Loading deepchem/feat/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -23,6 +23,8 @@ from deepchem.feat.molecule_featurizers import MolGraphConvFeaturizer from deepchem.feat.molecule_featurizers import CircularFingerprint from deepchem.feat.molecule_featurizers import CoulombMatrix from deepchem.feat.molecule_featurizers import CoulombMatrixEig from deepchem.feat.molecule_featurizers import MordredDescriptors from deepchem.feat.molecule_featurizers import Mol2VecFingerprint from deepchem.feat.molecule_featurizers import RawFeaturizer from deepchem.feat.molecule_featurizers import RDKitDescriptors from deepchem.feat.molecule_featurizers import SmilesToImage Loading deepchem/feat/molecule_featurizers/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -3,6 +3,8 @@ from deepchem.feat.molecule_featurizers.bp_symmetry_function_input import BPSymm from deepchem.feat.molecule_featurizers.circular_fingerprint import CircularFingerprint from deepchem.feat.molecule_featurizers.coulomb_matrices import CoulombMatrix from deepchem.feat.molecule_featurizers.coulomb_matrices import CoulombMatrixEig from deepchem.feat.molecule_featurizers.mordred_descriptors import MordredDescriptors from deepchem.feat.molecule_featurizers.mol2vec_fingerprint import Mol2VecFingerprint from deepchem.feat.molecule_featurizers.raw_featurizer import RawFeaturizer from deepchem.feat.molecule_featurizers.rdkit_descriptors import RDKitDescriptors from deepchem.feat.molecule_featurizers.smiles_to_image import SmilesToImage Loading deepchem/feat/molecule_featurizers/coulomb_matrices.py +22 −10 Original line number Diff line number Diff line Loading @@ -90,12 +90,16 @@ class CoulombMatrix(MolecularFeaturizer): ------- np.ndarray The coulomb matrices of the given molecule. The shape is `(num_confs, max_atoms, max_atoms)`. The default shape is `(num_confs, max_atoms, max_atoms)`. If num_confs == 1, the shape is `(max_atoms, max_atoms)`. """ features = self.coulomb_matrix(mol) if self.upper_tri: features = [f[np.triu_indices_from(f)] for f in features] features = np.asarray(features) if features.shape[0] == 1: # `(1, max_atoms, max_atoms)` -> `(max_atoms, max_atoms)` features = np.squeeze(features, axis=0) return features def coulomb_matrix(self, mol: RDKitMol) -> np.ndarray: Loading @@ -114,9 +118,16 @@ class CoulombMatrix(MolecularFeaturizer): """ try: from rdkit import Chem from rdkit.Chem import AllChem except ModuleNotFoundError: raise ValueError("This class requires RDKit to be installed.") # Check whether num_confs >=1 or not num_confs = len(mol.GetConformers()) if num_confs == 0: mol = Chem.AddHs(mol) AllChem.EmbedMolecule(mol, AllChem.ETKDG()) if self.remove_hydrogens: mol = Chem.RemoveHs(mol) n_atoms = mol.GetNumAtoms() Loading Loading @@ -203,8 +214,8 @@ class CoulombMatrixEig(CoulombMatrix): This featurizer computes the eigenvalues of the Coulomb matrices for provided molecules. Coulomb matrices are described in [1]_. Example ------- Examples -------- >>> featurizers = dc.feat.CoulombMatrixEig(max_atoms=23) >>> input_file = 'deepchem/feat/tests/data/water.sdf' # really backed by water.sdf.csv >>> tasks = ["atomization_energy"] Loading @@ -218,9 +229,6 @@ class CoulombMatrixEig(CoulombMatrix): processing systems. 2012. """ conformers = True name = 'coulomb_matrix' def __init__(self, max_atoms: int, remove_hydrogens: bool = False, Loading Loading @@ -266,10 +274,11 @@ class CoulombMatrixEig(CoulombMatrix): ------- np.ndarray The eigenvalues of Coulomb matrix for molecules. The shape is `(num_confs, max_atoms)`. The default shape is `(num_confs, max_atoms)`. If num_confs == 1, the shape is `(max_atoms,)`. """ cmat = self.coulomb_matrix(mol) features = [] features_list = [] for f in cmat: w, v = np.linalg.eig(f) w_abs = np.abs(w) Loading @@ -277,6 +286,9 @@ class CoulombMatrixEig(CoulombMatrix): sortidx = sortidx[::-1] w = w[sortidx] f = pad_array(w, self.max_atoms) features.append(f) features = np.asarray(features) features_list.append(f) features = np.asarray(features_list) if features.shape[0] == 1: # `(1, max_atoms)` -> `(max_atoms,)` features = np.squeeze(features, axis=0) return features deepchem/feat/molecule_featurizers/mol2vec_fingerprint.py +110 −0 Original line number Diff line number Diff line from os import path from typing import Optional import numpy as np from deepchem.utils import download_url, get_data_dir, untargz_file from deepchem.utils.typing import RDKitMol from deepchem.feat.base_classes import MolecularFeaturizer DEFAULT_PRETRAINED_MODEL_URL = 'https://deepchemdata.s3-us-west-1.amazonaws.com/trained_models/mol2vec_model_300dim.tar.gz' class Mol2VecFingerprint(MolecularFeaturizer): """Mol2Vec fingerprints. This class convert molecules to vector representations by using Mol2Vec. Mol2Vec is an unsupervised machine learning approach to learn vector representations of molecular substructures and the algorithm is based on Word2Vec, which is one of the most popular technique to learn word embeddings using neural network in NLP. Please see the details from [1]_. The Mol2Vec requires the pretrained model, so we use the model which is put on the mol2vec github repository [2]_. The default model was trained on 20 million compounds downloaded from ZINC using the following paramters. - radius 1 - UNK to replace all identifiers that appear less than 4 times - skip-gram and window size of 10 - embeddings size 300 References ---------- .. [1] Jaeger, Sabrina, Simone Fulle, and Samo Turk. "Mol2vec: unsupervised machine learning approach with chemical intuition." Journal of chemical information and modeling 58.1 (2018): 27-35. .. [2] https://github.com/samoturk/mol2vec/ Notes ----- This class requires mol2vec to be installed. """ def __init__(self, pretrain_model_path: Optional[str] = None, radius: int = 1, unseen: str = 'UNK', gather_method: str = 'sum'): """ Paremeters ---------- pretrain_file: str, optional The path for pretrained model. If this value is None, we use the model which is put on github repository (https://github.com/samoturk/mol2vec/tree/master/examples/models). The model is trained on 20 million compounds downloaded from ZINC. radius: int, optional (default 1) The fingerprint radius. The default value was used to train the model which is put on github repository. unseen: str, optional (default 'UNK') The string to used to replace uncommon words/identifiers while training. gather_method: str, optional (default 'sum') How to aggregate vectors of identifiers are extracted from Mol2vec. 'sum' or 'mean' is supported. """ try: from gensim.models import word2vec from mol2vec.features import mol2alt_sentence, sentences2vec except ModuleNotFoundError: raise ValueError("This class requires mol2vec to be installed.") self.radius = radius self.unseen = unseen self.gather_method = gather_method self.sentences2vec = sentences2vec self.mol2alt_sentence = mol2alt_sentence if pretrain_model_path is None: data_dir = get_data_dir() pretrain_model_path = path.join(data_dir, 'mol2vec_model_300dim.pkl') if not path.exists(pretrain_model_path): targz_file = path.join(data_dir, 'mol2vec_model_300dim.tar.gz') if not path.exists(targz_file): download_url(DEFAULT_PRETRAINED_MODEL_URL, data_dir) untargz_file( path.join(data_dir, 'mol2vec_model_300dim.tar.gz'), data_dir) # load pretrained models self.model = word2vec.Word2Vec.load(pretrain_model_path) def _featurize(self, mol: RDKitMol) -> np.ndarray: """ Calculate Mordred descriptors. Parameters ---------- mol: rdkit.Chem.rdchem.Mol RDKit Mol object Returns ------- np.ndarray 1D array of mol2vec fingerprint. The default length is 300. """ sentence = self.mol2alt_sentence(mol, self.radius) vec_identifiers = self.sentences2vec( sentence, self.model, unseen=self.unseen) if self.gather_method == 'sum': feature = np.sum(vec_identifiers, axis=0) elif self.gather_method == 'mean': feature = np.mean(vec_identifiers, axis=0) else: raise ValueError( 'Not supported gather_method type. Please set "sum" or "mean"') return feature deepchem/feat/molecule_featurizers/mordred_descriptors.py +67 −0 Original line number Diff line number Diff line import numpy as np from deepchem.utils.typing import RDKitMol from deepchem.feat.base_classes import MolecularFeaturizer class MordredDescriptors(MolecularFeaturizer): """Mordred descriptors. This class comptues a list of chemical descriptors using Mordred. Please see the details about all descripors from [1]_, [2]_. Attributes ---------- descriptors: List[str] List of RDKit descriptor names used in this class. References ---------- .. [1] Moriwaki, Hirotomo, et al. "Mordred: a molecular descriptor calculator." Journal of cheminformatics 10.1 (2018): 4. .. [2] http://mordred-descriptor.github.io/documentation/master/descriptors.html Notes ----- This class requires Mordred to be installed. """ def __init__(self, ignore_3D: bool = True): """ Paremeters ---------- ignore_3D: bool, optional (default True) Whether to use 3D information or not. """ try: from mordred import Calculator, descriptors, is_missing except ModuleNotFoundError: raise ValueError("This class requires RDKit to be installed.") self.calc = Calculator(descriptors, ignore_3D=ignore_3D) self.is_missing = is_missing self.descriptors = list(descriptors.__all__) def _featurize(self, mol: RDKitMol) -> np.ndarray: """ Calculate Mordred descriptors. Parameters ---------- mol: rdkit.Chem.rdchem.Mol RDKit Mol object Returns ------- np.ndarray 1D array of Mordred descriptors for `mol`. If ignore_3D is True, the length is 1613. If ignore_3D is False, the length is 1826. """ feature = self.calc(mol) # convert errors to zero feature = [ 0.0 if self.is_missing(val) or isinstance(val, str) else val for val in feature ] return np.asarray(feature) Loading
deepchem/feat/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -23,6 +23,8 @@ from deepchem.feat.molecule_featurizers import MolGraphConvFeaturizer from deepchem.feat.molecule_featurizers import CircularFingerprint from deepchem.feat.molecule_featurizers import CoulombMatrix from deepchem.feat.molecule_featurizers import CoulombMatrixEig from deepchem.feat.molecule_featurizers import MordredDescriptors from deepchem.feat.molecule_featurizers import Mol2VecFingerprint from deepchem.feat.molecule_featurizers import RawFeaturizer from deepchem.feat.molecule_featurizers import RDKitDescriptors from deepchem.feat.molecule_featurizers import SmilesToImage Loading
deepchem/feat/molecule_featurizers/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -3,6 +3,8 @@ from deepchem.feat.molecule_featurizers.bp_symmetry_function_input import BPSymm from deepchem.feat.molecule_featurizers.circular_fingerprint import CircularFingerprint from deepchem.feat.molecule_featurizers.coulomb_matrices import CoulombMatrix from deepchem.feat.molecule_featurizers.coulomb_matrices import CoulombMatrixEig from deepchem.feat.molecule_featurizers.mordred_descriptors import MordredDescriptors from deepchem.feat.molecule_featurizers.mol2vec_fingerprint import Mol2VecFingerprint from deepchem.feat.molecule_featurizers.raw_featurizer import RawFeaturizer from deepchem.feat.molecule_featurizers.rdkit_descriptors import RDKitDescriptors from deepchem.feat.molecule_featurizers.smiles_to_image import SmilesToImage Loading
