Files
rdkit/Code/GraphMol/Atom.cpp
Yakov Pechersky 0d886b9d08 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
2026-04-30 14:17:18 +02:00

1054 lines
31 KiB
C++

//
// Copyright (C) 2001-2024 Greg Landrum and other RDKit contributors
//
// @@ 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 <cmath>
#include "ROMol.h"
#include "Atom.h"
#include "PeriodicTable.h"
#include "SanitException.h"
#include "QueryOps.h"
#include "MonomerInfo.h"
#include <RDGeneral/Invariant.h>
#include <RDGeneral/RDLog.h>
#include <RDGeneral/types.h>
#include <RDGeneral/Dict.h>
namespace RDKit {
bool isAromaticAtom(const Atom &atom) {
if (atom.getIsAromatic()) {
return true;
}
if (atom.hasOwningMol()) {
for (const auto &bond : atom.getOwningMol().atomBonds(&atom)) {
if (bond->getIsAromatic() ||
bond->getBondType() == Bond::BondType::AROMATIC) {
return true;
}
}
}
return false;
}
unsigned int getEffectiveAtomicNum(const Atom &atom, bool checkValue) {
const auto *periodicTable = PeriodicTable::getTable();
auto effectiveAtomicNum = atom.getAtomicNum() - atom.getFormalCharge();
if (checkValue &&
(effectiveAtomicNum < 0 ||
effectiveAtomicNum >
static_cast<int>(periodicTable->getMaxAtomicNumber()))) {
throw AtomValenceException("Effective atomic number out of range",
atom.getIdx());
}
effectiveAtomicNum = std::clamp(
effectiveAtomicNum, 0,
static_cast<int>(periodicTable->getMaxAtomicNumber()));
return static_cast<unsigned int>(effectiveAtomicNum);
}
// Determine whether or not an element is to the left of carbon.
bool isEarlyAtom(int atomicNum) {
static const bool table[119] = {
false, // #0 *
false, // #1 H
false, // #2 He
true, // #3 Li
true, // #4 Be
true, // #5 B
false, // #6 C
false, // #7 N
false, // #8 O
false, // #9 F
false, // #10 Ne
true, // #11 Na
true, // #12 Mg
true, // #13 Al
false, // #14 Si
false, // #15 P
false, // #16 S
false, // #17 Cl
false, // #18 Ar
true, // #19 K
true, // #20 Ca
true, // #21 Sc
true, // #22 Ti
false, // #23 V
false, // #24 Cr
false, // #25 Mn
false, // #26 Fe
false, // #27 Co
false, // #28 Ni
false, // #29 Cu
true, // #30 Zn
true, // #31 Ga
true, // #32 Ge see github #2606
false, // #33 As
false, // #34 Se
false, // #35 Br
false, // #36 Kr
true, // #37 Rb
true, // #38 Sr
true, // #39 Y
true, // #40 Zr
true, // #41 Nb
false, // #42 Mo
false, // #43 Tc
false, // #44 Ru
false, // #45 Rh
false, // #46 Pd
false, // #47 Ag
true, // #48 Cd
true, // #49 In
true, // #50 Sn see github #2606
true, // #51 Sb see github #2775
false, // #52 Te
false, // #53 I
false, // #54 Xe
true, // #55 Cs
true, // #56 Ba
true, // #57 La
true, // #58 Ce
true, // #59 Pr
true, // #60 Nd
true, // #61 Pm
false, // #62 Sm
false, // #63 Eu
false, // #64 Gd
false, // #65 Tb
false, // #66 Dy
false, // #67 Ho
false, // #68 Er
false, // #69 Tm
false, // #70 Yb
false, // #71 Lu
true, // #72 Hf
true, // #73 Ta
false, // #74 W
false, // #75 Re
false, // #76 Os
false, // #77 Ir
false, // #78 Pt
false, // #79 Au
true, // #80 Hg
true, // #81 Tl
true, // #82 Pb see github #2606
true, // #83 Bi see github #2775
false, // #84 Po
false, // #85 At
false, // #86 Rn
true, // #87 Fr
true, // #88 Ra
true, // #89 Ac
true, // #90 Th
true, // #91 Pa
true, // #92 U
true, // #93 Np
false, // #94 Pu
false, // #95 Am
false, // #96 Cm
false, // #97 Bk
false, // #98 Cf
false, // #99 Es
false, // #100 Fm
false, // #101 Md
false, // #102 No
false, // #103 Lr
true, // #104 Rf
true, // #105 Db
true, // #106 Sg
true, // #107 Bh
true, // #108 Hs
true, // #109 Mt
true, // #110 Ds
true, // #111 Rg
true, // #112 Cn
true, // #113 Nh
true, // #114 Fl
true, // #115 Mc
true, // #116 Lv
true, // #117 Ts
true, // #118 Og
};
return ((unsigned int)atomicNum < 119) && table[atomicNum];
}
Atom::Atom() : RDProps() {
d_atomicNum = 0;
initAtom();
}
Atom::Atom(unsigned int num) : RDProps() {
d_atomicNum = num;
initAtom();
};
Atom::Atom(const std::string &what) : RDProps() {
d_atomicNum = PeriodicTable::getTable()->getAtomicNumber(what);
initAtom();
};
void Atom::initFromOther(const Atom &other) {
RDProps::operator=(other);
// NOTE: we do *not* copy ownership!
