def add_predict_args(parser: ArgumentParser):
"""
Adds predict arguments to an ArgumentParser.
:param parser: An ArgumentParser.
"""
parser.add_argument('--gpu', type=int,
choices=list(range(torch.cuda.device_count())),
help='Which GPU to use')
parser.add_argument('--test_path', type=str,
help='Path to CSV file containing testing data for which predictions will be made')
parser.add_argument('--use_compound_names', action='store_true', default=False,
help='Use when test data file contains compound names in addition to SMILES strings')
parser.add_argument('--preds_path', type=str,
help='Path to CSV file where predictions will be saved')
parser.add_argument('--checkpoint_dir', type=str,
help='Directory from which to load model checkpoints'
'(walks directory and ensembles all models that are found)')
parser.add_argument('--checkpoint_path', type=str, default=None,
help='Path to model checkpoint (.pt file)')
parser.add_argument('--batch_size', type=int, default=50,
help='Batch size')
parser.add_argument('--no_cuda', action='store_true', default=False,
help='Turn off cuda')
parser.add_argument('--features_generator', type=str, nargs='*',
choices=get_available_features_generators(),
help='Method of generating additional features')
parser.add_argument('--features_path', type=str, nargs='*',
help='Path to features to use in FNN (instead of features_generator)')
parser.add_argument('--no_features_scaling', action='store_true', default=False,
help='Turn off scaling of features')
parser.add_argument('--max_data_size', type=int,
help='Maximum number of data points to load')
def configure(self) -> None:
self.add_argument(
"--gpu", choices=list(range(torch.cuda.device_count()))
)
self.add_argument(
"--features_generator", choices=get_available_features_generators()
)
def test_save_features(self):
try:
parser = ArgumentParser()
parser.add_argument('--data_path', type=str,
help='Path to data CSV')
parser.add_argument('--features_generator', type=str,
choices=get_available_features_generators(),
help='Type of features to generate')
parser.add_argument('--save_path', type=str,
help='Path to .npz file where features will be saved as a numpy compressed archive')
parser.add_argument('--save_frequency', type=int, default=10000,
help='Frequency with which to save the features')
parser.add_argument('--restart', action='store_true', default=False,
help='Whether to not load partially complete featurization and instead start from scratch')
parser.add_argument('--max_data_size', type=int,
help='Maximum number of data points to load')
parser.add_argument('--sequential', action='store_true', default=False,
help='Whether to run sequentially rather than in parallel')
args = parser.parse_args([])
args.data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'delaney_toy.csv')
args.save_path = NamedTemporaryFile(suffix='.npz').name
args.restart = True
args.features_generator = 'morgan_count'
generate_and_save_features(args)
os.remove(args.save_path)
except:
self.fail('save_features')
def add_arguments(self) -> None:
self.add_argument('--gpu',
choices=list(range(torch.cuda.device_count())))
self.add_argument('--features_generator',
choices=get_available_features_generators())
def add_train_args(parser: ArgumentParser):
"""
Adds training arguments to an ArgumentParser.
:param parser: An ArgumentParser.
"""
# General arguments
parser.add_argument('--gpu',
type=int,
choices=list(range(torch.cuda.device_count())),
help='Which GPU to use')
parser.add_argument('--data_path',
type=str,
help='Path to data CSV file',
default='M_CYP1A2I_I.csv')
parser.add_argument(
'--use_compound_names',
action='store_true',
default=False,
help=
'Use when test data file contains compound names in addition to SMILES strings'
)
parser.add_argument('--max_data_size',
type=int,
help='Maximum number of data points to load')
parser.add_argument(
'--test',
action='store_true',
default=False,
help='Whether to skip training and only test the model')
parser.add_argument(
'--features_only',
action='store_true',
default=False,
help='Use only the additional features in an FFN, no graph network')
parser.add_argument('--features_generator',
type=str,
nargs='*',
choices=get_available_features_generators(),
help='Method of generating additional features')
parser.add_argument(
'--features_path',
type=str,
nargs='*',
help='Path to features to use in FNN (instead of features_generator)')
parser.add_argument('--save_dir',
type=str,
default='./ckpt',
help='Directory where model checkpoints will be saved')
parser.add_argument(
'--save_smiles_splits',
action='store_true',
default=False,
help=
'Save smiles for each train/val/test splits for prediction convenience later'
)
parser.add_argument(
'--checkpoint_dir',
type=str,
default=None,
help='Directory from which to load model checkpoints'
'(walks directory and ensembles all models that are found)')
parser.add_argument('--checkpoint_path',
type=str,
default=None,
help='Path to model checkpoint (.pt file)')
parser.add_argument(
'--dataset_type',
type=str,
choices=['classification', 'regression', 'multiclass'],
help='Type of dataset, e.g. classification or regression.'