deepchem/feat/molecule_featurizers/coulomb_matrices.py +22 −10 Original line number Diff line number Diff line Loading @@ -90,12 +90,16 @@ class CoulombMatrix(MolecularFeaturizer): ------- np.ndarray The coulomb matrices of the given molecule. The shape is `(num_confs, max_atoms, max_atoms)`. The default shape is `(num_confs, max_atoms, max_atoms)`. If num_confs == 1, the shape is `(max_atoms, max_atoms)`. """ features = self.coulomb_matrix(mol) if self.upper_tri: features = [f[np.triu_indices_from(f)] for f in features] features = np.asarray(features) if features.shape[0] == 1: # `(1, max_atoms, max_atoms)` -> `(max_atoms, max_atoms)` features = np.squeeze(features, axis=0) return features def coulomb_matrix(self, mol: RDKitMol) -> np.ndarray: Loading @@ -114,9 +118,16 @@ class CoulombMatrix(MolecularFeaturizer): """ try: from rdkit import Chem from rdkit.Chem import AllChem except ModuleNotFoundError: raise ValueError("This class requires RDKit to be installed.") # Check whether num_confs >=1 or not num_confs = len(mol.GetConformers()) if num_confs == 0: mol = Chem.AddHs(mol) AllChem.EmbedMolecule(mol, AllChem.ETKDG()) if self.remove_hydrogens: mol = Chem.RemoveHs(mol) n_atoms = mol.GetNumAtoms() Loading Loading @@ -203,8 +214,8 @@ class CoulombMatrixEig(CoulombMatrix): This featurizer computes the eigenvalues of the Coulomb matrices for provided molecules. Coulomb matrices are described in [1]_. Example ------- Examples -------- >>> featurizers = dc.feat.CoulombMatrixEig(max_atoms=23) >>> input_file = 'deepchem/feat/tests/data/water.sdf' # really backed by water.sdf.csv >>> tasks = ["atomization_energy"] Loading @@ -218,9 +229,6 @@ class CoulombMatrixEig(CoulombMatrix): processing systems. 2012. """ conformers = True name = 'coulomb_matrix' def __init__(self, max_atoms: int, remove_hydrogens: bool = False, Loading Loading @@ -266,10 +274,11 @@ class CoulombMatrixEig(CoulombMatrix): ------- np.ndarray The eigenvalues of Coulomb matrix for molecules. The shape is `(num_confs, max_atoms)`. The default shape is `(num_confs, max_atoms)`. If num_confs == 1, the shape is `(max_atoms,)`. """ cmat = self.coulomb_matrix(mol) features = [] features_list = [] for f in cmat: w, v = np.linalg.eig(f) w_abs = np.abs(w) Loading @@ -277,6 +286,9 @@ class CoulombMatrixEig(CoulombMatrix): sortidx = sortidx[::-1] w = w[sortidx] f = pad_array(w, self.max_atoms) features.append(f) features = np.asarray(features) features_list.append(f) features = np.asarray(features_list) if features.shape[0] == 1: # `(1, max_atoms)` -> `(max_atoms,)` features = np.squeeze(features, axis=0) return features
deepchem/feat/molecule_featurizers/mol2vec_fingerprint.py +110 −0 Original line number Diff line number Diff line from os import path from typing import Optional import numpy as np from deepchem.utils import download_url, get_data_dir, untargz_file from deepchem.utils.typing import RDKitMol from deepchem.feat.base_classes import MolecularFeaturizer DEFAULT_PRETRAINED_MODEL_URL = 'https://deepchemdata.s3-us-west-1.amazonaws.com/trained_models/mol2vec_model_300dim.tar.gz' class Mol2VecFingerprint(MolecularFeaturizer): """Mol2Vec fingerprints. This class convert molecules to vector representations by using Mol2Vec. Mol2Vec is an unsupervised machine learning approach to learn vector representations of molecular substructures and the algorithm is based on Word2Vec, which is one of the most popular technique to learn word embeddings using neural network in NLP. Please see the details from [1]_. The Mol2Vec requires the pretrained model, so we use the model which is put on the mol2vec github repository [2]_. The default model was trained on 20 million compounds downloaded from ZINC using the following paramters. - radius 1 - UNK to replace all identifiers that appear less than 4 times - skip-gram and window size of 10 - embeddings size 300 References ---------- .. [1] Jaeger, Sabrina, Simone Fulle, and Samo Turk. "Mol2vec: unsupervised machine learning approach with chemical intuition." Journal of chemical information and modeling 58.1 (2018): 27-35. .. [2] https://github.com/samoturk/mol2vec/ Notes ----- This class requires mol2vec to be installed. """ def __init__(self, pretrain_model_path: Optional[str] = None, radius: int = 1, unseen: str = 'UNK', gather_method: str = 'sum'): """ Paremeters ---------- pretrain_file: str, optional The path for pretrained model. If this value is None, we use the model which is put on github repository (https://github.com/samoturk/mol2vec/tree/master/examples/models). The model is trained on 20 million compounds downloaded from ZINC. radius: int, optional (default 1) The fingerprint radius. The default value was used to train the model which is put on github repository. unseen: str, optional (default 'UNK') The string to used to replace uncommon words/identifiers while training. gather_method: str, optional (default 'sum') How to aggregate vectors of identifiers are extracted from Mol2vec. 'sum' or 'mean' is supported. """ try: from gensim.models import word2vec from mol2vec.features import mol2alt_sentence, sentences2vec except ModuleNotFoundError: raise ValueError("This class requires mol2vec to be installed.") self.radius = radius self.unseen = unseen self.gather_method = gather_method self.sentences2vec = sentences2vec self.mol2alt_sentence = mol2alt_sentence if pretrain_model_path is None: data_dir = get_data_dir() pretrain_model_path = path.join(data_dir, 'mol2vec_model_300dim.pkl') if not path.exists(pretrain_model_path): targz_file = path.join(data_dir, 'mol2vec_model_300dim.tar.gz') if not path.exists(targz_file): download_url(DEFAULT_PRETRAINED_MODEL_URL, data_dir) untargz_file( path.join(data_dir, 'mol2vec_model_300dim.tar.gz'), data_dir) # load pretrained models self.model = word2vec.Word2Vec.load(pretrain_model_path) def _featurize(self, mol: RDKitMol) -> np.ndarray: """ Calculate Mordred descriptors. Parameters ---------- mol: rdkit.Chem.rdchem.Mol RDKit Mol object Returns ------- np.ndarray 1D array of mol2vec fingerprint. The default length is 300. """ sentence = self.mol2alt_sentence(mol, self.radius) vec_identifiers = self.sentences2vec( sentence, self.model, unseen=self.unseen) if self.gather_method == 'sum': feature = np.sum(vec_identifiers, axis=0) elif self.gather_method == 'mean': feature = np.mean(vec_identifiers, axis=0) else: raise ValueError( 'Not supported gather_method type. Please set "sum" or "mean"') return feature
deepchem/feat/molecule_featurizers/mordred_descriptors.py +67 −0 Original line number Diff line number Diff line import numpy as np from deepchem.utils.typing import RDKitMol from deepchem.feat.base_classes import MolecularFeaturizer class MordredDescriptors(MolecularFeaturizer): """Mordred descriptors. This class comptues a list of chemical descriptors using Mordred. Please see the details about all descripors from [1]_, [2]_. Attributes ---------- descriptors: List[str] List of RDKit descriptor names used in this class. References ---------- .. [1] Moriwaki, Hirotomo, et al. "Mordred: a molecular descriptor calculator." Journal of cheminformatics 10.1 (2018): 4. .. [2] http://mordred-descriptor.github.io/documentation/master/descriptors.html Notes ----- This class requires Mordred to be installed. """ def __init__(self, ignore_3D: bool = True): """ Paremeters ---------- ignore_3D: bool, optional (default True) Whether to use 3D information or not. """ try: from mordred import Calculator, descriptors, is_missing except ModuleNotFoundError: raise ValueError("This class requires RDKit to be installed.") self.calc = Calculator(descriptors, ignore_3D=ignore_3D) self.is_missing = is_missing self.descriptors = list(descriptors.__all__) def _featurize(self, mol: RDKitMol) -> np.ndarray: """ Calculate Mordred descriptors. Parameters ---------- mol: rdkit.Chem.rdchem.Mol RDKit Mol object Returns ------- np.ndarray 1D array of Mordred descriptors for `mol`. If ignore_3D is True, the length is 1613. If ignore_3D is False, the length is 1826. """ feature = self.calc(mol) # convert errors to zero feature = [ 0.0 if self.is_missing(val) or isinstance(val, str) else val for val in feature ] return np.asarray(feature)
