def test_valid_unique(self):
disable_rdkit_log()
mols = ['CCNC', 'CCC', 'INVALID', 'CCC']
assert np.allclose(fraction_valid(mols), 3 / 4), "Failed valid"
assert np.allclose(fraction_unique(mols, check_validity=False),
3 / 4), "Failed unique"
assert np.allclose(fraction_unique(mols, k=2), 1), "Failed unique"
mols = [Chem.MolFromSmiles(x) for x in mols]
assert np.allclose(fraction_valid(mols), 3 / 4), "Failed valid"
assert np.allclose(fraction_unique(mols, check_validity=False),
3 / 4), "Failed unique"
assert np.allclose(fraction_unique(mols, k=2), 1), "Failed unique"
enable_rdkit_log()
def get_all_metrics(test, gen, k=None, n_jobs=1, device='cpu',
batch_size=512, test_scaffolds=None,
ptest=None, ptest_scaffolds=None,
pool=None, gpu=None, train=None):
"""
Computes all available metrics between test (scaffold test)
and generated sets of SMILES.
Parameters:
test: list of test SMILES
gen: list of generated SMILES
k: int or list with values for unique@k. Will calculate number of
unique molecules in the first k molecules. Default [1000, 10000]
n_jobs: number of workers for parallel processing
device: 'cpu' or 'cuda:n', where n is GPU device number
batch_size: batch size for FCD metric
test_scaffolds: list of scaffold test SMILES
Will compute only on the general test set if not specified
ptest: dict with precalculated statistics of the test set
ptest_scaffolds: dict with precalculated statistics
of the scaffold test set
pool: optional multiprocessing pool to use for parallelization
gpu: deprecated, use `device`
train: list of train SMILES
Available metrics:
* %valid
* %unique@k
* Frechet ChemNet Distance (FCD)
* Fragment similarity (Frag)
* Scaffold similarity (Scaf)
* Similarity to nearest neighbour (SNN)
* Internal diversity (IntDiv)
* Internal diversity 2: using square root of mean squared
Tanimoto similarity (IntDiv2)
* %passes filters (Filters)
* Distribution difference for logP, SA, QED, NP, weight
* Novelty (molecules not present in train)
"""
if k is None:
k = [1000, 10000]
disable_rdkit_log()
metrics = {}
if gpu is not None:
warnings.warn(
"parameter `gpu` is deprecated. Use `device`",
DeprecationWarning
)
if gpu == -1:
device = 'cpu'
else:
device = 'cuda:{}'.format(gpu)
close_pool = False
if pool is None:
if n_jobs != 1:
pool = Pool(n_jobs)
close_pool = True
else:
pool = 1
metrics['valid'] = fraction_valid(gen, n_jobs=pool)
gen = remove_invalid(gen, canonize=True)
if not isinstance(k, (list, tuple)):
k = [k]
for _k in k:
metrics['unique@{}'.format(_k)] = fraction_unique(gen, _k, pool)
if ptest is None:
ptest = compute_intermediate_statistics(test, n_jobs=n_jobs,
device=device,
batch_size=batch_size,
pool=pool)
if test_scaffolds is not None and ptest_scaffolds is None:
ptest_scaffolds = compute_intermediate_statistics(
test_scaffolds, n_jobs=n_jobs,
device=device, batch_size=batch_size,
pool=pool
)
mols = mapper(pool)(get_mol, gen)
kwargs = {'n_jobs': pool, 'device': device, 'batch_size': batch_size}
kwargs_fcd = {'n_jobs': n_jobs, 'device': device, 'batch_size': batch_size}
metrics['FCD/Test'] = FCDMetric(**kwargs_fcd)(gen=gen, pref=ptest['FCD'])
metrics['SNN/Test'] = SNNMetric(**kwargs)(gen=mols, pref=ptest['SNN'])
metrics['Frag/Test'] = FragMetric(**kwargs)(gen=mols, pref=ptest['Frag'])
metrics['Scaf/Test'] = ScafMetric(**kwargs)(gen=mols, pref=ptest['Scaf'])
if ptest_scaffolds is not None:
metrics['FCD/TestSF'] = FCDMetric(**kwargs_fcd)(
gen=gen, pref=ptest_scaffolds['FCD']
)
metrics['SNN/TestSF'] = SNNMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['SNN']
)
metrics['Frag/TestSF'] = FragMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['Frag']
)
metrics['Scaf/TestSF'] = ScafMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['Scaf']
)
metrics['IntDiv'] = internal_diversity(mols, pool, device=device)
metrics['IntDiv2'] = internal_diversity(mols, pool, device=device, p=2)
metrics['Filters'] = fraction_passes_filters(mols, pool)
# Properties
for name, func in [('logP', logP), ('SA', SA),
('QED', QED), ('NP', NP),
('weight', weight)]:
metrics[name] = FrechetMetric(func, **kwargs)(gen=mols,
pref=ptest[name])
if train is not None:
metrics['Novelty'] = novelty(mols, train, pool)
enable_rdkit_log()
if close_pool:
pool.close()
pool.join()
return metrics
help='Path to the config csv with `name` and `path` columns. '
'`name` is a model name, and '
'`path` is a path to generated samples`')
parser.add_argument('--n_jobs',
type=int,
default=1,
help='number of processes to use')
parser.add_argument('--img_folder',
type=str,
default='images/',
help='Store images in this folder')
return parser
if __name__ == "__main__":
disable_rdkit_log()
parser = get_parser()
config, unknown = parser.parse_known_args()
if len(unknown) != 0:
raise ValueError("Unknown argument " + unknown[0])
os.makedirs(config.img_folder, exist_ok=True)
generated = OrderedDict(
{'MOSES': pd.DataFrame({'SMILES': get_dataset('test')})})
models = pd.read_csv(config.config)
for path, name in zip(models['path'], models['name']):
generated[name] = pd.read_csv(path)
metrics = {'weight': weight, 'logP': logP, 'SA': SA, 'QED': QED}
def get_all_metrics(test,
gen,
k=[1000, 10000],
n_jobs=1,
gpu=-1,
batch_size=512,
test_scaffolds=None,
ptest=None,
ptest_scaffolds=None):
'''
Computes all available metrics between test (scaffold test) and generated sets of SMILES.