dp_mol = nullptr;
d_atomicNum = other.d_atomicNum;
d_index = 0;
d_formalCharge = other.d_formalCharge;
df_noImplicit = other.df_noImplicit;
df_isAromatic = other.df_isAromatic;
d_numExplicitHs = other.d_numExplicitHs;
d_numRadicalElectrons = other.d_numRadicalElectrons;
d_isotope = other.d_isotope;
// d_pos = other.d_pos;
d_chiralTag = other.d_chiralTag;
d_hybrid = other.d_hybrid;
d_implicitValence = other.d_implicitValence;
d_explicitValence = other.d_explicitValence;
if (other.dp_monomerInfo) {
dp_monomerInfo = other.dp_monomerInfo->copy();
} else {
dp_monomerInfo = nullptr;
}
d_flags = other.d_flags;
}
Atom::Atom(const Atom &other) : RDProps() { initFromOther(other); }
Atom &Atom::operator=(const Atom &other) {
if (this == &other) {
return *this;
}
initFromOther(other);
return *this;
}
void Atom::initAtom() {
df_isAromatic = false;
df_noImplicit = false;
d_numExplicitHs = 0;
d_numRadicalElectrons = 0;
d_formalCharge = 0;
d_index = 0;
d_isotope = 0;
d_chiralTag = CHI_UNSPECIFIED;
d_hybrid = UNSPECIFIED;
dp_mol = nullptr;
dp_monomerInfo = nullptr;
d_implicitValence = -1;
d_explicitValence = -1;
}
Atom::~Atom() { delete dp_monomerInfo; }
Atom *Atom::copy() const {
auto *res = new Atom(*this);
return res;
}
void Atom::setOwningMol(ROMol *other) {
// NOTE: this operation does not update the topology of the owning
// molecule (i.e. this atom is not added to the graph). Only
// molecules can add atoms to themselves.
