def test_smiles_parser_target_index(mol_smiles, mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles,
return_smiles=True,
target_index=[0, 2],
return_is_successful=True)
dataset = result['dataset']
smiles = result['smiles']
assert len(dataset) == 2
is_successful = result['is_successful']
assert numpy.alltrue(is_successful)
assert len(is_successful) == 2
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
expect = preprocessor.get_input_features(mols[0])
check_input_features(dataset[0], expect)
expect = preprocessor.get_input_features(mols[2])
check_input_features(dataset[1], expect)
# check smiles array
assert type(smiles) == numpy.ndarray
assert smiles.ndim == 1
assert len(smiles) == len(dataset)
assert smiles[0] == 'CN=C=O'
assert smiles[1] == 'CC1=CC2CC(CC1)O2'
def test_smiles_parser_target_index(mol_smiles, mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles, return_smiles=True, target_index=[0, 2],
return_is_successful=True)
dataset = result['dataset']
smiles = result['smiles']
assert len(dataset) == 2
is_successful = result['is_successful']
assert numpy.alltrue(is_successful)
assert len(is_successful) == 2
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
expect = preprocessor.get_input_features(mols[0])
check_input_features(dataset[0], expect)
expect = preprocessor.get_input_features(mols[2])
check_input_features(dataset[1], expect)
# check smiles array
assert type(smiles) == numpy.ndarray
assert smiles.ndim == 1
assert len(smiles) == len(dataset)
assert smiles[0] == 'CN=C=O'
assert smiles[1] == 'CC1=CC2CC(CC1)O2'
def test_smiles_parser_not_return_smiles(mol_smiles, mols):
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles, return_smiles=False)
dataset = result['dataset']
smiles = result['smiles']
is_successful = result['is_successful']
assert len(dataset) == 3
assert smiles is None
assert is_successful is None
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
def test_smiles_parser_not_return_smiles(mol_smiles, mols):
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles, return_smiles=False)
dataset = result['dataset']
smiles = result['smiles']
is_successful = result['is_successful']
assert len(dataset) == 3
assert smiles is None
assert is_successful is None
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
def test_atomic_number_preprocessor_default():
preprocessor = AtomicNumberPreprocessor()
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])['dataset']
index = numpy.random.choice(len(dataset), None)
atoms, = dataset[index]
assert atoms.ndim == 1
assert atoms.dtype == numpy.int32
def test_smiles_parser_return_is_successful(mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
mol_smiles_with_invalid = [
'var', 'CN=C=O', 'hoge', 'Cc1ccccc1', 'CC1=CC2CC(CC1)O2']
result = parser.parse(mol_smiles_with_invalid, return_smiles=True,
return_is_successful=True)
dataset = result['dataset']
assert len(dataset) == 3
is_successful = result['is_successful']
assert len(is_successful) == 5
assert numpy.alltrue(is_successful[[1, 3, 4]])
assert numpy.alltrue(~is_successful[[0, 2]])
# We assume NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
def test_smiles_parser_return_smiles(mol_smiles, mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles, return_smiles=True)
dataset = result['dataset']
smiles = result['smiles']
assert len(dataset) == 3
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
# check smiles array
assert type(smiles) == numpy.ndarray
assert smiles.ndim == 1
assert len(smiles) == len(dataset)
assert smiles[0] == 'CN=C=O'
assert smiles[1] == 'Cc1ccccc1'
assert smiles[2] == 'CC1=CC2CC(CC1)O2'
def test_smiles_parser_return_is_successful(mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
mol_smiles_with_invalid = [
'var', 'CN=C=O', 'hoge', 'Cc1ccccc1', 'CC1=CC2CC(CC1)O2'
]
result = parser.parse(mol_smiles_with_invalid,
return_smiles=True,
return_is_successful=True)
dataset = result['dataset']
assert len(dataset) == 3
is_successful = result['is_successful']
assert len(is_successful) == 5
assert numpy.alltrue(is_successful[[1, 3, 4]])
assert numpy.alltrue(~is_successful[[0, 2]])
# We assume NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
def test_smiles_parser_return_smiles(mol_smiles, mols):
"""test `labels` option and retain_smiles=True."""
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
result = parser.parse(mol_smiles, return_smiles=True)
dataset = result['dataset']
smiles = result['smiles']
