std::vector<double> MolData3Ddescriptors::GetRelativeRcov(const RDKit::ROMol& mol){
int numAtoms= mol.getNumAtoms();
double* rcov=data3D.getRCOV();
std::vector<double> wroc(numAtoms, 0.0);
for( int i=0; i<numAtoms; ++i){
wroc[i]=rcov[mol.getAtomWithIdx(i)->getAtomicNum()-1]/rcov[5];
}
return wroc;
}
std::vector<double> MolData3Ddescriptors::GetRelativeENeg(const RDKit::ROMol& mol){
int numAtoms= mol.getNumAtoms();
double* relativeNeg=data3D.getNEG();
std::vector<double> neg(numAtoms, 0.0);
for( int i=0; i<numAtoms; ++i){
neg[i]=relativeNeg[mol.getAtomWithIdx(i)->getAtomicNum()-1];
}
return neg;
}
std::vector<double> MolData3Ddescriptors::GetRelativeMW(const RDKit::ROMol& mol){
double* relativeMw=data3D.getMW();
int numAtoms= mol.getNumAtoms();
std::vector<double> pol(numAtoms, 0.0);
for( int i=0; i<numAtoms; ++i){
pol[i]=relativeMw[mol.getAtomWithIdx(i)->getAtomicNum()-1];
}
return pol;
}
std::vector<double> MolData3Ddescriptors::GetRelativeIonPol(const RDKit::ROMol& mol){
int numAtoms= mol.getNumAtoms();
double* absionpol=data3D.getIonPOL();
std::vector<double> ionpols(numAtoms, 0.0);
for( int i=0; i<numAtoms; ++i){
ionpols[i]=absionpol[mol.getAtomWithIdx(i)->getAtomicNum()-1];
}
return ionpols;
}
std::vector<double> MolData3Ddescriptors::GetRelativeVdW(const RDKit::ROMol& mol){
int numAtoms= mol.getNumAtoms();
double* relativeVdW=data3D.getVDW();
std::vector<double> vdw(numAtoms, 0.0);
for( int i=0; i<numAtoms; ++i){
vdw[i]=relativeVdW[mol.getAtomWithIdx(i)->getAtomicNum()-1];
}
return vdw;
}
double calcSASA(const RDKit::ROMol &mol, const std::vector<double> &radii,
int confIdx, const RDKit::QueryAtom *query,
const SASAOpts &opts) {
double result = internalCalcSASA(mol, radii, confIdx, opts);
if (query) {
result = 0.0f;
for (ROMol::ConstQueryAtomIterator at = mol.beginQueryAtoms(query);
at != mol.endQueryAtoms(); ++at) {
const Atom *atom = *at;
result += atom->getProp<double>("SASA");
}
}
return result;
}
unsigned int Compute2DCoords(RDKit::ROMol &mol, bool canonOrient,
bool clearConfs, python::dict &coordMap,
unsigned int nFlipsPerSample = 3,
unsigned int nSamples = 100, int sampleSeed = 100,
bool permuteDeg4Nodes = false,
double bondLength = -1.0) {
RDGeom::INT_POINT2D_MAP cMap;
cMap.clear();
python::list ks = coordMap.keys();
for (unsigned int i = 0;
i < python::extract<unsigned int>(ks.attr("__len__")()); i++) {
unsigned int id = python::extract<unsigned int>(ks[i]);
if (id >= mol.getNumAtoms()) {
throw_value_error("atom index out of range");
}
cMap[id] = python::extract<RDGeom::Point2D>(coordMap[id]);
}
double oBondLen = RDDepict::BOND_LEN;
if (bondLength > 0) {
RDDepict::BOND_LEN = bondLength;
}
unsigned int res;
res = RDDepict::compute2DCoords(mol, &cMap, canonOrient, clearConfs,
nFlipsPerSample, nSamples, sampleSeed,
permuteDeg4Nodes);
if (bondLength > 0) {
RDDepict::BOND_LEN = oBondLen;
}
return res;
}
// adaptation from EState.py
// we need the Is value only there
std::vector<double> MolData3Ddescriptors::GetEState(const RDKit::ROMol &mol){
int numAtoms = mol.getNumAtoms();
std::vector<double> Is =GetIState(mol);
double tmp,p;
double *dist = RDKit::MolOps::getDistanceMat(mol,false,false);
double accum[numAtoms];
for (int i=0;i<numAtoms;i++) {
