def test_match_conformers(self):
matches = MoleculeStore._match_conformers(
"[Cl:1][H:2]",
db_conformers=[
DBConformerRecord(
coordinates=numpy.array([[-1.0, 0.0, 0.0], [1.0, 0.0, 0.0]])
),
DBConformerRecord(
coordinates=numpy.array([[-2.0, 0.0, 0.0], [2.0, 0.0, 0.0]])
),
],
conformers=[
ConformerRecord(
coordinates=numpy.array([[0.0, -2.0, 0.0], [0.0, 2.0, 0.0]]),
partial_charges=[],
bond_orders=[],
),
ConformerRecord(
coordinates=numpy.array([[0.0, -2.0, 0.0], [0.0, 3.0, 0.0]]),
partial_charges=[],
bond_orders=[],
),
ConformerRecord(
coordinates=numpy.array([[0.0, 0.0, 0.0], [-2.0, 0.0, 0.0]]),
partial_charges=[],
bond_orders=[],
),
],
)
assert matches == {0: 1, 2: 0}
def test_average_partial_charges(self):
record = MoleculeRecord(
smiles="[C:1]([H:2])([H:3])([H:4])",
conformers=[
ConformerRecord(
coordinates=numpy.ones((4, 3)),
partial_charges=[
PartialChargeSet(method="am1", values=[0.1, 0.2, 0.3, 0.4]),
],
),
ConformerRecord(
coordinates=numpy.zeros((4, 3)),
partial_charges=[
PartialChargeSet(method="am1", values=[0.3, 0.4, 0.5, 0.6]),
],
),
],
)
average_charges = record.average_partial_charges("am1")
assert isinstance(average_charges, tuple)
assert len(average_charges) == 4
assert numpy.allclose(average_charges, (0.2, 0.3, 0.4, 0.5))
def test_store_bond_order_data(self, tmp_path):
store = MoleculeStore(f"{tmp_path}.sqlite")
store.store(
MoleculeRecord(
smiles="[Cl:1][H:2]",
conformers=[
ConformerRecord(
coordinates=numpy.arange(6).reshape((2, 3)),
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 0.5)])
],
)
],
)
)
assert len(store) == 1
with pytest.raises(
RuntimeError, match=re.escape("am1 WBOs already stored for [Cl:1][H:2]")
):
store.store(
MoleculeRecord(
smiles="[Cl:2][H:1]",
conformers=[
ConformerRecord(
coordinates=numpy.arange(6).reshape((2, 3)),
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 0.5)])
],
)
],
)
)
store.store(
MoleculeRecord(
smiles="[Cl:2][H:1]",
conformers=[
ConformerRecord(
coordinates=numpy.zeros((2, 3)),
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 0.5)])
],
)
],
)
)
assert len(store) == 1
assert {*store.wbo_methods} == {"am1"}
record = store.retrieve()[0]
assert len(record.conformers) == 2
def test_reorder(self):
original_coordinates = numpy.arange(6).reshape((2, 3))
original_record = MoleculeRecord(
smiles="[Cl:2][H:1]",
conformers=[
ConformerRecord(
coordinates=original_coordinates,
partial_charges=[PartialChargeSet(method="am1", values=[0.5, 1.5])],
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 0.2)])
],
)
],
)
reordered_record = original_record.reorder("[Cl:1][H:2]")
assert reordered_record.smiles == "[Cl:1][H:2]"
reordered_conformer = reordered_record.conformers[0]
assert numpy.allclose(
reordered_conformer.coordinates, numpy.flipud(original_coordinates)
)
assert numpy.allclose(reordered_conformer.partial_charges[0].values, [1.5, 0.5])
assert numpy.allclose(reordered_conformer.bond_orders[0].values, [(1, 0, 0.2)])
def test_validate_coordinates(self, value, expected_raises):
with expected_raises:
record = ConformerRecord(coordinates=value)
assert isinstance(record.coordinates, numpy.ndarray)
assert record.coordinates.flags.writeable is False
def test_bond_orders_by_method(self):
record = ConformerRecord(
coordinates=numpy.ones((2, 3)),
