a complete graph. In particular, the transitive closure of a connected undirected graph is a complete graph. The transitive closure of a directed acyclic graph is a directed acyclic graph representing the full partial order.

EXAMPLES:

sage: g=graphs.PathGraph(4)
sage: g.transitive_closure()
Transitive closure of Path Graph: Graph on 4 vertices
sage: g.transitive_closure()==graphs.CompleteGraph(4)
True
sage: g=DiGraph({0:[1,2], 1:[3], 2:[4,5]})
sage: g.transitive_closure().edges(labels=False)
[(0, 1), (0, 2), (0, 3), (0, 4), (0, 5), (1, 3), (2, 4), (2, 5)]
transitive_reduction()

Returns a transitive reduction of a graph. The original graph is not modified.

A transitive reduction H of G has a path from x to y if and only if there was a path from x to y in G. Deleting any edge of H destroys this property. A transitive reduction is not unique in general. A transitive reduction has the same transitive closure as the original graph.

A transitive reduction of a complete graph is a tree. A transitive reduction of a tree is itself.

EXAMPLES:

sage: g=graphs.PathGraph(4)
sage: g.transitive_reduction()==g
True
sage: g=graphs.CompleteGraph(5)
sage: edges = g.transitive_reduction().edges(); len(edges)
4
sage: g=DiGraph({0:[1,2], 1:[2,3,4,5], 2:[4,5]})
sage: g.transitive_reduction().size()
5
traveling_salesman_problem(use_edge_labels=False, solver=None, constraint_generation=None, verbose=0, verbose_constraints=False)

Solves the traveling salesman problem (TSP)

Given a graph (resp. a digraph) G with weighted edges, the traveling salesman problem consists in finding a Hamiltonian cycle (resp. circuit) of the graph of minimum cost.

This TSP is one of the most famous NP-Complete problems, this function can thus be expected to take some time before returning its result.

INPUT:

  • use_edge_labels (boolean) – whether to consider the weights of the edges.

    • If set to False (default), all edges are assumed to weight 1
    • If set to True, the weights are taken into account, and the circuit returned is the one minimizing the sum of the weights.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.

  • constraint_generation (boolean) – whether to use constraint generation when solving the Mixed Integer Linear Program.

    When constraint_generation = None, constraint generation is used whenever the graph has a density larger than 70%.

  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

  • verbose_constraints – whether to display which constraints are being generated.

OUTPUT:

A solution to the TSP, as a Graph object whose vertex set is V(G), and whose edges are only those of the solution.

ALGORITHM:

This optimization problem is solved through the use of Linear Programming.

NOTE:

  • This function is correctly defined for both graph and digraphs. In the second case, the returned cycle is a circuit of optimal cost.

EXAMPLES:

The Heawood graph is known to be Hamiltonian:

sage: g = graphs.HeawoodGraph()
sage: tsp = g.traveling_salesman_problem()
sage: tsp
TSP from Heawood graph: Graph on 14 vertices

The solution to the TSP has to be connected

sage: tsp.is_connected()
True

It must also be a 2-regular graph:

sage: tsp.is_regular(k=2)
True

And obviously it is a subgraph of the Heawood graph:

sage: all([ e in g.edges() for e in tsp.edges()])
True

On the other hand, the Petersen Graph is known not to be Hamiltonian:

sage: g = graphs.PetersenGraph()
sage: tsp = g.traveling_salesman_problem()
Traceback (most recent call last):
...
ValueError: The given graph is not hamiltonian

One easy way to change is is obviously to add to this graph the edges corresponding to a Hamiltonian cycle.

If we do this by setting the cost of these new edges to 2, while the others are set to 1, we notice that not all the edges we added are used in the optimal solution

sage: for u, v in g.edges(labels = None):
...      g.set_edge_label(u,v,1)

sage: cycle = graphs.CycleGraph(10)
sage: for u,v in cycle.edges(labels = None):
...      if not g.has_edge(u,v):
...          g.add_edge(u,v)
...      g.set_edge_label(u,v,2)

sage: tsp = g.traveling_salesman_problem(use_edge_labels = True)
sage: sum( tsp.edge_labels() ) < 2*10
True

If we pick 1/2 instead of 2 as a cost for these new edges, they clearly become the optimal solution:

sage: for u,v in cycle.edges(labels = None):
...      g.set_edge_label(u,v,1/2)

sage: tsp = g.traveling_salesman_problem(use_edge_labels = True)
sage: sum( tsp.edge_labels() ) == (1/2)*10
True

