mirror of
https://github.com/rdkit/rdkit.git
synced 2026-06-04 21:54:27 +08:00
8609cd4883
* StructChecker changes. Initial commit. First implementation. Added some tests. * StructChecker: add GoodAtoms and AcidicAtoms. new updates * StructChecker: add new tests * StructChecker: added TransformAugmentedAtoms() * StructCheck: add structCheck to GraphMol. Fix compilation errors. * StructChecker: add stereo verification and some utilities. * StructChecker: function FixDubious3DMolecule was added * StructChecker: checkStereo added. done with stereo. * StructChecker: add StripSmallFragments() * StructChecker: add AtomClash() function. Some cosmetic + tests * StructChecker: checkAtoms() was started * StructChecker: checkAtoms is ready * StructChecker: user RingInfo from RDkit. Start regarge * StructChecker: ReCharge molecule method prototype * StructChecker: updates for ReCharge. Almost finished * StructChecker: all ReCharge is done except external data tables loading * StructChecker: add path tables into API. ReCharge completed * Adds augmented atom data Signed-off-by: Brian Kelley <brian.kelley@novartis.com> * Removes extra files Signed-off-by: Brian Kelley <brian.kelley@novartis.com> * Adds path to test data via RDBASE environment Signed-off-by: Brian Kelley <brian.kelley@novartis.com> * Revert "Struct checker apr15" * StructChecker: add missing tautomer tests * Updates test to use RDBASE * Adds initialization of data from data section * Adds Python API and tests * Fixes namespace for enum * StructChecker: update/imporve strip small fragments * StructChecker: fix acidic atoms (but logic does not work) * StructChecker: fix match issue for CheckAtoms * Adds macro guards * Adds loading API and proper constructor * Fixes tests, adds stereo test * Fixes crash bug, matches[0] was being accessed from an empty match vector * Reverts crash fix - conflicts with previous * Adds the rest of the structure checker options * StructChecker: fix atom matching for aromatic rings * StructChecker: add tautomers checks. Update some tests * StructChecker: stereo fixes. Add some tests * StructChecker: fix check atoms. Start ligand symbol list * StructChecker: fix some check atoms validation. Add Tranform to query lists. Start correct loading augmented atoms * update * another set of fixes * StructChecker: fix loadDefaultAugmentedAtoms. Some changes in CheckAtom + tests + debug conditional breakpoints (TEMP operators) * StructChecker: rewrited RecMatch() to sequential. Changed bond matching algorithm. small bug fixes * Adds better logging of mismatched atoms * Removes duplicated negative charge * Fixes charges * Adds nitro group test * StructChecker: add better logging * remove double logging * Reformats code using RDKit's clang-format style * StructChecker: Fix charge reformat using RDKit format. * StructChecker: compilation restore after merge * restore bond matching * Removes the same fragments that strucheck does in case of ties * Don't resanitize - this adds aromaticity which mucks things up * Adds empty molecule checks * Fixes atom clashes. * Removes debug printing * Removes debug logging info * First pass at stereo fixes * Fixes off by one error for dubious stereo fix * Fixes more off by one errors * Fixes more off by one errors * More off by one fixes. * Another off by one * Fixes chiral flag set in molfile check * Copies chiral flag over to largest fragment if necessary * Poor man’s parity check. * Find unspecified chiral centers ala Avalon. * StructChecker: fix recursive match. Fix transformations * StructChecker: fix transformation for atom list (using query atoms) * Fixes checks && to & * StructChecker: fix carboxylic acids tranform issue. Atom list is changed only if different * StructChecker: documentation was updated * Fixes snprintf and silences some warnings * Adds Get/Set StructCheckerOptions * Adds default AugmentedAtomTransforms
80 lines
2.9 KiB
C++
80 lines
2.9 KiB
C++
//
|
|
// Copyright (C) 2016 Novartis Institutes for BioMedical Research
|
|
//
|
|
// @@ 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 "../Substruct/SubstructMatch.h"
|
|
#include "StructChecker.h"
|
|
#include "Pattern.h"
|
|
#include "Tautomer.h"
|
|
|
|
namespace RDKit {
|
|
namespace StructureCheck {
|
|
|
|
bool StructCheckTautomer::applyTautomer(unsigned it) {
|
|
if (Options.FromTautomer.size() <= it || Options.ToTautomer.size() <= it) {
|
|
if (Options.Verbose)
|
|
BOOST_LOG(rdInfoLog) << "ERROR: incorrect Tautomer index it=" << it
|
|
<< "\n";
|
|
return false;
|
|
}
|
|
const ROMol &fromTautomer = *Options.FromTautomer[it];
|
|
const ROMol &toTautomer = *Options.ToTautomer[it];
|
|
if (toTautomer.getNumAtoms() != fromTautomer.getNumAtoms()) {
|
|
if (Options.Verbose)
|
|
BOOST_LOG(rdInfoLog) << "ERROR: incorrect data toTautomer.getNumAtoms() "
|
|
"!= fromTautomer.getNumAtoms()\n";
|
|
// incorrect data
|
|
// throw(.....);
|
|
return false;
|
|
}
|
|
const unsigned nta = toTautomer.getNumAtoms();
|
|
const unsigned ntb = toTautomer.getNumBonds();
|
|
MatchVectType match; // The format is (queryAtomIdx, molAtomIdx)
|
|
std::vector<unsigned> atomIdxMap(
|
|
Mol.getNumAtoms()); // matched tau atom indeces
|
|
|
|
if (!SubstructMatch(Mol, *Options.FromTautomer[it],
|
|
match)) // SSMatch(mp, from_tautomer, SINGLE_MATCH);
|
|
return false;
|
|
if (Options.Verbose)
|
|
BOOST_LOG(rdInfoLog) << "found match for from_tautomer with " << nta
|
|
<< " atoms\n";
|
|
// init
|
|
for (unsigned i = 0; i < Mol.getNumAtoms(); i++) atomIdxMap[i] = -1;
|
|
for (MatchVectType::const_iterator mit = match.begin(); mit != match.end();
|
|
mit++) {
|
|
unsigned tai = mit->first; // From and To Tautomer Atom index
|
|
unsigned mai = mit->second; // Mol Atom index
|
|
atomIdxMap[mai] = tai;
|
|
}
|
|
// scan for completely mapped bonds and replace bond order with mapped bond
|
|
// from to_tautomer
|
|
for (RDKit::BondIterator_ bond = Mol.beginBonds(); bond != Mol.endBonds();
|
|
bond++) {
|
|
unsigned ti = atomIdxMap[(*bond)->getBeginAtomIdx()];
|
|
unsigned tj = atomIdxMap[(*bond)->getEndAtomIdx()];
|
|
if (-1 == ti || -1 == tj) continue;
|
|
const Bond *tb = toTautomer.getBondBetweenAtoms(ti, tj);
|
|
if (tb && (*bond)->getBondType() != tb->getBondType()) {
|
|
(*bond)->setBondType(tb->getBondType());
|
|
}
|
|
}
|
|
// apply charge/radical fixes if any
|
|
for (unsigned i = 0; i < match.size(); i++) {
|
|
Atom &atom = *Mol.getAtomWithIdx(match[i].second);
|
|
const Atom &ta = *toTautomer.getAtomWithIdx(match[i].first);
|
|
atom.setFormalCharge(ta.getFormalCharge());
|
|
atom.setNumRadicalElectrons(ta.getNumRadicalElectrons());
|
|
}
|
|
|
|
return true;
|
|
}
|
|
|
|
} // namespace StructureCheck
|
|
} // namespace RDKit
|