Just one more thing (general confusion on my side): how come PseudoAtoms are allowed into the data module? And they are parsed by the current MDL reader.
Technically, aren't those things 'query' elements as well? Or are they abstract on another level..


On Thu, Dec 16, 2010 at 2:55 PM, Mark Rijnbeek <markr@ebi.ac.uk> wrote:
I guess I have to get more clued up on the builder architecture, thanks
for the pointers. It would be good to not copy/paste/duplicate all the
code from readMolecule(). The parsing is the same, just differs at the
point where bond types are dealt with.
Indeed, but we do need a new IQueryBond implementation, being the
MDLQueryBond class.

I will help. I have a very long standing wish to make the isomorphism
module independent of the data module, which the new builder
architecture makes possible, in the same way as I outlined in my
previous reply.



Mark Rijnbeek
Senior Software Engineer
Chemoinformatics and Metabolism Team
European Bioinformatics Institute - EMBL
Tel. +44 (0)1223 494421