Ticket #9910: trac-9910_reviewer.patch

File trac-9910_reviewer.patch, 21.4 KB (added by mvngu, 10 years ago)

  • sage/graphs/generic_graph.py 

    # HG changeset patch
# User Minh Van Nguyen <nguyenminh2@gmail.com>
# Date 1290885462 28800
# Node ID 9a5f569d00d280dfd09ba4f2417ab8745d5c6b3c
# Parent  9c9495ebab27b9b708c83c50bf7b3d88c0a903db
#9910: reviewer patch: longest path
Changes include:

* Many typo fixes.
* Proper ReST formatting.
* Use `range()` instead of `xrange()` to conform to Python 3.x.
* Use the function `print()` instead of the statement `print`.
* Make the doctests and code conform to Python coding conventions.
* When comparing an object with `None`, use the idiom "is not" instead
  of "!=".
* Doctests to take care of the case of the trivial graph.
* Remove superfluous trailing white spaces. These are mostly harmless,
  but in some cases they can result in hard to detect bugs.
* Handle the case where the argument `algorithm` is not assigned either
  "backtrack" or "MILP".
* Some general clean-ups.

diff --git a/sage/graphs/generic_graph.py b/sage/graphs/generic_graph.py
    a b  
    43744374
    43754375            return val
    43764376
    4377     def longest_path(self, s = None, t = None, weighted = False, algorithm = "MILP", solver = None, verbose = 0):
     4377    def longest_path(self, s=None, t=None, weighted=False, algorithm="MILP", solver=None, verbose=0):
    43784378        r"""
    43794379        Returns a longest path of ``self``.
    43804380
    … …  
    43864386        - ``t`` (vertex) -- forces the destination of the path. Set to
    43874387          ``None`` by default.
    43884388
    4389         - ``weighted`` (boolean) -- whether the label on the edges are
    4390           tobe considered as weights (a label set to ``None`` or
     4389        - ``weighted`` (boolean) -- whether the labels on the edges are
     4390          to be considered as weights (a label set to ``None`` or
    43914391          ``{}`` being considered as a weight of `1`). Set to
    43924392          ``False`` by default.
    43934393
    43944394        - ``algorithm`` -- one of ``"MILP"`` (default) or
    4395           ``"backtrack"``. Two remarks on this respect :
     4395          ``"backtrack"``. Two remarks on this respect:
    43964396
    43974397              * While the MILP formulation returns an exact answer,
    43984398                the backtrack algorithm is a randomized heuristic.
    43994399
    44004400              * As the backtrack algorithm does not support edge
    4401                 weighting, setting ``weighted`` to ``True`` will force
     4401                weighting, setting ``weighted=True`` will force
    44024402                the use of the MILP algorithm.
    44034403
    4404         - ``solver`` -- (default: ``None``) Specify a Linear Program (LP)
     4404        - ``solver`` -- (default: ``None``) Specify the Linear Program (LP)
    44054405          solver to be used. If set to ``None``, the default one is used. For
    44064406          more information on LP solvers and which default solver is used, see
    44074407          the method
    … …  
    44414441        The heuristic totally agrees::
    44424442
    44434443            sage: g = graphs.PetersenGraph()
    4444             sage: g.longest_path(algorithm = "backtrack").edges()
     4444            sage: g.longest_path(algorithm="backtrack").edges()
    44454445            [(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)]
    44464446
    4447         Let us compute longest path on random graphs with random
    4448         weights. We each time ensure the given graph is indeed a
     4447        Let us compute longest paths on random graphs with random
     4448        weights. Each time, we ensure the resulting graph is indeed a
    44494449        path::
    44504450
    4451             sage: for i in xrange(20):
    4452             ...       g = graphs.RandomGNP(15,.3)
    4453             ...       for u,v in g.edges(labels = False):
    4454             ...          g.set_edge_label(u,v,random())
     4451            sage: for i in range(20):
     4452            ...       g = graphs.RandomGNP(15, 0.3)
     4453            ...       for u, v in g.edges(labels=False):
     4454            ...           g.set_edge_label(u, v, random())
    44554455            ...       lp = g.longest_path()
    44564456            ...       if (not lp.is_forest() or
    44574457            ...           not max(lp.degree()) <= 2 or
    44584458            ...           not lp.is_connected()):
    4459             ...           print "Error !"
