Loading deepchem/feat/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -5,6 +5,7 @@ from deepchem.feat.base_classes import Featurizer from deepchem.feat.base_classes import MolecularFeaturizer from deepchem.feat.base_classes import StructureFeaturizer from deepchem.feat.base_classes import CompositionFeaturizer from deepchem.feat.base_classes import ReactionFeaturizer from deepchem.feat.base_classes import ComplexFeaturizer from deepchem.feat.base_classes import UserDefinedFeaturizer from deepchem.feat.graph_features import ConvMolFeaturizer Loading @@ -21,6 +22,7 @@ from deepchem.feat.raw_featurizer import RawFeaturizer from deepchem.feat.raw_featurizer import RawReactionFeaturizer from deepchem.feat.atomic_coordinates import AtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListComplexAtomicCoordinates from deepchem.feat.atomic_coordinates import AtomicConvFeaturizer from deepchem.feat.adjacency_fingerprints import AdjacencyFingerprint from deepchem.feat.smiles_featurizers import SmilesToSeq, SmilesToImage from deepchem.feat.materials_featurizers import ElementPropertyFingerprint, SineCoulombMatrix, StructureGraphFeaturizer deepchem/feat/atomic_coordinates.py +46 −16 Original line number Diff line number Diff line Loading @@ -3,7 +3,8 @@ Atomic coordinate featurizer. """ import logging import numpy as np from deepchem.feat import Featurizer import logging from deepchem.feat import MolecularFeaturizer from deepchem.feat import ComplexFeaturizer from deepchem.utils import rdkit_util, pad_array from deepchem.utils.rdkit_util import MoleculeLoadException Loading @@ -11,7 +12,7 @@ from deepchem.utils.rdkit_util import MoleculeLoadException logger = logging.getLogger(__name__) class AtomicCoordinates(Featurizer): class AtomicCoordinates(MolecularFeaturizer): """ Nx3 matrix of Cartesian coordinates [Angstrom] """ Loading @@ -26,7 +27,6 @@ class AtomicCoordinates(Featurizer): mol : RDKit Mol Molecule. """ N = mol.GetNumAtoms() coords = np.zeros((N, 3)) Loading @@ -47,7 +47,7 @@ class AtomicCoordinates(Featurizer): return coords def compute_neighbor_list(coords, neighbor_cutoff, max_num_neighbors, def _compute_neighbor_list(coords, neighbor_cutoff, max_num_neighbors, periodic_box_size): """Computes a neighbor list from atom coordinates.""" N = coords.shape[0] Loading Loading @@ -92,7 +92,7 @@ def get_coords(mol): return coords class NeighborListAtomicCoordinates(Featurizer): class NeighborListAtomicCoordinates(MolecularFeaturizer): """ Adjacency List of neighbors in 3-space Loading Loading @@ -138,7 +138,7 @@ class NeighborListAtomicCoordinates(Featurizer): # TODO(rbharath): Should this return a list? bohr_coords = self.coordinates_featurizer._featurize(mol)[0] coords = get_coords(mol) neighbor_list = compute_neighbor_list(coords, self.neighbor_cutoff, neighbor_list = _compute_neighbor_list(coords, self.neighbor_cutoff, self.max_num_neighbors, self.periodic_box_size) return (bohr_coords, neighbor_list) Loading Loading @@ -178,13 +178,13 @@ class NeighborListComplexAtomicCoordinates(ComplexFeaturizer): protein_coords, protein_mol = rdkit_util.load_molecule(protein_pdb_file) system_coords = rdkit_util.merge_molecules_xyz([mol_coords, protein_coords]) system_neighbor_list = compute_neighbor_list( system_neighbor_list = _compute_neighbor_list( system_coords, self.neighbor_cutoff, self.max_num_neighbors, None) return (system_coords, system_neighbor_list) class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): class AtomicConvFeaturizer(ComplexFeaturizer): """This class computes the featurization that corresponds to AtomicConvModel. This class computes featurizations needed for AtomicConvModel. Loading @@ -206,22 +206,35 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): """ def __init__(self, frag1_num_atoms, frag2_num_atoms, frag_num_atoms, complex_num_atoms, max_num_neighbors, neighbor_cutoff, strip_hydrogens=True): self.frag1_num_atoms = frag1_num_atoms self.frag2_num_atoms = frag2_num_atoms self.complex_num_atoms = complex_num_atoms """Initialize an AtomicConvFeaturizer object. Parameters ---------- frag_num_atoms: list[int] List of the number of atoms in each fragment. max_num_neighbors: int The maximum number of neighbors allowed neighbor_cutoff: float The distance in angstroms after which neighbors are cutoff. strip_hydrogens: bool, optional If true, remove hydrogens before featurizing. """ # TODO(rbharath): extend to more fragments if len(frag_num_atoms) != 2: raise ValueError("Currently only supports two fragments") self.complex_num_atoms = sum(frag_num_atoms) self.max_num_neighbors = max_num_neighbors self.neighbor_cutoff = neighbor_cutoff self.strip_hydrogens = strip_hydrogens self.neighborlist_featurizer = NeighborListComplexAtomicCoordinates( self.max_num_neighbors, self.neighbor_cutoff) def _featurize(self, mol_pdb_file, protein_pdb_file): frag_coords = [] frag_mols = [] try: frag1_coords, frag1_mol = rdkit_util.load_molecule( mol_pdb_file, is_protein=False, sanitize=True, add_hydrogens=False) Loading Loading @@ -265,7 +278,7 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): def featurize_mol(self, coords, mol, max_num_atoms): logging.info("Featurizing molecule of size: %d", len(mol.GetAtoms())) neighbor_list = compute_neighbor_list(coords, self.neighbor_cutoff, neighbor_list = _compute_neighbor_list(coords, self.neighbor_cutoff, self.max_num_neighbors, None) z = self.get_Z_matrix(mol, max_num_atoms) z = pad_array(z, max_num_atoms) Loading Loading @@ -304,3 +317,20 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): mol = MoleculeShim(atomic_numbers) coords = coords[indexes_to_keep] return coords, mol ############################# Deprecation warning for old name of AtomicConvFeaturizer ############################### DEPRECATION = "{} is deprecated and has been renamed to {} and will be removed in DeepChem 3.0." class ComplexNeighborListFragmentAtomicCoordinates(AtomicConvFeaturizer): def __init__(self, *args, **kwargs): warnings.warn( DEPRECATION.format("ComplexNeighborListFragmentAtomicCoordinates", "AtomicConvFeaturizer"), FutureWarning) super(ComplexNeighborListFragmentAtomicCoordinates, self).__init__( *args, **kwargs) deepchem/feat/graph_features.py +0 −133 Original line number Diff line number Diff line Loading @@ -732,136 +732,3 @@ class WeaveFeaturizer(MolecularFeaturizer): graph_distance=self.graph_distance) return WeaveMol(nodes, pairs) class AtomicConvFeaturizer(ComplexNeighborListFragmentAtomicCoordinates): """This class computes the Atomic Convolution features""" # TODO (VIGS25): Complete the description name = ['atomic_conv'] def __init__(self, labels, neighbor_cutoff, frag1_num_atoms=70, frag2_num_atoms=634, complex_num_atoms=701, max_num_neighbors=12, batch_size=24, atom_types=[ 6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53., -1. ], radial=[[ 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0 ], [0.0, 4.0, 8.0], [0.4]], layer_sizes=[32, 32, 16], strip_hydrogens=True, learning_rate=0.001, epochs=10): """ Parameters labels: numpy.ndarray Labels which we want to predict using the model neighbor_cutoff: int TODO (VIGS25): Add description frag1_num_atoms: int Number of atoms in first fragment frag2_num_atoms: int Number of atoms in second fragment complex_num_atoms: int TODO (VIGS25) : Add description max_num_neighbors: int Maximum number of neighbors possible for an atom batch_size: int Batch size used for training and evaluation atom_types: list List of atoms recognized by model. Atoms are indicated by their nuclear numbers. radial: list TODO (VIGS25): Add description layer_sizes: list List of layer sizes for the AtomicConvolutional Network strip_hydrogens: bool Whether to remove hydrogens while computing neighbor features learning_rate: float Learning rate for training the model epochs: int Number of epochs to train the model for """ self.atomic_conv_model = dc.models.atomic_conv.AtomicConvModel( frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms, max_num_neighbors=max_num_neighbors, batch_size=batch_size, atom_types=atom_types, radial=radial, layer_sizes=layer_sizes, learning_rate=learning_rate) super(AtomicConvFeaturizer, self).