dp_mol = other;
}
std::string Atom::getSymbol() const {
std::string res;
// handle dummies differently:
if (d_atomicNum != 0 ||
!getPropIfPresent<std::string>(common_properties::dummyLabel, res)) {
res = PeriodicTable::getTable()->getElementSymbol(d_atomicNum);
}
return res;
}
unsigned int Atom::getDegree() const {
return dp_mol ? getOwningMol().getAtomDegree(this) : 0;
}
unsigned int Atom::getTotalDegree() const {
unsigned int res = this->getTotalNumHs(false) + this->getDegree();
return res;
}
//
// If includeNeighbors is set, we'll loop over our neighbors
// and include any of them that are Hs in the count here
//
unsigned int Atom::getTotalNumHs(bool includeNeighbors) const {
int res = getNumExplicitHs() + getNumImplicitHs();
if (includeNeighbors && dp_mol) {
auto nbrs = dp_mol->atomNeighbors(this);
res += std::count_if(nbrs.begin(), nbrs.end(), [](const auto nbr) {
return (nbr->getAtomicNum() == 1);
});
}
return res;
}
unsigned int Atom::getNumImplicitHs() const {
if (df_noImplicit) {
return 0;
}
PRECONDITION(d_implicitValence > -1,
"getNumImplicitHs() called without preceding call to "
"calcImplicitValence()");
return getValence(ValenceType::IMPLICIT);
}
int Atom::getExplicitValence() const {
return getValence(ValenceType::EXPLICIT);
}
int Atom::getImplicitValence() const {
return getValence(ValenceType::IMPLICIT);
}
unsigned int Atom::getValence(ValenceType which) const {
if (!dp_mol) {
return 0;
}
PRECONDITION(
(which == ValenceType::IMPLICIT || d_explicitValence > -1),
"getValence(ValenceType::EXPLICIT) called without call to calcExplicitValence()");
PRECONDITION(
(which == ValenceType::EXPLICIT || df_noImplicit ||
d_implicitValence > -1),
"getValence(ValenceType::IMPLICIT) called without call to calcImplicitValence()");
if (which == ValenceType::EXPLICIT) {
return d_explicitValence;
} else {
return df_noImplicit ? 0 : d_implicitValence;
}
}
unsigned int Atom::getTotalValence() const {
return getValence(ValenceType::EXPLICIT) + getValence(ValenceType::IMPLICIT);
}
namespace {
bool canBeHypervalent(const Atom &atom, unsigned int effectiveAtomicNum) {
return (effectiveAtomicNum > 16 &&
(atom.getAtomicNum() == 15 || atom.getAtomicNum() == 16)) ||
(effectiveAtomicNum > 34 &&
(atom.getAtomicNum() == 33 || atom.getAtomicNum() == 34));
}
int calculateExplicitValence(const Atom &atom, bool strict, bool checkIt) {
const auto *periodicTable = PeriodicTable::getTable();
// FIX: contributions of bonds to valence are being done at best
// approximately
double accum = 0;
for (const auto bnd : atom.getOwningMol().atomBonds(&atom)) {
accum += bnd->getValenceContrib(&atom);
}
accum += atom.getNumExplicitHs();
const auto &ovalens = periodicTable->getValenceList(atom.getAtomicNum());
// if we start with an atom that doesn't have specified valences, we stick
// with that. otherwise we will use the effective valence
unsigned int effectiveAtomicNum = atom.getAtomicNum();
if (ovalens.size() > 1 || ovalens[0] != -1) {
effectiveAtomicNum = getEffectiveAtomicNum(atom, checkIt);
}
unsigned int dv = periodicTable->getDefaultValence(effectiveAtomicNum);
const auto &valens = periodicTable->getValenceList(effectiveAtomicNum);
if (accum > dv && isAromaticAtom(atom)) {
// this needs some explanation : if the atom is aromatic and
// accum > dv we assume that no hydrogen can be added
// to this atom. We set x = (v + chr) such that x is the
// closest possible integer to "accum" but less than
// "accum".
//
// "v" here is one of the allowed valences. For example:
// sulfur here : O=c1ccs(=O)cc1
// nitrogen here : c1cccn1C
int pval = dv;
for (auto val : valens) {
if (val == -1) {
break;
}
if (val > accum) {
break;
} else {
pval = val;
}
}
// if we're within 1.5 of the allowed valence, go ahead and take it.
// this reflects things like the N in c1cccn1C, which starts with
// accum of 4, but which can be kekulized to C1=CC=CN1C, where
// the valence is 3 or the bridging N in c1ccn2cncc2c1, which starts
// with a valence of 4.5, but can be happily kekulized down to a valence
// of 3
if (accum - pval <= 1.5) {
accum = pval;
}
}
// despite promising to not to blame it on him - this a trick Greg
// came up with: if we have a bond order sum of x.5 (i.e. 1.5, 2.5
// etc) we would like it to round to the higher integer value --
// 2.5 to 3 instead of 2 -- so we will add 0.1 to accum.
