def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = list()
self._uf = UnionFind()
self._pq = PriorityQueue()
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes():
self.hamiltonian_cycle.add_node(node)
self._uf = UnionFind()
self._pq = PriorityQueue()
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = list()
self._uf = UnionFind()
self._pq = PriorityQueue()
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes():
self.hamiltonian_cycle.add_node(node)
self._uf = UnionFind()
self._pq = PriorityQueue()
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
class SortedEdgeTSPWithGraph:
"""The sorted edge algorithm for TSP.
Attributes
----------
graph : input weighted complete graph
hamiltonian_cycle : cycle graph
_uf : disjoint-set data structure, private
_pq : priority queue, private
"""
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes():
self.hamiltonian_cycle.add_node(node)
self._uf = UnionFind()
self._pq = PriorityQueue()
def run(self, source=None):
"""Executable pseudocode."""
for node in self.graph.iternodes():
self._uf.create(node)
for edge in self.graph.iteredges():
self._pq.put((edge.weight, edge))
while not self._pq.empty():
_, edge = self._pq.get()
degree1 = self.hamiltonian_cycle.degree(edge.source)
degree2 = self.hamiltonian_cycle.degree(edge.target)
# Zabezpieczam przed skrzyzowaniami T, X, itp.
if degree1 == 2 or degree2 == 2:
continue
if degree1 == 0 or degree2 == 0:
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.add_edge(edge)
continue
# Here degree1 = degree2 = 1.
if self._uf.find(edge.source) != self._uf.find(edge.target):
# Join pieces.
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.add_edge(edge)
elif self.hamiltonian_cycle.e() == self.hamiltonian_cycle.v()-1:
# Close the Hamiltonian cycle.
self.hamiltonian_cycle.add_edge(edge)
class KruskalMSTSorted:
"""Kruskal's algorithm for finding a minimum spanning tree.
Attributes
----------
graph : input undirected graph or multigraph
mst : graph (MST)
_uf : disjoint-set data structure, private
Examples
--------
>>> from graphtheory.structures.edges import Edge
>>> from graphtheory.structures.graphs import Graph
>>> from graphtheory.spanningtrees.kruskal import KruskalMSTSorted
>>> G = Graph(n=10, False) # an exemplary undirected graph
# Add nodes and edges here.
>>> algorithm = KruskalMSTSorted(G) # initialization
>>> algorithm.run() # calculations
>>> algorithm.mst.show() # MST as an undirected graph
Notes
-----
Based on:
Cormen, T. H., Leiserson, C. E., Rivest, R. L., and Stein, C., 2009,
Introduction to Algorithms, third edition, The MIT Press,
Cambridge, London.
https://en.wikipedia.org/wiki/Kruskal's_algorithm
J. B. Kruskal, On the shortest spanning subtree of a graph
and the traveling salesman problem,
Proc. Amer. Math. Soc. 7, 48-50 (1956).
D. Eppstein, Kruskal's algorithm for minimum spanning trees, 2006,
https://www.ics.uci.edu/~eppstein/PADS/MinimumSpanningTree.py
"""
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
def run(self):
"""Finding MST."""
for node in self.graph.iternodes():
self._uf.create(node)
#for edge in sorted(self.graph.iteredges()): # very slow
#for (_, edge) in sorted((edge.weight, edge) for edge in self.graph.iteredges()):
from operator import attrgetter
for edge in sorted(self.graph.iteredges(), key=attrgetter("weight")):
if self._uf.find(edge.source) != self._uf.find(edge.target):
self._uf.union(edge.source, edge.target)
self.mst.add_edge(edge)
def to_tree(self):
"""Compatibility with other classes."""
return self.mst
class KruskalMSTSorted:
"""Kruskal's algorithm for finding a minimum spanning tree.