TESTS:

Comparing the results returned according to the value of constraint_generation. First, for graphs:

sage: from operator import itemgetter
sage: n = 20
sage: for i in range(20):
...       g = Graph()
...       g.allow_multiple_edges(False)
...       for u,v in graphs.RandomGNP(n,.2).edges(labels = False):
...            g.add_edge(u,v,round(random(),5))
...       for u,v in graphs.CycleGraph(n).edges(labels = False):
...            if not g.has_edge(u,v):
...                g.add_edge(u,v,round(random(),5))
...       v1 = g.traveling_salesman_problem(constraint_generation = False, use_edge_labels = True)
...       v2 = g.traveling_salesman_problem(use_edge_labels = True)
...       c1 = sum(map(itemgetter(2), v1.edges()))
...       c2 = sum(map(itemgetter(2), v2.edges()))
...       if c1 != c2:
...           print "Error !",c1,c2
...           break

Then for digraphs:

sage: from operator import itemgetter
sage: set_random_seed(0)
sage: n = 20
sage: for i in range(20):
...       g = DiGraph()
...       g.allow_multiple_edges(False)
...       for u,v in digraphs.RandomDirectedGNP(n,.2).edges(labels = False):
...            g.add_edge(u,v,round(random(),5))
...       for u,v in digraphs.Circuit(n).edges(labels = False):
...            if not g.has_edge(u,v):
...                g.add_edge(u,v,round(random(),5))
...       v2 = g.traveling_salesman_problem(use_edge_labels = True)
...       v1 = g.traveling_salesman_problem(constraint_generation = False, use_edge_labels = True)
...       c1 = sum(map(itemgetter(2), v1.edges()))
...       c2 = sum(map(itemgetter(2), v2.edges()))
...       if c1 != c2:
...           print "Error !",c1,c2
...           print "With constraint generation :",c2
...           print "Without constraint generation :",c1
...           break
triangles_count(algorithm='iter')

Returns the number of triangles in the (di)graph.

For digraphs, we count the number of directed circuit of length 3.

INPUT:

  • algorithm – (default: 'matrix') specifies the algorithm to use among:

    • 'matrix' uses the trace of the cube of the adjacency matrix.
    • 'iter' iterates over the pairs of neighbors of each vertex. This is faster for sparse graphs.

EXAMPLES:

The Petersen graph is triangle free and thus:

sage: G = graphs.PetersenGraph()
sage: G.triangles_count()
0

Any triple of vertices in the complete graph induces a triangle so we have:

sage: G = graphs.CompleteGraph(150)
sage: G.triangles_count() == binomial(150,3)
True

The 2-dimensional DeBruijn graph of 2 symbols has 2 directed C3:

sage: G = digraphs.DeBruijn(2,2)
sage: G.triangles_count()
2

The directed n-cycle is trivially triangle free for n > 3:

sage: G = digraphs.Circuit(10)
sage: G.triangles_count()
0

TESTS:

Comparison on algorithms:

sage: for i in xrange(10): # long test
...       G = graphs.RandomBarabasiAlbert(50,2)
...       tm = G.triangles_count(algorithm='matrix')
...       te = G.triangles_count(algorithm='iter')
...       if tm!=te:
...          print "That's not good!"

Asking for an unknown algorithm:

sage: G = Graph()
sage: G.triangles_count(algorithm='tip top')
Traceback (most recent call last):
...
ValueError: Algorithm 'tip top' not yet implemented. Please contribute.
union(other)

Returns the union of self and other.

If the graphs have common vertices, the common vertices will be identified.

EXAMPLES:

sage: G = graphs.CycleGraph(3)
sage: H = graphs.CycleGraph(4)
sage: J = G.union(H); J
Graph on 4 vertices
sage: J.vertices()
[0, 1, 2, 3]
sage: J.edges(labels=False)
[(0, 1), (0, 2), (0, 3), (1, 2), (2, 3)]
vertex_boundary(vertices1, vertices2=None)

Returns a list of all vertices in the external boundary of vertices1, intersected with vertices2. If vertices2 is None, then vertices2 is the complement of vertices1. This is much faster if vertices1 is smaller than vertices2.

The external boundary of a set of vertices is the union of the neighborhoods of each vertex in the set. Note that in this implementation, since vertices2 defaults to the complement of vertices1, if a vertex v has a loop, then vertex_boundary(v) will not contain v.

