Skip to content

CPP.md

Latest commit

 

History

History
117 lines (87 loc) · 3.7 KB

CPP.md

File metadata and controls

117 lines (87 loc) · 3.7 KB

SMSD Pro C++ Guide

SMSD Pro’s C++ layer is header-only and provides the native implementations for MolGraph construction, substructure search, MCS, fingerprints, SMARTS matching, molfile I/O, stereo/CIP assignment, and layout utilities.

Include

#include "smsd/smsd.hpp"

Core Use

#include "smsd/smsd.hpp"

auto q = smsd::parseSMILES("c1ccccc1");
auto t = smsd::parseSMILES("c1ccc(O)cc1");

smsd::ChemOptions chem;
smsd::MCSOptions mcsOpts;
mcsOpts.timeout_ms = 10000;

bool hit = smsd::isSubstructure(q, t, chem, 10000);
auto mcs = smsd::findMCS(q, t, chem, mcsOpts);

Fingerprints

The canonical C++ fingerprint API lives in smsd::batch::detail and is verified byte-identical with the Java and Python engines in 7.1.1 (same bits for the same molecules at every radius).

#include "smsd/batch.hpp"

auto q = smsd::parseSMILES("c1ccc(O)cc1");

// Circular (ECFP / FCFP) — binary and count-based
auto ecfp  = smsd::batch::detail::computeCircularFingerprintECFP(q, 2, 2048);
auto fcfp  = smsd::batch::detail::computeCircularFingerprintFCFP(q, 2, 2048);
auto ecfpc = smsd::batch::detail::computeCircularFingerprintECFPCount(q, 2, 2048);
auto fcfpc = smsd::batch::detail::computeCircularFingerprintFCFPCount(q, 2, 2048);

// Path / topological torsion / MACCS
auto pathFP  = smsd::batch::detail::computePathFingerprint(q, 7, 2048);
auto torsion = smsd::batch::detail::computeTopologicalTorsion(q, 2048);
auto maccs   = smsd::batch::detail::computeMACCSKeys(q);

// Similarity
double tani = smsd::batch::detail::tanimoto(ecfp, fcfp);
double dice = smsd::batch::detail::dice(ecfp, fcfp);

Note. The pre-7.1.1 fp/mol/circular.hpp, fp/mol/path.hpp, fp/mol/pharmacophore.hpp, and fp/mol/torsion.hpp headers were unmaintained shims that drifted from the real Python/Java bit pattern. They have been removed. Use the smsd::batch::detail::* entry points documented above — these are the exact functions the Python binding and the Java FingerprintEngine are tested against.

Public MCS / Substructure Entry Points

The high-level entry points are smsd::findMCS(), smsd::findSubstructure() and smsd::isSubstructure() declared in smsd/mcs.hpp and smsd/substructure.hpp. The internal solver headers (smsd/clique_solver.hpp, the partition-refinement backtracker, edge-growth refinement) are private implementation details whose signatures may change between minor releases — do not depend on them in out-of-tree code.

#include "smsd/mcs.hpp"
#include "smsd/substructure.hpp"

auto mapping     = smsd::findMCS(g1, g2, smsd::ChemOptions{}, smsd::MCSOptions{});
auto sub_mapping = smsd::findSubstructure(query, target, smsd::ChemOptions{});
bool contained   = smsd::isSubstructure(query, target, smsd::ChemOptions{});

Scaffold Library (7.1.0)

#include "smsd/scaffold_library.hpp"
auto scaffold = smsd::scaffold::murckoScaffold(mol);

Hungarian Algorithm (7.1.0)

Optimal assignment solver for atom matching cost matrices.

#include "smsd/hungarian.hpp"
auto assignment = smsd::hungarian::solve(costMatrix);

SMARTS and CIP

auto query = smsd::parseSMARTS("[#6]~[#7]");
auto rs = smsd::cip::assignRSAll(q);
auto ez = smsd::cip::assignEZAll(q);

Native MOL/SDF I/O

auto mol = smsd::readMolBlock(molBlockText);
std::string v2000 = smsd::writeMolBlock(mol);
std::string v3000 = smsd::writeMolBlockV3000(mol);
std::string sdf = smsd::writeSDFRecord(mol);

The native I/O path in 7.1.0 covers practical V2000/V3000 graph round-trip, metadata, SDF properties, atom maps/classes, and patent-style R# handling. The most exotic MDL query chemistry features are still intentionally documented as out of scope until they are implemented natively.