     4459            ...           print("Error!")
    44604460            ...           break
    4461             ...
    4462             sage: print "Test finished !"
    4463             Test finished !
    4464 
    4465         TESTS::
     4461
     4462        TESTS:
     4463
     4464        The argument ``algorithm`` must be either ``'backtrack'`` or
     4465        ``'MILP'``::
     4466
     4467            sage: graphs.PetersenGraph().longest_path(algorithm="abc")
     4468            Traceback (most recent call last):
     4469            ...
     4470            ValueError: algorithm must be either 'backtrack' or 'MILP'
    44664471
    44674472        Disconnected graphs not weighted::
    44684473       
    … …  
    44764481        Disconnected graphs weighted::
    44774482
    44784483            sage: g1 = graphs.PetersenGraph()
    4479             sage: for u,v in g.edges(labels = False):
    4480             ...       g.set_edge_label(u,v,random())
     4484            sage: for u,v in g.edges(labels=False):
     4485            ...       g.set_edge_label(u, v, random())
    44814486            sage: g2 = 2 * g1
    4482             sage: lp1 = g1.longest_path(weighted = True)
    4483             sage: lp2 = g2.longest_path(weighted = True)
     4487            sage: lp1 = g1.longest_path(weighted=True)
     4488            sage: lp2 = g2.longest_path(weighted=True)
    44844489            sage: lp1[0] == lp2[0]
    44854490            True
    44864491
    … …  
    44884493
    44894494            sage: Graph().longest_path()
    44904495            Graph on 0 vertices
    4491             sage: Graph().longest_path(weighted = True)
     4496            sage: Graph().longest_path(weighted=True)
    44924497            [0, Graph on 0 vertices]
     4498            sage: graphs.EmptyGraph().longest_path()
     4499            Graph on 0 vertices
     4500            sage: graphs.EmptyGraph().longest_path(weighted=True)
     4501            [0, Graph on 0 vertices]
     4502
     4503        Trivial graphs::
     4504
     4505            sage: G = Graph()
     4506            sage: G.add_vertex(0)
     4507            sage: G.longest_path()
     4508            Graph on 0 vertices
     4509            sage: G.longest_path(weighted=True)
     4510            [0, Graph on 0 vertices]
     4511            sage: graphs.CompleteGraph(1).longest_path()
     4512            Graph on 0 vertices
     4513            sage: graphs.CompleteGraph(1).longest_path(weighted=True)
     4514            [0, Graph on 0 vertices]
    44934515
    44944516        Random test for digraphs::
    44954517
    4496             sage: for i in xrange(20):
    4497             ...       g = digraphs.RandomDirectedGNP(15,.3)
    4498             ...       for u,v in g.edges(labels = False):
    4499             ...          g.set_edge_label(u,v,random())
     4518            sage: for i in range(20):
     4519            ...       g = digraphs.RandomDirectedGNP(15, 0.3)
     4520            ...       for u, v in g.edges(labels=False):
     4521            ...           g.set_edge_label(u, v, random())
    45004522            ...       lp = g.longest_path()
    45014523            ...       if (not lp.is_directed_acyclic() or
    45024524            ...           not max(lp.out_degree()) <= 1 or
    45034525            ...           not max(lp.in_degree()) <= 1 or
    45044526            ...           not lp.is_connected()):
    4505             ...           print "Error !"
     4527            ...           print("Error!")
    45064528            ...           break
    4507             ...
    4508             sage: print "Test finished !"
