Whilst loading some mol files from MetaCyc i got a NoSuchAtomType exception.
Investigating it seems that the CDKAtomTypeMatcher defines two types 'Se.2' and "Co.plus" that are present in the matcher but missing in the "cdk-atom-types.owl'. I've added the Co.plus type to the owl file (see. patch) but am not 100% sure on the Se.2 type.
From looking at the naming convention the number seems to indicate the number of formal neighbours for that atom type. However Se does not seem to follow this rule, It might actually be that the 'perceiveSelenium()' is incorrect in providing the atom type identifier but i'm not sure.
J
Git Patch
Patch applied. Se.2 I need to look at. The .2 actually follows the Sybyl scheme here, and refers to sp2...
But, importantly, the CDK atom type identifiers do not follow a naming scheme, unlike several people seem to expect... maybe a feature request for master then?
Closing this one, as the Se.2 part is a duplicate of:
https://sourceforge.net/tracker/?func=detail&aid=3483198&group_id=20024&atid=120024
Oh, and I was wrong about Se.2 ... that's no longer like S.2 where the .2 indicated sp2... not sure what it means in Gilleain's patch...