In a digraph, the external boundary of a vertex v are those vertices u with an arc (v, u).

EXAMPLES:

sage: G = graphs.CubeGraph(4)
sage: l = ['0111', '0000', '0001', '0011', '0010', '0101', '0100', '1111', '1101', '1011', '1001']
sage: G.vertex_boundary(['0000', '1111'], l)
['0111', '0001', '0010', '0100', '1101', '1011']
sage: D = DiGraph({0:[1,2], 3:[0]})
sage: D.vertex_boundary([0])
[1, 2]
vertex_connectivity(value_only=True, sets=False, solver=None, verbose=0)

Returns the vertex connectivity of the graph. For more information, see the Wikipedia article on connectivity.

Note

  • When the graph is a directed graph, this method actually computes the strong connectivity, (i.e. a directed graph is strongly k-connected if there are k disjoint paths between any two vertices u, v). If you do not want to consider strong connectivity, the best is probably to convert your DiGraph object to a Graph object, and compute the connectivity of this other graph.
  • By convention, a complete graph on n vertices is n-1 connected. In this case, no certificate can be given as there is no pair of vertices split by a cut of size k-1. For this reason, the certificates returned in this situation are empty.

INPUT:

  • value_only – boolean (default: True)
    • When set to True (default), only the value is returned.
    • When set to False , both the value and a minimum vertex cut are returned.
  • sets – boolean (default: False)
    • When set to True, also returns the two sets of vertices that are disconnected by the cut. Implies value_only=False
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

EXAMPLES:

A basic application on a PappusGraph:

sage: g=graphs.PappusGraph()
sage: g.vertex_connectivity()
3

In a grid, the vertex connectivity is equal to the minimum degree, in which case one of the two sets is of cardinality 1:

sage: g = graphs.GridGraph([ 3,3 ])
sage: [value, cut, [ setA, setB ]] = g.vertex_connectivity(sets=True)
sage: len(setA) == 1 or len(setB) == 1
True

A vertex cut in a tree is any internal vertex:

sage: g = graphs.RandomGNP(15,.5)
sage: tree = Graph()
sage: tree.add_edges(g.min_spanning_tree())
sage: [val, [cut_vertex]] = tree.vertex_connectivity(value_only=False)
sage: tree.degree(cut_vertex) > 1
True

When value_only = True, this function is optimized for small connectivity values and does not need to build a linear program.

It is the case for connected graphs which are not connected:

sage: g = 2 * graphs.PetersenGraph()
sage: g.vertex_connectivity()
0

Or if they are just 1-connected:

sage: g = graphs.PathGraph(10)
sage: g.vertex_connectivity()
1

For directed graphs, the strong connectivity is tested through the dedicated function:

sage: g = digraphs.ButterflyGraph(3)
sage: g.vertex_connectivity()
0

A complete graph on 10 vertices is 9-connected:

sage: g = graphs.CompleteGraph(10)
sage: g.vertex_connectivity()
9

A complete digraph on 10 vertices is 9-connected:

sage: g = DiGraph(graphs.CompleteGraph(10))
sage: g.vertex_connectivity()
9
vertex_cut(s, t, value_only=True, vertices=False, solver=None, verbose=0)

Returns a minimum vertex cut between non-adjacent vertices s and t represented by a list of vertices.

A vertex cut between two non-adjacent vertices is a set U of vertices of self such that the graph obtained by removing U from self is disconnected. For more information, see the Wikipedia article on cuts.

INPUT:

  • value_only – boolean (default: True). When set to True, only the size of the minimum cut is returned.
  • vertices – boolean (default: False). When set to True, also returns the two sets of vertices that are disconnected by the cut. Implies value_only set to False.
  • solver – (default: None) Specify a Linear Program (LP) solver to be used. If set to None, the default one is used. For more information on LP solvers and which default solver is used, see the method solve of the class MixedIntegerLinearProgram.
  • verbose – integer (default: 0). Sets the level of verbosity. Set to 0 by default, which means quiet.

OUTPUT:

Real number or tuple, depending on the given arguments (examples are given below).

