def cpp_test(G, v_I, v_W, v_F, min_cost):
res = Mstar(G, v_I, v_W, v_F, False).solve()
if res[1] < min_cost:
print(res[1])
assert res[1] < min_cost or math.isclose(res[1], min_cost)
paths = res[0]
for i in range(len(v_I)):
assert v_I[i] == list(paths[i][0])
for i in range(len(v_W)):
if len(v_W[i]) > 0:
for waypoint in v_W[i]:
if waypoint != [-1, -1]:
assert tuple(waypoint) in paths[i]
for i in range(len(v_F)):
assert v_F[i] == list(paths[i][-1])
@pytest.mark.parametrize(
"test_id",
get_all_benchmarks(without=[10, 19, 21, 23, 24, 54, 58, 59, 61]))
def test_cpp_benchmark(test_id):
benchmarker = MapfwBenchmarker("42cf6ce8D2A5B954", test_id, "M*", "Test",
True)
for problem in benchmarker:
cpp_waypoints = problem.waypoints
for i in range(len(cpp_waypoints)):
if len(cpp_waypoints[i]) == 0:
cpp_waypoints[i] = [[-1, -1]]
cpp_test(problem.grid, problem.starts, cpp_waypoints, problem.goals,
min_cost[test_id])
neighbours = []
if i != 0 and problem.grid[i - 1][j] == 0:
neighbours.append((j, i - 1))
if j != 0 and problem.grid[i][j - 1] == 0:
neighbours.append((j - 1, i))
if i != problem.height - 1 and problem.grid[i + 1][j] == 0:
neighbours.append((j, i + 1))
if j != problem.width - 1 and problem.grid[i][j + 1] == 0:
neighbours.append((j + 1, i))
grap_new[current] = neighbours
# Create V_I and v_F
v_I = tuple(tuple(start) for start in problem.starts)
v_W = []
for agent_waypoints in problem.waypoints:
if len(agent_waypoints) > 0:
v_W_i = []
for waypoint in agent_waypoints:
v_W_i.append(tuple(tuple(waypoint)))
v_W.append(v_W_i)
else:
v_W.append(())
v_W = tuple(v_W)
v_F = tuple(tuple(target) for target in problem.goals)
return grap_new, v_I, v_W, v_F
@pytest.mark.parametrize("test_id", get_all_benchmarks(without=[10, 19, 21, 23, 24, 54, 58, 59, 61]))
def test_python_benchmark(test_id):
benchmarker = MapfwBenchmarker("42cf6ce8D2A5B954", test_id, "M*", "Test", True)
for problem in benchmarker:
graph, v_I, v_W, v_F = setup_benchmark(problem)
python_test(graph, v_I, v_W, v_F, min_cost[test_id])
77, 54, 112000, 18, np.nan, 210, 7030, 88, 857, np.nan, 4280, 92, 11990,
13760, 516, np.nan, 1980, np.nan, 7, 6660, 154
]
MLA_times = [
1, 1, 6, np.nan, 53, 1, 35, 134, 228, 385, 7, 12, 2, 1, 13, 17, 8, 3, 861,
np.nan, 20, 23, 45, 244, 14, 9, 27, np.nan, np.nan, 104, np.nan, 246,
np.nan, 6, np.nan
]
BCP_times = [
11, 172, 13, 45, 341, 28, np.nan, np.nan, np.nan, 30940, 17, 49, 30,
np.nan, 30, 27, 385, 24, np.nan, 15, 38, 9, 922, 1460, 494, 14, 2320, 8710,
67, 2549000, 247, 653000, 12, 599, 6930
]
y_axis = ["WM*", "inflated WM*", "CBS", "MLA*", "A*+OD+ID", "BCP"]
x_axis = sorted(get_all_benchmarks(without=[54, 58, 77, 78]))
zipper = zip(Mstar_times, Mstar_inflated_times, CBS_times, MLA_times,
Astar_times, BCP_times)
data = []
remove = []
for i in range(len(y_axis)):
data.append([])
for i, benchmark in enumerate(zipper):
optimal = np.nanmin(benchmark)
if all(val == optimal or np.isnan(val) for val in benchmark):
remove.append(i)
else:
for i in range(len(benchmark)):
if (benchmark[i] - optimal) == 0:
data[i].append(1)