def __cal_coul_mat(self, mol):
"""
Parameters
----------
mol: molecule object
Returns
-------
"""
if isinstance(mol, Molecule):
if mol.xyz is None:
msg = "The molecule must be a chemml.chem.Molecule object with xyz information."
raise ValueError(msg)
else:
msg = "The molecule must be a chemml.chem.Molecule object."
raise ValueError(msg)
mol = np.append(mol.xyz.atomic_numbers, mol.xyz.geometry, axis=1)
cm = []
for i in range(len(mol)):
vect = []
for k in range(0, i):
vect.append(cm[k][i])
for j in range(i, len(mol)):
if i == j:
vect.append(0.5 * mol[i, 0]**2.4)
else:
vect.append((mol[i, 0] * mol[j, 0] * self.const) /
np.linalg.norm(mol[i, 1:] - mol[j, 1:]))
for m in range(len(mol), self.max_n_atoms_):
vect.append(0.0)
cm.append(vect)
# pad with zero values
if self.max_n_atoms_ > len(mol):
cm = padaxis(np.array(cm), self.max_n_atoms_, 0, 0)
return np.array(cm)[:self.max_n_atoms_, :
self.max_n_atoms_] #shape nAtoms*nAtoms
def concat_mol_tensors(mol_tensors_list, match_degree=True, match_max_atoms=False):
"""Concatenates a list of molecule tensors
# Arguments:
mol_tensor_list: list of molecule tensors (e.g. list of
`(atoms, bonds, edges)`-triplets)
match_degree: bool, if True, the degrees of all tensors should match,
if False, unmatching degrees will be padded to align them.
match_max_atoms: bool, simular to match_degree but for max_atoms
# Retuns:
a single molecule tensor (as returned by `tensorise_smiles`)
"""
assert isinstance(mol_tensors_list, (tuple, list)), 'Provide a list or tuple of molecule tensors to concatenate'
# get max_atoms (#1) of atoms (#0) tensor of first batch (#0)
# and max_degree (#2) of bonds (#1) tensor of first batch (#0)
max_atoms = mol_tensors_list[0][0].shape[1]
max_degree = mol_tensors_list[0][1].shape[2]
# Obtain the max_degree and max_atoms of all tensors in the list
for atoms, bonds, edges in mol_tensors_list:
assert bonds.shape[0] == edges.shape[0] == atoms.shape[0], "batchsize doesn't match within tensor"
assert bonds.shape[1] == edges.shape[1] == atoms.shape[1], "max_atoms doesn't match within tensor"
assert bonds.shape[2] == edges.shape[2], "degree doesn't match within tensor"
if match_max_atoms:
assert max_atoms == atoms.shape[1], '`max_atoms` of molecule tensors does not match, set `match_max_atoms` to False to adjust'
else:
max_atoms = max(max_atoms, atoms.shape[1])
if match_degree:
assert max_degree == bonds.shape[2], '`degree` of molecule tensors does not match, set `match_degree` to False to adjust'
else:
max_degree = max(max_degree, bonds.shape[2])
# Pad if necessary and separate tensors
atoms_list = []
bonds_list = []
edges_list = []
for atoms, bonds, edges in mol_tensors_list:
atoms = padaxis(atoms, max_atoms, axis=1)
bonds = padaxis(bonds, max_atoms, axis=1)
edges = padaxis(edges, max_atoms, axis=1, pad_value=-1)
bonds = padaxis(bonds, max_degree, axis=2)
edges = padaxis(edges, max_degree, axis=2, pad_value=-1)
atoms_list.append(atoms)
bonds_list.append(bonds)
edges_list.append(edges)
#stack along batch-size axis
atoms = np.concatenate(atoms_list, axis=0)
bonds = np.concatenate(bonds_list, axis=0)
edges = np.concatenate(edges_list, axis=0)
return atoms, bonds, edges
def tensorise_molecules_singlecore(molecules, max_degree=5, max_atoms=None):
"""
Takes a list of molecules and provides tensor representation of atom and bond features.
Parameters
----------
molecules : chemml.chem.Molecule object or array
If list, it must be a list of chemml.chem.Molecule objects, otherwise we raise a ValueError.
In addition, all the molecule objects must provide the SMILES representation.
We try to create the SMILES representation if it's not available.
max_degree : int, optional (default=5)
The maximum number of neighbour per atom that each molecule can have
(to which all molecules will be padded), use 'None' for auto
max_atoms : int, optional (default=None)
The maximum number of atoms per molecule (to which all
molecules will be padded), use 'None' for auto
Notes
-----
It is not recommended to set max_degree to `None`/auto when
using `NeuralGraph` layers. Max_degree determines the number of
trainable parameters and is essentially a hyperparameter.