    4509             Test finished !
    45104529        """
    45114530        if weighted:
    45124531            algorithm = "MILP"
    4513 
     4532        if algorithm not in ("backtrack", "MILP"):
     4533            raise ValueError("algorithm must be either 'backtrack' or 'MILP'")
    45144534
    45154535        # Quick improvement
    45164536        if not self.is_connected():
    45174537            if weighted:
    4518                 return max( g.longest_path(s = s, t = t,
    4519                                            weighted = weighted,
    4520                                            algorithm = algorithm)
    4521                             for g in self.connected_components_subgraphs() )
    4522             else:
    4523                 return max( (g.longest_path(s = s, t = t,
    4524                                             weighted = weighted,
    4525                                             algorithm = algorithm)
    4526                             for g in self.connected_components_subgraphs() ),
    4527                             key = lambda x : x.order() )
     4538                return max(g.longest_path(s=s, t=t,
     4539                                          weighted=weighted,
     4540                                          algorithm=algorithm)
     4541                           for g in self.connected_components_subgraphs())
     4542            else:
     4543                return max((g.longest_path(s=s, t=t,
     4544                                           weighted=weighted,
     4545                                           algorithm=algorithm)
     4546                            for g in self.connected_components_subgraphs()),
     4547                           key=lambda x: x.order())
    45284548
    45294549        # Stupid cases
    4530 
    4531         # - Graph having less than 1 vertex
     4550        # - Graph having <= 1 vertex.
    45324551        #
    45334552        # - The source has outdegree 0 in a directed graph, or
    4534         #   degree 0, or is not a vertex of the graph
     4553        #   degree 0, or is not a vertex of the graph.
    45354554        #
    45364555        # - The destination has indegree 0 in a directed graph, or
    4537         #   degree 0, or is not a vertex of the graph
     4556        #   degree 0, or is not a vertex of the graph.
    45384557        #
    45394558        # - Both s and t are specified, but there is no path between
    4540         #   the two in a directed graph (the graph is connected)
    4541 
     4559        #   the two in a directed graph (the graph is connected).
    45424560        if (self.order() <= 1 or
    4543             (s != None and (
     4561            (s is not None and (
    45444562                    (s not in self) or
    45454563                    (self._directed and self.degree_out(s) == 0) or
    45464564                    (not self._directed and self.degree(s) == 0))) or
    4547 
    4548             (t != None and (
     4565            (t is not None and (
    45494566                    (t not in self) or
    45504567                    (self._directed and self.degree_in(t) == 0) or
    45514568                    (not self._directed and self.degree(t) == 0))) or
    4552 
    4553             (self._directed and s != None and t != None and
    4554              len(self.shortest_path(s,t) == 0))):
    4555 
     4569            (self._directed and (s is not None) and (t is not None) and
     4570             len(self.shortest_path(s, t) == 0))):
    45564571            if self._directed:
    45574572                from sage.graphs.all import DiGraph
    45584573                return [0, DiGraph()] if weighted else DiGraph()
    4559             else:
    4560                 from sage.graphs.all import Graph
    4561                 return [0, Graph()] if weighted else Graph()
    4562 
     4574            from sage.graphs.all import Graph
     4575            return [0, Graph()] if weighted else Graph()
    45634576
    45644577        # Calling the backtrack heuristic if asked
    45654578        if algorithm == "backtrack":
    45664579            from sage.graphs.generic_graph_pyx import find_hamiltonian as fh
    4567             x = fh( self, find_path = True )[1]
    4568             return self.subgraph(vertices = x,