Attributes
----------
graph : input undirected graph or multigraph
mst : graph (MST)
_uf : disjoint-set data structure, private
Examples
--------
>>> from graphtheory.structures.edges import Edge
>>> from graphtheory.structures.graphs import Graph
>>> from graphtheory.spanningtrees.kruskal import KruskalMSTSorted
>>> G = Graph(n=10, False) # an exemplary undirected graph
# Add nodes and edges here.
>>> algorithm = KruskalMSTSorted(G) # initialization
>>> algorithm.run() # calculations
>>> algorithm.mst.show() # MST as an undirected graph
Notes
-----
Based on:
Cormen, T. H., Leiserson, C. E., Rivest, R. L., and Stein, C., 2009,
Introduction to Algorithms, third edition, The MIT Press,
Cambridge, London.
https://en.wikipedia.org/wiki/Kruskal's_algorithm
J. B. Kruskal, On the shortest spanning subtree of a graph
and the traveling salesman problem,
Proc. Amer. Math. Soc. 7, 48-50 (1956).
D. Eppstein, Kruskal's algorithm for minimum spanning trees, 2006,
https://www.ics.uci.edu/~eppstein/PADS/MinimumSpanningTree.py
"""
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
def run(self):
"""Finding MST."""
for node in self.graph.iternodes():
self._uf.create(node)
#for edge in sorted(self.graph.iteredges()): # very slow
#for (_, edge) in sorted((edge.weight, edge) for edge in self.graph.iteredges()):
from operator import attrgetter
for edge in sorted(self.graph.iteredges(), key=attrgetter("weight")):
if self._uf.find(edge.source) != self._uf.find(edge.target):
self._uf.union(edge.source, edge.target)
self.mst.add_edge(edge)
def to_tree(self):
"""Compatibility with other classes."""
return self.mst
class BoruvkaMST:
"""Boruvka's algorithm for finding a minimum spanning tree.
Attributes
----------
graph : input undirected graph or multigraph
mst : graph (MST)
_uf : disjoint-set data structure, private
Examples
--------
>>> from graphtheory.structures.edges import Edge
>>> from graphtheory.structures.graphs import Graph
>>> from graphtheory.spanningtrees.boruvka import BoruvkaMST
>>> G = Graph(n=10, False) # an exemplary undirected graph
# Add nodes and edges here.
>>> algorithm = BoruvkaMST(G)
>>> algorithm.run() # calculations
>>> algorithm.mst.show() # MST as an undirected graph
Notes
-----
Based on pseudocode from Wikipedia:
https://en.wikipedia.org/wiki/Boruvka's_algorithm
https://en.wikipedia.org/wiki/Minimum_spanning_tree
"""
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
def run(self):
"""Finding MST."""
for node in self.graph.iternodes():
self._uf.create(node)
forest = set(node for node in self.graph.iternodes())
dummy_edge = Edge(None, None, float("inf"))
new_len = len(forest)
old_len = new_len + 1
while old_len > new_len:
old_len = new_len
min_edges = dict(((node, dummy_edge) for node in forest))
# Finding the cheapest eadges.
for edge in self.graph.iteredges(): # O(E) time
source = self._uf.find(edge.source)
target = self._uf.find(edge.target)
if source != target: # different components
if edge < min_edges[source]:
min_edges[source] = edge
if edge < min_edges[target]:
min_edges[target] = edge
# Connecting components, total time is O(V).
forest = set()
for edge in min_edges.itervalues():
if edge is dummy_edge: # a disconnected graph
continue
source = self._uf.find(edge.source)
target = self._uf.find(edge.target)
if source != target: # different components
self._uf.union(source, target)
forest.add(source)
self.mst.add_edge(edge)
# Remove duplicates, total time is O(V).
forest = set(self._uf.find(node) for node in forest)
new_len = len(forest)
if new_len == 1: # a connected graph
break
def to_tree(self):
"""Return the minimum spanning tree."""
return self.mst
def test_unionfind(self):
algorithm = UnionFind()
for node in range(10):
algorithm.create(node)
pairs = [(0, 1), (2, 3), (4, 5), (6, 7), (8, 9), (3, 1)]
for a, b in pairs:
algorithm.union(a, b)
self.assertTrue(algorithm.find(1) == algorithm.find(2))
self.assertTrue(algorithm.find(7) != algorithm.find(9))
pairs = [(7, 5), (3, 7), (8, 7)]
for a, b in pairs:
algorithm.union(a, b)
self.assertTrue(algorithm.find(0) == algorithm.find(6))
self.assertTrue(algorithm.find(1) == algorithm.find(9))
class SortedEdgeTSPWithEdges:
"""The sorted edge algorithm (a.k.a. cheapest link) for TSP.
