def test_multi_file_multi_mol(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [
get_data_file_path('input_five_confs_flexible.sdf'),
get_data_file_path('input_eight_stereoisomers.sdf')
]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == [
'BBB-00000-00.sdf', # The first input has 5 confs of the same molecule
'BBB-00000-01.sdf',
'BBB-00000-02.sdf',
'BBB-00000-03.sdf',
'BBB-00000-04.sdf',
'BBB-00001-00.sdf', # The there are 8 different stereoisomers with 1 conf each
'BBB-00002-00.sdf',
'BBB-00003-00.sdf',
'BBB-00004-00.sdf',
'BBB-00005-00.sdf',
'BBB-00006-00.sdf',
'BBB-00007-00.sdf',
'BBB-00008-00.sdf',
]
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == []
def test_multi_file_multi_mol_duplicates(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [
get_data_file_path('input_one_stereoisomer.sdf'),
get_data_file_path('input_eight_stereoisomers.sdf')
]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == [
'BBB-00000-00.sdf',
'BBB-00001-00.sdf',
'BBB-00002-00.sdf',
'BBB-00003-00.sdf',
'BBB-00004-00.sdf',
'BBB-00005-00.sdf',
'BBB-00006-00.sdf',
'BBB-00007-00.sdf',
]
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
assert len(error_files) == 1
def test_multiple_files(n_procs):
molecules = [Molecule.from_file(get_data_file_path('1-validate_and_assign_graphs_and_confs/BBB-00000-00.sdf'), "sdf"),
Molecule.from_file(get_data_file_path('1-validate_and_assign_graphs_and_confs/BBB-00001-00.sdf'), "sdf")]
coverage, success_mols, error_mols = generate_coverage_report(input_molecules=molecules,
forcefield_name='openff_unconstrained-1.3.0.offxml',
processors=n_procs)
assert len(success_mols) == 2
assert len(error_mols) == 0
assert coverage["passed_unique_molecules"] == 2
assert coverage["total_unique_molecules"] == 2
def test_cli_add_molecules_error(tmpdir):
"""Make sure when adding error molecules the coverage report is not changed and the molecule is put in the error mols folder"""
with tmpdir.as_cwd():
test_dir = '3-coverage_report'
input_dir = "1-validate_and_assign_graphs_and_confs"
# copy all files to a local folder
shutil.copytree(get_data_file_path(input_dir), input_dir)
# run once to get the coverage report
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"-o", test_dir,
input_dir],
catch_exceptions=False)
# count the number of files in the output
n_out_mols = len(os.listdir(test_dir))
# get the coverage report
with open(os.path.join(test_dir, "coverage_report.json")) as report:
old_report = json.load(report)
# now add a new molecule to dir
mol = Molecule.from_file(get_data_file_path("missing_valence_params.sdf"))
mol.properties["group_name"] = "BBB"
mol.properties["molecule_index"] = "99999"
mol.to_file(os.path.join(input_dir, "BBB-99999-00.sdf"), "sdf")
# run again with add
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"--add",
"-o", test_dir,
input_dir],
catch_exceptions=False)
# make sure no new molecules were added
n_new_out_mols = len(os.listdir(test_dir))
assert n_new_out_mols == n_out_mols
# make sure the molecule is in the error folder
assert len(glob.glob(os.path.join(test_dir, "error_mols", "*.sdf"))) == 1
# get the new coverage report and make sure it has been updated
with open(os.path.join(test_dir, "coverage_report.json")) as report:
new_report = json.load(report)
assert new_report.pop("total_unique_molecules") > old_report.pop("total_unique_molecules")
assert new_report.pop("passed_unique_molecules") == old_report.pop("passed_unique_molecules")
assert new_report.pop("forcefield_name") == old_report.pop("forcefield_name")
# now we only have parameter counts left, make sure they have not changed
assert new_report == old_report
def test_cli_add_no_molecules(tmpdir):
"""Make sure that the cli exits if users run coverage report with add but no new molecules are found."""