__init__( frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms, max_num_neighbors=max_num_neighbors, neighbor_cutoff=neighbor_cutoff, strip_hydrogens=strip_hydrogens) self.epochs = epochs self.labels = labels def featurize(self, mol_files, protein_files): features = [] failures = [] for i, (mol_file, protein_pdb) in enumerate(zip(mol_files, protein_files)): logging.info("Featurizing %d / %d" % (i, len(mol_files))) new_features = self._featurize(mol_file, protein_pdb) # Handle loading failures which return None if new_features is not None: features.append(new_features) else: failures.append(ind) features = np.asarray(features) labels = np.delete(self.labels, failures) dataset = DiskDataset.from_numpy(features, labels) # Fit atomic conv model self.atomic_conv_model.fit(dataset, nb_epoch=self.epochs) # Add the Atomic Convolution layers to fetches layers_to_fetch = [ self.atomic_conv_model._frag1_conv, self.atomic_conv_model._frag2_conv, self.atomic_conv_model._complex_conv ] # Extract the atomic convolution features atomic_conv_features = list() batch_generator = self.atomic_conv_model.default_generator( dataset=dataset, epochs=1) for X, y, w in batch_generator: frag1_conv, frag2_conv, complex_conv = self.atomic_conv_model.predict_on_generator( [(X, y, w)], outputs=layers_to_fetch) concatenated = np.concatenate( [frag1_conv, frag2_conv, complex_conv], axis=1) atomic_conv_features.append(concatenated) batch_size = self.atomic_conv_model.batch_size if len(features) % batch_size != 0: num_batches = (len(features) // batch_size) + 1 num_to_skip = num_batches * batch_size - len(features) else: num_to_skip = 0 atomic_conv_features = np.asarray(atomic_conv_features) atomic_conv_features = atomic_conv_features[-num_to_skip:] atomic_conv_features = np.squeeze(atomic_conv_features) return atomic_conv_features, failures deepchem/feat/rdkit_grid_featurizer.py +55 −269 File changed.Preview size limit exceeded, changes collapsed. Show changes deepchem/feat/tests/test_atomic_coordinates.py +36 −38 Original line number Diff line number Diff line Loading @@ -10,7 +10,7 @@ from deepchem.feat.atomic_coordinates import get_coords from deepchem.feat.atomic_coordinates import AtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListAtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListComplexAtomicCoordinates from deepchem.feat.atomic_coordinates import ComplexNeighborListFragmentAtomicCoordinates from deepchem.feat.atomic_coordinates import AtomicConvFeaturizer logger = logging.getLogger(__name__) Loading Loading @@ -158,41 +158,39 @@ class TestAtomicCoordinates(unittest.TestCase): for atom in range(N): assert len(system_neighbor_list[atom]) <= max_num_neighbors def test_full_complex_featurization(self): """Unit test for AtomicConvFeaturizer.""" dir_path = os.path.dirname(os.path.realpath(__file__)) ligand_file = os.path.join(dir_path, "data/3zso_ligand_hyd.pdb") protein_file = os.path.join(dir_path, "data/3zso_protein.pdb") # Pulled from PDB files. For larger datasets with more PDBs, would use # max num atoms instead of exact. frag1_num_atoms = 44 # for ligand atoms frag2_num_atoms = 2336 # for protein atoms complex_num_atoms = 2380 # in total max_num_neighbors = 4 # Cutoff in angstroms neighbor_cutoff = 4 complex_featurizer = AtomicConvFeaturizer( frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, neighbor_cutoff) (frag1_coords, frag1_neighbor_list, frag1_z, frag2_coords, frag2_neighbor_list, frag2_z, complex_coords, complex_neighbor_list, complex_z) = complex_featurizer._featurize_complex( ligand_file, protein_file) # TODO(rbharath): This test will be uncommented in the next PR up on the docket. # def test_full_complex_featurization(self): # """Unit test for ComplexNeighborListFragmentAtomicCoordinates.""" # dir_path = os.path.dirname(os.path.realpath(__file__)) # ligand_file = os.path.join(dir_path, "data/3zso_ligand_hyd.pdb") # protein_file = os.path.join(dir_path, "data/3zso_protein.pdb") # # Pulled from PDB files. For larger datasets with more PDBs, would use # # max num atoms instead of exact. # frag1_num_atoms = 44 # for ligand atoms # frag2_num_atoms = 2336 # for protein atoms # complex_num_atoms = 2380 # in total # max_num_neighbors = 4 # # Cutoff in angstroms # neighbor_cutoff = 4 # complex_featurizer = ComplexNeighborListFragmentAtomicCoordinates( # frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, # neighbor_cutoff) # (frag1_coords, frag1_neighbor_list, frag1_z, frag2_coords, # frag2_neighbor_list, frag2_z, complex_coords, # complex_neighbor_list, complex_z) = complex_featurizer._featurize_complex( # ligand_file, protein_file) # # assert frag1_coords.shape == (frag1_num_atoms, 3) # self.assertEqual( # sorted(list(frag1_neighbor_list.keys())), list(range(frag1_num_atoms))) # self.assertEqual(frag1_z.shape, (frag1_num_atoms,)) # # self.assertEqual(frag2_coords.shape, (frag2_num_atoms, 3)) # self.assertEqual( # sorted(list(frag2_neighbor_list.keys())), list(range(frag2_num_atoms))) # self.assertEqual(frag2_z.shape, (frag2_num_atoms,)) # # self.assertEqual(complex_coords.shape, (complex_num_atoms, 3)) # self.assertEqual( # sorted(list(complex_neighbor_list.keys())), # list(range(complex_num_atoms))) # self.assertEqual(complex_z.shape, (complex_num_atoms,)) self.assertEqual(frag1_coords.shape, (frag1_num_atoms, 3)) self.assertEqual( sorted(list(frag1_neighbor_list.keys())), list(range(frag1_num_atoms))) self.assertEqual(frag1_z.shape, (frag1_num_atoms,)) self.assertEqual(frag2_coords.shape, (frag2_num_atoms, 3)) self.assertEqual( sorted(list(frag2_neighbor_list.keys())), list(range(frag2_num_atoms))) self.assertEqual(frag2_z.shape, (frag2_num_atoms,)) self.assertEqual(complex_coords.shape, (complex_num_atoms, 3)) self.assertEqual( sorted(list(complex_neighbor_list.keys())), list(range(complex_num_atoms))) self.assertEqual(complex_z.shape, (complex_num_atoms,)) Loading
deepchem/feat/__init__.py +2 −0 Original line number Diff line number Diff line Loading @@ -5,6 +5,7 @@ from deepchem.feat.base_classes import Featurizer from deepchem.feat.base_classes import MolecularFeaturizer from deepchem.feat.base_classes import StructureFeaturizer from deepchem.feat.base_classes import CompositionFeaturizer from deepchem.feat.base_classes import ReactionFeaturizer from deepchem.feat.base_classes import ComplexFeaturizer from deepchem.feat.base_classes import UserDefinedFeaturizer from deepchem.feat.graph_features import ConvMolFeaturizer Loading @@ -21,6 +22,7 @@ from deepchem.feat.raw_featurizer import RawFeaturizer from deepchem.feat.raw_featurizer import RawReactionFeaturizer from deepchem.feat.atomic_coordinates import AtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListComplexAtomicCoordinates from deepchem.feat.atomic_coordinates import AtomicConvFeaturizer from deepchem.feat.adjacency_fingerprints import AdjacencyFingerprint from deepchem.feat.smiles_featurizers import SmilesToSeq, SmilesToImage from deepchem.feat.materials_featurizers import ElementPropertyFingerprint, SineCoulombMatrix, StructureGraphFeaturizer
deepchem/feat/atomic_coordinates.py +46 −16 Original line number Diff line number Diff line Loading @@ -3,7 +3,8 @@ Atomic coordinate featurizer. """ import logging import numpy as np from deepchem.feat import Featurizer import logging from deepchem.feat import MolecularFeaturizer from deepchem.feat import ComplexFeaturizer from deepchem.utils import rdkit_util, pad_array from deepchem.utils.rdkit_util import MoleculeLoadException Loading @@ -11,7 +12,7 @@ from deepchem.utils.rdkit_util import MoleculeLoadException logger = logging.getLogger(__name__) class AtomicCoordinates(Featurizer): class AtomicCoordinates(MolecularFeaturizer): """ Nx3 matrix of Cartesian coordinates [Angstrom] """ Loading @@ -26,7 +27,6 @@ class AtomicCoordinates(Featurizer): mol : RDKit Mol Molecule. """ N = mol.GetNumAtoms() coords = np.zeros((N, 3)) Loading @@ -47,7 +47,7 @@ class AtomicCoordinates(Featurizer): return coords def compute_neighbor_list(coords, neighbor_cutoff, max_num_neighbors, def _compute_neighbor_list(coords, neighbor_cutoff, max_num_neighbors, periodic_box_size): """Computes a neighbor list from atom coordinates.""" N = coords.shape[0] Loading Loading @@ -92,7 +92,7 @@ def get_coords(mol): return coords class NeighborListAtomicCoordinates(Featurizer): class NeighborListAtomicCoordinates(MolecularFeaturizer): """ Adjacency List of neighbors in 3-space Loading Loading @@ -138,7 +138,7 @@ class NeighborListAtomicCoordinates(Featurizer): # TODO(rbharath): Should this return a list? bohr_coords = self.coordinates_featurizer._featurize(mol)[0] coords = get_coords(mol) neighbor_list = compute_neighbor_list(coords, self.neighbor_cutoff, neighbor_list = _compute_neighbor_list(coords, self.neighbor_cutoff, self.max_num_neighbors, self.periodic_box_size) return (bohr_coords, neighbor_list) Loading Loading @@ -178,13 +178,13 @@ class NeighborListComplexAtomicCoordinates(ComplexFeaturizer): protein_coords, protein_mol = rdkit_util.load_molecule(protein_pdb_file) system_coords = rdkit_util.merge_molecules_xyz([mol_coords, protein_coords]) system_neighbor_list = compute_neighbor_list( system_neighbor_list = _compute_neighbor_list( system_coords, self.neighbor_cutoff, self.max_num_neighbors, None) return (system_coords, system_neighbor_list) class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): class AtomicConvFeaturizer(ComplexFeaturizer): """This class computes the featurization that corresponds to AtomicConvModel. This class computes featurizations needed for AtomicConvModel. Loading @@ -206,22 +206,35 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): """ def __init__(self, frag1_num_atoms, frag2_num_atoms, frag_num_atoms, complex_num_atoms, max_num_neighbors, neighbor_cutoff, strip_hydrogens=True): self.frag1_num_atoms = frag1_num_atoms self.frag2_num_atoms = frag2_num_atoms self.complex_num_atoms = complex_num_atoms """Initialize an AtomicConvFeaturizer object. Parameters ---------- frag_num_atoms: list[int] List of the number of atoms in each fragment. max_num_neighbors: int The maximum number of neighbors allowed neighbor_cutoff: float The distance in angstroms after which neighbors are cutoff. strip_hydrogens: bool, optional If true, remove hydrogens before featurizing. """ # TODO(rbharath): extend to more fragments if len(frag_num_atoms) != 2: raise ValueError("Currently only supports two fragments") self.complex_num_atoms = sum(frag_num_atoms) self.max_num_neighbors = max_num_neighbors self.neighbor_cutoff = neighbor_cutoff self.strip_hydrogens = strip_hydrogens self.neighborlist_featurizer = NeighborListComplexAtomicCoordinates( self.max_num_neighbors, self.neighbor_cutoff) def _featurize(self, mol_pdb_file, protein_pdb_file): frag_coords = [] frag_mols = [] try: frag1_coords, frag1_mol = rdkit_util.load_molecule( mol_pdb_file, is_protein=False, sanitize=True, add_hydrogens=False) Loading Loading @@ -265,7 +278,7 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): def featurize_mol(self, coords, mol, max_num_atoms): logging.info("Featurizing molecule of size: %d", len(mol.GetAtoms())) neighbor_list = compute_neighbor_list(coords, self.neighbor_cutoff, neighbor_list = _compute_neighbor_list(coords, self.neighbor_cutoff, self.max_num_neighbors, None) z = self.get_Z_matrix(mol, max_num_atoms) z = pad_array(z, max_num_atoms) Loading Loading @@ -304,3 +317,20 @@ class ComplexNeighborListFragmentAtomicCoordinates(ComplexFeaturizer): mol = MoleculeShim(atomic_numbers) coords = coords[indexes_to_keep] return coords, mol ############################# Deprecation warning for old name of AtomicConvFeaturizer ############################### DEPRECATION = "{} is deprecated and has been renamed to {} and will be removed in DeepChem 3.0." class ComplexNeighborListFragmentAtomicCoordinates(AtomicConvFeaturizer): def __init__(self, *args, **kwargs): warnings.warn( DEPRECATION.format("ComplexNeighborListFragmentAtomicCoordinates", "AtomicConvFeaturizer"), FutureWarning) super(ComplexNeighborListFragmentAtomicCoordinates, self).__init__( *args, **kwargs)
