[Cdk-devel] VF lib

[Mark R. <ma...@eb...>]

>Yes, a contribution like that would be much appreciated.
>Let me know if I/we can assist you with integrating your code into the
>CDK, for example, regarding practicalities about SVN access, using
>branches, etc.

hi Egon,

Rajarshi's code was pretty much working, I made a few fixes, added an extra sorting class to be able to sort the query atom container, and now I'm writing a JUnit test class. The test is built from some smiles tests and from the current UIT test cases.
I think it's not too much work to wrap it up and put it into the CDK, providing the UIT's JUnit tests can be mapped succesfully. Stephan suggested I'd then zip the source and send it to the list for code review.


There's room for further code extensions. Suggestions from Rajarshi : 
- ideally, the API for this would be identical to that of the UIT
- we could have SubgraphIsomorphism be the gateway class to all isomorphism codes (more extensive refactoring)

But I think we should just release this as it is, and maybe extend functionality later on. The UIT is a big class with many methods (for isomorphism search and  maximum common substructure search). The new classes are more narrow in scope and focus on subgraph searching. Perhaps we should not try to blend these classes into a common API, and stick the new classes in their own package "org.openscience.cdk.subgraph" or so.

thanks,
Mark