def test():
timings = []
repetitions = []
are_the_same = []
instances = tools.load_instances("../data/new_instances_25_30_35_40.json")
limit = 50
for i, instance in enumerate(instances['easy40'][:limit], start=1):
puzzle, dataset_solution = instance['puzzle'], instance['solution']
possibilities = _precompute_possibilities(puzzle)
print(f'Starting instance #{i}')
start = time.time()
dfs_solution = _dfs(puzzle, possibilities)
end = time.time()
time_taken = end - start
timings.append(time_taken)
reps = _validate(dfs_solution)
repetitions.append(reps)
same = (dfs_solution == dataset_solution).all()
are_the_same.append(same)
print(
f'Solution found by DFS after {time_taken} seconds, {reps} repetitions, are solutions the same: {same}'
)
print(' Timings:', timings)
print(' Repetitions:', repetitions)
print('Are the same:', are_the_same)
def run(cfg: config.DFSConfig) -> None:
sudoku = cfg.sudoku_instance
possibilities = _precompute_possibilities(sudoku)
start = time.time()
solution = _dfs(sudoku, possibilities)
end = time.time()
print('solution found by DFS in {} seconds:'.format(end - start))
print(solution)
print('repetitions:', _validate(solution))
print('solution from dataset:')
print(cfg.sudoku_solution)
print('are they the same?', solution == cfg.sudoku_solution)
def final_tests():
instances = tools.load_instances('../data/new_instances_25_30_35_40.json')
instances['01_easy'] = instances['easy40']
del instances['easy40']
instances['02_medium'] = instances['medium35']
del instances['medium35']
instances['03_advanced'] = instances['advanced30']
del instances['advanced30']
instances['04_hard'] = instances['hard25']
del instances['hard25']
instances_to_test = {
'01_easy': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14],
'02_medium':
[50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64],
'03_advanced': [100, 101, 102, 103, 104, 105, 106, 107, 108],
'04_hard': [150, 151, 152, 153, 154, 155, 156, 157, 158, 159],
}
sublist = lambda data, ix: [data[index] for index in ix]
time_results = {
'01_easy': {},
'02_medium': {},
'03_advanced': {},
'04_hard': {},
}
for difficulty_level, instances_ids in instances_to_test.items():
for instance_true_id in instances_ids:
index = instance_true_id % len(instances[difficulty_level])
instance = instances[difficulty_level][index]
puzzle = instance['puzzle']
print(f'Starting instance #{instance_true_id}')
start = time.time()
puzzle, possibilities = _preprocess_puzzle(puzzle)
dfs_solution = _dfs(puzzle, possibilities)
end = time.time()
time_taken = end - start
time_results[difficulty_level][instance_true_id] = time_taken
reps = _validate(dfs_solution)
if reps > 0:
print(f'instance #{instance_true_id} was not solved properly!')
print(dfs_solution)
with open('dfs_report.yaml', 'w') as f:
yaml.dump(time_results, f)
def run(cfg: config.DFSConfig) -> None:
sudoku = cfg.sudoku_instance
print('not preprocessed sudoku:', np.count_nonzero(sudoku == 0),
'slots to fill')
preproccessed_sudoku, _ = _preprocess_puzzle(sudoku)
print('preprocessed sudoku:', np.count_nonzero(preproccessed_sudoku == 0),
'slots to fill')
return
start = time.time()
sudoku, possibilities = _preprocess_puzzle(sudoku)
solution = _dfs(sudoku, possibilities)
end = time.time()
print('solution found by DFS in {} seconds:'.format(end - start))
print(solution)
print('repetitions:', _validate(solution))
print('solution from dataset:')
print(cfg.sudoku_solution)
print('are they the same?', solution == cfg.sudoku_solution)
def test_evaluation_correct(self):
self.assertEqual(_validate(self.correct_sudoku), 0)
def test_evaluation_bad_columns(self):
self.assertEqual(_validate(self.sudoku_with_bad_columns),
9 * self.row_error + 9 * self.square_error)