assert len(dataset) == 3
# As we want test CSVFileParser, we assume
# NFPPreprocessor works as documented.
for i in range(3):
expect = preprocessor.get_input_features(mols[i])
check_input_features(dataset[i], expect)
# check smiles array
assert type(smiles) == numpy.ndarray
assert smiles.ndim == 1
assert len(smiles) == len(dataset)
assert smiles[0] == 'CN=C=O'
assert smiles[1] == 'Cc1ccccc1'
assert smiles[2] == 'CC1=CC2CC(CC1)O2'
def test_rsgcn_preprocessor_default():
preprocessor = RSGCNPreprocessor()
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])['dataset']
index = numpy.random.choice(len(dataset), None)
atoms, adjacency = dataset[index]
assert atoms.ndim == 1 # (atom, )
assert atoms.dtype == numpy.int32
assert adjacency.ndim == 2
assert adjacency.dtype == numpy.float32
def test_nfp_preprocessor_default():
preprocessor = NFPPreprocessor()
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])['dataset']
index = numpy.random.choice(len(dataset), None)
atoms, adjs = dataset[index]
assert atoms.ndim == 1 # (atom, )
assert atoms.dtype == numpy.int32
# (atom from, atom to)
assert adjs.ndim == 2
assert adjs.dtype == numpy.float32
def test_relgcn_preprocessor_kekulize():
preprocessor = RelGCNPreprocessor(kekulize=True)
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])["dataset"]
atoms1, adjs1 = dataset[1]
assert numpy.allclose(
atoms1, numpy.array([6, 6, 6, 7, 6, 6, 8, 7, 6], numpy.int32))
# NOT include aromatic bond (ch=3)
expect_adjs = numpy.array([[[0., 1., 0., 0., 0., 0., 0., 0., 0.],
[1., 0., 0., 0., 0., 0., 0., 1., 0.],
[0., 0., 0., 1., 0., 0., 0., 0., 0.],
[0., 0., 1., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 1., 0., 1., 0.],
[0., 0., 0., 0., 1., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 1., 0., 0., 1., 0., 0., 0., 1.],
[0., 0., 0., 0., 0., 0., 0., 1., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 1., 0., 0., 0., 0., 0., 0.],
[0., 1., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 1., 0., 0., 0., 0.],
[0., 0., 0., 1., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 1., 0., 0.],
[0., 0., 0., 0., 0., 1., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]]],
dtype=numpy.float32)
assert numpy.allclose(adjs1, expect_adjs)
def test_weave_preprocessor(max_atoms, use_fixed_atom_feature):
preprocessor = WeaveNetPreprocessor(
max_atoms=max_atoms, use_fixed_atom_feature=use_fixed_atom_feature)
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])["dataset"]
index = numpy.random.choice(len(dataset), None)
atoms, adjs = dataset[index]
if use_fixed_atom_feature:
assert atoms.ndim == 2 # (atom, ch)
assert atoms.dtype == numpy.float32
else:
assert atoms.ndim == 1 # (atom, )
assert atoms.dtype == numpy.int32
# (atom from * atom to, ch)
assert adjs.ndim == 2
assert adjs.shape[0] == max_atoms * max_atoms
assert adjs.dtype == numpy.float32
# TODO(nakago): test feature extraction behavior...
atoms0, adjs0 = dataset[0]
def test_smiles_parser_extract_total_num(mol_smiles):
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
num = parser.extract_total_num(mol_smiles)
assert num == 3
def test_relgcn_preprocessor():
preprocessor = RelGCNPreprocessor()
dataset = SmilesParser(preprocessor).parse(
['C#N', 'Cc1cnc(C=O)n1C', 'c1ccccc1'])["dataset"]
index = numpy.random.choice(len(dataset), None)
atoms, adjs = dataset[index]
assert atoms.ndim == 1 # (atom, )
assert atoms.dtype == numpy.int32
# (edge_type, atom from, atom to)
assert adjs.ndim == 3
assert adjs.dtype == numpy.float32
atoms0, adjs0 = dataset[0]
assert numpy.allclose(atoms0, numpy.array([6, 7], numpy.int32))
expect_adjs = numpy.array([[[0., 0.], [0., 0.]], [[0., 0.], [0., 0.]],
[[0., 1.], [1., 0.]], [[0., 0.], [0., 0.]]],
dtype=numpy.float32)
assert numpy.allclose(adjs0, expect_adjs)
atoms1, adjs1 = dataset[1]
assert numpy.allclose(
atoms1, numpy.array([6, 6, 6, 7, 6, 6, 8, 7, 6], numpy.int32))
# include aromatic bond (ch=3)
expect_adjs = numpy.array([[[0., 1., 0., 0., 0., 0., 0., 0., 0.],
[1., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 1., 0., 0., 0.],
[0., 0., 0., 0., 1., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 1.],
[0., 0., 0., 0., 0., 0., 0., 1., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 1., 0., 0.],
[0., 0., 0., 0., 0., 1., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]],
[[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 1., 0., 0., 0., 0., 1., 0.],
[0., 1., 0., 1., 0., 0., 0., 0., 0.],
[0., 0., 1., 0., 1., 0., 0., 0., 0.],
[0., 0., 0., 1., 0., 0., 0., 1., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 1., 0., 0., 1., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0.]]],
dtype=numpy.float32)
assert numpy.allclose(adjs1, expect_adjs)
def test_smiles_parser_extract_total_num(mol_smiles):
preprocessor = NFPPreprocessor()
parser = SmilesParser(preprocessor)
num = parser.extract_total_num(mol_smiles)
assert num == 3