Re: [Cdk-devel] extending createAnyAtomContainer()

[Kai H. <Kai.Hartmann@Uni-Koeln.De>]

Rajarshi Guha schrieb:
> Hi, I've been using the isomorphism code and I see that there are a
> number of methods to create query molecules.
> 
> However I would like to be able to have a query atom container that has
> arbitrary atoms *and* arbitrary bonds.
> 
> So by analogy with createAnyAtomContainer() I made a method called
> createAnyAtomAnyBondContainer() which just replaces OrderQueryBond with
> AnyOrderQueryBond.
> 
> Thus I would expect that if my target molecule is O1C=CC=C1 and 
> my query fragment is C1CCCC1 then
> 
> UniversalIsomorphismTester.isSubgraph(target, query) == TRUE
> 
> and 
> 
> UniversalIsomorphismTester.isIsomorph(target, query) == TRUE
> 
> However when I run it, it turns out that the former is false and the
> latter is true. In both cases I make sure that I don't check for
> aromaticity.
> 
> I would have thought that both the above would be true. 
> 
> Any pointers would be appreciated

if (!testSubgraphHeuristics(g1, g2)) return false;

It seems with the introduction of the QueryAtomContainer into
UniversalIsomorphismTester, this method has not been updated. Please
enter a bug report. A junit test would be very much appreciated.

Kai