Parameters:
test: list of test SMILES
gen: list of generated SMILES
k: list with values for unique@k.
Will calculate number of unique molecules in the first k molecules.
n_jobs: number of workers for parallel processing
gpu: index of GPU for FCD metric and internal diversity, -1 means use CPU
batch_size: batch size for FCD metric
test_scaffolds: list of scaffold test SMILES
Will compute only on the general test set if not specified
ptest: dict with precalculated statistics of the test set
ptest_scaffolds: dict with precalculated statistics of the scaffold test set
Available metrics:
* %valid
* %unique@k
* Frechet ChemNet Distance (FCD)
* Fragment similarity (Frag)
* Scaffold similarity (Scaf)
* Similarity to nearest neighbour (SNN)
* Internal diversity (IntDiv)
* Internal diversity 2: using square root of mean squared Tanimoto similarity (IntDiv2)
* %passes filters (Filters)
* Distribution difference for logP, SA, QED, NP, weight
'''
disable_rdkit_log()
metrics = {}
if n_jobs != 1:
pool = Pool(n_jobs)
else:
pool = 1
metrics['valid'] = fraction_valid(gen, n_jobs=n_jobs)
gen = remove_invalid(gen, canonize=True)
if not isinstance(k, (list, tuple)):
k = [k]
for _k in k:
metrics['unique@{}'.format(_k)] = fraction_unique(gen, _k, pool)
if ptest is None:
ptest = compute_intermediate_statistics(test,
n_jobs=n_jobs,
gpu=gpu,
batch_size=batch_size)
if test_scaffolds is not None and ptest_scaffolds is None:
ptest_scaffolds = compute_intermediate_statistics(
test_scaffolds, n_jobs=n_jobs, gpu=gpu, batch_size=batch_size)
mols = mapper(pool)(get_mol, gen)
kwargs = {'n_jobs': pool, 'gpu': gpu, 'batch_size': batch_size}
metrics['FCD/Test'] = FCDMetric(**kwargs)(gen=gen, ptest=ptest['FCD'])
metrics['SNN/Test'] = SNNMetric(**kwargs)(gen=mols, ptest=ptest['SNN'])
metrics['Frag/Test'] = FragMetric(**kwargs)(gen=mols, ptest=ptest['Frag'])
metrics['Scaf/Test'] = ScafMetric(**kwargs)(gen=mols, ptest=ptest['Scaf'])
if ptest_scaffolds is not None:
metrics['FCD/TestSF'] = FCDMetric(**kwargs)(
gen=gen, ptest=ptest_scaffolds['FCD'])
metrics['SNN/TestSF'] = SNNMetric(**kwargs)(
gen=mols, ptest=ptest_scaffolds['SNN'])
metrics['Frag/TestSF'] = FragMetric(**kwargs)(
gen=mols, ptest=ptest_scaffolds['Frag'])
metrics['Scaf/TestSF'] = ScafMetric(**kwargs)(
gen=mols, ptest=ptest_scaffolds['Scaf'])
metrics['IntDiv'] = internal_diversity(mols, pool, gpu=gpu)
metrics['IntDiv2'] = internal_diversity(mols, pool, gpu=gpu, p=2)
metrics['Filters'] = fraction_passes_filters(mols, pool)
# Properties
for name, func in [('logP', logP), ('SA', SA), ('QED', QED), ('NP', NP),
('weight', weight)]:
metrics[name] = FrechetMetric(func, **kwargs)(gen=mols,
ptest=ptest[name])
enable_rdkit_log()
if n_jobs != 1:
pool.terminate()
return metrics
def get_all_metrics(gen,
k=None,
n_jobs=1,
device='cpu',
batch_size=512,
pool=None,
test=None,
test_scaffolds=None,
ptest=None,
ptest_scaffolds=None,
train=None):
"""
Computes all available metrics between test (scaffold test)
and generated sets of SMILES.