bond_orders=[WibergBondOrderSet(method="am1", values=[(0, 1, 0.1)])],
)
assert record.bond_orders_by_method == {"am1": ((0, 1, 0.1),)}
def tmp_molecule_store(tmp_path) -> MoleculeStore:
store = MoleculeStore(f"{tmp_path}.sqlite")
expected_records = [
MoleculeRecord(
smiles="[Ar:1]",
conformers=[
ConformerRecord(
coordinates=numpy.array([[0.0, 0.0, 0.0]]),
partial_charges=[PartialChargeSet(method="am1", values=[0.5])],
bond_orders=[],
)
],
),
MoleculeRecord(
smiles="[He:1]",
conformers=[
ConformerRecord(
coordinates=numpy.array([[0.0, 0.0, 0.0]]),
partial_charges=[PartialChargeSet(method="am1bcc", values=[-0.5])],
bond_orders=[],
)
],
),
MoleculeRecord(
smiles="[Cl:1][Cl:2]",
conformers=[
ConformerRecord(
coordinates=numpy.array([[-1.0, 0.0, 0.0], [1.0, 0.0, 0.0]]),
partial_charges=[
PartialChargeSet(method="am1", values=[0.5, -0.5]),
PartialChargeSet(method="am1bcc", values=[0.75, -0.75]),
],
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 1.2)])
],
)
],
),
]
store.store(*expected_records)
return store
def test_validate_partial_charges(self, value, expected_raises):
with expected_raises:
record = ConformerRecord(
coordinates=numpy.ones((2, 3)), partial_charges=value
)
assert isinstance(record.partial_charges, tuple)
assert len(record.partial_charges) == len(value)
def test_partial_charges_by_method(self):
record = ConformerRecord(
coordinates=numpy.ones((4, 3)),
partial_charges=[
PartialChargeSet(method="am1", values=[0.1, 0.2, 0.3, 0.4]),
PartialChargeSet(method="am1bcc", values=[1.0, 2.0, 3.0, 4.0]),
],
)
assert record.partial_charges_by_method == {
"am1": (0.1, 0.2, 0.3, 0.4),
"am1bcc": (1.0, 2.0, 3.0, 4.0),
}
def test_data_set_from_molecule_stores(tmpdir):
molecule_store = MoleculeStore(os.path.join(tmpdir, "store.sqlite"))
molecule_store.store(
MoleculeRecord(
smiles="[Cl:1]-[H:2]",
conformers=[
ConformerRecord(
coordinates=numpy.array([[-1.0, 0.0, 0.0], [1.0, 0.0,
0.0]]),
partial_charges=[
PartialChargeSet(method="am1", values=[0.1, -0.1])
],
bond_orders=[
WibergBondOrderSet(method="am1", values=[(0, 1, 1.1)])
],
)
],
))
data_set = DGLMoleculeDataset.from_molecule_stores(molecule_store, "am1",
"am1",
[AtomConnectivity()],
[BondIsInRing()])
assert len(data_set) == 1
assert data_set.n_features == 4
dgl_molecule, labels = data_set[0]
assert isinstance(dgl_molecule, DGLMolecule)
assert dgl_molecule.n_atoms == 2
assert "am1-charges" in labels
assert labels["am1-charges"].numpy().shape == (2, )
assert "am1-wbo" in labels
assert labels["am1-wbo"].numpy().shape == (1, )
def mock_data_store(self, tmpdir) -> str:
store_path = os.path.join(tmpdir, "store.sqlite")
store = MoleculeStore(store_path)
store.store(
MoleculeRecord(
smiles="[Cl:1][Cl:2]",
conformers=[
ConformerRecord(
coordinates=numpy.array([[0.0, 0.0, 0.0],
[1.0, 0.0, 0.0]]),
partial_charges=[
PartialChargeSet(method="am1bcc",
values=[1.0, -1.0])
],
bond_orders=[
WibergBondOrderSet(method="am1",
values=[(0, 1, 1.0)])
],
)
],
))
return store_path
def label_molecule(
molecule: Union[str, "Molecule"],
guess_stereochemistry: bool,
partial_charge_methods: List[ChargeMethod],
bond_order_methods: List[WBOMethod],
n_conformers: int = 500,
rms_cutoff: float = 0.05,
) -> MoleculeRecord:
"""Computes sets of partial charges and bond orders for an input molecule.
