Files
rdkit/Code/GraphMol/TautomerQuery/TautomerQuery.cpp
Yakov Pechersky c6cabf4153 Speed-up tautomer canonicalization, no API changes (#9134)
* Speed up tautomer canonicalization by deferring on SSSR calc

* Lazy kekulization for tautomer enumeration

Defer kekulization of tautomers until they are actually needed for
transform matching. This avoids creating kekulized copies for:
1. The initial tautomer (until first iteration)
2. New tautomers that may never be processed (if enumeration ends early)

The Tautomer class now supports lazy initialization of the kekulized
form via getKekulized() method.

Performance improvement: ~7% additional speedup (total ~22-24% from baseline)

* Use count-only substructure matching in tautomer scoring

* Add SubstructMatchCount regression test

* MolStandardize: reduce enumerate overhead

* MolStandardize: avoid per-tautomer ring recomputation

* Atom: cache PeriodicTable pointer in valence calcs

* Atom: reuse PeriodicTable in getEffectiveAtomicNum

* PeriodicTable: add atomic fast path for getTable

* GraphMol: reduce ROMol copy reallocations

* MolStandardize: use quickCopy for per-match product copies

Use RWMol(*kmol, true) in tautomer enumeration to avoid copying properties/bookmarks/conformers for each candidate. This reduces deep-copy overhead without changing chemistry.

* MolStandardize: pre-filter scoring patterns by element/connectivity

For tautomer scoring, pre-compute which SubstructTerms are relevant for
a given input molecule. Since tautomerization only moves H atoms and
changes bond orders (never creates/destroys heavy-atom bonds), patterns
requiring missing elements or connectivity can be skipped for all
tautomers of that molecule.

Two-stage filtering:
1. Element check: skip patterns requiring atoms not in the molecule
2. Connectivity check: skip patterns whose bond-order-agnostic structure
   doesn't match the input molecule's connectivity

This reduces the number of VF2 substructure calls per tautomer from 12
to typically 3-5, depending on the molecule's composition.

* MolStandardize: preserve molecule properties for canonical tautomer

Copy molecule properties from the original input to the canonical tautomer
result. Since quickCopy during enumeration skips d_props to avoid overhead,
extended SMILES data like link nodes (LN) was lost. This restores them
on the final result.

* TautomerQuery: preserve molecule properties (e.g. link nodes) in tautomers

TautomerQuery::fromMol() uses TautomerEnumerator::enumerate() which uses
quickCopy for performance. This doesn't copy molecule properties like
_molLinkNodes. Without this fix, XQMol output would lose link node
extensions in the SMILES.

Copy properties from the original query molecule to all enumerated
tautomers before constructing the TautomerQuery. This preserves extended
SMILES data without impacting enumeration performance.

* MolStandardize: use parallel iteration and cache bond lookups

Replace O(n) getAtomWithIdx/getBondWithIdx calls with parallel iteration
over atom/bond ranges in canonicalizeInPlace and enumerate. Cache bond
lookups in setTautomerStereoAndIsoHs to avoid repeated O(n) searches.

* perf: add specialized matchers for simple tautomer scoring patterns

Replace VF2 graph matching with O(n) loops for 6 simple patterns:
- countDoubleOrAromaticBonds: C=O, N=O, P=O patterns
- countMethyls: [CX4H3] methyl groups
- countCarbonDoubleHetero: [C]=[/home/dcvuser/rdkit;Code/GraphMol/MolStandardize/Tautomer.h] aliphatic C=hetero
- countAromaticCarbonExocyclicN: [c]=aromatic C=exocyclic N
Complex patterns (benzoquinone, oxim, guanidine, aci-nitro) still use VF2.
Combined with the pre-filtering optimization, this achieves ~3.7x speedup
(~2500ms vs ~9300ms original) for tautomer canonicalization.

* Fix tautomer canonicalize dropping conformers from quickCopy

quickCopy (RWMol(*mol, true)) skips conformers, so tautomer
enumeration products lose 2D/3D coordinates. This causes InChI
generation to omit the /b (double bond E/Z stereo) layer, since
E/Z is derived from atomic coordinates.

Fix: copy conformers from the original molecule onto the canonical
tautomer after pickCanonical in TautomerEnumerator::canonicalize().

Tests: SMILES-based E/Z check in testTautomer.cpp, molblock-based
conformer preservation check in catch_tests.cpp.

