def pyScorePair(at1, at2, dist, atomCodes=None):
""" Returns a score for an individual atom pair.
>>> m = Chem.MolFromSmiles('CCCCC')
>>> c1 = Utils.GetAtomCode(m.GetAtomWithIdx(0))
>>> c2 = Utils.GetAtomCode(m.GetAtomWithIdx(1))
>>> c3 = Utils.GetAtomCode(m.GetAtomWithIdx(2))
>>> t = 1 | min(c1,c2)<<numPathBits | max(c1,c2)<<(rdMolDescriptors.AtomPairsParameters.codeSize+numPathBits)
>>> pyScorePair(m.GetAtomWithIdx(0),m.GetAtomWithIdx(1),1)==t
1
>>> pyScorePair(m.GetAtomWithIdx(1),m.GetAtomWithIdx(0),1)==t
1
>>> t = 2 | min(c1,c3)<<numPathBits | max(c1,c3)<<(rdMolDescriptors.AtomPairsParameters.codeSize+numPathBits)
>>> pyScorePair(m.GetAtomWithIdx(0),m.GetAtomWithIdx(2),2)==t
1
>>> pyScorePair(m.GetAtomWithIdx(0),m.GetAtomWithIdx(2),2,
... atomCodes=(Utils.GetAtomCode(m.GetAtomWithIdx(0)),Utils.GetAtomCode(m.GetAtomWithIdx(2))))==t
1
"""
if not atomCodes:
code1 = Utils.GetAtomCode(at1)
code2 = Utils.GetAtomCode(at2)
else:
code1, code2 = atomCodes
accum = int(dist) % _maxPathLen
accum |= min(code1, code2) << numPathBits
accum |= max(code1, code2) << (
rdMolDescriptors.AtomPairsParameters.codeSize + numPathBits)
return accum
def ExplainPairScore(score):
"""
>>> m = Chem.MolFromSmiles('C=CC')
>>> score = pyScorePair(m.GetAtomWithIdx(0),m.GetAtomWithIdx(1),1)
>>> ExplainPairScore(score)
(('C', 1, 1), 1, ('C', 2, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(0),m.GetAtomWithIdx(2),2)
>>> ExplainPairScore(score)
(('C', 1, 0), 2, ('C', 1, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(1),m.GetAtomWithIdx(2),1)
>>> ExplainPairScore(score)
(('C', 1, 0), 1, ('C', 2, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(2),m.GetAtomWithIdx(1),1)
>>> ExplainPairScore(score)
(('C', 1, 0), 1, ('C', 2, 1))
"""
codeMask = (1 << rdMolDescriptors.AtomPairsParameters.codeSize) - 1
pathMask = (1 << numPathBits) - 1
dist = score & pathMask
score = score >> numPathBits
code1 = score & codeMask
score = score >> rdMolDescriptors.AtomPairsParameters.codeSize
code2 = score & codeMask
res = Utils.ExplainAtomCode(code1), dist, Utils.ExplainAtomCode(code2)
return res
def testGithub334(self):
m1 = Chem.MolFromSmiles('N#C')
self.assertEqual(Utils.NumPiElectrons(m1.GetAtomWithIdx(0)), 2)
self.assertEqual(Utils.NumPiElectrons(m1.GetAtomWithIdx(1)), 2)
m1 = Chem.MolFromSmiles('N#[CH]')
self.assertEqual(Utils.NumPiElectrons(m1.GetAtomWithIdx(0)), 2)
self.assertEqual(Utils.NumPiElectrons(m1.GetAtomWithIdx(1)), 2)
def ExplainPairScore(score, includeChirality=False):
"""
>>> from rdkit import Chem
>>> m = Chem.MolFromSmiles('C=CC')
>>> score = pyScorePair(m.GetAtomWithIdx(0), m.GetAtomWithIdx(1), 1)
>>> ExplainPairScore(score)
(('C', 1, 1), 1, ('C', 2, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(0), m.GetAtomWithIdx(2), 2)
>>> ExplainPairScore(score)
(('C', 1, 0), 2, ('C', 1, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(1), m.GetAtomWithIdx(2), 1)
>>> ExplainPairScore(score)
(('C', 1, 0), 1, ('C', 2, 1))
>>> score = pyScorePair(m.GetAtomWithIdx(2), m.GetAtomWithIdx(1), 1)
>>> ExplainPairScore(score)
(('C', 1, 0), 1, ('C', 2, 1))
We can optionally deal with chirality too
