def run(source, destination, show, prediction, seed):
(graph, ways, data) = read_xml(config.osm_path, config.elev_path)
source = sys.argv[1].lower()
dest = sys.argv[2].lower()
try:
source = int(source)
assert source in graph
except:
if source in ways:
source = ways[source][0]
else:
print('Source must be a valid node ID or a street name')
return
try:
dest = int(dest)
assert dest in graph
except:
if dest in ways:
dest = ways[dest][0]
else:
print('Destination must be a valid node ID or a street name')
return
costfunc = None
heuristic = None
if prediction == 'toblers':
costfunc = astar.toblers
heuristic = astar.toblers_heuristic
if prediction in ('linear', 'nearest'):
walks = models.read_walk_data(config.walk_data_path, graph, data)
training_walks, test_walks = models.partition_walks(walks, seed=seed)
train = models.build_dist_elev_examples(training_walks, data)
test = models.build_dist_elev_examples(test_walks, data)
if prediction == 'linear':
model = models.linear_model(train)
costfunc = lambda a, b: model([astar.euclidean(a, b), b.z_m - a.z_m])
heuristic = costfunc
elif prediction == 'nearest':
model = models.nearest_neighbor_model(train)
costfunc = lambda a, b: model([astar.euclidean(a, b), b.z_m - a.z_m])
heuristic = lambda a, b: 0
result = astar.astar(costfunc, heuristic, graph, data, source, dest)
if result is None:
print('No path to destination found')
pathstr = ', '.join((str(nd) for nd in result[0]))
print('\npath: {}\n'.format(pathstr))
print('time: {:.2f} minutes\n'.format(result[1]))
#walk_data = models.read_walk_data(config.walk_data_path, graph, data)
#print(walk_data)
#if seed == None:
# seed = random.random()
#models.compare_models(walk_data, data, seed)
if show:
graphics.display(graph, data, result[0], result[1])
def path_dist(path, id_to_data):
pathdata = [id_to_data[nid] for nid in path]
xydistsum = sum([astar.euclidean(l, r) for l, r in zip(pathdata, pathdata[1:])])
return xydistsum