EXAMPLE:

A basic application in the Pappus graph:

sage: g = graphs.PappusGraph()
sage: g.vertex_cut(1, 16, value_only=True)
3

In the bipartite complete graph K_{2,8}, a cut between the two vertices in the size 2 part consists of the other 8 vertices:

sage: g = graphs.CompleteBipartiteGraph(2, 8)
sage: [value, vertices] = g.vertex_cut(0, 1, value_only=False)
sage: print value
8
sage: vertices == range(2,10)
True

Clearly, in this case the two sides of the cut are singletons

sage: [value, vertices, [set1, set2]] = g.vertex_cut(0,1, vertices=True)
sage: len(set1) == 1
True
sage: len(set2) == 1
True
vertex_disjoint_paths(s, t)

Returns a list of vertex-disjoint paths between two vertices as given by Menger’s theorem.

The vertex version of Menger’s theorem asserts that the size of the minimum vertex cut between two vertices s and`t` (the minimum number of vertices whose removal disconnects s and t) is equal to the maximum number of pairwise vertex-independent paths from s to t.

This function returns a list of such paths.

EXAMPLE:

In a complete bipartite graph

sage: g = graphs.CompleteBipartiteGraph(2,3)
sage: g.vertex_disjoint_paths(0,1)
[[0, 2, 1], [0, 3, 1], [0, 4, 1]]
vertex_iterator(vertices=None)

Returns an iterator over the given vertices.

Returns False if not given a vertex, sequence, iterator or None. None is equivalent to a list of every vertex. Note that for v in G syntax is allowed.

INPUT:

  • vertices - iterated vertices are these intersected with the vertices of the (di)graph

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: for v in P.vertex_iterator():
...    print v
...
0
1
2
...
8
9
sage: G = graphs.TetrahedralGraph()
sage: for i in G:
...    print i
0
1
2
3

Note that since the intersection option is available, the vertex_iterator() function is sub-optimal, speed-wise, but note the following optimization:

sage: timeit V = P.vertices()                   # not tested
100000 loops, best of 3: 8.85 [micro]s per loop
sage: timeit V = list(P.vertex_iterator())      # not tested
100000 loops, best of 3: 5.74 [micro]s per loop
sage: timeit V = list(P._nxg.adj.iterkeys())    # not tested
100000 loops, best of 3: 3.45 [micro]s per loop

In other words, if you want a fast vertex iterator, call the dictionary directly.

vertices(key=None, boundary_first=False)

Return a list of the vertices.

INPUT:

  • key - default: None - a function that takes a vertex as its one argument and returns a value that can be used for comparisons in the sorting algorithm.
  • boundary_first - default: False - if True, return the boundary vertices first.

OUTPUT:

The vertices of the list.

Warning

There is always an attempt to sort the list before returning the result. However, since any object may be a vertex, there is no guarantee that any two vertices will be comparable. With default objects for vertices (all integers), or when all the vertices are of the same simple type, then there should not be a problem with how the vertices will be sorted. However, if you need to guarantee a total order for the sort, use the key argument, as illustrated in the examples below.

EXAMPLES:

sage: P = graphs.PetersenGraph()
sage: P.vertices()
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

If you do not care about sorted output and you are concerned about the time taken to sort, consider the following alternatives. The moral is: if you want a fast vertex iterator, call the dictionary directly.

sage: timeit V = P.vertices()                     # not tested
100000 loops, best of 3: 8.85 [micro]s per loop
sage: timeit V = list(P.vertex_iterator())        # not tested
100000 loops, best of 3: 5.74 [micro]s per loop
sage: timeit V = list(P._nxg.adj.iterkeys())      # not tested
100000 loops, best of 3: 3.45 [micro]s per loop

We illustrate various ways to use a key to sort the list:

sage: H=graphs.HanoiTowerGraph(3,3,labels=False)
sage: H.vertices()
[0, 1, 2, 3, 4, ... 22, 23, 24, 25, 26]
sage: H.vertices(key=lambda x: -x)
[26, 25, 24, 23, 22, ... 4, 3, 2, 1, 0]
sage: G=graphs.HanoiTowerGraph(3,3)
sage: G.vertices()
[(0, 0, 0), (0, 0, 1), (0, 0, 2), (0, 1, 0), ... (2, 2, 1), (2, 2, 2)]
sage: G.vertices(key = lambda x: (x[1], x[2], x[0]))
[(0, 0, 0), (1, 0, 0), (2, 0, 0), (0, 0, 1), ... (1, 2, 2), (2, 2, 2)]

The discriminant of a polynomial is a function that returns an integer. We build a graph whose vertices are polynomials, and use the discriminant function to provide an ordering. Note that since functions are first-class objects in Python, we can specify precisely the function from the Sage library that we wish to use as the key.

