Thank you John and Jonas for your answers.
One big issue is I still don't have a good grasp of how CDK does things. The
second is that I'm doing it through Python and the Pype bridge.
The third is that I last looked at this part of the code over a year ago, and
wrote most of the code about 4 years ago.
> On Jun 24, 2025, at 17:34, John Mayfield <john.wilkinson...@gmail.com> wrote:
> First off for the SMARTS matcher you can turn off the "prepare" or use the
> lower level APIs and work on the input aromaticity.
>
> IChemObjectBuilder bldr = SilentChemObjectBuilder.getInstance();
Is there a reason for using SilentChemObjectBuilder instead of what I use,
which is:
cdk.DefaultChemObjectBuilder.getInstance()
? I see cinfony does what you suggest, as does Jonas:
> SmartsPattern pat = SmartsPattern.create("C=CC=N");
> pat.setPrepare(false); // turn off auto ring+arom perception
Jonas also mentioned the prepare method:
> //prevent the SMARTS pattern from perceiving aromaticity
> pattern.setPrepare(false);
I've never used this method.
With the default of true, does each SMARTS match re-perceive aromaticity each
time?
John:
> Cycles.markRingAtomsAndBonds(mol);
> Aromaticity.apply(Aromaticity.Model.Daylight, mol);
Hmmm. It looks like I don't understand who is supposed to be in charge of doing
perception, or what the processing steps to get a fully prepared structure.
What I've been doing is using SmilesParser(_default_builder).parseSmiles() and
assuming the molecule was in the right state.
I then use one of the fingerprinters, or do the SMARTS matches for a couple of
my own fingerprint types.
Am I always supposed to perceive rings and aromaticity if I use SmilesParser?
Is there any reason to not use the same aromcity perception steps in CDK
Depict, using Daylight aromaticity?
What about if I use MDLV2000Reader/MDLV3000Reader? Or IteratingSDFReader or
IteratingSMILESReader with hasNext()/next() to get the molecules? Do I need to
perceive those too?
Also, I'm looking at SubstructureFingerprinter.java and see:
SmartsPattern.prepare(atomContainer)
Do I need this too? Jonas wrote "SmartPattern.matchAll() is called in the web
app, which internally calls SmartsPattern.prepare", so I don't think I need it.
John:
> I'm not sure how you got that output:
Because I was confused when I wrote the code in the first place?
I can spend some time pulling the CDK-specific code out of chemfp to get a
stand-alone reproducible, but it's probably a better use of my time to just get
the processing steps done correctly.
Andrew
da...@dalkescientific.com
_______________________________________________
Cdk-user mailing list
Cdk-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/cdk-user