// this plays a role in the number of hydrogen that are implicitly
// added. This will only happen when the accum is a non-integer
// value and less than the default valence (otherwise the above if
// statement should have caught it). An example of where this can
// happen is the following smiles:
// C1ccccC1
// Daylight accepts this smiles and we should be able to Kekulize
// correctly.
accum += 0.1;
auto res = static_cast<int>(std::round(accum));
if (strict || checkIt) {
int maxValence = valens.back();
int offset = 0;
// we have to include a special case here for negatively charged P, S, As,
// and Se, which all support "hypervalent" forms, but which can be
// isoelectronic to Cl/Ar or Br/Kr, which do not support hypervalent forms.
if (canBeHypervalent(atom, effectiveAtomicNum)) {
maxValence = ovalens.back();
offset -= atom.getFormalCharge();
}
// we have historically accepted two-coordinate [H-] as a valid atom. This
// is highly questionable, but changing it requires some thought. For now we
// will just keep accepting it
if (atom.getAtomicNum() == 1 && atom.getFormalCharge() == -1) {
maxValence = 2;
}
// maxValence == -1 signifies that we'll take anything at the high end
if (maxValence >= 0 && ovalens.back() >= 0 && (res + offset) > maxValence) {
// the explicit valence is greater than any
// allowed valence for the atoms
if (strict) {
// raise an error
std::ostringstream errout;
errout << "Explicit valence for atom # " << atom.getIdx() << " "
<< periodicTable->getElementSymbol(atom.getAtomicNum())
<< ", " << res << ", is greater than permitted";
std::string msg = errout.str();
BOOST_LOG(rdErrorLog) << msg << std::endl;
throw AtomValenceException(msg, atom.getIdx());
} else {
return -1;
}
}
}
return res;
}
} // namespace
// NOTE: this uses the explicitValence, so it will call
// calculateExplicitValence if it is not set on the given atom
int calculateImplicitValence(const Atom &atom, bool strict, bool checkIt) {
if (atom.df_noImplicit) {
return 0;
}
const auto *periodicTable = PeriodicTable::getTable();
auto explicitValence = atom.d_explicitValence;
if (explicitValence == -1) {
explicitValence = calculateExplicitValence(atom, strict, checkIt);
}
// special cases
auto atomicNum = atom.d_atomicNum;
if (atomicNum == 0) {
return 0;
}
for (const auto bnd : atom.getOwningMol().atomBonds(&atom)) {
if (QueryOps::hasComplexBondTypeQuery(*bnd)) {
return 0;
}
}
auto formalCharge = atom.d_formalCharge;
auto numRadicalElectrons = atom.d_numRadicalElectrons;
if (explicitValence == 0 && numRadicalElectrons == 0 && atomicNum == 1) {
if (formalCharge == 1 || formalCharge == -1) {
return 0;
} else if (formalCharge == 0) {
return 1;
} else {
if (strict) {
std::ostringstream errout;
errout << "Unreasonable formal charge on atom # " << atom.getIdx()
<< ".";
std::string msg = errout.str();
BOOST_LOG(rdErrorLog) << msg << std::endl;
throw AtomValenceException(msg, atom.getIdx());
} else if (checkIt) {
return -1;
} else {
return 0;
}
}
}
int explicitPlusRadV = atom.d_explicitValence + atom.d_numRadicalElectrons;
const auto &ovalens = periodicTable->getValenceList(atom.d_atomicNum);
// if we start with an atom that doesn't have specified valences, we stick
// with that. otherwise we will use the effective valence for the rest of
// this.
unsigned int effectiveAtomicNum = atom.d_atomicNum;
if (ovalens.size() > 1 || ovalens[0] != -1) {
effectiveAtomicNum = getEffectiveAtomicNum(atom, checkIt);
}
if (effectiveAtomicNum == 0) {
return 0;
}
// this is basically the difference between the allowed valence of
// the atom and the explicit valence already specified - tells how
// many Hs to add
//
// The d-block and f-block of the periodic table (i.e. transition metals,
// lanthanoids and actinoids) have no default valence.
int dv = periodicTable->getDefaultValence(effectiveAtomicNum);
if (dv == -1) {
return 0;
}
// here is how we are going to deal with the possibility of
// multiple valences
// - check the explicit valence "ev"
// - if it is already equal to one of the allowed valences for the
// atom return 0
// - otherwise take return difference between next larger allowed
// valence and "ev"
// if "ev" is greater than all allowed valences for the atom raise an
// exception
// finally aromatic cases are dealt with differently - these atoms are allowed
// only default valences
// we have to include a special case here for negatively charged P, S, As,
// and Se, which all support "hypervalent" forms, but which can be
// isoelectronic to Cl/Ar or Br/Kr, which do not support hypervalent forms.
