def test_factory_cmiles():
"""
Test the basic factories ability to make cmiles attributes for the molecules.
"""
factory = BasicDatasetFactory()
mol = Molecule.from_smiles("CC")
cmiles_factory = factory.create_cmiles_metadata(mol)
# now make our own cmiles
test_cmiles = {
"canonical_smiles":
mol.to_smiles(isomeric=False, explicit_hydrogens=False, mapped=False),
"canonical_isomeric_smiles":
mol.to_smiles(isomeric=True, explicit_hydrogens=False, mapped=False),
"canonical_explicit_hydrogen_smiles":
mol.to_smiles(isomeric=False, explicit_hydrogens=True, mapped=False),
"canonical_isomeric_explicit_hydrogen_smiles":
mol.to_smiles(isomeric=True, explicit_hydrogens=True, mapped=False),
"canonical_isomeric_explicit_hydrogen_mapped_smiles":
mol.to_smiles(isomeric=True, explicit_hydrogens=True, mapped=True),
"molecular_formula":
mol.hill_formula,
"standard_inchi":
mol.to_inchi(fixed_hydrogens=False),
"inchi_key":
mol.to_inchikey(fixed_hydrogens=False),
"fixed_hydrogen_inchi":
mol.to_inchi(fixed_hydrogens=True),
"fixed_hydrogen_inchi_key":
mol.to_inchikey(fixed_hydrogens=True),
"unique_fixed_hydrogen_inchi_keys": {'OTMSDBZUPAUEDD-UHFFFAOYNA-N'},
}
assert test_cmiles == cmiles_factory
def test_scf_properties_assignment():
"""Test adding different scf_properties and make sure they are validated correctly."""
factory = BasicDatasetFactory()
# incorrect spellings
with pytest.raises(DatasetInputError):
factory.scf_properties = ["diapole", "qudrupole"]
def test_adding_removing_scf_properties():
"""
Test adding different scf_properties which should be passed through validation.
"""
factory = BasicDatasetFactory()
# test strange caps
factory.scf_properties = ["QuaDruPole"]
factory.remove_scf_property(scf_property="QuaDruPole")
# test adding new property
factory.add_scf_property(scf_property="mulliken_charges")
# add the wrong property
with pytest.raises(DatasetInputError):
factory.add_scf_property(scf_property="FakeProperty")
assert factory.scf_properties == ["mulliken_charges"]
def test_factory_round_trip(file_type, tmpdir):
"""
Test round tripping a factory to file with a workflow.
"""
with tmpdir.as_cwd():
factory = BasicDatasetFactory(driver="energy", maxiter=1)
efilter = workflow_components.ElementFilter()
weight = workflow_components.MolecularWeightFilter()
conformer = workflow_components.StandardConformerGenerator()
factory.add_workflow_components(efilter, weight, conformer)
file_name = "test." + file_type
factory.export(file_name)
factory2 = BasicDatasetFactory.from_file(file_name)
assert factory2.driver == factory.driver
assert factory2.workflow == factory.workflow
def test_basic_submissions_single_spec(fractal_compute_server, specification):
"""Test submitting a basic dataset to a snowflake server."""
client = FractalClient(fractal_compute_server)
qc_spec, driver = specification
program = qc_spec["program"]
if not has_program(program):
pytest.skip(f"Program '{program}' not found.")