Notes:
Conformers will be pruned using the ELF10 method provided by the OpenFF toolkit
Args:
molecule: The molecule (or SMILES representation of the molecule) to label.
guess_stereochemistry: Whether to guess the stereochemistry of the SMILES
representation of the molecule if provided and if the stereochemistry of
some atoms / bonds is not fully defined.
partial_charge_methods: The methods to compute the partial charges using.
bond_order_methods: The methods to compute the bond orders using.
n_conformers: The *maximum* number of conformers to compute partial charge and
bond orders using.
rms_cutoff: The RMS cutoff [Å] to use when generating the conformers.
Returns:
The labelled molecule stored in a record object
"""
from simtk import unit
if isinstance(molecule, str):
molecule = smiles_to_molecule(
molecule, guess_stereochemistry=guess_stereochemistry)
else:
molecule = copy.deepcopy(molecule)
# Generate a diverse set of ELF10 conformers
molecule.generate_conformers(n_conformers=n_conformers,
rms_cutoff=rms_cutoff * unit.angstrom)
molecule.apply_elf_conformer_selection()
conformer_records = []
for conformer in molecule.conformers:
charge_sets = []
for charge_method in partial_charge_methods:
molecule.assign_partial_charges(
_OPENFF_CHARGE_METHODS[charge_method],
use_conformers=[conformer])
charge_sets.append(
PartialChargeSet(
method=charge_method,
values=[
atom.partial_charge.value_in_unit(
unit.elementary_charge) for atom in molecule.atoms
],
))
bond_order_sets = []
for bond_order_method in bond_order_methods:
molecule.assign_fractional_bond_orders(
_OPENFF_WBO_METHODS[bond_order_method],
use_conformers=[conformer])
bond_order_sets.append(
WibergBondOrderSet(
method=bond_order_method,
values=[(
bond.atom1_index,
bond.atom2_index,
bond.fractional_bond_order,
) for bond in molecule.bonds],
))
conformer_records.append(
ConformerRecord(
coordinates=conformer.value_in_unit(unit.angstrom),
partial_charges=charge_sets,
bond_orders=bond_order_sets,
))
return MoleculeRecord(
smiles=molecule.to_smiles(isomeric=True, mapped=True),
conformers=conformer_records,
)
def test_store_partial_charge_data(self, tmp_path):
store = MoleculeStore(f"{tmp_path}.sqlite")
store.store(
MoleculeRecord(
smiles="[Cl:1][H:2]",
conformers=[
ConformerRecord(
coordinates=numpy.arange(6).reshape((2, 3)),
partial_charges=[
PartialChargeSet(method="am1", values=[0.50, 1.50])
],
)
],
)
)
assert len(store) == 1
store.store(
MoleculeRecord(
smiles="[Cl:2][H:1]",
conformers=[
ConformerRecord(
coordinates=numpy.flipud(numpy.arange(6).reshape((2, 3))),
partial_charges=[
PartialChargeSet(method="am1bcc", values=[0.25, 0.75])
],
)
],
)
)
assert len(store) == 1
assert {*store.charge_methods} == {"am1", "am1bcc"}
record = store.retrieve()[0]
assert len(record.conformers) == 1
with pytest.raises(
RuntimeError,
match=re.escape("am1bcc charges already stored for [Cl:1][H:2]"),
):
store.store(
MoleculeRecord(
smiles="[Cl:2][H:1]",
conformers=[
ConformerRecord(
coordinates=numpy.arange(6).reshape((2, 3)),
partial_charges=[
PartialChargeSet(method="am1bcc", values=[0.25, 0.75])
],
)
],
)
)
assert len(store) == 1
assert {*store.charge_methods} == {"am1", "am1bcc"}
record = store.retrieve()[0]
assert len(record.conformers) == 1
def test_validate_bond_orders(self, value, expected_raises):
with expected_raises:
record = ConformerRecord(coordinates=numpy.ones((2, 3)), bond_orders=value)
assert isinstance(record.bond_orders, tuple)
assert len(record.bond_orders) == len(value)