if (canBeHypervalent(atom, effectiveAtomicNum)) {
effectiveAtomicNum = atomicNum;
explicitPlusRadV -= atom.d_formalCharge;
}
const auto &valens = periodicTable->getValenceList(effectiveAtomicNum);
int res = 0;
// if we have an aromatic case treat it differently
if (isAromaticAtom(atom)) {
if (explicitPlusRadV <= dv) {
res = dv - explicitPlusRadV;
} else {
// As we assume when finding the explicitPlusRadValence if we are
// aromatic we should not be adding any hydrogen and already
// be at an accepted valence state,
// FIX: this is just ERROR checking and probably moot - the
// explicitPlusRadValence function called above should assure us that
// we satisfy one of the accepted valence states for the
// atom. The only diff I can think of is in the way we handle
// formal charge here vs the explicit valence function.
bool satis = false;
for (auto vi = valens.begin(); vi != valens.end() && *vi > 0; ++vi) {
if (explicitPlusRadV == *vi) {
satis = true;
break;
}
}
if (!satis && (strict || checkIt)) {
if (strict) {
std::ostringstream errout;
errout << "Explicit valence for aromatic atom # " << atom.getIdx()
<< " not equal to any accepted valence\n";
std::string msg = errout.str();
BOOST_LOG(rdErrorLog) << msg << std::endl;
throw AtomValenceException(msg, atom.getIdx());
} else {
return -1;
}
}
res = 0;
}
} else {
// non-aromatic case we are allowed to have non default valences
// and be able to add hydrogens
res = -1;
for (auto vi = valens.begin(); vi != valens.end() && *vi >= 0; ++vi) {
int tot = *vi;
if (explicitPlusRadV <= tot) {
res = tot - explicitPlusRadV;
break;
}
}
if (res < 0) {
if ((strict || checkIt) && valens.back() != -1 && ovalens.back() > 0) {
// this means that the explicit valence is greater than any
// allowed valence for the atoms
if (strict) {
// raise an error
std::ostringstream errout;
errout << "Explicit valence for atom # " << atom.getIdx() << " "
<< PeriodicTable::getTable()->getElementSymbol(atomicNum)
<< " greater than permitted";
std::string msg = errout.str();
BOOST_LOG(rdErrorLog) << msg << std::endl;
throw AtomValenceException(msg, atom.getIdx());
} else {
return -1;
}
} else {
res = 0;
}
}
}
return res;
}
int Atom::calcExplicitValence(bool strict) {
bool checkIt = false;
d_explicitValence = calculateExplicitValence(*this, strict, checkIt);
return d_explicitValence;
}
int Atom::calcImplicitValence(bool strict) {
if (d_explicitValence == -1) {
calcExplicitValence(strict);
}
bool checkIt = false;
d_implicitValence = calculateImplicitValence(*this, strict, checkIt);
return d_implicitValence;
}
void Atom::setMonomerInfo(AtomMonomerInfo *info) {
delete dp_monomerInfo;
dp_monomerInfo = info;
}
void Atom::setIsotope(unsigned int what) { d_isotope = what; }
double Atom::getMass() const {
if (d_isotope) {
double res =
PeriodicTable::getTable()->getMassForIsotope(d_atomicNum, d_isotope);
if (d_atomicNum != 0 && res == 0.0) {
res = d_isotope;
}
return res;
} else {
return PeriodicTable::getTable()->getAtomicWeight(d_atomicNum);
}
}
bool Atom::hasValenceViolation() const {
// Ignore dummy atoms, query atoms, or atoms attached to query bonds
auto bonds = getOwningMol().atomBonds(this);
auto is_query = [](auto b) { return b->hasQuery(); };
if (getAtomicNum() == 0 || hasQuery() ||
std::any_of(bonds.begin(), bonds.end(), is_query)) {
return false;
}
