def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
"""
#mol_graph,featureでやってることを書く
type_check_num_atoms(mol, self.max_atoms)
smiles = Chem.MolToSmiles(mol)
ecfp_array = array_rep_from_smiles(smiles, False)
fcfp_array = array_rep_from_smiles(smiles, True)
atom_array1 = ecfp_array['atom_features'].astype(np.float32)
atom_array2 = fcfp_array['atom_features'].astype(np.float32)
degree = construct_adj_matrix(mol, out_size=self.out_size)
self.degree_num = [int(np.sum(x)) for x in degree]
adj_array = ecfp_array['bond_features']
#import pdb;pdb.set_trace()
return atom_array1, atom_array2, adj_array
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
num_atoms = mol.GetNumAtoms()
# Construct the atom array and adjacency matrix.
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
# Adjust the adjacency matrix.
degree_vec = numpy.sum(adj_array[:num_atoms], axis=1)
degree_sqrt_inv = 1. / numpy.sqrt(degree_vec)
adj_array[:num_atoms, :num_atoms] *= numpy.broadcast_to(
degree_sqrt_inv[:, None], (num_atoms, num_atoms))
adj_array[:num_atoms, :num_atoms] *= numpy.broadcast_to(
degree_sqrt_inv[None, :], (num_atoms, num_atoms))
return atom_array, adj_array
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
num_atoms = mol.GetNumAtoms()
# Construct the atom array and adjacency matrix.
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
# Adjust the adjacency matrix.
degree_vec = numpy.sum(adj_array[:num_atoms], axis=1)
degree_sqrt_inv = 1. / numpy.sqrt(degree_vec)
adj_array[:num_atoms, :num_atoms] *= numpy.broadcast_to(
degree_sqrt_inv[:, None], (num_atoms, num_atoms))
adj_array[:num_atoms, :num_atoms] *= numpy.broadcast_to(
degree_sqrt_inv[None, :], (num_atoms, num_atoms))
super_node_x = construct_supernode_feature(mol, atom_array, adj_array, out_size=self.out_size_super)
return atom_array, adj_array, super_node_x
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_tensor = construct_discrete_edge_matrix(mol,
out_size=self.out_size)
return relGCN_pre((atom_array, adj_tensor), out_size=self.out_size)
def get_input_features(self, mol):
type_check_num_atoms(mol, self.max_atoms)
atoms = self.create_atoms(mol)
i_jbond_dict = self.create_ijbonddict(mol)
subgraph_array = self.extract_subgraph(atoms, i_jbond_dict,
self.radius)
adj_array = construct_discrete_edge_matrix(mol)
return subgraph_array, adj_array
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
return atom_array
def get_input_features(self, mol):
"""get input features
Args:
mol (rdkit.Chem.rdchem.Mol)
Returns:
input features
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = self.atom_feature_func(mol, out_size=self.out_size)
adj_tensor = construct_discrete_edge_matrix(mol,
out_size=self.out_size,
kekulize=self.kekulize)
return (atom_array, adj_tensor)
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol): Molecule input
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_discrete_edge_matrix(mol, out_size=self.out_size)
return atom_array, adj_array
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
super_node_x = construct_supernode_feature(mol, atom_array, adj_array, out_size=self.out_size_super)
return atom_array, adj_array, super_node_x
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
dist_array = construct_distance_matrix(mol,
out_size=self.out_size,
contain_Hs=self.add_Hs)
return atom_array, dist_array
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
print ("atom_array : ", atom_array)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
print ("adj_array : ", adj_array)
return atom_array, adj_array
def get_input_features(self, mol):
"""get input features from mol object
Args:
mol (Mol):
"""
type_check_num_atoms(mol, self.max_atoms)
atom_feature = construct_atom_feature(mol, self.use_all_feature,
self.atom_list,
self.include_unknown_atom)
pair_feature, bond_idx = construct_pair_feature(mol,
self.use_all_feature)
global_feature = construct_global_state_feature(mol)
return atom_feature, pair_feature, global_feature, bond_idx
def get_input_features(self, mol):
"""get input features
Args:
mol (Mol):
Returns:
"""
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
# adjust adjacent matrix
degree_vec = numpy.sum(adj_array, axis=1)
degree_vec = numpy.sqrt(degree_vec)
adj_array *= numpy.broadcast_to(degree_vec[:, None], adj_array.shape)
adj_array *= numpy.broadcast_to(degree_vec[None, :], adj_array.shape)
return atom_array, adj_array
def get_input_features(self, mol):
"""get input features for WeaveNet
WeaveNetPreprocessor automatically add `H` to `mol`
Args:
mol (Mol):
"""
type_check_num_atoms(mol, self.max_atoms)
if self.use_fixed_atom_feature:
# original paper feature extraction
atom_array = construct_atom_feature(mol, self.add_Hs,
self.max_atoms, self.atom_list,
self.include_unknown_atom)
else:
# embed id of atomic numbers
atom_array = construct_atomic_number_array(mol, self.max_atoms)
pair_feature = construct_pair_feature(mol,
num_max_atoms=self.max_atoms)
return atom_array, pair_feature
def get_input_features(self, mol):
"""get input features for WeaveNet
WeaveNetPreprocessor automatically add `H` to `mol`
Args:
mol (Mol):
"""
type_check_num_atoms(mol, self.max_atoms)
if self.use_fixed_atom_feature:
# original paper feature extraction
atom_array = construct_atom_feature(mol, self.add_Hs,
self.max_atoms, self.atom_list,
self.include_unknown_atom)
else:
# embed id of atomic numbers
atom_array = construct_atomic_number_array(mol, self.max_atoms)
pair_feature = construct_pair_feature(mol,
num_max_atoms=self.max_atoms)
return atom_array, pair_feature
def get_input_features(self, mol):
type_check_num_atoms(mol, self.max_atoms)
atom_array = construct_atomic_number_array(mol, out_size=self.out_size)
adj_array = construct_adj_matrix(mol, out_size=self.out_size)
return atom_array, adj_array
