void BuildSimpleMolecule() {
// build the molecule: C/C=C\C
RWMol *mol = new RWMol();
// add atoms and bonds:
mol->addAtom(new Atom(6)); // atom 0
mol->addAtom(new Atom(6)); // atom 1
mol->addAtom(new Atom(6)); // atom 2
mol->addAtom(new Atom(6)); // atom 3
mol->addBond(0, 1, Bond::SINGLE); // bond 0
mol->addBond(1, 2, Bond::DOUBLE); // bond 1
mol->addBond(2, 3, Bond::SINGLE); // bond 2
// setup the stereochem:
mol->getBondWithIdx(0)->setBondDir(Bond::ENDUPRIGHT);
mol->getBondWithIdx(2)->setBondDir(Bond::ENDDOWNRIGHT);
// do the chemistry perception:
RDKit::MolOps::sanitizeMol(*mol);
// Get the canonical SMILES, include stereochemistry:
std::string smiles;
smiles = MolToSmiles(*(static_cast<ROMol *>(mol)), true);
BOOST_LOG(rdInfoLog) << " sample 1 SMILES: " << smiles << std::endl;
}
// Test the following GitHub issue: https://github.com/rdkit/rdkit/issues/114
void testIssue114(std::string rdbase) {
BOOST_LOG(rdInfoLog) << "-----------------------------------" << std::endl;
BOOST_LOG(rdInfoLog) << "-- testing GitHub issue #114 --" << std::endl;
BOOST_LOG(rdInfoLog) << "-----------------------------------" << std::endl;
std::string fName = rdbase + "/Code/GraphMol/FileParsers/test_data/EZ_mol2_issue114.mol2";
RWMol *mol = Mol2FileToMol(fName);
TEST_ASSERT(mol);
TEST_ASSERT(mol->getBondWithIdx(1)->getStereo() == Bond::STEREOZ);
delete mol;
BOOST_LOG(rdInfoLog) << "------------------------------------" << std::endl;
BOOST_LOG(rdInfoLog) << "-- DONE --" << std::endl;
BOOST_LOG(rdInfoLog) << "------------------------------------" << std::endl;
}
ROMol *fragmentOnBonds(
const ROMol &mol, const std::vector<unsigned int> &bondIndices,
bool addDummies,
const std::vector<std::pair<unsigned int, unsigned int> > *dummyLabels,
const std::vector<Bond::BondType> *bondTypes,
std::vector<unsigned int> *nCutsPerAtom) {
PRECONDITION((!dummyLabels || dummyLabels->size() == bondIndices.size()),
"bad dummyLabel vector");
PRECONDITION((!bondTypes || bondTypes->size() == bondIndices.size()),
"bad bondType vector");
PRECONDITION((!nCutsPerAtom || nCutsPerAtom->size() == mol.getNumAtoms()),
"bad nCutsPerAtom vector");
if (nCutsPerAtom) {
BOOST_FOREACH (unsigned int &nCuts, *nCutsPerAtom) { nCuts = 0; }
}
RWMol *res = new RWMol(mol);
if (!mol.getNumAtoms()) return res;
std::vector<Bond *> bondsToRemove;
bondsToRemove.reserve(bondIndices.size());
BOOST_FOREACH (unsigned int bondIdx, bondIndices) {
bondsToRemove.push_back(res->getBondWithIdx(bondIdx));
}
bool kekulizeWorker(RWMol &mol, const INT_VECT &allAtms,
boost::dynamic_bitset<> dBndCands,
boost::dynamic_bitset<> dBndAdds, INT_VECT done,
unsigned int maxBackTracks) {
INT_DEQUE astack;
INT_INT_DEQ_MAP options;
int lastOpt = -1;
boost::dynamic_bitset<> localBondsAdded(mol.getNumBonds());
// ok the algorithm goes something like this
// - start with an atom that has been marked aromatic before
// - check if it can have a double bond
// - add its neighbors to the stack
// - check if one of its neighbors can also have a double bond
// - if yes add a double bond.