molecules = Molecule.from_file(get_data("butane_conformers.pdb"), "pdb")
factory = BasicDatasetFactory(driver=driver)
factory.add_qc_spec(**qc_spec,
spec_name="default",
spec_description="testing the single points",
overwrite=True)
dataset = factory.create_dataset(
dataset_name=f"Test single points info {program}, {driver}",
molecules=molecules,
description="Test basics dataset",
tagline="Testing single point datasets",
)
# force a metadata validation error
dataset.metadata.long_description = None
with pytest.raises(DatasetInputError):
dataset.submit(client=client)
# re-add the description so we can submit the data
dataset.metadata.long_description = "Test basics dataset"
# now submit again
dataset.submit(client=client)
fractal_compute_server.await_results()
# make sure of the results are complete
ds = client.get_collection("Dataset", dataset.dataset_name)
# check the metadata
meta = Metadata(**ds.data.metadata)
assert meta == dataset.metadata
assert ds.data.description == dataset.description
assert ds.data.tagline == dataset.dataset_tagline
assert ds.data.tags == dataset.dataset_tags
# check the provenance
assert dataset.provenance == ds.data.provenance
# check the qc spec
assert ds.data.default_driver == dataset.driver
# get the last ran spec
for specification in ds.data.history:
driver, program, method, basis, spec_name = specification
spec = dataset.qc_specifications[spec_name]
assert driver == dataset.driver
assert program == spec.program
assert method == spec.method
assert basis == spec.basis
break
else:
raise RuntimeError(
f"The requested compute was not found in the history {ds.data.history}"
)
for spec in dataset.qc_specifications.values():
query = ds.get_records(
method=spec.method,
basis=spec.basis,
program=spec.program,
)
# make sure all of the conformers were submitted
assert len(query.index) == len(molecules)
for index in query.index:
result = query.loc[index].record
assert result.status.value.upper() == "COMPLETE"
assert result.error is None
assert result.return_result is not None
def test_basic_submissions_wavefunction(fractal_compute_server):
"""
Test submitting a basic dataset with a wavefunction protocol and make sure it is executed.
"""
# only a psi4 test
if not has_program("psi4"):
pytest.skip("Program psi4 not found.")
client = FractalClient(fractal_compute_server)
molecules = Molecule.from_file(get_data("butane_conformers.pdb"), "pdb")
factory = BasicDatasetFactory(driver="energy")
factory.clear_qcspecs()
factory.add_qc_spec(method="hf",
basis="sto-6g",
program="psi4",
spec_name="default",
spec_description="wavefunction spec",
store_wavefunction="orbitals_and_eigenvalues")
dataset = factory.create_dataset(
dataset_name="Test single points with wavefunction",
molecules=molecules,
description="Test basics dataset",
tagline="Testing single point datasets with wavefunction",
)
# submit the dataset
# now submit again
dataset.submit(client=client)
fractal_compute_server.await_results()
# make sure of the results are complete
ds = client.get_collection("Dataset", dataset.dataset_name)
# check the metadata
meta = Metadata(**ds.data.metadata)
assert meta == dataset.metadata
assert ds.data.description == dataset.description
assert ds.data.tagline == dataset.dataset_tagline
assert ds.data.tags == dataset.dataset_tags
# check the provenance
assert dataset.provenance == ds.data.provenance
# check the qc spec
assert ds.data.default_driver == dataset.driver
# get the last ran spec
for specification in ds.data.history:
driver, program, method, basis, spec_name = specification
spec = dataset.qc_specifications[spec_name]
assert driver == dataset.driver
assert program == spec.program
assert method == spec.method
assert basis == spec.basis
for spec in dataset.qc_specifications.values():
query = ds.get_records(
method=spec.method,
basis=spec.basis,
program=spec.program,
)
for index in query.index:
result = query.loc[index].record
assert result.status.value.upper() == "COMPLETE"
assert result.error is None
assert result.return_result is not None
basis = result.get_wavefunction("basis")
assert basis.name.lower() == "sto-6g"
orbitals = result.get_wavefunction("orbitals_a")
assert orbitals.shape is not None
def test_basic_submissions_single_pcm_spec(fractal_compute_server):
"""Test submitting a basic dataset to a snowflake server with pcm water in the specification."""
client = FractalClient(fractal_compute_server)
program = "psi4"
if not has_program(program):
pytest.skip(f"Program '{program}' not found.")