deepchem/feat/graph_features.py +0 −133 Original line number Diff line number Diff line Loading @@ -732,136 +732,3 @@ class WeaveFeaturizer(MolecularFeaturizer): graph_distance=self.graph_distance) return WeaveMol(nodes, pairs) class AtomicConvFeaturizer(ComplexNeighborListFragmentAtomicCoordinates): """This class computes the Atomic Convolution features""" # TODO (VIGS25): Complete the description name = ['atomic_conv'] def __init__(self, labels, neighbor_cutoff, frag1_num_atoms=70, frag2_num_atoms=634, complex_num_atoms=701, max_num_neighbors=12, batch_size=24, atom_types=[ 6, 7., 8., 9., 11., 12., 15., 16., 17., 20., 25., 30., 35., 53., -1. ], radial=[[ 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0 ], [0.0, 4.0, 8.0], [0.4]], layer_sizes=[32, 32, 16], strip_hydrogens=True, learning_rate=0.001, epochs=10): """ Parameters labels: numpy.ndarray Labels which we want to predict using the model neighbor_cutoff: int TODO (VIGS25): Add description frag1_num_atoms: int Number of atoms in first fragment frag2_num_atoms: int Number of atoms in second fragment complex_num_atoms: int TODO (VIGS25) : Add description max_num_neighbors: int Maximum number of neighbors possible for an atom batch_size: int Batch size used for training and evaluation atom_types: list List of atoms recognized by model. Atoms are indicated by their nuclear numbers. radial: list TODO (VIGS25): Add description layer_sizes: list List of layer sizes for the AtomicConvolutional Network strip_hydrogens: bool Whether to remove hydrogens while computing neighbor features learning_rate: float Learning rate for training the model epochs: int Number of epochs to train the model for """ self.atomic_conv_model = dc.models.atomic_conv.AtomicConvModel( frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms, max_num_neighbors=max_num_neighbors, batch_size=batch_size, atom_types=atom_types, radial=radial, layer_sizes=layer_sizes, learning_rate=learning_rate) super(AtomicConvFeaturizer, self).__init__( frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms, max_num_neighbors=max_num_neighbors, neighbor_cutoff=neighbor_cutoff, strip_hydrogens=strip_hydrogens) self.epochs = epochs self.labels = labels def featurize(self, mol_files, protein_files): features = [] failures = [] for i, (mol_file, protein_pdb) in enumerate(zip(mol_files, protein_files)): logging.info("Featurizing %d / %d" % (i, len(mol_files))) new_features = self._featurize(mol_file, protein_pdb) # Handle loading failures which return None if new_features is not None: features.append(new_features) else: failures.append(ind) features = np.asarray(features) labels = np.delete(self.labels, failures) dataset = DiskDataset.from_numpy(features, labels) # Fit atomic conv model self.atomic_conv_model.fit(dataset, nb_epoch=self.epochs) # Add the Atomic Convolution layers to fetches layers_to_fetch = [ self.atomic_conv_model._frag1_conv, self.atomic_conv_model._frag2_conv, self.atomic_conv_model._complex_conv ] # Extract the atomic convolution features atomic_conv_features = list() batch_generator = self.atomic_conv_model.default_generator( dataset=dataset, epochs=1) for X, y, w in batch_generator: frag1_conv, frag2_conv, complex_conv = self.atomic_conv_model.predict_on_generator( [(X, y, w)], outputs=layers_to_fetch) concatenated = np.concatenate( [frag1_conv, frag2_conv, complex_conv], axis=1) atomic_conv_features.append(concatenated) batch_size = self.atomic_conv_model.batch_size if len(features) % batch_size != 0: num_batches = (len(features) // batch_size) + 1 num_to_skip = num_batches * batch_size - len(features) else: num_to_skip = 0 atomic_conv_features = np.asarray(atomic_conv_features) atomic_conv_features = atomic_conv_features[-num_to_skip:] atomic_conv_features = np.squeeze(atomic_conv_features) return atomic_conv_features, failures