// - if multiple neighbors can have double bonds - add them to a
// options stack we may have to retrace out path if we chose the
// wrong neighbor to add the double bond
// - if double bond added update the candidates for double bond
// - move to the next atom on the stack and repeat the process
// - if an atom that can have multiple a double bond has no
// neighbors that can take double bond - we made a mistake
// earlier by picking a wrong candidate for double bond
// - in this case back track to where we made the mistake
int curr;
INT_DEQUE btmoves;
unsigned int numBT = 0; // number of back tracks so far
while ((done.size() < allAtms.size()) || (astack.size() > 0)) {
// pick a curr atom to work with
if (astack.size() > 0) {
curr = astack.front();
astack.pop_front();
} else {
for (int allAtm : allAtms) {
if (std::find(done.begin(), done.end(), allAtm) == done.end()) {
curr = allAtm;
break;
}
}
}
done.push_back(curr);
// loop over the neighbors if we can add double bonds or
// simply push them onto the stack
INT_DEQUE opts;
bool cCand = false;
if (dBndCands[curr]) {
cCand = true;
}
int ncnd;
// if we are here because of backtracking
if (options.find(curr) != options.end()) {
opts = options[curr];
CHECK_INVARIANT(opts.size() > 0, "");
} else {
RWMol::ADJ_ITER nbrIdx, endNbrs;
boost::tie(nbrIdx, endNbrs) =
mol.getAtomNeighbors(mol.getAtomWithIdx(curr));
while (nbrIdx != endNbrs) {
// ignore if the neighbor has already been dealt with before
if (std::find(done.begin(), done.end(), static_cast<int>(*nbrIdx)) !=
done.end()) {
++nbrIdx;
continue;
}
// ignore if the neighbor is not part of the fused system
if (std::find(allAtms.begin(), allAtms.end(),
static_cast<int>(*nbrIdx)) == allAtms.end()) {
++nbrIdx;
continue;
}
// if the neighbor is not on the stack add it
if (std::find(astack.begin(), astack.end(),
static_cast<int>(*nbrIdx)) == astack.end()) {
astack.push_back(rdcast<int>(*nbrIdx));
}
// check if the neighbor is also a candidate for a double bond
// the refinement that we'll make to the candidate check we've already
// done is to make sure that the bond is either flagged as aromatic
// or involves a dummy atom. This was Issue 3525076.
// This fix is not really 100% of the way there: a situation like
// that for Issue 3525076 but involving a dummy atom in the cage
// could lead to the same failure. The full fix would require
// a fairly detailed analysis of all bonds in the molecule to determine
// which of them is eligible to be converted.
if (cCand && dBndCands[*nbrIdx] &&
(mol.getBondBetweenAtoms(curr, *nbrIdx)->getIsAromatic() ||
mol.getAtomWithIdx(curr)->getAtomicNum() == 0 ||
mol.getAtomWithIdx(*nbrIdx)->getAtomicNum() == 0)) {
opts.push_back(rdcast<int>(*nbrIdx));
} // end of curr atoms can have a double bond
++nbrIdx;
} // end of looping over neighbors
}
// now add a double bond from current to one of the neighbors if we can
if (cCand) {
if (opts.size() > 0) {
ncnd = opts.front();
opts.pop_front();
Bond *bnd = mol.getBondBetweenAtoms(curr, ncnd);
bnd->setBondType(Bond::DOUBLE);
// remove current and the neighbor from the dBndCands list
dBndCands[curr] = 0;
dBndCands[ncnd] = 0;
// add them to the list of bonds to which have been made double
dBndAdds[bnd->getIdx()] = 1;
localBondsAdded[bnd->getIdx()] = 1;
// if this is an atom we previously visted and picked we
// simply tried a different option now, overwrite the options
// stored for this atoms
if (options.find(curr) != options.end()) {
if (opts.size() == 0) {
options.erase(curr);
btmoves.pop_back();
if (btmoves.size() > 0) {
lastOpt = btmoves.back();
} else {
lastOpt = -1;
}
} else {
options[curr] = opts;
}
} else {
// this is new atoms we are trying and have other
// neighbors as options to add double bond store this to
// the options stack, we may have made a mistake in
// which one we chose and have to return here
if (opts.size() > 0) {
lastOpt = curr;
btmoves.push_back(lastOpt);
options[curr] = opts;
}
}
} // end of adding a double bond
else {
// we have an atom that should be getting a double bond
// but none of the neighbors can take one. Most likely
// because of a wrong choice earlier so back track
if ((lastOpt >= 0) && (numBT < maxBackTracks)) {
// std::cerr << "PRE BACKTRACK" << std::endl;
// mol.debugMol(std::cerr);
backTrack(mol, options, lastOpt, done, astack, dBndCands, dBndAdds);
// std::cerr << "POST BACKTRACK" << std::endl;
// mol.debugMol(std::cerr);
numBT++;
} else {
// undo any remaining changes we made while here
// this was github #962
for (unsigned int bidx = 0; bidx < mol.getNumBonds(); ++bidx) {
if (localBondsAdded[bidx]) {
mol.getBondWithIdx(bidx)->setBondType(Bond::SINGLE);
}
}
return false;
}
} // end of else try to backtrack
} // end of curr atom atom being a cand for double bond
} // end of while we are not done with all atoms
return true;
}