    4569                                  edges = zip(x[:-1], x[1:]))
     4580            x = fh(self, find_path=True)[1]
     4581            return self.subgraph(vertices=x, edges=zip(x[:-1], x[1:]))
    45704582
    45714583        ##################
    45724584        # LP Formulation #
    … …  
    45784590
    45794591        # Associating a weight to a label
    45804592        if weighted:
    4581             weight = lambda x : x if (x is not None and x != {}) else 1
    4582         else:
    4583             weight = lambda x : 1
     4593            weight = lambda x: x if (x is not None and x != {}) else 1
     4594        else:
     4595            weight = lambda x: 1
    45844596
    45854597        from sage.numerical.mip import MixedIntegerLinearProgram, Sum
    4586 
    45874598        p = MixedIntegerLinearProgram()
    45884599       
    45894600        # edge_used[(u,v)] == 1 if (u,v) is used
    4590         edge_used = p.new_variable(binary = True)
     4601        edge_used = p.new_variable(binary=True)
    45914602
    45924603        # relaxed version of the previous variable, to prevent the
    45934604        # creation of cycles
    45944605        r_edge_used = p.new_variable()
    45954606
    45964607        # vertex_used[v] == 1 if vertex v is used
    4597         vertex_used = p.new_variable(binary = True)
    4598 
    4599         if self._directed:
    4600 
     4608        vertex_used = p.new_variable(binary=True)
     4609
     4610        if self._directed:
    46014611            # if edge uv is used, vu can not be
    4602             for u,v in self.edges(labels = False):
    4603                 if self.has_edge(v,u):
    4604                     p.add_constraint(edge_used[(u,v)] + edge_used[(v,u)], max = 1)
    4605            
    4606            
     4612            for u, v in self.edges(labels=False):
     4613                if self.has_edge(v, u):
     4614                    p.add_constraint(edge_used[(u,v)] + edge_used[(v,u)], max=1)
    46074615            # A vertex is used if one of its incident edges is
    46084616            for v in self:
    4609                 for e in self.incoming_edges(labels = False):
    4610                     p.add_constraint(vertex_used[v] - edge_used[e], min = 0)
    4611                 for e in self.outgoing_edges(labels = False):
    4612                     p.add_constraint(vertex_used[v] - edge_used[e], min = 0)
    4613 
     4617                for e in self.incoming_edges(labels=False):
     4618                    p.add_constraint(vertex_used[v] - edge_used[e], min=0)
     4619                for e in self.outgoing_edges(labels=False):
     4620                    p.add_constraint(vertex_used[v] - edge_used[e], min=0)
    46144621            # A path is a tree. If n vertices are used, at most n-1 edges are
    4615             p.add_constraint(Sum( vertex_used[v] for v in self)
    4616                              - Sum( edge_used[e] for e in self.edges(labels = False)),
    4617                              min = 1, max = 1)
    4618 
     4622            p.add_constraint(
     4623                Sum(vertex_used[v] for v in self)
     4624                - Sum(edge_used[e] for e in self.edges(labels=False)),
     4625                min=1, max=1)
    46194626            # A vertex has at most one incoming used edge and at most
    46204627            # one outgoing used edge
    46214628            for v in self:
    4622                 p.add_constraint( Sum( edge_used[(u,v)] for u in self.neighbors_in(v) ), max = 1)
    4623                 p.add_constraint( Sum( edge_used[(v,u)] for u in self.neighbors_out(v) ), max = 1)
    4624 
     4629                p.add_constraint(
     4630                    Sum(edge_used[(u,v)] for u in self.neighbors_in(v)),
     4631                    max=1)
     4632                p.add_constraint(
     4633                    Sum(edge_used[(v,u)] for u in self.neighbors_out(v)),
     4634                    max=1)
    46254635            # r_edge_used is "more" than edge_used, though it ignores
    46264636            # the direction
    4627             for u,v in self.edges(labels = False):
    4628                 p.add_constraint(  r_edge_used[(u,v)]
    4629                                  + r_edge_used[(v,u)]
    4630                                  - edge_used[(u,v)],