Attributes
----------
graph : input weighted complete graph
hamiltonian_cycle : list of edges
_uf : disjoint-set data structure, private
_pq : priority queue, private
"""
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = list()
self._uf = UnionFind()
self._pq = PriorityQueue()
def run(self, source=None):
"""Executable pseudocode."""
degree_dict = dict((node, 0) for node in self.graph.iternodes())
for node in self.graph.iternodes():
self._uf.create(node)
for edge in self.graph.iteredges():
self._pq.put((edge.weight, edge))
while not self._pq.empty():
_, edge = self._pq.get()
degree1 = degree_dict[edge.source]
degree2 = degree_dict[edge.target]
# Zabezpieczam przed skrzyzowaniami T, X, itp.
if degree1 == 2 or degree2 == 2:
continue
if degree1 == 0 or degree2 == 0:
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
continue
# Here degree1 = degree2 = 1.
if self._uf.find(edge.source) != self._uf.find(edge.target):
# Join pieces.
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
elif len(self.hamiltonian_cycle) == self.graph.v() - 1:
# Close the Hamiltonian cycle.
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
# Create the corrected (directed) list of edges.
# Przeksztalcam cykl Hamiltona do postaci poprawnej listy krawedzi.
# Na razie krawedzie moga miec zla kolejnosc i kierunek.
# Tworze slownik z krawedziami w obie strony.
edge_dict = dict((node, []) for node in self.graph.iternodes())
for edge in self.hamiltonian_cycle: # O(V) time
edge_dict[edge.source].append(edge)
edge_dict[edge.target].append(~edge)
edge = self.hamiltonian_cycle[0]
self.hamiltonian_cycle = [edge]
# Kompletuje kolejne krawedzie.
for step in xrange(self.graph.v() - 1): # O(V) time
edge1, edge2 = edge_dict[edge.target]
if edge1.target == edge.source:
self.hamiltonian_cycle.append(edge2)
edge = edge2
else:
self.hamiltonian_cycle.append(edge1)
edge = edge1
def test_unionfind(self):
algorithm = UnionFind()
for node in range(10):
algorithm.create(node)
pairs = [(0, 1), (2, 3), (4, 5), (6, 7), (8, 9), (3, 1)]
for a, b in pairs:
algorithm.union(a, b)
self.assertTrue(algorithm.find(1) == algorithm.find(2))
self.assertTrue(algorithm.find(7) != algorithm.find(9))
pairs = [(7, 5), (3, 7), (8, 7)]
for a, b in pairs:
algorithm.union(a, b)
self.assertTrue(algorithm.find(0) == algorithm.find(6))
self.assertTrue(algorithm.find(1) == algorithm.find(9))
class BoruvkaMST:
"""Boruvka's algorithm for finding a minimum spanning tree.
Attributes
----------
graph : input undirected graph or multigraph
mst : graph (MST)
_uf : disjoint-set data structure, private
Examples
--------
>>> from graphtheory.structures.edges import Edge
>>> from graphtheory.structures.graphs import Graph
>>> from graphtheory.spanningtrees.boruvka import BoruvkaMST
>>> G = Graph(n=10, False) # an exemplary undirected graph
# Add nodes and edges here.
>>> algorithm = BoruvkaMST(G)
>>> algorithm.run() # calculations
>>> algorithm.mst.show() # MST as an undirected graph
Notes
-----
Based on pseudocode from Wikipedia:
https://en.wikipedia.org/wiki/Boruvka's_algorithm
https://en.wikipedia.org/wiki/Minimum_spanning_tree
"""
def __init__(self, graph):
"""The algorithm initialization.