with tmpdir.as_cwd():
test_dir = '3-coverage_report'
input_folder = get_data_file_path('1-validate_and_assign_graphs_and_confs')
# run once to get the coverage report
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"-o", test_dir,
input_folder],
catch_exceptions=False)
# count the number of files in the output
n_out_mols = len(os.listdir(test_dir))
# now run again with the add flag
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"-o", test_dir,
"--add",
input_folder],
catch_exceptions=False)
assert response.output == f"No new files found in {input_folder}, the coverage report was not changed.\n"
# make sure the number of output files has not changed
n_out_add_mols = len(os.listdir(test_dir))
assert n_out_mols == n_out_add_mols
def test_cli_move_all_confs(tmpdir):
"""
Make sure that if a molecule passes all conformers are also moved.
"""
with tmpdir.as_cwd():
test_dir = '3-coverage_report'
input_folder = get_data_file_path('1-validate_and_assign_graphs_and_confs')
# get the number of input molecules and conformers
n_input_moles = len(glob.glob(os.path.join(input_folder, "*.sdf")))
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"-o", test_dir,
input_folder],
catch_exceptions=False)
n_out_mols = len(glob.glob(os.path.join(test_dir, "*.sdf")))
# assuming no molecules fail
assert n_input_moles == n_out_mols
n_error_mols = len(glob.glob(os.path.join(test_dir, "error_mols", "*.sdf")))
assert n_error_mols == 0
# load the coverage report and make sure the unique mols is correct
with open(os.path.join(test_dir, "coverage_report.json"), "r") as data:
report = json.load(data)
assert report["passed_unique_molecules"] == 5
assert report["total_unique_molecules"] == 5
def test_generate_conformers(tmpdir):
with tmpdir.as_cwd():
# test_name = inspect.stack()[0].function
input_dir = get_data_file_path(
'1-validate_and_assign_graphs_and_confs')
output_dir = '2-generate_conformers'
# generate_conformers(input_dir, output_dir)
response = runner.invoke(
cli,
["preprocess", "generate-conformers", "-o", output_dir, input_dir],
catch_exceptions=False)
## BBB-00000 starts with two conformers, so many more conformers should be created
bbb0_confs = glob.glob(os.path.join(output_dir, 'BBB-00000-*.sdf'))
assert len(bbb0_confs) > 3
## BBB-00001 starts with a one conformer, so many more conformers should be created
bbb1_confs = glob.glob(os.path.join(output_dir, 'BBB-00001-*.sdf'))
assert len(bbb1_confs) > 2
## BBB-00002 starts with one conformer.
# It is rigid so only one conformer should be created
bbb2_confs = glob.glob(os.path.join(output_dir, 'BBB-00002-*.sdf'))
assert len(bbb2_confs) == 1
## BBB-00003 starts with 12 conformers.
# We should see 12 output confs here, since we NEVER delete user confs
bbb3_confs = glob.glob(os.path.join(output_dir, 'BBB-00003-*.sdf'))
assert len(bbb3_confs) == 12
def test_single_file_single_mol(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [get_data_file_path('input_single_mol_rigid.sdf')]
input_mols = [
os.path.abspath(input_mol) for input_mol in input_mols
]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert 'BBB-00000-00.sdf' in output_files
assert len(output_files) == 1
file_text = open(os.path.join(test_dir, 'BBB-00000-00.sdf')).read()
assert """
> <group_name> (1)
BBB""" in file_text
assert """
> <molecule_index> (1)
0""" in file_text
assert """
> <conformer_index> (1)
0""" in file_text
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == []
def test_generate_conformers_add(tmpdir):
with tmpdir.as_cwd():
# test_name = inspect.stack()[0].function
input_dir = '1-validate_and_assign_graphs_and_confs'
# Make a copy of this directory, otherwise we'll contaminate the original when we add a new mol for the test
shutil.copytree(get_data_file_path(input_dir), input_dir)
output_dir = '2-generate_conformers'
# generate_conformers(input_dir, output_dir)
response = runner.invoke(
cli,
["preprocess", "generate-conformers", "-o", output_dir, input_dir],
catch_exceptions=False)
initial_confs = glob.glob(os.path.join(output_dir, 'BBB-*.sdf'))
initial_confs = [
os.path.basename(filename) for filename in initial_confs
]
# now add a new ridiculously flexible molecule to dir
mol = Molecule.from_smiles('CCCCC[C@H](COCOC)COCCOCCCCCCC')