    4631                                    min = 0)
    4632 
     4637            for u, v in self.edges(labels=False):
     4638                p.add_constraint(r_edge_used[(u,v)]
     4639                                 + r_edge_used[(v,u)]
     4640                                 - edge_used[(u,v)],
     4641                                 min=0)
    46334642            # No cycles
    46344643            for v in self:
    4635                 p.add_constraint( Sum( r_edge_used[(u,v)] for u in self.neighbors(v)), max = 1-epsilon)
    4636 
     4644                p.add_constraint(
     4645                    Sum(r_edge_used[(u,v)] for u in self.neighbors(v)),
     4646                    max=1-epsilon)
    46374647            # Enforcing the source if asked.. If s is set, it has no
    46384648            # incoming edge and exactly one son
    4639             if s!=None:
    4640                 p.add_constraint( Sum( edge_used[(u,s)] for u in self.neighbors_in(s)), max = 0, min = 0)
    4641                 p.add_constraint( Sum( edge_used[(s,u)] for u in self.neighbors_out(s)), min = 1, max = 1)
    4642 
     4649            if s is not None:
     4650                p.add_constraint(
     4651                    Sum(edge_used[(u,s)] for u in self.neighbors_in(s)),
     4652                    max=0, min=0)
     4653                p.add_constraint(
     4654                    Sum(edge_used[(s,u)] for u in self.neighbors_out(s)),
     4655                    min=1, max=1)
    46434656            # Enforcing the destination if asked.. If t is set, it has
    46444657            # no outgoing edge and exactly one parent
    4645             if t!=None:
    4646                 p.add_constraint( Sum( edge_used[(u,t)] for u in self.neighbors_in(t)), min = 1, max = 1)
    4647                 p.add_constraint( Sum( edge_used[(t,u)] for u in self.neighbors_out(t)), max = 0, min = 0)
    4648 
    4649                
     4658            if t is not None:
     4659                p.add_constraint(
     4660                    Sum(edge_used[(u,t)] for u in self.neighbors_in(t)),
     4661                    min=1, max=1)
     4662                p.add_constraint(
     4663                    Sum(edge_used[(t,u)] for u in self.neighbors_out(t)),
     4664                    max=0, min=0)
    46504665            # Defining the objective
    4651             p.set_objective( Sum( weight(l) * edge_used[(u,v)] for u,v,l in self.edges() ) )
    4652            
    4653         else:
    4654 
     4666            p.set_objective(
     4667                Sum(weight(l) * edge_used[(u,v)] for u, v, l in self.edges()))
     4668        else:
    46554669            # f_edge_used calls edge_used through reordering u and v
    46564670            # to avoid having two different variables
    4657             f_edge_used = lambda u,v : edge_used[ (u,v) if hash(u) < hash(v) else (v,u) ]
    4658 
     4671            f_edge_used = lambda u, v: edge_used[
     4672                (u,v) if hash(u) < hash(v) else (v,u)]
    46594673            # A vertex is used if one of its incident edges is
    46604674            for v in self:
    46614675                for u in self.neighbors(v):
    4662                     p.add_constraint(vertex_used[v] - f_edge_used(u,v), min = 0)
    4663 
     4676                    p.add_constraint(vertex_used[v] - f_edge_used(u,v), min=0)
    46644677            # A path is a tree. If n vertices are used, at most n-1 edges are
    4665             p.add_constraint( Sum( vertex_used[v] for v in self)
    4666                               - Sum( f_edge_used(u,v) for u,v in self.edges(labels = False)),
    4667                               min = 1, max = 1)
    4668 
     4678            p.add_constraint(
     4679                Sum(vertex_used[v] for v in self)
     4680                - Sum(f_edge_used(u,v) for u, v in self.edges(labels=False)),
     4681                min=1, max=1)
    46694682            # A vertex has at most two incident edges used
    46704683            for v in self:
    4671                 p.add_constraint( Sum( f_edge_used(u,v) for u in self.neighbors(v) ), max = 2)
    4672 
     4684                p.add_constraint(
     4685                    Sum(f_edge_used(u,v) for u in self.neighbors(v)), max=2)
    46734686            # r_edge_used is "more" than edge_used