sage: t = polygen(QQ, 't')
sage: K = Graph({5*t:[t^2], t^2:[t^2+2], t^2+2:[4*t^2-6], 4*t^2-6:[5*t]})
sage: dsc = sage.rings.polynomial.polynomial_rational_flint.Polynomial_rational_flint.discriminant
sage: verts = K.vertices(key=dsc)
sage: verts
[t^2 + 2, t^2, 5*t, 4*t^2 - 6]
sage: [x.discriminant() for x in verts]
[-8, 0, 1, 96]

If boundary vertices are requested first, then they are sorted separately from the remainder (which are also sorted).

sage: P = graphs.PetersenGraph()
sage: P.set_boundary((5..9))
sage: P.vertices(boundary_first=True)
[5, 6, 7, 8, 9, 0, 1, 2, 3, 4]
sage: P.vertices(boundary_first=True, key=lambda x: -x)
[9, 8, 7, 6, 5, 4, 3, 2, 1, 0]
weighted(new=None)

Whether the (di)graph is to be considered as a weighted (di)graph.

Note that edge weightings can still exist for (di)graphs G where G.weighted() is False.

EXAMPLES:

Here we have two graphs with different labels, but weighted() is False for both, so we just check for the presence of edges:

sage: G = Graph({0:{1:'a'}}, sparse=True)
sage: H = Graph({0:{1:'b'}}, sparse=True)
sage: G == H
True

Now one is weighted and the other is not, and thus the graphs are not equal:

sage: G.weighted(True)
sage: H.weighted()
False
sage: G == H
False

However, if both are weighted, then we finally compare ‘a’ to ‘b’:

sage: H.weighted(True)
sage: G == H
False

TESTS:

Ensure that ticket #10490 is fixed: allows a weighted graph to be set as unweighted.

sage: G = Graph({1:{2:3}})
sage: G.weighted()
False
sage: G.weighted('a')
sage: G.weighted(True)
sage: G.weighted()
True
sage: G.weighted('a')
sage: G.weighted()
True
sage: G.weighted(False)
sage: G.weighted()
False
sage: G.weighted('a')
sage: G.weighted()
False
sage: G.weighted(True)
sage: G.weighted()
True
weighted_adjacency_matrix(sparse=True, boundary_first=False)

Returns the weighted adjacency matrix of the graph.

Each vertex is represented by its position in the list returned by the vertices() function.

EXAMPLES:

sage: G = Graph(sparse=True, weighted=True)
sage: G.add_edges([(0,1,1),(1,2,2),(0,2,3),(0,3,4)])
sage: M = G.weighted_adjacency_matrix(); M
[0 1 3 4]
[1 0 2 0]
[3 2 0 0]
[4 0 0 0]
sage: H = Graph(data=M, format='weighted_adjacency_matrix', sparse=True)
sage: H == G
True

The following doctest verifies that #4888 is fixed:

sage: G = DiGraph({0:{}, 1:{0:1}, 2:{0:1}}, weighted = True,sparse=True)
sage: G.weighted_adjacency_matrix()
[0 0 0]
[1 0 0]
[1 0 0]
wiener_index(G)

Returns the Wiener index of the graph.

The Wiener index of a graph G can be defined in two equivalent ways [KRG96b] :

  • W(G) = \frac 1 2 \sum_{u,v\in G} d(u,v) where d(u,v) denotes the distance between vertices u and v.
  • Let \Omega be a set of \frac {n(n-1)} 2 paths in G such that \Omega contains exactly one shortest u-v path for each set \{u,v\} of vertices in G. Besides, \forall e\in E(G), let \Omega(e) denote the paths from \Omega containing e. We then have W(G) = \sum_{e\in E(G)}|\Omega(e)|.

EXAMPLE:

From [GYLL93c], cited in [KRG96b]:

sage: g=graphs.PathGraph(10)
sage: w=lambda x: (x*(x*x -1)/6)
sage: g.wiener_index()==w(10)
True
sage.graphs.generic_graph.graph_isom_equivalent_non_edge_labeled_graph(g, partition=None, standard_label=None, return_relabeling=False, return_edge_labels=False, inplace=False, ignore_edge_labels=False)

Helper function for canonical labeling of edge labeled (di)graphs.

Translates to a bipartite incidence-structure type graph appropriate for computing canonical labels of edge labeled and/or multi-edge graphs. Note that this is actually computationally equivalent to implementing a change on an inner loop of the main algorithm- namely making the refinement procedure sort for each label.

