SMARTS export

Java can write a portable SMARTS query from an MCS match or a whole molecular graph. SmartsWriter and SearchResult.Match.toSmarts are new in 8.1.0, while the published Java artefacts stay at 8.0.3 in this release, so compile this example against a Java source build:

import com.bioinception.smsd.core.MCS;
import com.bioinception.smsd.core.MolGraph;
import com.bioinception.smsd.core.SearchResult;
import com.bioinception.smsd.core.SmartsWriter;
import com.bioinception.smsd.core.Substructure;
import org.openscience.cdk.silent.SilentChemObjectBuilder;
import org.openscience.cdk.smiles.SmilesParser;

public class SmartsExportExample {
    public static void main(String[] args) throws Exception {
        SmilesParser parser = new SmilesParser(SilentChemObjectBuilder.getInstance());
        MolGraph query = new MolGraph(parser.parseSmiles("CCO"));
        MolGraph target = new MolGraph(parser.parseSmiles("CCCO"));
        MolGraph libraryMolecule = new MolGraph(parser.parseSmiles("CCOCC"));
        MCS.Options options = new MCS.Options();
        SearchResult common = MCS.find(query).withOptions(options).match(target);
        if (!common.found()) throw new IllegalStateException("No common substructure");
        String pattern = common.matches().get(0).toSmarts(query, options.chemistry);
        SearchResult hit = Substructure.smarts(pattern).match(libraryMolecule);
        if (!hit.found()) throw new IllegalStateException("Exported pattern did not match");
        System.out.println(pattern);
        System.out.println(SmartsWriter.toSmarts(query, options.chemistry));
    }
}

SmartsWriter.toSmarts(graph, chemistry) writes every atom and bond. A match writes only its selected atoms and common bonds, including an MCES match that leaves out a source bond. Ring and aromaticity flags retain their original graph context. The atom order follows the source indices; the text is deterministic for that numbering and is not canonical.

The writer supports element wildcards, formal charge including neutral atoms, isotope wildcards, strict or flexible aromaticity, bond-order modes and ring membership. As in molecule matching, isotope zero matches any isotope on the other side. The pattern is a subgraph query; it does not encode the induced MCIS objective, a search budget or an optimality proof.

Export throws IllegalArgumentException for stereo, tautomer-aware matching, complete-ring matching and strict ring-fusion matching. With flexible aromaticity and strict bond order, it also rejects a non-aromatic single or double bond between two aromatic atoms: the conditional resonance rule has no direct portable bond expression. Use supported options explicitly rather than silently dropping these constraints.

Portable output assumes molecular targets with consistent aromatic bond flags and bond orders. Contradictory source annotations supplied through a graph builder are rejected. Java SMARTS target preparation applies Daylight aromaticity. Strict export therefore also rejects selected aromatic flags that change under that model, including an uppercase Kekule benzene graph whose original flags are non-aromatic. This check uses the full source graph, preserving the context of selected atoms and bonds. Strict predicate parity with molecule matching requires library targets whose aromatic flags also agree with that preparation. At most 99 Java ring closures can be open at once; closure numbers are reused after they close.