accum[i]=0.0;
}
for (int i=0;i<numAtoms;i++) {
for (int j=i+1;j<numAtoms;j++) {
p = dist[i * numAtoms + j]+1;
if (p < 1e6) {
tmp = (Is[i] - Is[j]) / (p * p);
accum[i] += tmp;
accum[j] -= tmp;
}
}
}
for (int i=0;i<numAtoms;i++) {
Is[i]+=accum[i];
}
return Is;
}
std::string MolpherMol::MolpherMolImpl::asMolBlock(bool include_locks) const {
RDKit::ROMol* temp = asRDMol(include_locks);
std::ostringstream os;
os << RDKit::MolToMolBlock(*temp) << std::endl;
if (include_locks) {
RDKit::STR_VECT prop_names = temp->getPropList();
for (MolpherAtom::LockingMask lock : MolpherAtom::atom_locks) {
std::string lock_prop_name("MOLPHER_" + MolpherAtom::lockToString(lock));
if (std::find(prop_names.begin(), prop_names.end(), lock_prop_name) != prop_names.end()) {
os << "> <" << lock_prop_name << ">" << std::endl;
os << temp->getProp<std::string>(lock_prop_name) << std::endl;
os << std::endl;
}
}
os << "$$$$" << std::endl;
}
delete temp;
return os.str();
}
std::vector<double> MolData3Ddescriptors::GetIState(const RDKit::ROMol &mol){
int numAtoms = mol.getNumAtoms();
std::vector<double> Is;
for (int i = 0; i < numAtoms; ++i) {
const RDKit::Atom * atom= mol.getAtomWithIdx(i);
int atNum=atom->getAtomicNum();
int degree = atom->getDegree();
if (degree>0 and atNum>1) {
int h = atom->getTotalNumHs();
int Zv = RDKit::PeriodicTable::getTable()->getNouterElecs(atNum);
double dv =(double) Zv-h;
dv = dv / (double) (atNum-Zv-1);
int N = GetPrincipalQuantumNumber(atNum);
Is.push_back(round(1000*(4.0/(N*N)*dv+1.0)/degree)/1000); // WHIM-P5.pdf paper 1997 => +7 & NoHydrogens is used!
}
else Is.push_back(0);
}
return Is;
}
void testautocorrelation() {
std::cout << "=>start test rdf\n";
std::string pathName = getenv("RDBASE");
std::string sdfName =
pathName + "/Code/GraphMol/Descriptors/test_data/chlorobenzene.sdf";
RDKit::SDMolSupplier reader(sdfName, true, false);
int nDone = 0;
while (!reader.atEnd()) {
++nDone;
RDKit::ROMol *m = reader.next();
TEST_ASSERT(m);
std::string nm;
m->getProp("_Name",nm);
std::vector<double> dwhim;
//for (int i=1;i<11;i++) {
// std::cout << "i:" << 0.005*i << "\n";
dwhim = RDKit::Descriptors::AUTOCORR3D(*m, -1);
for (int j=0;j<80;j++) {
std::cout << dwhim[j] << ",";
}
std::cout << "\n";
//}
std::cout << "=>read molecule: " << nDone << std::endl;
delete m;
}
BOOST_LOG(rdErrorLog) << " done" << std::endl;
}
// modification of previous code to follow documentation from Padel code
std::vector<double> MolData3Ddescriptors::GetEState2(const RDKit::ROMol &mol){
int numAtoms = mol.getNumAtoms();
std::vector<double> Si =GetIState(mol);
// in WHIM definition it's write:
double tmp,p,d;
double *dist = RDKit::MolOps::getDistanceMat(mol,false,false);
double accum[numAtoms];
for (int i=0;i<numAtoms;i++) {
accum[i]=0.0;
}
for (int i=0;i<numAtoms;i++) {
for (int j=i+1;j<numAtoms;j++) {
d = dist[i * numAtoms + j];
p = dist[i * numAtoms + j]+1;
if (d == 1) {
tmp = (Si[i] - Si[j]) / (p * p);
accum[i] += tmp;
accum[j] -= tmp;
}
}
}
// add the Accum to the Si
// WHIM Si values
// electrotopological indices are scaled thus: Si'=Si + 7 => Si' > 0
// In this case, only the nonhydrogen atoms are considered,
// and the atomic electrotopological charge of each atom depends on its atom neighbor.