mol.generate_conformers()
mol.properties["group_name"] = "BBB"
mol.properties["molecule_index"] = "99999"
mol.to_file(os.path.join(input_dir, "BBB-99999-00.sdf"), "sdf")
response = runner.invoke(cli, [
"preprocess", "generate-conformers", "-o", output_dir, "--add",
input_dir
],
catch_exceptions=False)
final_confs = glob.glob(os.path.join(output_dir, 'BBB-*.sdf'))
final_confs = [os.path.basename(filename) for filename in final_confs]
assert 'BBB-99999-00.sdf' in final_confs
assert 'BBB-99999-09.sdf' in final_confs
assert len(final_confs) == len(initial_confs) + 10
def test_single_file(n_procs):
molecules = Molecule.from_file(get_data_file_path('1-validate_and_assign_graphs_and_confs/BBB-00000-00.sdf'), "sdf" )
coverage, success_mols, error_mols = generate_coverage_report(input_molecules=molecules,
forcefield_name='openff_unconstrained-1.3.0.offxml',
processors=n_procs)
assert len(success_mols) == 1, error_mols[0][1]
assert len(error_mols) == 0
def test_single_file_multi_mol(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [
get_data_file_path(
'input_one_stereoisomer_and_multi_conf_flexible.sdf')
]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == [
'BBB-00000-00.sdf', 'BBB-00001-00.sdf', 'BBB-00001-01.sdf',
'BBB-00001-02.sdf', 'BBB-00001-03.sdf', 'BBB-00001-04.sdf'
]
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == []
def test_cli_error_mol(tmpdir):
"""Make sure that molecules that fail are correctly put in error mols folder."""
with tmpdir.as_cwd():
test_dir = '3-coverage_report'
input_folder = "2-generate_conformers"
os.mkdir(input_folder)
mol = Molecule.from_file(get_data_file_path('missing_valence_params.sdf'), "sdf", allow_undefined_stereo=True)
mol.properties["group_name"] = "OFF"
mol.properties["molecule_index"] = "00001"
mol.to_file(os.path.join(input_folder, "OFF-00001-00.sdf"), "sdf")
response = runner.invoke(cli, ["preprocess", "coverage-report",
"-p", 1,
"-o", test_dir,
input_folder],
catch_exceptions=False)
n_out_mols = len(glob.glob(os.path.join(test_dir, "*.sdf")))
assert n_out_mols == 0
n_error_mols = len(glob.glob(os.path.join(test_dir, "error_mols", "*.sdf")))
assert n_error_mols == 1
# load the coverage report and make sure the unique mols is correct
with open(os.path.join(test_dir, "coverage_report.json"), "r") as data:
report = json.load(data)
assert report["passed_unique_molecules"] == 0
assert report["total_unique_molecules"] == 1
def test_error_uncovered_antechamber_param():
molecule = Molecule.from_file(get_data_file_path('sodium_carbide.sdf'), "sdf")
coverage, success_mols, error_mols = generate_coverage_report(input_molecules=molecule,
forcefield_name='openff_unconstrained-1.3.0.offxml')
assert len(success_mols) == 0
assert len(error_mols) == 1
assert coverage["passed_unique_molecules"] == 0
assert coverage["total_unique_molecules"] == 1
assert "Command '['antechamber'" in str(error_mols[0][1])
def test_error_missing_valence_param(n_procs):
molecules = Molecule.from_file(get_data_file_path('missing_valence_params.sdf'), "sdf")
coverage, success_mols, error_mols = generate_coverage_report(input_molecules=molecules,
forcefield_name='openff_unconstrained-1.3.0.offxml',
processors=n_procs)
assert len(success_mols) == 0
assert len(error_mols) == 1
assert coverage["passed_unique_molecules"] == 0
assert coverage["total_unique_molecules"] == 1
assert "BondHandler was not able to find parameters" in str(error_mols[0][1])
def test_add_and_delete_existing_error(tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [get_data_file_path('input_one_stereoisomer.sdf')]
with pytest.raises(Exception, match='Can not specify BOTH') as context:
response = runner.invoke(cli, [
"preprocess", "validate", "--add", "--delete-existing", "-g",
"BBB", "-o", test_dir, *input_mols
],
catch_exceptions=False)
def test_do_overwrite_output_directory(tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [get_data_file_path('input_single_mol_rigid.sdf')]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir, *input_mols
],
catch_exceptions=False)
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
"--delete-existing", *input_mols
],
catch_exceptions=False)
def test_double_smirks():
"""Test filtering based on 2 different smirks patterns."""