* add test on canonicalize losing stereo

* add regression test for exocyclic C=C tautomer canonicalization

The getTautomerStateKey() pre-filter (commit 2595ef748) can falsely
deduplicate distinct tautomers when their atom-index-ordered state
patterns happen to match, leading canonicalize() to pick the wrong
canonical form for molecules with STEREOTRANS-pinned exocyclic C=C
bonds after RemoveHs.

Test verifies that O=C(CC1=CC2=CC=COC2)NC1=O canonicalizes to the
exocyclic form O=C1CC(=CC2=CC=COC2)C(=O)N1, not the endocyclic form
O=C1C=C(C=C2CC=COC2)C(=O)N1.

Currently expected to FAIL until the state key dedup bug is fixed.

* MolStandardize: expand tautomer connectivity SMARTS

* MolStandardize: scope tautomer pattern enum

* MolStandardize: trim tautomer pattern enum

* MolStandardize: use symmetric ring scoring
2026-03-31 06:42:40 +02:00

346 lines
11 KiB
C++

//
// Created by Gareth Jones on 5/7/2020.
//
// Copyright 2020 Schrodinger, Inc
// @@ All Rights Reserved @@
// This file is part of the RDKit.
// The contents are covered by the terms of the BSD license
// which is included in the file license.txt, found at the root
// of the RDKit source tree.
#include "TautomerQuery.h"
#include <functional>
#include <set>
#include <utility>
#include <span>
#include <GraphMol/RDKitBase.h>
#include <GraphMol/MolStandardize/Tautomer.h>
#include <GraphMol/Bond.h>
#include <GraphMol/QueryOps.h>
#include <GraphMol/QueryAtom.h>
#include <GraphMol/QueryBond.h>
#include <GraphMol/Substruct/SubstructUtils.h>
#include <GraphMol/Fingerprints/Fingerprints.h>
// #define VERBOSE
#ifdef RDK_USE_BOOST_SERIALIZATION
#include <RDGeneral/BoostStartInclude.h>
#include <boost/archive/text_oarchive.hpp>
#include <boost/archive/text_iarchive.hpp>
#include <boost/archive/archive_exception.hpp>
#include <RDGeneral/BoostEndInclude.h>
#endif
#ifdef VERBOSE
#include <GraphMol/SmilesParse/SmilesWrite.h>
#endif
namespace {
using namespace RDKit;
// Adapted from Code/GraphMol/Substruct/SubstructUtils.cpp#removeDuplicates
void removeTautomerDuplicates(std::vector<MatchVectType> &matches,
std::vector<ROMOL_SPTR> *matchingTautomers,
int nAtoms) {
//
// This works by tracking the indices of the atoms in each match vector.
// This can lead to unexpected behavior when looking at rings and queries
// that don't specify bond orders. For example querying this molecule:
// C1CCC=1
// with the pattern constructed from SMARTS C~C~C~C will return a
// single match, despite the fact that there are 4 different paths
// when valence is considered. The defense of this behavior is
// that the 4 paths are equivalent in the semantics of the query.
// Also, OELib returns the same results
//
std::set<boost::dynamic_bitset<>> seen;
std::vector<MatchVectType> res;
res.reserve(matches.size());
for (auto &&match : matches) {
boost::dynamic_bitset<> val(nAtoms);
for (const auto &ci : match) {
val.set(ci.second);
}
auto pos = seen.lower_bound(val);
if (pos == seen.end() || *pos != val) {
res.push_back(std::move(match));
seen.insert(pos, std::move(val));
} else if (matchingTautomers) {
auto position = res.size();
matchingTautomers->erase(matchingTautomers->begin() + position);
}
}
res.shrink_to_fit();
matches = std::move(res);
}
} // namespace
namespace RDKit {
bool TautomerQueryCanSerialize() {
#ifdef RDK_USE_BOOST_SERIALIZATION
return true;
#else
return false;
#endif
}
class TautomerQueryMatcher {
private:
const TautomerQuery &d_tautomerQuery;
const SubstructMatchParameters &d_params;
std::vector<ROMOL_SPTR> *d_matchingTautomers;
public:
TautomerQueryMatcher(const TautomerQuery &tautomerQuery,