    4674             for u,v in self.edges(labels = False):
    4675                 p.add_constraint(  r_edge_used[(u,v)]
    4676                                  + r_edge_used[(v,u)]
    4677                                  - f_edge_used(u,v),
    4678                                    min = 0)
    4679 
     4687            for u, v in self.edges(labels=False):
     4688                p.add_constraint(r_edge_used[(u,v)]
     4689                                 + r_edge_used[(v,u)]
     4690                                 - f_edge_used(u,v),
     4691                                 min=0)
    46804692            # No cycles
    46814693            for v in self:
    4682                 p.add_constraint( Sum( r_edge_used[(u,v)] for u in self.neighbors(v)), max = 1-epsilon)
    4683 
     4694                p.add_constraint(
     4695                    Sum(r_edge_used[(u,v)] for u in self.neighbors(v)),
     4696                    max=1-epsilon)
    46844697            # Enforcing the destination if asked.. If s or t are set,
    46854698            # they have exactly one incident edge
    4686             if s != None:
    4687                 p.add_constraint( Sum( f_edge_used(s,u) for u in self.neighbors(s) ), max = 1, min = 1)
    4688             if t != None:
    4689                 p.add_constraint( Sum( f_edge_used(t,u) for u in self.neighbors(t) ), max = 1, min = 1)
    4690 
     4699            if s is not None:
     4700                p.add_constraint(
     4701                    Sum(f_edge_used(s,u) for u in self.neighbors(s)),
     4702                    max=1, min=1)
     4703            if t is not None:
     4704                p.add_constraint(
     4705                    Sum(f_edge_used(t,u) for u in self.neighbors(t)),
     4706                    max=1, min=1)
    46914707            # Defining the objective
    4692             p.set_objective( Sum( weight(l) * f_edge_used(u,v) for u,v,l in self.edges() ) )
    4693 
     4708            p.set_objective(Sum(weight(l) * f_edge_used(u,v)
     4709                                for u, v, l in self.edges()))
    46944710        # Computing the result. No exception has to be raised, as this
    46954711        # problem always has a solution (there is at least one edge,
    4696         # and a path from s to t if they are specified
    4697 
    4698         p.solve(solver = solver, log = verbose)
    4699 
     4712        # and a path from s to t if they are specified).
     4713        p.solve(solver=solver, log=verbose)
    47004714        edge_used = p.get_values(edge_used)
    47014715        vertex_used = p.get_values(vertex_used)
    4702 
    4703         if self._directed:
    4704             g = self.subgraph( vertices =
    4705                                   (v for v in self if vertex_used[v] >= .5),
    4706                                edges =
    4707                                   ( (u,v,l) for u,v,l in self.edges()
    4708                                     if edge_used[(u,v)] >= .5 ))
    4709         else:
    4710             g = self.subgraph( vertices =
    4711                                   (v for v in self if vertex_used[v] >= .5),
    4712                                edges =
    4713                                   ( (u,v,l) for u,v,l in self.edges()
    4714                                     if f_edge_used(u,v) >= .5 ))
    4715 
     4716        if self._directed:
     4717            g = self.subgraph(
     4718                vertices=(v for v in self if vertex_used[v] >= 0.5),
     4719                edges=((u,v,l) for u, v, l in self.edges()
     4720                       if edge_used[(u,v)] >= 0.5))
     4721        else:
     4722            g = self.subgraph(
     4723                vertices=(v for v in self if vertex_used[v] >= 0.5),
     4724                edges=((u,v,l) for u, v, l in self.edges()
     4725                       if f_edge_used(u,v) >= 0.5))
    47164726        if weighted:
    4717             return sum(map(weight,g.edge_labels())), g
     4727            return sum(map(weight, g.edge_labels())), g
    47184728        else:
    47194729            return g
    47204730
    4721 
    47224731    def traveling_salesman_problem(self, weighted = True, solver = None, verbose = 0):
    47234732        r"""
    47244733        Solves the traveling salesman problem (TSP)