def rdkit_2d_normalized_features_generator(mol):
smiles = Chem.MolToSmiles(mol,
isomericSmiles=True) if type(mol) != str else mol
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = generator.process(smiles)[1:]
return features
def smiles2rdkit2d(s):
try:
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = generator.process(s)[1:]
except:
print('descriptastorus not found this smiles: ' + s +
' convert to all 1 features')
features = np.ones((200, ))
return np.array(features)
def smiles2rdkit2d(s):
try:
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = np.array(generator.process(s)[1:])
NaNs = np.isnan(features)
features[NaNs] = 0
except:
print('descriptastorus not found this smiles: ' + s + ' convert to all 0 features')
features = np.zeros((200, ))
return np.array(features)
def generate_rdDescriptors(mol, Normalized=True):
smiles = Chem.MolToSmiles(
mol, isomericSmiles=True) if type(mol) != str else mol
from descriptastorus.descriptors import rdDescriptors, rdNormalizedDescriptors
if Normalized:
generator = rdNormalizedDescriptors.RDKit2DNormalized()
tors = generator.process(smiles)
else:
generator = rdDescriptors.RDKit2D()
tors = generator.process(smiles)
return tors[1:]
def rdkit_2d_features_normalized_generator(mol: Molecule) -> np.ndarray:
"""
Generates RDKit 2D normalized features for a molecule.
:param mol: A molecule (i.e. either a SMILES string or an RDKit molecule).
:return: A 1D numpy array containing the RDKit 2D normalized features.
"""
smiles = Chem.MolToSmiles(
mol, isomericSmiles=True) if type(mol) != str else mol
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = generator.process(smiles)[1:]
return features
def generate_rdDescriptorsSets(mols, Normalized=True):
from descriptastorus.descriptors import rdDescriptors, rdNormalizedDescriptors
if Normalized:
generator = rdNormalizedDescriptors.RDKit2DNormalized()
else:
generator = rdDescriptors.RDKit2D()
tors = []
for mol in mols:
smiles = Chem.MolToSmiles(
mol, isomericSmiles=True) if type(mol) != str else mol
tors.append(generator.process(smiles)[1:])
return np.asarray(tors)
import logging
from rdkit import Chem
try:
from descriptastorus.descriptors import rdNormalizedDescriptors
generator = rdNormalizedDescriptors.RDKit2DNormalized()
def rdkit_2d_normalized_features(smiles: str):
# the first element is true/false if the mol was properly computed
if type(smiles) == str:
return generator.process(smiles)[1:]
else:
# this is a bit of a waste, but the desciptastorus API is smiles
# based for normalization purposes
return generator.process(
Chem.MolToSmiles(smiles, isomericSmiles=True))[1:]
except ImportError:
logging.getLogger(__name__).warning(
"descriptastorus is not available, normalized descriptors are not available"
)
rdkit_2d_normalized_features = None
descriptors = calc.pandas(mols)
descriptors = descriptors.astype(str)
masks = descriptors.apply(lambda d: d.str.contains('[a-zA-Z]', na=False))
descriptors = descriptors[~masks]
descriptors = descriptors.astype(float)
y = pd.DataFrame(y, index=smiles, columns=[y_name])
if dataset_type == 3:
from descriptastorus.descriptors.DescriptorGenerator import MakeGenerator
from descriptastorus.descriptors import rdDescriptors
from descriptastorus.descriptors import rdNormalizedDescriptors
gen1 = MakeGenerator(('rdkit2d', 'Morgan3counts'))
gen2 = rdDescriptors.RDKit2D()
gen3 = rdNormalizedDescriptors.RDKit2DNormalized()
data1 = gen1.process(smiles)
data2 = gen2.process(smiles)
data3 = gen3.process(smiles)
for col in gen1.GetColumns():
y_name.append(col)
y = pd.DataFrame(y, index=smiles, columns=[y_name])
if dataset_type == 4: #3D Descriptors
from e3fp.fingerprint.generate import fp, fprints_dict_from_mol
from e3fp.conformer.generate import generate_conformers
mols = [Chem.MolFromSmiles(smile) for smile in smiles]
optimize_mols = []
for mol in mols:
mh = Chem.AddHs(mol)
def get_rdnorm_fp(smiles):
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = generator.process(smiles)[1:]
arr = np.array(features)
return arr
def preprocess(data):
feature_generator = rdNormalizedDescriptors.RDKit2DNormalized()
result = [[MoleculeDatapoint(line, feature_generator)] for line in data]
return result
def generate_descriptors(smi):
generator = rdNormalizedDescriptors.RDKit2DNormalized()
features = generator.process(smi)[1:]
return features[:114]