// So we should not use all the terms in the sum but only Adj matrix cases!
// Correct the Si adding the rescaling parameter for WHIM only
for (int i=0;i<numAtoms;i++) {
Si[i]+=accum[i]+7.0;
}
return Si;
}
unsigned int Compute2DCoordsMimicDistmat(
RDKit::ROMol &mol, python::object distMat, bool canonOrient,
bool clearConfs, double weightDistMat, unsigned int nFlipsPerSample,
unsigned int nSamples, int sampleSeed, bool permuteDeg4Nodes,
double bondLength = -1.0) {
PyObject *distMatPtr = distMat.ptr();
if (!PyArray_Check(distMatPtr)) {
throw_value_error("Argument isn't an array");
}
PyArrayObject *dmatrix = reinterpret_cast<PyArrayObject *>(distMatPtr);
unsigned int nitems = PyArray_DIM(dmatrix, 0);
unsigned int na = mol.getNumAtoms();
if (nitems != na * (na - 1) / 2) {
throw_value_error(
"The array size does not match the number of atoms in the molecule");
}
double *inData = reinterpret_cast<double *>(PyArray_DATA(dmatrix));
double *cData = new double[nitems];
memcpy(static_cast<void *>(cData), static_cast<const void *>(inData),
nitems * sizeof(double));
DOUBLE_SMART_PTR dmat(cData);
double oBondLen = RDDepict::BOND_LEN;
if (bondLength > 0) {
RDDepict::BOND_LEN = bondLength;
}
unsigned int res;
res = RDDepict::compute2DCoordsMimicDistMat(
mol, &dmat, canonOrient, clearConfs, weightDistMat, nFlipsPerSample,
nSamples, sampleSeed, permuteDeg4Nodes);
if (bondLength > 0) {
RDDepict::BOND_LEN = oBondLen;
}
return res;
}
std::vector<double> MolData3Ddescriptors::GetCharges(const RDKit::ROMol &mol) {
std::vector<double> charges(mol.getNumAtoms(), 0);
// use 12 iterations... can be more
RDKit::computeGasteigerCharges(mol, charges, 12, true);
return charges;
}
void getExperimentalTorsions(const RDKit::ROMol &mol, CrystalFFDetails &details,
bool useExpTorsions, bool useBasicKnowledge,
unsigned int version, bool verbose) {
unsigned int nb = mol.getNumBonds();
unsigned int na = mol.getNumAtoms();
if (!na) {
throw ValueErrorException("molecule has no atoms");
}
// check that vectors are empty
details.expTorsionAtoms.clear();
details.expTorsionAngles.clear();
details.improperAtoms.clear();
unsigned int aid1, aid2, aid3, aid4;
unsigned int bid2;
boost::dynamic_bitset<> doneBonds(nb);
if (useExpTorsions) {
// we set the torsion angles with experimental data
const ExpTorsionAngleCollection *params =
ExpTorsionAngleCollection::getParams(version);
// loop over patterns
for (const auto ¶m : *params) {
std::vector<MatchVectType> matches;
SubstructMatch(mol, *(param.dp_pattern.get()), matches, false, true);
// loop over matches
for (std::vector<MatchVectType>::const_iterator matchIt = matches.begin();
matchIt != matches.end(); ++matchIt) {
// get bond indices
aid1 = (*matchIt)[param.idx[0]].second;
aid2 = (*matchIt)[param.idx[1]].second;
aid3 = (*matchIt)[param.idx[2]].second;
aid4 = (*matchIt)[param.idx[3]].second;
// FIX: check if bond is NULL
bid2 = mol.getBondBetweenAtoms(aid2, aid3)->getIdx();
if (!doneBonds[bid2]) {
doneBonds[bid2] = 1;
std::vector<int> atoms(4);
atoms[0] = aid1;
atoms[1] = aid2;
atoms[2] = aid3;
atoms[3] = aid4;
details.expTorsionAtoms.push_back(atoms);
details.expTorsionAngles.push_back(
std::make_pair(param.signs, param.V));
if (verbose) {
std::cout << param.smarts << ": " << aid1 << " " << aid2 << " "
<< aid3 << " " << aid4 << ", (";
for (unsigned int i = 0; i < param.V.size() - 1; ++i) {
std::cout << param.V[i] << ", ";
}
std::cout << param.V[param.V.size() - 1] << ") " << std::endl;
}
} // if not donePaths
} // end loop over matches
} // end loop over patterns
}
// apply basic knowledge such as flat aromatic rings, other sp2-centers,
// straight triple bonds, etc.