unsigned int effectiveAtomicNum;
try {
bool checkIt = true;
effectiveAtomicNum = getEffectiveAtomicNum(*this, checkIt);
} catch (const AtomValenceException &) {
return true;
}
// special case for H:
if (getAtomicNum() == 1) {
if (getFormalCharge() > 1 || getFormalCharge() < -1) {
return true;
}
} else {
// Non-H checks for absurd charge values:
// 1. the formal charge is larger than the atomic number
// 2. the formal charge moves us to a different row of the periodic table
if (getFormalCharge() > getAtomicNum() ||
PeriodicTable::getTable()->getRow(d_atomicNum) !=
PeriodicTable::getTable()->getRow(effectiveAtomicNum)) {
return true;
}
}
bool strict = false;
bool checkIt = true;
if (calculateExplicitValence(*this, strict, checkIt) == -1 ||
calculateImplicitValence(*this, strict, checkIt) == -1) {
return true;
}
return false;
}
void Atom::setQuery(Atom::QUERYATOM_QUERY *) {
// Atoms don't have complex queries so this has to fail
PRECONDITION(0, "plain atoms have no Query");
}
Atom::QUERYATOM_QUERY *Atom::getQuery() const { return nullptr; };
void Atom::expandQuery(Atom::QUERYATOM_QUERY *, Queries::CompositeQueryType,
bool) {
PRECONDITION(0, "plain atoms have no Query");
}
bool Atom::Match(Atom const *what) const {
PRECONDITION(what, "bad query atom");
bool res = getAtomicNum() == what->getAtomicNum();
// special dummy--dummy match case:
// [*] matches [*],[1*],[2*],etc.
// [1*] only matches [*] and [1*]
if (res) {
if (!this->getAtomicNum()) {
// this is the new behavior, based on the isotopes:
int tgt = this->getIsotope();
int test = what->getIsotope();
if (tgt && test && tgt != test) {
res = false;
}
} else {
// standard atom-atom match: The general rule here is that if this atom
// has a property that
// deviates from the default, then the other atom should match that value.
if ((this->getFormalCharge() &&
this->getFormalCharge() != what->getFormalCharge()) ||
(this->getIsotope() && this->getIsotope() != what->getIsotope()) ||
(this->getNumRadicalElectrons() &&
this->getNumRadicalElectrons() != what->getNumRadicalElectrons())) {
res = false;
}
}
}
return res;
}
void Atom::updatePropertyCache(bool strict) {
calcExplicitValence(strict);
calcImplicitValence(strict);
}
bool Atom::needsUpdatePropertyCache() const {
return !(this->d_explicitValence >= 0 &&
(this->df_noImplicit || this->d_implicitValence >= 0));
}
void Atom::clearPropertyCache() {
d_explicitValence = -1;
d_implicitValence = -1;
}
// returns the number of swaps required to convert the ordering
// of the probe list to match the order of our incoming bonds:
//
// e.g. if our incoming bond order is: [0,1,2,3]:
// getPerturbationOrder([1,0,2,3]) = 1
// getPerturbationOrder([1,2,3,0]) = 3
// getPerturbationOrder([1,2,0,3]) = 2
int Atom::getPerturbationOrder(const INT_LIST &probe) const {
INT_LIST ref;
for (const auto bnd : getOwningMol().atomBonds(this)) {
ref.push_back(bnd->getIdx());
}
return static_cast<int>(countSwapsToInterconvert(probe, ref));
}
static const unsigned char octahedral_invert[31] = {
0, // 0 -> 0
2, // 1 -> 2
1, // 2 -> 1
16, // 3 -> 16
14, // 4 -> 14
15, // 5 -> 15
18, // 6 -> 18
17, // 7 -> 17
10, // 8 -> 10
11, // 9 -> 11
8, // 10 -> 8
9, // 11 -> 9
13, // 12 -> 13
12, // 13 -> 12
4, // 14 -> 4
5, // 15 -> 5
3, // 16 -> 3
7, // 17 -> 7
6, // 18 -> 6
24, // 19 -> 24
23, // 20 -> 23
22, // 21 -> 22