'This determines the loss function used during training.',
default='regression') # classification
parser.add_argument(
'--multiclass_num_classes',
type=int,
default=3,
help='Number of classes when running multiclass classification')
parser.add_argument('--separate_val_path',
type=str,
help='Path to separate val set, optional')
parser.add_argument('--separate_val_features_path',
type=str,
nargs='*',
help='Path to file with features for separate val set')
parser.add_argument('--separate_test_path',
type=str,
help='Path to separate test set, optional')
parser.add_argument(
'--separate_test_features_path',
type=str,
nargs='*',
help='Path to file with features for separate test set')
parser.add_argument(
'--split_type',
type=str,
default='random',
choices=[
'random', 'scaffold_balanced', 'predetermined', 'crossval',
'index_predetermined'
],
help='Method of splitting the data into train/val/test')
parser.add_argument(
'--split_sizes',
type=float,
nargs=3,
default=[0.8, 0.1, 0.1],
help='Split proportions for train/validation/test sets')
parser.add_argument(
'--num_folds',
type=int,
default=1,
help='Number of folds when performing cross validation')
parser.add_argument('--folds_file',
type=str,
default=None,
help='Optional file of fold labels')
parser.add_argument(
'--val_fold_index',
type=int,
default=None,
help='Which fold to use as val for leave-one-out cross val')
parser.add_argument(
'--test_fold_index',
type=int,
default=None,
help='Which fold to use as test for leave-one-out cross val')
parser.add_argument(
'--crossval_index_dir',
type=str,
help='Directory in which to find cross validation index files')
parser.add_argument('--crossval_index_file',
type=str,
help='Indices of files to use as train/val/test'
'Overrides --num_folds and --seed.')
parser.add_argument(
'--seed',
type=int,
default=0,
help='Random seed to use when splitting data into train/val/test sets.'
'When `num_folds` > 1, the first fold uses this seed and all'
'subsequent folds add 1 to the seed.')
parser.add_argument(
'--metric',
type=str,
default=None,
choices=[
'auc', 'prc-auc', 'rmse', 'mae', 'mse', 'r2', 'accuracy',
'cross_entropy'
],
help='Metric to use during evaluation.'
'Note: Does NOT affect loss function used during training'
'(loss is determined by the `dataset_type` argument).'
'Note: Defaults to "auc" for classification and "rmse" for regression.'
)
parser.add_argument('--quiet',
action='store_true',
default=False,
help='Skip non-essential print statements')
parser.add_argument(
'--log_frequency',
type=int,
default=10,
help='The number of batches between each logging of the training loss')
parser.add_argument('--no_cuda',
action='store_true',
default=False,
help='Turn off cuda')
parser.add_argument(
'--show_individual_scores',
action='store_true',
default=False,
help=
'Show all scores for individual targets, not just average, at the end')
parser.add_argument('--no_cache',
action='store_true',
default=False,
help='Turn off caching mol2graph computation')
parser.add_argument(
'--config_path',
type=str,
help=
'Path to a .json file containing arguments. Any arguments present in the config'
'file will override arguments specified via the command line or by the defaults.'
)
# Training arguments
parser.add_argument('--epochs',
type=int,
default=30,
help='Number of epochs to run')
parser.add_argument('--batch_size',
type=int,
default=50,
help='Batch size')
parser.add_argument(
'--warmup_epochs',
type=float,
default=2.0,
help=
'Number of epochs during which learning rate increases linearly from'
'init_lr to max_lr. Afterwards, learning rate decreases exponentially'
'from max_lr to final_lr.')