Parameters
----------
graph : undirected weighted graph or multigraph
"""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.mst = self.graph.__class__(self.graph.v())
for node in self.graph.iternodes(): # isolated nodes are possible
self.mst.add_node(node)
self._uf = UnionFind()
def run(self):
"""Finding MST."""
for node in self.graph.iternodes():
self._uf.create(node)
forest = set(self.graph.iternodes())
dummy_edge = Edge(None, None, float("inf"))
new_len = len(forest)
old_len = new_len + 1
while old_len > new_len:
old_len = new_len
min_edges = dict(((node, dummy_edge) for node in forest))
# Finding the cheapest edges.
for edge in self.graph.iteredges(): # O(E) time
source = self._uf.find(edge.source)
target = self._uf.find(edge.target)
if source != target: # different components
if edge < min_edges[source]:
min_edges[source] = edge
if edge < min_edges[target]:
min_edges[target] = edge
# Connecting components, total time is O(V).
for edge in min_edges.itervalues():
if edge is dummy_edge: # a disconnected graph
continue
source = self._uf.find(edge.source)
target = self._uf.find(edge.target)
if source != target: # different components
self._uf.union(source, target)
self.mst.add_edge(edge)
# Remove duplicates, total time is O(V).
forest = set(self._uf.find(node) for node in forest)
new_len = len(forest)
if new_len == 1: # a connected graph
break
def to_tree(self):
"""Return the minimum spanning tree."""
return self.mst
class SortedEdgeTSPWithEdges:
"""The sorted edge algorithm (a.k.a. cheapest link) for TSP.
Attributes
----------
graph : input weighted complete graph
hamiltonian_cycle : list of edges
_uf : disjoint-set data structure, private
_pq : priority queue, private
"""
def __init__(self, graph):
"""The algorithm initialization."""
if graph.is_directed():
raise ValueError("the graph is directed")
self.graph = graph
self.hamiltonian_cycle = list()
self._uf = UnionFind()
self._pq = PriorityQueue()
def run(self, source=None):
"""Executable pseudocode."""
degree_dict = dict((node, 0) for node in self.graph.iternodes())
for node in self.graph.iternodes():
self._uf.create(node)
for edge in self.graph.iteredges():
self._pq.put((edge.weight, edge))
while not self._pq.empty():
_, edge = self._pq.get()
degree1 = degree_dict[edge.source]
degree2 = degree_dict[edge.target]
# Zabezpieczam przed skrzyzowaniami T, X, itp.
if degree1 == 2 or degree2 == 2:
continue
if degree1 == 0 or degree2 == 0:
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
continue
# Here degree1 = degree2 = 1.
if self._uf.find(edge.source) != self._uf.find(edge.target):
# Join pieces.
self._uf.union(edge.source, edge.target)
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
elif len(self.hamiltonian_cycle) == self.graph.v()-1:
# Close the Hamiltonian cycle.
self.hamiltonian_cycle.append(edge)
degree_dict[edge.source] += 1
degree_dict[edge.target] += 1
# Create the corrected (directed) list of edges.
# Przeksztalcam cykl Hamiltona do postaci poprawnej listy krawedzi.
# Na razie krawedzie moga miec zla kolejnosc i kierunek.
# Tworze slownik z krawedziami w obie strony.
edge_dict = dict((node, []) for node in self.graph.iternodes())
for edge in self.hamiltonian_cycle: # O(V) time
edge_dict[edge.source].append(edge)
edge_dict[edge.target].append(~edge)
edge = self.hamiltonian_cycle[0]
self.hamiltonian_cycle = [edge]
# Kompletuje kolejne krawedzie.
for step in xrange(self.graph.v()-1): # O(V) time
edge1, edge2 = edge_dict[edge.target]
if edge1.target == edge.source:
self.hamiltonian_cycle.append(edge2)
edge = edge2
else:
self.hamiltonian_cycle.append(edge1)
edge = edge1
