def FullStandardization(smi):
try:
mol = Chem.MolFromSmiles(smi)
if mol == None:
# If rdkit could not parse the smiles, returns Error 1
return "Error 1"
else:
STD = Standardizer()
LFC = LargestFragmentChooser()
UC = Uncharger()
RI = Reionizer()
TC = TautomerCanonicalizer()
mol = STD(mol)
mol = LFC(mol)
allowed_elements = {
"H", "B", "C", "N", "O", "F", "Si", "P", "S", "Cl", "Se", "Br",
"I"
}
actual_elements = set(
[atom.GetSymbol() for atom in mol.GetAtoms()])
if len(actual_elements - allowed_elements) == 0:
mol = UC(mol)
mol = RI(mol)
RemoveStereochemistry(mol)
mol = TC(mol)
return Chem.MolToSmiles(mol)
else:
# If molecule contains other than the allowed elements, returns "Error 2"
return "Error 2"
except:
return "Check manually"
def getReionisedMolecule(mol):
reioniser = Reionizer()
return reioniser.reionize(mol)
def test_reionize():
"""Test reionizer moves proton to weaker acid."""
mol = Chem.MolFromSmiles('C1=C(C=CC(=C1)[S]([O-])=O)[S](O)(=O)=O')
r = Reionizer()
mol = r.reionize(mol)
assert Chem.MolToSmiles(mol) == 'O=S(O)c1ccc(S(=O)(=O)[O-])cc1'
def test_reionize3():
""""""
mol = Chem.MolFromSmiles('C[N+]1=C[CH-]N(C(=N)N)/C1=C/[N+](=O)[O-]')
r = Reionizer()
mol = r.reionize(mol)
assert Chem.MolToSmiles(mol) == 'C[N+]1=CCN(C(=N)N)C1=[C-][N+](=O)[O-]'
def test_reionize2():
"""Test charged carbon doesn't get recognised as alpha-carbon-hydrogen-keto."""
mol = Chem.MolFromSmiles('CCOC(=O)C(=O)[CH-]C#N')
r = Reionizer()
mol = r.reionize(mol)
assert Chem.MolToSmiles(mol) == 'CCOC(=O)C(=O)[CH-]C#N'