def compute_labels_from_problem(self, problem_fname):
""" Using the problem file, compute the labels """
object = io_utils.optObject().read_problem_from_file(problem_fname)
graph = object.get_graph()
return compute_solutions.get_all_optimal_tsp_solutions(graph)
from optlearn import experiment_utils
from optlearn import graph_utils
from optlearn import io_utils
from optlearn import cc_solve
from optlearn.mst import mst_model
files = experiment_utils.list_files_recursive("/home/james/Data/MATHILDA/tsp/")
model = mst_model.mstSparsifier()
ones, twos, threes, fours, fives = [], [], [], [], []
for num, file in enumerate(files):
object = io_utils.optObject().read_from_file(file)
graph = object.get_graph()
graph = graph_utils.delete_self_weights(graph)
graph = graph.to_undirected()
solution = cc_solve.solution_from_path(file)
edges = graph_utils.get_tour_edges(solution.tour)
if np.min(solution.tour) < 1 and np.min(graph.nodes) > 0:
edges += 1
edges = model.tuples_to_strings(edges)
features = model.run_sparsify(graph, 15)
features = [model.tuples_to_strings(item) for item in features]
one = np.sum([edge in features[0] for edge in edges])
two = np.sum([edge in features[1] for edge in edges]) + one
def load_object(self, problem_fname):
""" Load the problem file and store it as a hidden attribute """
self._object = io_utils.optObject()
self._object.read_problem_from_file(problem_fname)
self._graph = self._object.get_graph()
def load_problem(problem_path):
""" Given the path to a tsp problem, load it as a networkx complete graph """
object = io_utils.optObject().read_problem_from_file(problem_path)
return object.get_graph()