I have downloaded CDK a short time ago and I am currently
on my way to explore this nice package.
Among other things I'm interested in searching of substructures in
molecules. At this, it would be fine, to make a generalised search
like e.g. for all molecules that contain Hal-C=C-[O,S]- H. Hal is any
halogen and [O,S] is either oxygen or sulfur (i.e. something in the spirit of
Is such a functionality supported somehow by the CDK
or an easy way to implement such a functionality?
[One possibility seems to rename the atom symbols
(e.g. rename the symbols of "F", "Cl", "Br", "I" in the molecules to "Hal" and
search for "Hal-C=C-OH"
with the UniversalIsomorphismTester), but that is no universal solution.]
In the current version (cdk-20030412) the method "getFingerprint(AtomContainer
ac, int size)"
of the class "Fingerprinter" does not generate a fingerprint of a given size but
always a fingerprint of the
default size of 1024. It seems, that in the source code for the method
"getFingerprint(AtomContainer ac, int size)"
position = new java.util.Random(path.hashCode()).nextInt(defaultSize);
has to be replaced by the line
position = new java.util.Random(path.hashCode()).nextInt(size);
All the best,