>>> m = Chem.MolFromSmiles('C[C@H](F)Cl')
>>> score = pyScorePair(m.GetAtomWithIdx(0), m.GetAtomWithIdx(1), 1)
>>> ExplainPairScore(score)
(('C', 1, 0), 1, ('C', 3, 0))
>>> score = pyScorePair(m.GetAtomWithIdx(0), m.GetAtomWithIdx(1), 1, includeChirality=True)
>>> ExplainPairScore(score, includeChirality=True)
(('C', 1, 0, ''), 1, ('C', 3, 0, 'R'))
>>> m = Chem.MolFromSmiles('F[C@@H](Cl)[C@H](F)Cl')
>>> score = pyScorePair(m.GetAtomWithIdx(1), m.GetAtomWithIdx(3), 1, includeChirality=True)
>>> ExplainPairScore(score, includeChirality=True)
(('C', 3, 0, 'R'), 1, ('C', 3, 0, 'S'))
"""
codeSize = rdMolDescriptors.AtomPairsParameters.codeSize
if includeChirality:
codeSize += rdMolDescriptors.AtomPairsParameters.numChiralBits
codeMask = (1 << codeSize) - 1
pathMask = (1 << numPathBits) - 1
dist = score & pathMask
score = score >> numPathBits
code1 = score & codeMask
score = score >> codeSize
code2 = score & codeMask
return (Utils.ExplainAtomCode(code1,
includeChirality=includeChirality), dist,
Utils.ExplainAtomCode(code2, includeChirality=includeChirality))
def pyScorePath(mol, path, size, atomCodes=None):
""" Returns a score for an individual path.
>>> from rdkit import Chem
>>> m = Chem.MolFromSmiles('CCCCC')
>>> c1 = Utils.GetAtomCode(m.GetAtomWithIdx(0),1)
>>> c2 = Utils.GetAtomCode(m.GetAtomWithIdx(1),2)
>>> c3 = Utils.GetAtomCode(m.GetAtomWithIdx(2),2)
>>> c4 = Utils.GetAtomCode(m.GetAtomWithIdx(3),1)
>>> t = c1 | (c2 << rdMolDescriptors.AtomPairsParameters.codeSize) | (c3 << (rdMolDescriptors.AtomPairsParameters.codeSize*2)) | (c4 << (rdMolDescriptors.AtomPairsParameters.codeSize*3))
>>> pyScorePath(m,(0,1,2,3),4)==t
1
The scores are path direction independent:
>>> pyScorePath(m,(3,2,1,0),4)==t
1
>>> m = Chem.MolFromSmiles('C=CC(=O)O')
>>> c1 = Utils.GetAtomCode(m.GetAtomWithIdx(0),1)
>>> c2 = Utils.GetAtomCode(m.GetAtomWithIdx(1),2)
>>> c3 = Utils.GetAtomCode(m.GetAtomWithIdx(2),2)
>>> c4 = Utils.GetAtomCode(m.GetAtomWithIdx(4),1)
>>> t = c1 | (c2 << rdMolDescriptors.AtomPairsParameters.codeSize) | (c3 << (rdMolDescriptors.AtomPairsParameters.codeSize*2)) | (c4 << (rdMolDescriptors.AtomPairsParameters.codeSize*3))
>>> pyScorePath(m,(0,1,2,4),4)==t
1
"""
codes = [None] * size
for i in range(size):
if i == 0 or i == (size - 1):
sub = 1
else:
sub = 2
if not atomCodes:
codes[i] = Utils.GetAtomCode(mol.GetAtomWithIdx(path[i]), sub)
else:
base = atomCodes[path[i]]
codes[i] = base - sub
# "canonicalize" the code vector:
beg = 0
end = len(codes) - 1
while (beg < end):
if codes[beg] > codes[end]:
codes.reverse()
break
elif codes[beg] == codes[end]:
beg += 1
end -= 1
else:
break
accum = 0
for i in range(size):
accum |= codes[i] << (rdMolDescriptors.AtomPairsParameters.codeSize *
i)
return accum
def ExplainPathScore(score, size=4):
"""
>>> from rdkit import Chem
>>> m = Chem.MolFromSmiles('C=CC')
>>> score=pyScorePath(m, (0, 1, 2), 3)
>>> ExplainPathScore(score, 3)
(('C', 1, 0), ('C', 2, 1), ('C', 1, 1))
Again, it's order independent:
>>> score = pyScorePath(m, (2, 1, 0), 3)
>>> ExplainPathScore(score, 3)
(('C', 1, 0), ('C', 2, 1), ('C', 1, 1))
