Following on from some patches last year to make things work easier in JCP this adds better RGroup support to the core library. Specifically like with reactions with now have the ability to flatten/unflatten an RGroupQuery structure so you can work with and pass it around like a normal molecule. This is handled by the new RGroupQueryManipulator. Ideally this would be located along with the other manipulators but the use of the Query APIs makes this tricky - perhaps something to target for CDK 3.0.

Two of the existing RGroup test files are not actually valid (tested with BIOVIA tools) and I have added a strict mode for that.

I have also added support for reading/writing from CXSMILES and doing the layout depiction logic - this is already live on SiMolecule / cdkdepict. This requires handling implied attachment points and disambiguating nesting. For example consider this structure:

*c1ncccc1 |$R1$,RG:_R1={C(=O)* |$;;R2$,RG:_R2={OC},{N}|},{C(=N)* |$;;R2$,RG:_R2={OCC},{C(F)(F)F}|}|

Note there are 2 conflicting definitions for R2 depending on which substituent is loaded. Without care we would accidentally load the structure like this:

Instead what we do is relabel one of the R2's to allow an unambiguous depiction:

The existing RGroupQuery does not really support nesting like this but the depiction/layout is at least correct.

One slight quirk is due to dependency/packaging I can't currently access the RGroupManipulator from the SMILES writer so you need to manually convert it before generating. I think that is acceptable for now:

String dir = "/Users/john/Workspace/GitHub/cdk/storage/ctab/src/test/resources/org/openscience/cdk/io/";
SmilesGenerator smigen = new SmilesGenerator(SmiFlavor.Default+SmiFlavor.UseAromaticSymbols);
for (String fname : new String[]{"rgfile.1.mol", "rgfile.2.mol",
                                 "rgfile.3.mol", "rgfile.4.mol",
                                 "rgfile.5.mol", "rgfile.6.mol",
                                 "rgfile.7.mol"}) {
    try (RGroupQueryReader rdr = new RGroupQueryReader(new FileReader(new File(dir, fname)))) {
        // rdr.setReaderMode(IChemObjectReader.Mode.STRICT);
        RGroupQuery query = new RGroupQuery(SilentChemObjectBuilder.getInstance());
        query = rdr.read(query);
        System.out.println(smigen.create(query) + "\t" + fname);
    }
}

For reference here is the CXSMILES for our 7 rgfile tests:

C1CCCP(*)C1 |$;;;;;R1$,RG:_R1={CN* |$;;_AP1$|},{C(Br)O* |$;;;_AP1$|},{C(C)S* |$;;;_AP1$|}|	rgfile.1.mol
*[PbH]1C(C(C(C1)*)*N2C=CC=C2)* |$R11;;;;;;R2;R1;;;;;;R2$,LO:7:8.3,RG:_R1={C(SO*)* |$;;;_AP2;_AP1$|},{N(C)(O*)* |$;;;_AP2;_AP1$|},_R2={C1(CCS1)* |$;;;;_AP1$|},{[SiH]1(CC1)* |$;;;_AP1$|},_R11={[Pt](C)* |$;;_AP1$|}|	rgfile.2.mol
[Po]1*[Te]C2C1C(S*O2)* |$;R1;;;;;;R1;;R$,LO:1:2.0,7:6.8,RG:_R1={C(C(=O)*)(=O)* |$;;;_AP2;;_AP1$|},{C(N(CC)*)(=O)* |$;;;;_AP2;;_AP1$|}|	rgfile.3.mol
C1C=CC=C1.* |$;;;;;R1$,RG:_R1={[NH+]#[C-]},{N(=O)=O}|	rgfile.4.mol
N1([P@@]([Si@H]([N+]([Ge@@H]([GeH](N([Pb](P([SiH]1*)*)=*)*)*)*)=*)*)*)* |$;;;;;;;;;;R1;R1;R3;R2;R1;R2;R3;R4;R2;R1$,r,RG:_R1={CF},{CO},{[Pb]=O},_R2={Br},{P},{S},_R3={[SiH4]},{[SnH2]},{[BiH3]},_R4={O},{Cl}|	rgfile.5.mol (SOFT ERROR - attachments not defined in RGfile)
*C1C(C(C(C1)*)*N2C=CC=C2)* |$R11;;;;;;R2;R1;;;;;;R2$,LO:7:8.3,RG:_R1=_R2=_R11={[Pt](C)* |$;;_AP1$|}|	rgfile.6.mol
C1=S([Dy]*)*[Bi]1*(Cl)N |$;;;R32;R32;;R32$,LO:4:1.5,6:5.7.8,RG:_R32={[P](Br)(O)(*)* |$;;;_AP1;_AP2$|},{[GeH](F)(*)* |$;;_AP1;_AP2$|}|	rgfile.7.mol (ERROR - R group has 1,2, or 3 connections)