While models can be rebuilt using different `max_atoms`, they cannot
be rebuild for different values of `max_degree`, as the architecture
will be different.
For organic molecules `max_degree=5` is a good value (Duvenaud et. al, 2015)
Returns
-------
atoms : array
An atom feature array of shape (molecules, max_atoms, atom_features)
bonds : array
A bonds array of shape (molecules, max_atoms, max_degree)
edges : array
A connectivity array of shape (molecules, max_atoms, max_degree, bond_features)
"""
# TODO: Arguments for sparse vector encoding
# molecules
if isinstance(molecules, list) or isinstance(molecules, np.ndarray):
molecules = np.array(molecules)
elif isinstance(molecules, Molecule):
molecules = np.array([molecules])
else:
msg = "The input molecules must be a chemml.chem.Molecule object or a list of objects."
raise ValueError(msg)
# import sizes
n = len(molecules)
n_atom_features = num_atom_features()
n_bond_features = num_bond_features()
# preallocate atom tensor with 0's and bond tensor with -1 (because of 0 index)
# If max_degree or max_atoms is set to None (auto), initialise dim as small
# as possible (1)
atom_tensor = np.zeros((n, max_atoms or 1, n_atom_features))
bond_tensor = np.zeros((n, max_atoms or 1, max_degree or 1, n_bond_features))
edge_tensor = -np.ones((n, max_atoms or 1, max_degree or 1), dtype=int)
for mol_ix, mol in enumerate(molecules):
#load mol, atoms and bonds
if mol.rdkit_molecule is None:
try:
mol.to_smiles()
except:
msg = "The SMILES representation of the molecule %s can not be generated."%str(mol)
raise ValueError(msg)
atoms = mol.rdkit_molecule.GetAtoms()
bonds = mol.rdkit_molecule.GetBonds()
# If max_atoms is exceeded, resize if max_atoms=None (auto), else raise
if len(atoms) > atom_tensor.shape[1]:
assert max_atoms is None, 'too many atoms ({0}) in molecule: {1}'.format(len(atoms), str(mol))
atom_tensor = padaxis(atom_tensor, len(atoms), axis=1)
bond_tensor = padaxis(bond_tensor, len(atoms), axis=1)
edge_tensor = padaxis(edge_tensor, len(atoms), axis=1, pad_value=-1)
rdkit_ix_lookup = {}
connectivity_mat = {}
for atom_ix, atom in enumerate(atoms):
# write atom features
atom_tensor[mol_ix, atom_ix, : n_atom_features] = atom_features(atom)
# store entry in idx
rdkit_ix_lookup[atom.GetIdx()] = atom_ix
# preallocate array with neighbour lists (indexed by atom)
connectivity_mat = [ [] for _ in atoms]
for bond in bonds:
# lookup atom ids
a1_ix = rdkit_ix_lookup[bond.GetBeginAtom().GetIdx()]
a2_ix = rdkit_ix_lookup[bond.GetEndAtom().GetIdx()]
# lookup how many neighbours are encoded yet
a1_neigh = len(connectivity_mat[a1_ix])
a2_neigh = len(connectivity_mat[a2_ix])
# If max_degree is exceeded, resize if max_degree=None (auto), else raise
new_degree = max(a1_neigh, a2_neigh) + 1
if new_degree > bond_tensor.shape[2]:
assert max_degree is None, 'too many neighours ({0}) in molecule: {1}'.format(new_degree, mol)
bond_tensor = padaxis(bond_tensor, new_degree, axis=2)
edge_tensor = padaxis(edge_tensor, new_degree, axis=2, pad_value=-1)
# store bond features
bond_feature = np.array(bond_features(bond), dtype=int)
bond_tensor[mol_ix, a1_ix, a1_neigh, :] = bond_feature
bond_tensor[mol_ix, a2_ix, a2_neigh, :] = bond_feature
# add to connectivity matrix
connectivity_mat[a1_ix].append(a2_ix)
connectivity_mat[a2_ix].append(a1_ix)
# store connectivity matrix
for a1_ix, neighbours in enumerate(connectivity_mat):
degree = len(neighbours)
edge_tensor[mol_ix, a1_ix, : degree] = neighbours
return atom_tensor, bond_tensor, edge_tensor
