For feature request #3056330.
Please include sources (if any) for the SMARTS patterns
In the acidic groups descriptor it's not clear what $(n1nnnc1) and $(n1nncn1) are supposed to match. Could you include an example in the Javadocs (or bring the Javadocs in sync with the code as these are commented out)
A few more comprehensive (ie bigger target SMILES) tests would be useful
otherwise looks good
Rajarshi, thanx for your review. The SMARTS are derived from JOELib, per the feature request, but will add that information. I also left a question on BOx . Unfortunately, the JOELib JavaDoc is not more informative about the source of those SMARTS. I assume those two ring system are corner cases, and will fix the JavaDoc.
Regarding the larger molecule, I can add a e.g. this:
Is that what you have a in mind? A larger, curated test set would of course be ideal, but I am not aware of such.
wrt sources, I see. Well if we can determine the sources that's fine.
For the test case, yes, I was just thinking of a few bigger molecules - not necessarily a large curated set. Just something more meaty than HCN :)
has the patch been updated?
Halfway through the updates, but got a bit distracted with finishing up in Uppsala...
Any udates on this?
Rebased on top of cdk-1.4.x and with these fixes:
I think in due time this descriptor can be replaced by a better one, but for now it does implement the feature request.
the test class annotation on AcidicGroupCountDescriptor.java and BasicGroupCountDescriptor is wrong
is it actually useful to have multiple SQT's rather than loooping over the strings?
New patch attached:
Applied and pushed