Parameters:
gen: list of generated SMILES
k: int or list with values for unique@k. Will calculate number of
unique molecules in the first k molecules. Default [1000, 10000]
n_jobs: number of workers for parallel processing
device: 'cpu' or 'cuda:n', where n is GPU device number
batch_size: batch size for FCD metric
pool: optional multiprocessing pool to use for parallelization
test (None or list): test SMILES. If None, will load
a default test set
test_scaffolds (None or list): scaffold test SMILES. If None, will
load a default scaffold test set
ptest (None or dict): precalculated statistics of the test set. If
None, will load default test statistics. If you specified a custom
test set, default test statistics will be ignored
ptest_scaffolds (None or dict): precalculated statistics of the
scaffold test set If None, will load default scaffold test
statistics. If you specified a custom test set, default test
statistics will be ignored
train (None or list): train SMILES. If None, will load a default
train set
Available metrics:
* %valid
* %unique@k
* Frechet ChemNet Distance (FCD)
* Fragment similarity (Frag)
* Scaffold similarity (Scaf)
* Similarity to nearest neighbour (SNN)
* Internal diversity (IntDiv)
* Internal diversity 2: using square root of mean squared
Tanimoto similarity (IntDiv2)
* %passes filters (Filters)
* Distribution difference for logP, SA, QED, weight
* Novelty (molecules not present in train)
"""
if test is None:
if ptest is not None:
raise ValueError("You cannot specify custom test "
"statistics for default test set")
test = get_dataset('test')
ptest = get_statistics('test')
if test_scaffolds is None:
if ptest_scaffolds is not None:
raise ValueError("You cannot specify custom scaffold test "
"statistics for default scaffold test set")
test_scaffolds = get_dataset('test_scaffolds')
ptest_scaffolds = get_statistics('test_scaffolds')
train = train or get_dataset('train')
if k is None:
k = [1000, 10000]
disable_rdkit_log()
metrics = {}
close_pool = False
if pool is None:
if n_jobs != 1:
pool = Pool(n_jobs)
close_pool = True
else:
pool = 1
metrics['valid'] = fraction_valid(gen, n_jobs=pool)
gen = remove_invalid(gen, canonize=True)
if not isinstance(k, (list, tuple)):
k = [k]
for _k in k:
metrics['unique@{}'.format(_k)] = fraction_unique(gen, _k, pool)
if ptest is None:
ptest = compute_intermediate_statistics(test,
n_jobs=n_jobs,
device=device,
batch_size=batch_size,
pool=pool)
if test_scaffolds is not None and ptest_scaffolds is None:
ptest_scaffolds = compute_intermediate_statistics(
test_scaffolds,
n_jobs=n_jobs,
device=device,
batch_size=batch_size,
pool=pool)
mols = mapper(pool)(get_mol, gen)
kwargs = {'n_jobs': pool, 'device': device, 'batch_size': batch_size}
kwargs_fcd = {'n_jobs': n_jobs, 'device': device, 'batch_size': batch_size}
metrics['FCD/Test'] = FCDMetric(**kwargs_fcd)(gen=gen, pref=ptest['FCD'])
metrics['SNN/Test'] = SNNMetric(**kwargs)(gen=mols, pref=ptest['SNN'])
metrics['Frag/Test'] = FragMetric(**kwargs)(gen=mols, pref=ptest['Frag'])
metrics['Scaf/Test'] = ScafMetric(**kwargs)(gen=mols, pref=ptest['Scaf'])
if ptest_scaffolds is not None:
metrics['FCD/TestSF'] = FCDMetric(**kwargs_fcd)(
gen=gen, pref=ptest_scaffolds['FCD'])
metrics['SNN/TestSF'] = SNNMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['SNN'])
metrics['Frag/TestSF'] = FragMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['Frag'])
metrics['Scaf/TestSF'] = ScafMetric(**kwargs)(
gen=mols, pref=ptest_scaffolds['Scaf'])
metrics['IntDiv'] = internal_diversity(mols, pool, device=device)
metrics['IntDiv2'] = internal_diversity(mols, pool, device=device, p=2)
metrics['Filters'] = fraction_passes_filters(mols, pool)
# Properties
for name, func in [('logP', logP), ('SA', SA), ('QED', QED),
('weight', weight)]:
metrics[name] = WassersteinMetric(func, **kwargs)(gen=mols,
pref=ptest[name])
if train is not None:
metrics['Novelty'] = novelty(mols, train, pool)
enable_rdkit_log()
if close_pool:
pool.close()
pool.join()
return metrics