def constructGraph(self, v,w, D_T_T, D_F_F):
G = Graph()
V = {}
V['s'] = 0
i = 1
for v_i in v.getChildren():
V["v" + str(v_i.getHashNum())] = i
i += 1
V['e_i'] = i
i += 1
for w_i in w.getChildren():
V["w" + str(w_i.getHashNum())] = i
i += 1
V['e_j'] = i
i += 1
V['t'] = i
G.V = range(len(V))
G.capacity = numpy.zeros([len(V),len(V)])
for v_i in v.getChildren():
G.capacity[V['s'],V["v" + str(v_i.getHashNum())]] = 1
for w_i in w.getChildren():
G.capacity[V["v" + str(v_i.getHashNum())],V["w" + str(w_i.getHashNum())]] = 1
G.capacity[V["v" + str(v_i.getHashNum())],V['e_j']] = 1
for w_i in w.getChildren():
G.capacity[V["w" + str(w_i.getHashNum())], V['t']] = 1
G.capacity[V['e_i'],V["w" + str(w_i.getHashNum())]] = 1
G.capacity[V['s'],V['e_i']] = n_j = len(w.getChildren())
G.capacity[V['e_j'],V['t']] = n_i = len(v.getChildren())
G.capacity[V['e_i'],V['e_j']] = max(n_i, n_j) - min(n_i, n_j)
G.cost = numpy.zeros([len(V),len(V)])
for v_i in v.getChildren():
for w_i in w.getChildren():
G.cost[V["v" + str(v_i.getHashNum())], V["w" + str(w_i.getHashNum())]] = self.getConstrainedTED(v_i, w_i, D_T_T, D_F_F)
G.cost[V['e_i'], V["w" + str(w_i.getHashNum())]] = self.getConstrainedTED(Tree.getEmptyTree(), w_i, D_T_T, D_F_F)
G.cost[V["v" + str(v_i.getHashNum())], V['e_j']] = self.getConstrainedTED(v_i, Tree.getEmptyTree(), D_T_T, D_F_F)
G.init_edges()
G.init_cost_rev()
return (G,V)
def getConstrainedTED(self, T1, T2, D_T_T, D_F_F): n1 = n2 = 0 #hashnum of empty node D_T_T[n1,n2] = 0 D_F_F[n1,n2] = 0 if D_T_T[T1.getHashNum(), T2.getHashNum()] != -1: return D_T_T[T1.getHashNum(), T2.getHashNum()] for node in T1.getVertices(): d = 0 for c in node.getChildren(): D_T_T[c.getHashNum(), n2] = self.getConstrainedTED(c, Tree.getEmptyTree(), D_T_T, D_F_F) d += D_T_T[c.getHashNum(), n2] D_F_F[node.getHashNum(), n2] = d D_T_T[node.getHashNum(), n2] = D_F_F[node.getHashNum(), n2] + self.remove_cost for node in T2.getVertices(): d = 0 for c in node.getChildren(): D_T_T[n1, c.getHashNum()] = self.getConstrainedTED(Tree.getEmptyTree(), c, D_T_T, D_F_F) d += D_T_T[n2, c.getHashNum()] D_F_F[n1, node.getHashNum()] = d D_T_T[n1, node.getHashNum()] = D_F_F[n1, node.getHashNum()] + self.add_cost for v in T1.getVertices(): for w in T2.getVertices(): if v.getChildren() == [] and w.getChildren() == []: #D_F_F[v.getHashNum(), w.getHashNum()] = 0 D_F_F[v.getHashNum(), w.getHashNum()] = D_F_F[n1, n2] elif v.getChildren() == [] and w.getChildren() != []: #D_F_F[v.getHashNum(), w.getHashNum()] = sum([getConstrainedTED(Tree.getEmptyTree(), w_i, D_T_T, D_F_F) for w_i in w.getChildren()]) D_F_F[v.getHashNum(), w.getHashNum()] = D_F_F[n1, w.getHashNum()] elif v.getChildren() != [] and w.getChildren() == []: #D_F_F[v.getHashNum(), w.getHashNum()] = sum([getConstrainedTED(v_i, Tree.getEmptyTree(), D_T_T, D_F_F) for v_i in v.getChildren()]) D_F_F[v.getHashNum(), w.getHashNum()] = D_F_F[v.getHashNum(), n2] else: d1 = D_F_F[n1, w.getHashNum()] + min([D_F_F[v.getHashNum(), k.getHashNum()] - D_F_F[n1, k.getHashNum()] for k in w.getChildren()]) d2 = D_F_F[v.getHashNum(), n2] + min([D_F_F[k.getHashNum(), w.getHashNum()] - D_F_F[k.getHashNum(), n2] for k in v.getChildren()]) d3 = self.minMM(v,w, D_T_T, D_F_F) D_F_F[v.getHashNum(), w.getHashNum()] = min(d1,d2,d3) if w.getChildren() != []: d4 = self.getConstrainedTED(Tree.getEmptyTree(), w, D_T_T, D_F_F) + numpy.min([self.getConstrainedTED(v, k, D_T_T, D_F_F) - self.getConstrainedTED(Tree.getEmptyTree(), k, D_T_T, D_F_F) for k in w.getChildren()]) else: d4 = self.getConstrainedTED(Tree.getEmptyTree(), w, D_T_T, D_F_F) + self. getConstrainedTED(v, Tree.getEmptyTree(), D_T_T, D_F_F) if v.getChildren() != []: d5 = self.getConstrainedTED(v, Tree.getEmptyTree(), D_T_T, D_F_F) + numpy.min([self.getConstrainedTED(k, w, D_T_T, D_F_F) - self.getConstrainedTED(k, Tree.getEmptyTree(), D_T_T, D_F_F) for k in v.getChildren()]) else: d5 = self.getConstrainedTED(v, Tree.getEmptyTree(), D_T_T, D_F_F) + self.getConstrainedTED(Tree.getEmptyTree(), w, D_T_T, D_F_F) d6 = D_F_F[v.getHashNum(), w.getHashNum()] + self.getRelabelCost(v, w) D_T_T[v.getHashNum(), w.getHashNum()] = min(d4, d5, d6) return D_T_T[T1.getHashNum(), T2.getHashNum()]
