def generate_dataset_and_parser():
folder_name = ''.join(random.choice(string.ascii_lowercase) for _ in range(10))
generator = IndependentGenerator(alphabet_size=ALPHABET_SIZE, seed=0)
edit_dataset_name = folder_name + '/test_ed.pkl'
strings = [generate_random_dna(50)] + [generate_random_dna(random.randint(10, 50)) for _ in range(19)]
strings_dict = {'train': strings[:10], 'val': strings[10:15], 'test': strings[15:]}
edit_dataset = EditDistanceGenomicDatasetGenerator(strings=strings_dict)
edit_dataset.save_as_pickle(edit_dataset_name)
parser = general_arg_parser()
args = parser.parse_args()
args.data = edit_dataset_name
args.epochs = 2
args.print_every = 1
args.construct_msa_tree = 'True'
return folder_name, edit_dataset_name, args
def generate_dataset_and_parser():
folder_name = ''.join(
random.choice(string.ascii_lowercase) for _ in range(10))
generator = IndependentGenerator(alphabet_size=ALPHABET_SIZE, seed=0)
edit_dataset_name = folder_name + '/test_ed.pkl'
edit_dataset = EditDistanceDatasetGenerator(N_batches={
"train": 2,
"val": 2,
"test": 2
},
batch_size={
"train": 5,
"val": 3,
"test": 3
},
len_sequence={
"train": 10,
"val": 10,
"test": 10
},
max_changes={
"train": 2,
"val": 2,
"test": 2
},
string_generator=generator,
seed=0)
edit_dataset.save_as_pickle(edit_dataset_name)
hc_dataset_name = folder_name + '/test_hc.pkl'
hc_dataset = HierarchicalClusteringDatasetGenerator(
N_reference=3,
N_leaves=4,
len_sequence=10,
min_changes=2,
max_changes=4,
string_generator=generator,
seed=0)
hc_dataset.save_as_pickle(hc_dataset_name)
parser = general_arg_parser()
args = parser.parse_args()
args.data = edit_dataset_name
args.epochs = 2
args.print_every = 1
args.hierarchical_data_path = hc_dataset_name
return folder_name, edit_dataset_name, hc_dataset_name, args
def generate_dataset_and_parser():
folder_name = ''.join(random.choice(string.ascii_lowercase) for _ in range(10))
generator = IndependentGenerator(alphabet_size=ALPHABET_SIZE, seed=0)
dataset_name = folder_name + '/test_ed_model.pkl'
dataset = EditDistanceDatasetGenerator(
N_batches={"train": 2, "val": 2, "test": 2},
batch_size={"train": 5, "val": 3, "test": 3},
len_sequence={"train": 10, "val": 10, "test": 10},
max_changes={"train": 2, "val": 2, "test": 2},
string_generator=generator, seed=0)
dataset.save_as_pickle(dataset_name)
parser = general_arg_parser()
args = parser.parse_args()
args.data = dataset_name
args.epochs = 2
args.print_every = 1
args.distance = "euclidean"
return folder_name, dataset_name, args
from edit_distance.models.feedforward.model import MLPEncoder
from edit_distance.train import execute_train, general_arg_parser
from util.data_handling.data_loader import BOOL_CHOICE
parser = general_arg_parser()
parser.add_argument('--layers', type=int, default=1, help='Number of fully connected layers')
parser.add_argument('--hidden_size', type=int, default=100, help='Size of hidden layers')
parser.add_argument('--batch_norm', type=str, default='False', help='Batch normalization')
args = parser.parse_args()
assert args.batch_norm in BOOL_CHOICE, "Boolean values have to be either 'True' or 'False' "
execute_train(model_class=MLPEncoder,
model_args=dict(layers=args.layers,
hidden_size=args.hidden_size,
batch_norm=True if args.batch_norm == 'True' else False),
args=args)