const SubstructMatchParameters &params,
std::vector<ROMOL_SPTR> *matchingTautomers)
: d_tautomerQuery(tautomerQuery),
d_params(params),
d_matchingTautomers(matchingTautomers) {}
bool match(const ROMol &mol, const std::span<const unsigned int> &match) {
#ifdef VERBOSE
std::cout << "Checking template match" << std::endl;
#endif
for (auto tautomer : d_tautomerQuery.getTautomers()) {
#ifdef VERBOSE
std::cout << "Checking Tautomer " << MolToSmiles(*tautomer) << std::endl;
#endif
if (d_tautomerQuery.matchTautomer(mol, *tautomer, match, d_params)) {
auto matchingTautomer = d_params.extraFinalCheck
? d_params.extraFinalCheck(mol, match)
: true;
if (matchingTautomer) {
#ifdef VERBOSE
std::cout << "Got Match " << std::endl;
#endif
if (d_matchingTautomers) {
d_matchingTautomers->push_back(tautomer);
}
}
return matchingTautomer;
}
}
return false;
}
};
TautomerQuery::TautomerQuery(std::vector<ROMOL_SPTR> tautomers,
const ROMol *const templateMolecule,
std::vector<size_t> modifiedAtoms,
std::vector<size_t> modifiedBonds)
: d_tautomers(std::move(tautomers)),
d_templateMolecule(templateMolecule),
d_modifiedAtoms(std::move(modifiedAtoms)),
d_modifiedBonds(std::move(modifiedBonds)) {}
TautomerQuery *TautomerQuery::fromMol(
const ROMol &query, const std::string &tautomerTransformFile) {
auto tautomerParams = std::unique_ptr<MolStandardize::TautomerCatalogParams>(
new MolStandardize::TautomerCatalogParams(tautomerTransformFile));
MolStandardize::TautomerEnumerator tautomerEnumerator(
new MolStandardize::TautomerCatalog(tautomerParams.get()));
const auto res = tautomerEnumerator.enumerate(query);
std::vector<size_t> modifiedAtoms;
modifiedAtoms.reserve(res.modifiedAtoms().count());
for (size_t i = 0; i < query.getNumAtoms(); i++) {
if (res.modifiedAtoms().test(i)) {
modifiedAtoms.push_back(i);
}
}
std::vector<size_t> modifiedBonds;
modifiedBonds.reserve(res.modifiedBonds().count());
for (size_t i = 0; i < query.getNumBonds(); i++) {
if (res.modifiedBonds().test(i)) {
modifiedBonds.push_back(i);
}
}
auto templateMolecule = new RWMol(query);
for (auto idx : modifiedAtoms) {
const auto atom = query.getAtomWithIdx(idx);
const auto queryAtom = new QueryAtom(atom->getAtomicNum());
// Forward original queries
if (atom->hasQuery()) {
auto originalAtomQuery = static_cast<const QueryAtom *>(atom)->getQuery();
queryAtom->setQuery(originalAtomQuery->copy());
}
const bool updateLabel = false;
const bool preserveProps = true;
templateMolecule->replaceAtom(idx, queryAtom, updateLabel, preserveProps);
delete queryAtom;
}
for (auto idx : modifiedBonds) {
auto bondQuery = makeSingleOrDoubleOrAromaticBondQuery();
auto queryBond = new QueryBond();
queryBond->setQuery(bondQuery);
templateMolecule->replaceBond(idx, queryBond, true);
delete queryBond;
}
// Copy molecule properties from the original query to all tautomers.
// Tautomer enumeration uses quickCopy for performance, which doesn't copy
// molecule properties. We need to preserve them here for proper CXSMILES
// output (e.g. link nodes).
auto tautomers = res.tautomers();
for (auto &tautomer : tautomers) {
tautomer->updateProps(query);
}
return new TautomerQuery(std::move(tautomers),
static_cast<ROMol *>(templateMolecule),
modifiedAtoms, modifiedBonds);
}
bool TautomerQuery::matchTautomer(
const ROMol &mol, const ROMol &tautomer,
const std::span<const unsigned int> &match,
const SubstructMatchParameters &params) const {
for (auto idx : d_modifiedAtoms) {
const auto queryAtom = tautomer.getAtomWithIdx(idx);
const auto targetAtom = mol.getAtomWithIdx(match[idx]);
#ifdef VERBOSE
std::cout << "Query atom " << queryAtom->getSymbol() << " target atom "