21, // 22 -> 21
20, // 23 -> 20
19, // 24 -> 19
30, // 25 -> 30
29, // 26 -> 29
28, // 27 -> 28
27, // 28 -> 27
26, // 29 -> 26
25 // 30 -> 25
};
static const unsigned char trigonalbipyramidal_invert[21] = {
0, // 0 -> 0
2, // 1 -> 2
1, // 2 -> 1
4, // 3 -> 4
3, // 4 -> 3
6, // 5 -> 6
5, // 6 -> 5
8, // 7 -> 8
7, // 8 -> 7
11, // 9 -> 11
12, // 10 -> 12
9, // 11 -> 9
10, // 12 -> 10
14, // 13 -> 14
13, // 14 -> 13
20, // 15 -> 20
19, // 16 -> 19
18, // 17 -> 28
17, // 18 -> 17
16, // 19 -> 16
15 // 20 -> 15
};
bool Atom::invertChirality() {
unsigned int perm;
switch (getChiralTag()) {
case CHI_TETRAHEDRAL_CW:
setChiralTag(CHI_TETRAHEDRAL_CCW);
return true;
case CHI_TETRAHEDRAL_CCW:
setChiralTag(CHI_TETRAHEDRAL_CW);
return true;
case CHI_TETRAHEDRAL:
if (getPropIfPresent(common_properties::_chiralPermutation, perm)) {
if (perm == 1) {
perm = 2;
} else if (perm == 2) {
perm = 1;
} else {
perm = 0;
}
setProp(common_properties::_chiralPermutation, perm);
return perm != 0;
}
break;
case CHI_TRIGONALBIPYRAMIDAL:
if (getPropIfPresent(common_properties::_chiralPermutation, perm)) {
perm = (perm <= 20) ? trigonalbipyramidal_invert[perm] : 0;
setProp(common_properties::_chiralPermutation, perm);
return perm != 0;
}
break;
case CHI_OCTAHEDRAL:
if (getPropIfPresent(common_properties::_chiralPermutation, perm)) {
perm = (perm <= 30) ? octahedral_invert[perm] : 0;
setProp(common_properties::_chiralPermutation, perm);
return perm != 0;
}
break;
default:
break;
}
return false;
}
void setAtomRLabel(Atom *atm, int rlabel) {
PRECONDITION(atm, "bad atom");
// rlabel ==> n2 => 0..99
PRECONDITION(rlabel >= 0 && rlabel < 100,
"rlabel out of range for MDL files");
if (rlabel) {
atm->setProp(common_properties::_MolFileRLabel,
static_cast<unsigned int>(rlabel));
} else {
atm->clearProp(common_properties::_MolFileRLabel);
}
}
//! Gets the atom's RLabel
int getAtomRLabel(const Atom *atom) {
PRECONDITION(atom, "bad atom");
unsigned int rlabel = 0;
atom->getPropIfPresent(common_properties::_MolFileRLabel, rlabel);
return static_cast<int>(rlabel);
}
void setAtomAlias(Atom *atom, const std::string &alias) {
PRECONDITION(atom, "bad atom");
if (alias != "") {
atom->setProp(common_properties::molFileAlias, alias);
} else {
atom->clearProp(common_properties::molFileAlias);
}
}
std::string getAtomAlias(const Atom *atom) {
PRECONDITION(atom, "bad atom");
std::string alias;
atom->getPropIfPresent(common_properties::molFileAlias, alias);
return alias;
}
void setAtomValue(Atom *atom, const std::string &value) {
PRECONDITION(atom, "bad atom");
if (value != "") {
atom->setProp(common_properties::molFileValue, value);
} else {
atom->clearProp(common_properties::molFileValue);
}
}
std::string getAtomValue(const Atom *atom) {
PRECONDITION(atom, "bad atom");
std::string value;
atom->getPropIfPresent(common_properties::molFileValue, value);
return value;
}
void setSupplementalSmilesLabel(Atom *atom, const std::string &label) {
PRECONDITION(atom, "bad atom");
if (label != "") {
atom->setProp(common_properties::_supplementalSmilesLabel, label);
} else {
atom->clearProp(common_properties::_supplementalSmilesLabel);
}
}
std::string getSupplementalSmilesLabel(const Atom *atom) {
PRECONDITION(atom, "bad atom");
std::string label;
atom->getPropIfPresent(common_properties::_supplementalSmilesLabel, label);