>>> m = Chem.MolFromSmiles('C=CO')
>>> score=pyScorePath(m, (0, 1, 2), 3)
>>> ExplainPathScore(score, 3)
(('C', 1, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('OC=CO')
>>> score=pyScorePath(m, (0, 1, 2, 3), 4)
>>> ExplainPathScore(score, 4)
(('O', 1, 0), ('C', 2, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('CC=CO')
>>> score=pyScorePath(m, (0, 1, 2, 3), 4)
>>> ExplainPathScore(score, 4)
(('C', 1, 0), ('C', 2, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('C=CC(=O)O')
>>> score=pyScorePath(m, (0, 1, 2, 3), 4)
>>> ExplainPathScore(score, 4)
(('C', 1, 1), ('C', 2, 1), ('C', 3, 1), ('O', 1, 1))
>>> score=pyScorePath(m, (0, 1, 2, 4), 4)
>>> ExplainPathScore(score, 4)
(('C', 1, 1), ('C', 2, 1), ('C', 3, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('OOOO')
>>> score=pyScorePath(m, (0, 1, 2), 3)
>>> ExplainPathScore(score, 3)
(('O', 1, 0), ('O', 2, 0), ('O', 2, 0))
>>> score=pyScorePath(m, (0, 1, 2, 3), 4)
>>> ExplainPathScore(score, 4)
(('O', 1, 0), ('O', 2, 0), ('O', 2, 0), ('O', 1, 0))
"""
codeSize: int = rdMolDescriptors.AtomPairsParameters.codeSize
codeMask = (1 << codeSize) - 1
res = [None] * size
for i in range(size):
if i == 0 or i == size - 1:
sub = 1
else:
sub = 2
code = score & codeMask
score = score >> codeSize
symb, nBranch, nPi = Utils.ExplainAtomCode(code)
res[i] = (symb, nBranch + sub, nPi)
return tuple(res)
def ExplainPathScore(score, size=4):
"""
>>> m = Chem.MolFromSmiles('C=CC')
>>> score=pyScorePath(m,(0,1,2),3)
>>> ExplainPathScore(score,3)
(('C', 1, 0), ('C', 2, 1), ('C', 1, 1))
Again, it's order independent:
>>> score=pyScorePath(m,(2,1,0),3)
>>> ExplainPathScore(score,3)
(('C', 1, 0), ('C', 2, 1), ('C', 1, 1))
>>> m = Chem.MolFromSmiles('C=CO')
>>> score=pyScorePath(m,(0,1,2),3)
>>> ExplainPathScore(score,3)
(('C', 1, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('OC=CO')
>>> score=pyScorePath(m,(0,1,2,3),4)
>>> ExplainPathScore(score,4)
(('O', 1, 0), ('C', 2, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('CC=CO')
>>> score=pyScorePath(m,(0,1,2,3),4)
>>> ExplainPathScore(score,4)
(('C', 1, 0), ('C', 2, 1), ('C', 2, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('C=CC(=O)O')
>>> score=pyScorePath(m,(0,1,2,3),4)
>>> ExplainPathScore(score,4)
(('C', 1, 1), ('C', 2, 1), ('C', 3, 1), ('O', 1, 1))
>>> score=pyScorePath(m,(0,1,2,4),4)
>>> ExplainPathScore(score,4)
(('C', 1, 1), ('C', 2, 1), ('C', 3, 1), ('O', 1, 0))
>>> m = Chem.MolFromSmiles('OOOO')
>>> score=pyScorePath(m,(0,1,2),3)
>>> ExplainPathScore(score,3)
(('O', 1, 0), ('O', 2, 0), ('O', 2, 0))
>>> score=pyScorePath(m,(0,1,2,3),4)
>>> ExplainPathScore(score,4)
(('O', 1, 0), ('O', 2, 0), ('O', 2, 0), ('O', 1, 0))
"""
codeMask = (1 << rdMolDescriptors.AtomPairsParameters.codeSize) - 1
res = [None] * size
#print '>>>>>>>>>>>',score,size,codeMask
for i in range(size):
if i == 0 or i == (size - 1):
sub = 1
else:
sub = 2
code = score & codeMask
#print i,code,score
score = score >> rdMolDescriptors.AtomPairsParameters.codeSize
symb, nBranch, nPi = Utils.ExplainAtomCode(code)
expl = symb, nBranch + sub, nPi
res[i] = expl
return tuple(res)