<< targetAtom->getSymbol() << std::endl;
#endif
if (!atomCompat(queryAtom, targetAtom, params)) {
#ifdef VERBOSE
std::cout << "Atom incompatibility" << std::endl;
#endif
return false;
}
}
for (auto idx : d_modifiedBonds) {
const auto queryBond = tautomer.getBondWithIdx(idx);
const auto beginIdx = queryBond->getBeginAtomIdx();
const auto endIdx = queryBond->getEndAtomIdx();
const auto targetBeginIdx = match[beginIdx];
const auto targetEndIdx = match[endIdx];
const auto targetBond =
mol.getBondBetweenAtoms(targetBeginIdx, targetEndIdx);
#ifdef VERBOSE
std::cout << "Query bond " << queryBond->getBondType() << " target bond "
<< targetBond->getBondType() << std::endl;
#endif
if (!bondCompat(queryBond, targetBond, params)) {
#ifdef VERBOSE
std::cout << "Bond incompatibility" << std::endl;
#endif
return false;
}
}
#ifdef VERBOSE
std::cout << "Tautomer match" << std::endl;
#endif
return true;
}
std::vector<MatchVectType> TautomerQuery::substructOf(
const ROMol &mol, const SubstructMatchParameters &params,
std::vector<ROMOL_SPTR> *matchingTautomers) const {
if (matchingTautomers) {
matchingTautomers->clear();
}
#ifdef VERBOSE
std::cout << "Tautomer search with query " << MolToSmiles(*d_templateMolecule)
<< " on " << MolToSmiles(mol) << " max matches "
<< params.maxMatches << std::endl;
#endif
SubstructMatchParameters templateParams(params);
// need to check all mappings of template to target
templateParams.uniquify = false;
TautomerQueryMatcher tautomerQueryMatcher(*this, params, matchingTautomers);
// use this functor as a final check to see if any tautomer matches the target
auto checker = [&tautomerQueryMatcher](
const ROMol &mol,
const std::span<const unsigned int> &match) mutable {
return tautomerQueryMatcher.match(mol, match);
};
templateParams.extraFinalCheck = checker;
auto matches =
RDKit::SubstructMatch(mol, *d_templateMolecule, templateParams);
#ifdef VERBOSE
std::cout << "Found " << matches.size() << " matches " << std::endl;
#endif
if (params.uniquify && matches.size() > 1) {
removeTautomerDuplicates(matches, matchingTautomers, mol.getNumAtoms());
#ifdef VERBOSE
std::cout << "After removing duplicates " << matches.size() << " matches "
<< std::endl;
#endif
}
return matches;
}
bool TautomerQuery::isSubstructOf(const ROMol &mol,
const SubstructMatchParameters &params) {
SubstructMatchParameters params2(params);
params2.maxMatches = 1;
auto matches = substructOf(mol, params2);
return matches.size() > 0;
}
ExplicitBitVect *TautomerQuery::patternFingerprintTemplate(
unsigned int fpSize) const {
return PatternFingerprintMol(*d_templateMolecule, fpSize, nullptr, nullptr,
true);
}
ExplicitBitVect *TautomerQuery::patternFingerprintTarget(const ROMol &target,
unsigned int fpSize) {
return PatternFingerprintMol(target, fpSize, nullptr, nullptr, true);
}
std::vector<MatchVectType> SubstructMatch(
const ROMol &mol, const TautomerQuery &query,
const SubstructMatchParameters &params) {
return query.substructOf(mol, params);
}
void TautomerQuery::toStream(std::ostream &ss) const {
#ifndef RDK_USE_BOOST_SERIALIZATION
PRECONDITION(0, "Boost SERIALIZATION is not enabled")
#else
boost::archive::text_oarchive ar(ss);
ar << *this;
#endif
}
std::string TautomerQuery::serialize() const {
std::stringstream ss;
toStream(ss);
return ss.str();
}
void TautomerQuery::initFromStream(std::istream &ss) {
#ifndef RDK_USE_BOOST_SERIALIZATION
PRECONDITION(0, "Boost SERIALIZATION is not enabled")
#else
boost::archive::text_iarchive ar(ss);
ar >> *this;
#endif
}
void TautomerQuery::initFromString(const std::string &text) {
std::stringstream ss(text);
initFromStream(ss);
}
} // namespace RDKit