return label;
}
unsigned int numPiElectrons(const Atom &atom) {
unsigned int res = 0;
if (atom.getIsAromatic()) {
res = 1;
} else if (atom.getHybridization() != Atom::SP3) {
auto val =
static_cast<unsigned int>(atom.getValence(Atom::ValenceType::EXPLICIT));
unsigned int physical_bonds = atom.getNumExplicitHs();
const auto &mol = atom.getOwningMol();
for (const auto bond : mol.atomBonds(&atom)) {
if (bond->getValenceContrib(&atom) != 0.0) {
++physical_bonds;
}
}
CHECK_INVARIANT(val >= physical_bonds,
"explicit valence exceeds atom degree");
res = val - physical_bonds;
}
return res;
}
} // namespace RDKit
namespace {
constexpr const char *hybridizationToString(
RDKit::Atom::HybridizationType type) {
switch (type) {
case RDKit::Atom::HybridizationType::UNSPECIFIED:
return "";
case RDKit::Atom::HybridizationType::S:
return "S";
case RDKit::Atom::HybridizationType::SP:
return "SP";
case RDKit::Atom::HybridizationType::SP2:
return "SP2";
case RDKit::Atom::HybridizationType::SP3:
return "SP3";
case RDKit::Atom::HybridizationType::SP3D:
return "SP3D";
case RDKit::Atom::HybridizationType::SP2D:
return "SP2D";
case RDKit::Atom::HybridizationType::SP3D2:
return "SP3D2";
case RDKit::Atom::HybridizationType::OTHER:
return "OTHER";
}
return "";
}
constexpr const char *chiralityToString(RDKit::Atom::ChiralType type) {
switch (type) {
case RDKit::Atom::ChiralType::CHI_UNSPECIFIED:
return "Unspecified";
case RDKit::Atom::ChiralType::CHI_TETRAHEDRAL_CW:
return "CW";
case RDKit::Atom::ChiralType::CHI_TETRAHEDRAL_CCW:
return "CCW";
case RDKit::Atom::ChiralType::CHI_OTHER:
return "Other";
case RDKit::Atom::ChiralType::CHI_TETRAHEDRAL:
return "Td";
case RDKit::Atom::ChiralType::CHI_ALLENE:
return "Allene";
case RDKit::Atom::ChiralType::CHI_SQUAREPLANAR:
return "SqP";
case RDKit::Atom::ChiralType::CHI_TRIGONALBIPYRAMIDAL:
return "Tbp";
case RDKit::Atom::ChiralType::CHI_OCTAHEDRAL:
return "Oh";
}
return "";
}
} // namespace
std::ostream &operator<<(std::ostream &target, const RDKit::Atom &at) {
target << at.getIdx() << " " << at.getAtomicNum() << " " << at.getSymbol();
target << " chg: " << at.getFormalCharge();
target << " deg: " << at.getDegree();
target << " exp: ";
target << (at.d_explicitValence >= 0 ? std::to_string(at.d_explicitValence)
: "N/A");
target << " imp: ";
if (at.df_noImplicit) {
target << "0";
} else {
target << (at.d_implicitValence >= 0 ? std::to_string(at.d_implicitValence)
: "N/A");
}
target << " hyb: " << hybridizationToString(at.getHybridization());
if (at.getIsAromatic()) {
target << " arom?: " << at.getIsAromatic();
}
if (at.getChiralTag() != RDKit::Atom::CHI_UNSPECIFIED) {
target << " chi: " << chiralityToString(at.getChiralTag());
int perm;
if (at.getPropIfPresent(RDKit::common_properties::_chiralPermutation,
perm)) {
target << "(" << perm << ")";
}
target << " nbrs:[";
bool first = true;
for (const auto nbr : at.getOwningMol().atomNeighbors(&at)) {
if (!first) {
target << " ";
} else {
first = false;
}
target << nbr->getIdx();
}
target << "]";
}
if (at.getNumRadicalElectrons()) {
target << " rad: " << at.getNumRadicalElectrons();
}
if (at.getIsotope()) {
target << " iso: " << at.getIsotope();
}
if (at.getAtomMapNum()) {
target << " mapno: " << at.getAtomMapNum();
}
if (at.hasQuery()) {
target << " query: " << at.getQuery()->getDescription();
}
return target;
};