deepchem/feat/rdkit_grid_featurizer.py +55 −269 File changed.Preview size limit exceeded, changes collapsed. Show changes
deepchem/feat/tests/test_atomic_coordinates.py +36 −38 Original line number Diff line number Diff line Loading @@ -10,7 +10,7 @@ from deepchem.feat.atomic_coordinates import get_coords from deepchem.feat.atomic_coordinates import AtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListAtomicCoordinates from deepchem.feat.atomic_coordinates import NeighborListComplexAtomicCoordinates from deepchem.feat.atomic_coordinates import ComplexNeighborListFragmentAtomicCoordinates from deepchem.feat.atomic_coordinates import AtomicConvFeaturizer logger = logging.getLogger(__name__) Loading Loading @@ -158,41 +158,39 @@ class TestAtomicCoordinates(unittest.TestCase): for atom in range(N): assert len(system_neighbor_list[atom]) <= max_num_neighbors def test_full_complex_featurization(self): """Unit test for AtomicConvFeaturizer.""" dir_path = os.path.dirname(os.path.realpath(__file__)) ligand_file = os.path.join(dir_path, "data/3zso_ligand_hyd.pdb") protein_file = os.path.join(dir_path, "data/3zso_protein.pdb") # Pulled from PDB files. For larger datasets with more PDBs, would use # max num atoms instead of exact. frag1_num_atoms = 44 # for ligand atoms frag2_num_atoms = 2336 # for protein atoms complex_num_atoms = 2380 # in total max_num_neighbors = 4 # Cutoff in angstroms neighbor_cutoff = 4 complex_featurizer = AtomicConvFeaturizer( frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, neighbor_cutoff) (frag1_coords, frag1_neighbor_list, frag1_z, frag2_coords, frag2_neighbor_list, frag2_z, complex_coords, complex_neighbor_list, complex_z) = complex_featurizer._featurize_complex( ligand_file, protein_file) # TODO(rbharath): This test will be uncommented in the next PR up on the docket. # def test_full_complex_featurization(self): # """Unit test for ComplexNeighborListFragmentAtomicCoordinates.""" # dir_path = os.path.dirname(os.path.realpath(__file__)) # ligand_file = os.path.join(dir_path, "data/3zso_ligand_hyd.pdb") # protein_file = os.path.join(dir_path, "data/3zso_protein.pdb") # # Pulled from PDB files. For larger datasets with more PDBs, would use # # max num atoms instead of exact. # frag1_num_atoms = 44 # for ligand atoms # frag2_num_atoms = 2336 # for protein atoms # complex_num_atoms = 2380 # in total # max_num_neighbors = 4 # # Cutoff in angstroms # neighbor_cutoff = 4 # complex_featurizer = ComplexNeighborListFragmentAtomicCoordinates( # frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, # neighbor_cutoff) # (frag1_coords, frag1_neighbor_list, frag1_z, frag2_coords, # frag2_neighbor_list, frag2_z, complex_coords, # complex_neighbor_list, complex_z) = complex_featurizer._featurize_complex( # ligand_file, protein_file) # # assert frag1_coords.shape == (frag1_num_atoms, 3) # self.assertEqual( # sorted(list(frag1_neighbor_list.keys())), list(range(frag1_num_atoms))) # self.assertEqual(frag1_z.shape, (frag1_num_atoms,)) # # self.assertEqual(frag2_coords.shape, (frag2_num_atoms, 3)) # self.assertEqual( # sorted(list(frag2_neighbor_list.keys())), list(range(frag2_num_atoms))) # self.assertEqual(frag2_z.shape, (frag2_num_atoms,)) # # self.assertEqual(complex_coords.shape, (complex_num_atoms, 3)) # self.assertEqual( # sorted(list(complex_neighbor_list.keys())), # list(range(complex_num_atoms))) # self.assertEqual(complex_z.shape, (complex_num_atoms,)) self.assertEqual(frag1_coords.shape, (frag1_num_atoms, 3)) self.assertEqual( sorted(list(frag1_neighbor_list.keys())), list(range(frag1_num_atoms))) self.assertEqual(frag1_z.shape, (frag1_num_atoms,)) self.assertEqual(frag2_coords.shape, (frag2_num_atoms, 3)) self.assertEqual( sorted(list(frag2_neighbor_list.keys())), list(range(frag2_num_atoms))) self.assertEqual(frag2_z.shape, (frag2_num_atoms,)) self.assertEqual(complex_coords.shape, (complex_num_atoms, 3)) self.assertEqual( sorted(list(complex_neighbor_list.keys())), list(range(complex_num_atoms))) self.assertEqual(complex_z.shape, (complex_num_atoms,))
