Here's a fun one.
I was writing up an article about how to generate a SMILES with an arbitrary
atom ordering by using a lot of dot disconnects.
I had to work around what I think is a problem with RDKit's SMILES parser
Here's a minimal example of the failure
>>> mol = Chem.MolFromSmiles("C1.C12.C2")
>>> mol = Chem.MolFromSmiles("C1.C11.C1")
[04:17:18]
****
Pre-condition Violation
attempt to add self-bond
Violation occurred on line 349 in file
/Users/dalke/ftps/RDKit_Q22010_1/Code/GraphMol/ROMol.cpp
Failed Expression: bond_pin->getBeginAtomIdx()!=bond_pin->getEndAtomIdx()
****
>>>
You can see that I used the same closure number twice for the second atom: once
to close the open closure, and once to start a new one.
This is legitimate, and RDKit supports it
>>> from rdkit import Chem
>>> mol = Chem.MolFromSmiles("C1CCC11CCC1")
but once I put the dot disconnect in, blam!
>>> mol = Chem.MolFromSmiles("C1C.CC11CCC1")
[04:19:31]
****
Pre-condition Violation
attempt to add self-bond
Violation occurred on line 349 in file
/Users/dalke/ftps/RDKit_Q22010_1/Code/GraphMol/ROMol.cpp
Failed Expression: bond_pin->getBeginAtomIdx()!=bond_pin->getEndAtomIdx()
****
while of course
>>> mol = Chem.MolFromSmiles("C1C.CC12CCC2")
has no problem.
Andrew
[email protected]
------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and,
should the need arise, upgrade to a full multi-node Oracle RAC database
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
Rdkit-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