parser.add_argument('--init_lr',
type=float,
default=1e-4,
help='Initial learning rate')
parser.add_argument('--max_lr',
type=float,
default=1e-3,
help='Maximum learning rate')
parser.add_argument('--final_lr',
type=float,
default=1e-4,
help='Final learning rate')
parser.add_argument('--no_features_scaling',
action='store_true',
default=False,
help='Turn off scaling of features')
# Model arguments
parser.add_argument('--ensemble_size',
type=int,
default=1,
help='Number of models in ensemble')
parser.add_argument('--hidden_size',
type=int,
default=300,
help='Dimensionality of hidden layers in MPN')
parser.add_argument('--bias',
action='store_true',
default=False,
help='Whether to add bias to linear layers')
parser.add_argument('--depth',
type=int,
default=3,
help='Number of message passing steps')
parser.add_argument('--dropout',
type=float,
default=0.0,
help='Dropout probability')
parser.add_argument(
'--activation',
type=str,
default='ReLU',
choices=['ReLU', 'LeakyReLU', 'PReLU', 'tanh', 'SELU', 'ELU'],
help='Activation function')
parser.add_argument(
'--undirected',
action='store_true',
default=False,
help='Undirected edges (always sum the two relevant bond vectors)')
parser.add_argument(
'--ffn_hidden_size',
type=int,
default=None,
help='Hidden dim for higher-capacity FFN (defaults to hidden_size)')
parser.add_argument('--ffn_num_layers',
type=int,
default=2,
help='Number of layers in FFN after MPN encoding')
parser.add_argument(
'--atom_messages',
action='store_true',
default=False,
help='Use messages on atoms instead of messages on bonds')
except OverflowError:
print(
'Features array is too large to save as a single file. Instead keeping features as a directory of files.'
)
if __name__ == '__main__':
parser = ArgumentParser()
parser.add_argument('--data_path',
type=str,
required=True,
help='Path to data CSV')
parser.add_argument('--features_generator',
type=str,
required=True,
choices=get_available_features_generators(),
help='Type of features to generate')
parser.add_argument(
'--save_path',
type=str,
required=True,
help=
'Path to .npz file where features will be saved as a compressed numpy archive'
)
parser.add_argument('--save_frequency',
type=int,
default=10000,
help='Frequency with which to save the features')
parser.add_argument(
'--restart',
action='store_true',
def add_arguments(self) -> None:
self.add_argument('--features_generator',
choices=get_available_features_generators())
def add_predict_args(parser: ArgumentParser):
"""
Adds predict arguments to an ArgumentParser.
:param parser: An ArgumentParser.
"""
parser.add_argument('--gpu',
type=int,
choices=list(range(torch.cuda.device_count())),
help='Which GPU to use')
parser.add_argument(
'--data_format',
type=str,
default=None,
help='SSPFF12 example for 2 smiles, prop, followed by feats')
parser.add_argument(
'--test_path',
type=str,
help=
'Path to CSV file containing testing data for which predictions will be made'
)
parser.add_argument(
'--use_compound_names',
action='store_true',
default=False,
help=
'Use when test data file contains compound names in addition to SMILES strings'
)
parser.add_argument(
'--preds_path',
type=str,
help='Path to CSV file where predictions will be saved')
parser.add_argument(
'--checkpoint_dir',
type=str,
help='Directory from which to load model checkpoints'
'(walks directory and ensembles all models that are found)')
parser.add_argument('--checkpoint_path',
type=str,
default=None,
help='Path to model checkpoint (.pt file)')
parser.add_argument('--batch_size',
type=int,
default=50,
help='Batch size')
parser.add_argument('--no_cuda',
action='store_true',
default=False,
help='Turn off cuda')
parser.add_argument('--features_generator',
type=str,
nargs='*',
choices=get_available_features_generators(),
help='Method of generating additional features')
parser.add_argument(
'--features_path',
type=str,
nargs='*',
help='Path to features to use in FNN (instead of features_generator)')
parser.add_argument('--no_features_scaling',
action='store_true',
default=False,
help='Turn off scaling of features')
parser.add_argument('--max_data_size',
type=int,
help='Maximum number of data points to load')
parser.add_argument(
'--config_path',
type=str,
help=
'Path to a .json file containing arguments. Any arguments present in the config'
'file will override arguments specified via the command line or by the defaults.'
)
parser.add_argument('--ops',
type=str,
default='concat',
choices=['plus', 'minus', 'concat'],
help='Operation for embeddings')