molecules = []
for i in [0, 1, 2, 3, 5]:
molecules.append(
Molecule.from_file(
get_data_file_path(
f'1-validate_and_assign_graphs_and_confs/BBB-0000{i}-00.sdf'
), "sdf"))
# filter P should only be one molecule, and F should also be one molecule
result = smirks_filter(input_molecules=molecules,
filtered_smirks=["[P:1]", "[F:1]"],
processors=1)
assert result.n_filtered == 2
assert result.n_molecules == 3
def test_multi_file_single_mol_redundant_conf(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [
get_data_file_path('input_single_mol_rigid.sdf'),
get_data_file_path('input_single_mol_rigid_translated.sdf')
]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == ['BBB-00000-00.sdf']
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == ['error_mol_0.sdf']
error_txts = glob.glob(
os.path.join(test_dir, 'error_mols', '*.txt'))
error_txts = [open(fname).read() for fname in error_txts]
assert 'Duplicate molecule conformer input detected' in error_txts[
0]
def test_cli_move_molecules(tmpdir):
"""Make sure that the cli can correctly move the molecules to the passed and fail directories."""
with tmpdir.as_cwd():
input_folder = get_data_file_path(
'1-validate_and_assign_graphs_and_confs')
n_input_moles = len(glob.glob(os.path.join(input_folder, "*.sdf")))
test_dir = '5-smirks_filter'
response = runner.invoke(cli, [
"filter", "smirks", input_folder, test_dir, "-p", 1, "-s", "[P:1]"
],
catch_exceptions=False)
n_out_mols = len(glob.glob(os.path.join(test_dir, "*.sdf")))
# this should only remove 1 molecule with 2 conformers
assert n_out_mols == n_input_moles - 2
n_error_mols = len(
glob.glob(os.path.join(test_dir, "error_mols", "*.sdf")))
assert n_error_mols == 2
def test_dont_overwrite_output_directory(tmpdir):
with tmpdir.as_cwd():
test_name = inspect.stack()[0].function
input_dir = get_data_file_path(
'1-validate_and_assign_graphs_and_confs')
output_dir = os.path.join(test_name, '2-generate_conformers')
response = runner.invoke(
cli,
["preprocess", "generate-conformers", "-o", output_dir, input_dir],
catch_exceptions=False)
with pytest.raises(Exception,
match='Specify `--delete-existing` to remove'):
response = runner.invoke(cli, [
"preprocess", "generate-conformers", "-o", output_dir,
input_dir
],
catch_exceptions=False)
def test_bad_macrocycle(tmpdir):
with tmpdir.as_cwd():
# test_name = inspect.stack()[0].function
input_dir = get_data_file_path(
'1-validate_and_assign_graphs_and_confs_bad_macrocycle')
output_dir = '2-generate_conformers'
# generate_conformers(input_dir, output_dir)
response = runner.invoke(
cli,
["preprocess", "generate-conformers", "-o", output_dir, input_dir],
catch_exceptions=False)
