def test10NormalizeFromData(self):
data = """// Name SMIRKS
Nitro to N+(O-)=O [N,P,As,Sb;X3:1](=[O,S,Se,Te:2])=[O,S,Se,Te:3]>>[*+1:1]([*-1:2])=[*:3]
Sulfone to S(=O)(=O) [S+2:1]([O-:2])([O-:3])>>[S+0:1](=[O-0:2])(=[O-0:3])
Pyridine oxide to n+O- [n:1]=[O:2]>>[n+:1][O-:2]
// Azide to N=N+=N- [*,H:1][N:2]=[N:3]#[N:4]>>[*,H:1][N:2]=[N+:3]=[N-:4]
"""
normalizer1 = rdMolStandardize.Normalizer()
params = rdMolStandardize.CleanupParameters()
normalizer2 = rdMolStandardize.NormalizerFromData(data, params)
imol = Chem.MolFromSmiles("O=N(=O)CCN=N#N", sanitize=False)
mol1 = normalizer1.normalize(imol)
mol2 = normalizer2.normalize(imol)
self.assertEqual(Chem.MolToSmiles(imol), "N#N=NCCN(=O)=O")
self.assertEqual(Chem.MolToSmiles(mol1), "[N-]=[N+]=NCC[N+](=O)[O-]")
self.assertEqual(Chem.MolToSmiles(mol2), "N#N=NCC[N+](=O)[O-]")
Sulfoxide to -S+(O-) [!O:1][S+0;D3:2](=[O:3])[!O:4]>>[*:1][S+1:2]([O-:3])[*:4]
// this form addresses a pathological case that came up a few times in testing:
Sulfoxide to -S+(O-) 2 [!O:1][SH1+1;D3:2](=[O:3])[!O:4]>>[*:1][S+1:2]([O-:3])[*:4]
Trivalent S [O:1]=[S;D2;+0:2]-[#6:3]>>[*:1]=[*+1:2]-[*:3]
// Note that the next one doesn't work propertly because repeated appplications
// don't carry the cations from the previous rounds through. This should be
// fixed by implementing single-molecule transformations, but that's a longer-term
// project
//Alkaline oxide to ions [Li,Na,K;+0:1]-[O+0:2]>>([*+1:1].[O-:2])
Bad amide tautomer1 [C:1]([OH1;D1:2])=;!@[NH1:3]>>[C:1](=[OH0:2])-[NH2:3]
Bad amide tautomer2 [C:1]([OH1;D1:2])=;!@[NH0:3]>>[C:1](=[OH0:2])-[NH1:3]
Halogen with no neighbors [F,Cl,Br,I;X0;+0:1]>>[*-1:1]
Odd pyridine/pyridazine oxide structure [C,N;-;D2,D3:1]-[N+2;D3:2]-[O-;D1:3]>>[*-0:1]=[*+1:2]-[*-:3]
"""
_normalizer_params = rdMolStandardize.CleanupParameters()
_normalizer = rdMolStandardize.NormalizerFromData(_normalization_transforms,
_normalizer_params)
_alkoxide_pattern = Chem.MolFromSmarts('[Li,Na,K;+0]-[#7,#8;+0]')
def normalize_mol(m):
"""
"""
Chem.FastFindRings(m)
if m.HasSubstructMatch(_alkoxide_pattern):
m = Chem.RWMol(m)
for match in m.GetSubstructMatches(_alkoxide_pattern):
m.RemoveBond(match[0], match[1])
m.GetAtomWithIdx(match[0]).SetFormalCharge(1)
m.GetAtomWithIdx(match[1]).SetFormalCharge(-1)
res = _normalizer.normalize(m)