If the graph is a multigraph, it is translated to a non-multigraph, where each edge is labeled with a dictionary describing how many edges of each label were originally there. Then in either case we are working on a graph without multiple edges. At this point, we create another (bipartite) graph, whose left vertices are the original vertices of the graph, and whose right vertices represent the edges. We partition the left vertices as they were originally, and the right vertices by common labels: only automorphisms taking edges to like-labeled edges are allowed, and this additional partition information enforces this on the bipartite graph.

INPUT:

  • g – Graph or DiGraph
  • partition – (default:None) if given, the partition of the vertices is as well relabeled
  • standard_label – (default:None) the standard label is not considered to be changed
  • return_relabeling – (defaut:False) if True, a dictionary containing the relabeling is returned
  • return_edge_labels – (defaut:False) if True, the different edge_labels are returned (useful if inplace is True)
  • inplace – (default:False) if True, g is modified, otherwise the result is returned. Note that attributes of g are not copied for speed issues, only edges and vertices.

OUTPUT:

  • if not inplace: the unlabeled graph without multiple edges
  • the partition of the vertices
  • if return_relabeling: a dictionary containing the relabeling
  • if return_edge_labels: the list of (former) edge labels is returned

EXAMPLES:

sage: from sage.graphs.generic_graph import graph_isom_equivalent_non_edge_labeled_graph

sage: G = Graph(multiedges=True,sparse=True)
sage: G.add_edges( (0,1,i) for i in range(10) )
sage: G.add_edge(1,2,'string')
sage: G.add_edge(2,123)
sage: g = graph_isom_equivalent_non_edge_labeled_graph(G, partition=[[0,123],[1,2]]); g
[Graph on 6 vertices, [[0, 3], [1, 2], [4], [5]]]

sage: g = graph_isom_equivalent_non_edge_labeled_graph(G); g
[Graph on 6 vertices, [[0, 1, 2, 3], [4], [5]]]
sage: g[0].edges()
[(0, 4, None), (1, 4, None), (1, 5, None), (2, 3, None), (2, 5, None)]

sage: g = graph_isom_equivalent_non_edge_labeled_graph(G,standard_label='string',return_edge_labels=True); g
[Graph on 6 vertices, [[0, 1, 2, 3], [5], [4]], [[[None, 1]], [[0, 1], [1, 1], [2, 1], [3, 1], [4, 1], [5, 1], [6, 1], [7, 1], [8, 1], [9, 1]], [['string', 1]]]]
sage: g[0].edges()
[(0, 4, None), (1, 2, None), (1, 4, None), (2, 5, None), (3, 5, None)]

sage: graph_isom_equivalent_non_edge_labeled_graph(G,inplace=True)
[[[0, 1, 2, 3], [4], [5]]]
sage: G.edges()
[(0, 4, None), (1, 4, None), (1, 5, None), (2, 3, None), (2, 5, None)]

Ensure that #14108 is fixed:

sage: G=DiGraph([[0,0],[0,0],[0,0],[1,1],[1,1],[1,1]])
sage: H=DiGraph([[0,0],[0,0],[0,0],[0,0],[1,1],[1,1]])
sage: G.is_isomorphic(H)
False
sage: H=DiGraph([[0,0],[0,0],[0,0],[0,0],[0,0],[1,1],[1,1]])
sage: HH=DiGraph([[0,0],[0,0],[0,0],[0,0],[1,1],[1,1],[1,1]])
sage: H.is_isomorphic(HH)
False
sage: H.is_isomorphic(HH, edge_labels=True)
False
sage.graphs.generic_graph.tachyon_vertex_plot(g, bgcolor=(1, 1, 1), vertex_colors=None, vertex_size=0.06, pos3d=None, **kwds)

Helper function for plotting graphs in 3d with Tachyon. Returns a plot containing only the vertices, as well as the 3d position dictionary used for the plot.

INPUT:
  • pos3d - a 3D layout of the vertices
  • various rendering options

EXAMPLES:

sage: G = graphs.TetrahedralGraph()
sage: from sage.graphs.generic_graph import tachyon_vertex_plot
sage: T,p = tachyon_vertex_plot(G, pos3d = G.layout(dim=3))
sage: type(T)
<class 'sage.plot.plot3d.tachyon.Tachyon'>
sage: type(p)
<type 'dict'>

Table Of Contents

Previous topic

Graph Theory

Next topic

Undirected graphs

This Page