# JAN_00203 has a macrocycle that RDKit generates bad conformers for. These conformers
# have twisted double bonds and can't be parsed by subsequent processing steps.
jan_203_confs = glob.glob(os.path.join(output_dir, 'JAN-00203-*.sdf'))
assert len(jan_203_confs) == 1
def test_add_doesnt_overwrite_error_mols(self, tmpdir):
"""
Run add multiple times, such that error mols are generated by two separate invocations. Then, make sure
that the error outputs don't overwrite each other
"""
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [get_data_file_path('input_single_mol_rigid.sdf')]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
response = runner.invoke(cli, [
"preprocess", "validate", '--add', "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
response = runner.invoke(cli, [
"preprocess", "validate", '--add', "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == ['BBB-00000-00.sdf']
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert sorted(error_files) == [
'error_mol_0.sdf', 'error_mol_1.sdf'
]
error_txts = sorted(
glob.glob(os.path.join(test_dir, 'error_mols', '*.txt')))
error_txts = [open(fname).read() for fname in error_txts]
assert "Input molecule graph is already present in output" in error_txts[
0]
assert "Input molecule graph is already present in output" in error_txts[
1]
def test_add(self, tmpdir):
with tmpdir.as_cwd():
test_dir = '1-validate_and_assign'
input_mols = [get_data_file_path('input_one_stereoisomer.sdf')]
response = runner.invoke(cli, [
"preprocess", "validate", "-g", "BBB", "-o", test_dir,
*input_mols
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == ['BBB-00000-00.sdf']
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == []
# Test that output names were correctly assigned
output_name_assignments = []
with open(os.path.join(test_dir, 'name_assignments.csv')) as of:
csv_reader = csv.reader(of)
for row in csv_reader:
new_row = [os.path.basename(i) for i in row]
output_name_assignments.append(new_row)
assert output_name_assignments == [
['orig_name', 'orig_file', 'orig_file_index', 'out_file_name'],
[
'PDB_DB00136_99', 'input_one_stereoisomer.sdf', '0',
'BBB-00000-00'
],
]
# Run again, with a partially overlapping set of molecules
input_mols2 = [get_data_file_path('input_eight_stereoisomers.sdf')]
response = runner.invoke(cli, [
"preprocess", "validate", "--add", "-g", "BBB", "-o", test_dir,
*input_mols2
],
catch_exceptions=False)
output_files = glob.glob(os.path.join(test_dir, '*.sdf'))
output_files = [os.path.basename(fname) for fname in output_files]
assert sorted(output_files) == [
'BBB-00000-00.sdf',
'BBB-00001-00.sdf',
'BBB-00002-00.sdf',
'BBB-00003-00.sdf',
'BBB-00004-00.sdf',
'BBB-00005-00.sdf',
'BBB-00006-00.sdf',
'BBB-00007-00.sdf',
]
error_files = glob.glob(
os.path.join(test_dir, 'error_mols', '*.sdf'))
error_files = [os.path.basename(fname) for fname in error_files]
assert error_files == ['error_mol_0.sdf']
error_txts = glob.glob(
os.path.join(test_dir, 'error_mols', '*.txt'))
error_txts = [open(fname).read() for fname in error_txts]
assert "Input molecule graph is already present in output" in error_txts[
0]
# Test that output names were correctly assigned
output_name_assignments = []
with open(os.path.join(test_dir, 'name_assignments.csv')) as of:
csv_reader = csv.reader(of)
for row in csv_reader:
new_row = [os.path.basename(i) for i in row]
output_name_assignments.append(new_row)
assert output_name_assignments == [
['orig_name', 'orig_file', 'orig_file_index', 'out_file_name'],
[
'PDB_DB00136_99', 'input_one_stereoisomer.sdf', '0',
'BBB-00000-00'
],
[
'PDB_DB00136_01', 'input_eight_stereoisomers.sdf', '1',
'BBB-00001-00'
],
[
'PDB_DB00136_02', 'input_eight_stereoisomers.sdf', '2',
'BBB-00002-00'
],
[
'PDB_DB00136_03', 'input_eight_stereoisomers.sdf', '3',
'BBB-00003-00'
],
[
'PDB_DB00136_04', 'input_eight_stereoisomers.sdf', '4',
'BBB-00004-00'
],
[
'PDB_DB00136_05', 'input_eight_stereoisomers.sdf', '5',
'BBB-00005-00'
],
[
'PDB_DB00136_06', 'input_eight_stereoisomers.sdf', '6',
'BBB-00006-00'
],
[
'PDB_DB00136_07', 'input_eight_stereoisomers.sdf', '7',
'BBB-00007-00'
],
]