molecules = Molecule.from_file(get_data("butane_conformers.pdb"), "pdb")
factory = BasicDatasetFactory(driver="energy")
factory.add_qc_spec(method="hf",
basis="sto-3g",
program=program,
spec_name="default",
spec_description="testing the single points with pcm",
implicit_solvent=PCMSettings(units="au",
medium_Solvent="water"),
overwrite=True)
# only use one molecule due to the time it takes to run with pcm
dataset = factory.create_dataset(
dataset_name="Test single points with pcm water",
molecules=molecules[0],
description="Test basics dataset with pcm water",
tagline="Testing single point datasets with pcm water",
)
# force a metadata validation error
dataset.metadata.long_description = None
with pytest.raises(DatasetInputError):
dataset.submit(client=client)
# re-add the description so we can submit the data
dataset.metadata.long_description = "Test basics dataset"
# now submit again
dataset.submit(client=client)
fractal_compute_server.await_results()
# make sure of the results are complete
ds = client.get_collection("Dataset", dataset.dataset_name)
# check the metadata
meta = Metadata(**ds.data.metadata)
assert meta == dataset.metadata
assert ds.data.description == dataset.description
assert ds.data.tagline == dataset.dataset_tagline
assert ds.data.tags == dataset.dataset_tags
# check the provenance
assert dataset.provenance == ds.data.provenance
# check the qc spec
assert ds.data.default_driver == dataset.driver
# get the last ran spec
for specification in ds.data.history:
driver, program, method, basis, spec_name = specification
spec = dataset.qc_specifications[spec_name]
assert driver == dataset.driver
assert program == spec.program
assert method == spec.method
assert basis == spec.basis
break
else:
raise RuntimeError(
f"The requested compute was not found in the history {ds.data.history}"
)
for spec in dataset.qc_specifications.values():
query = ds.get_records(
method=spec.method,
basis=spec.basis,
program=spec.program,
)
for index in query.index:
result = query.loc[index].record
assert result.status.value.upper() == "COMPLETE"
assert result.error is None
assert result.return_result is not None
# make sure the PCM result was captured
assert result.extras["qcvars"]["PCM POLARIZATION ENERGY"] < 0
def test_basic_submissions_multiple_spec(fractal_compute_server):
"""Test submitting a basic dataset to a snowflake server with multiple qcspecs."""
client = FractalClient(fractal_compute_server)
qc_specs = [{
"method": "openff-1.0.0",
"basis": "smirnoff",
"program": "openmm",
"spec_name": "openff"
}, {
"method": "gaff-2.11",
"basis": "antechamber",
"program": "openmm",
"spec_name": "gaff"
}]
molecules = Molecule.from_file(get_data("butane_conformers.pdb"), "pdb")
factory = BasicDatasetFactory(driver="energy")
factory.clear_qcspecs()
for spec in qc_specs:
factory.add_qc_spec(**spec,
spec_description="testing the single points")
dataset = factory.create_dataset(
dataset_name="Test single points multiple specs",
molecules=molecules,
description="Test basics dataset",
tagline="Testing single point datasets",
)
# force a metadata validation error
dataset.metadata.long_description = None
with pytest.raises(DatasetInputError):
dataset.submit(client=client)
# re-add the description so we can submit the data
dataset.metadata.long_description = "Test basics dataset"
# now submit again
dataset.submit(client=client)
fractal_compute_server.await_results()
# make sure of the results are complete
ds = client.get_collection("Dataset", dataset.dataset_name)
# check the metadata
meta = Metadata(**ds.data.metadata)
assert meta == dataset.metadata
assert ds.data.description == dataset.description
assert ds.data.tagline == dataset.dataset_tagline
assert ds.data.tags == dataset.dataset_tags
# check the provenance
assert dataset.provenance == ds.data.provenance
# check the qc spec
assert ds.data.default_driver == dataset.driver
# get the last ran spec
for specification in ds.data.history:
driver, program, method, basis, spec_name = specification
spec = dataset.qc_specifications[spec_name]
assert driver == dataset.driver
assert program == spec.program
assert method == spec.method
assert basis == spec.basis
for spec in dataset.qc_specifications.values():
query = ds.get_records(
method=spec.method,
basis=spec.basis,
program=spec.program,
)
# make sure all conformers are submitted
assert len(query.index) == len(molecules)
for index in query.index:
result = query.loc[index].record
assert result.status.value.upper() == "COMPLETE"
assert result.error is None
assert result.return_result is not None
def get_molecule_cmiles(molecule: off.Molecule) -> MoleculeAttributes:
"""
Generate the molecule cmiles data.
"""
factory = BasicDatasetFactory()
return factory.create_cmiles_metadata(molecule)