H atoms in SGroups cause RDKit to clear SGroups after parsing

[ricrogz]

A simple example (in this case, the H atom is not directly in the SGroup, but in the crossing bond):

from rdkit import Chem
mol = Chem.MolFromMolBlock("""
                    3D

  2  1  0  0  0  0            999 V2000
   -2.3620   -1.8090    1.7866 C   0  0  0  0  0  0
   -2.9887   -1.8667    2.6766 H   0  0  0  0  0  0
  1  2  1  0  0  0
M  STY  1   1 SRU
M  SAL   1  1   1
M  SBL   1  1   1
M  SCN  1   1 HT
M  END""")
print(len(Chem.GetMolSubstanceGroups(mol)))

This will print "0". Looked a bit at it in the debugger, and checked that the SGroup does actually get parsed, but is deleted before asking for the SGroups again. I'm suspecting it has to do with the H atom being turned "explicitH", so I changed it in the input to be an F atom, and in this case, the SGroup is not deleted:

In [9]: mol2 = Chem.MolFromMolBlock("""
   ...:                     3D
   ...: 
   ...:   2  1  0  0  0  0            999 V2000
   ...:    -2.3620   -1.8090    1.7866 C   0  0  0  0  0  0
   ...:    -2.9887   -1.8667    2.6766 F   0  0  0  0  0  0
   ...:   1  2  1  0  0  0
   ...: M  STY  1   1 SRU
   ...: M  SAL   1  1   1
   ...: M  SBL   1  1   1
   ...: M  SCN  1   1 HT
   ...: M  END""")
   ...: print(len(Chem.GetMolSubstanceGroups(mol2)))
   ...: 
   ...: 
1

I'm planning on submitting a patch to keep these H atoms as explicit nodes if they are included in any SGroups atoms, patoms, or bonds list (and possibly other SGroup features, if required).