if (useBasicKnowledge) {
boost::dynamic_bitset<> doneAtoms(na);
ROMol::ADJ_ITER nbrIdx;
ROMol::ADJ_ITER endNbrs;
// inversion terms (improper torsions / out-of-plane bends / inversion)
// loop over atoms
for (aid2 = 0; aid2 < na; ++aid2) {
if (!(doneAtoms[aid2])) {
std::vector<int> atoms(4, -1);
atoms[1] = aid2;
const Atom *atom2 = mol.getAtomWithIdx(atoms[1]);
int at2AtomicNum = atom2->getAtomicNum();
// if atom is a N,O or C and SP2-hybridized
if (((at2AtomicNum == 6) || (at2AtomicNum == 7) ||
(at2AtomicNum == 8)) &&
(atom2->getHybridization() == Atom::SP2)) {
// get neighbors
boost::tie(nbrIdx, endNbrs) = mol.getAtomNeighbors(atom2);
// check if enough neighbours
if (mol.getAtomDegree(atom2) != 3) {
continue;
}
unsigned int i = 0;
unsigned int isBoundToSP2O = 0; // false
for (; nbrIdx != endNbrs; ++nbrIdx) {
const Atom *atomX = mol[*nbrIdx];
atoms[i] = atomX->getIdx();
// if the central atom is sp2 carbon and is bound to sp2 oxygen, set
// a flag
if (!isBoundToSP2O) {
isBoundToSP2O =
((at2AtomicNum == 6) && (atomX->getAtomicNum() == 8) &&
(atomX->getHybridization() == Atom::SP2));
}
if (!i) {
++i;
}
++i;
}
atoms.push_back(at2AtomicNum);
atoms.push_back(isBoundToSP2O);
details.improperAtoms.push_back(atoms);
/*if (verbose) {
std::cout << "out-of-plane bend: " << atoms[0] << " " << atoms[1] <<
" "
<< atoms[2] << " " << atoms[3] << std::endl;
}*/
}
} // if atom is a N,O or C and SP2-hybridized
}
// torsions for flat rings
const RingInfo *rinfo =
mol.getRingInfo(); // FIX: make sure we have ring info
CHECK_INVARIANT(rinfo, "");
const VECT_INT_VECT &atomRings = rinfo->atomRings();
for (const auto &atomRing : atomRings) {
unsigned int rSize = atomRing.size();
// we don't need to deal with 3 membered rings
// and we do not treat rings greater than 6
if (rSize < 4 || rSize > 6) {
continue;
}
// loop over ring atoms
for (unsigned int i = 0; i < rSize; ++i) {
// proper torsions
aid1 = atomRing[i];
aid2 = atomRing[(i + 1) % rSize];
aid3 = atomRing[(i + 2) % rSize];
aid4 = atomRing[(i + 3) % rSize];
bid2 = mol.getBondBetweenAtoms(aid2, aid3)->getIdx();
// if all 4 atoms are SP2, add torsion
if (!(doneBonds[bid2]) &&
(mol.getAtomWithIdx(aid1)->getHybridization() == Atom::SP2) &&
(mol.getAtomWithIdx(aid2)->getHybridization() == Atom::SP2) &&
(mol.getAtomWithIdx(aid3)->getHybridization() == Atom::SP2) &&
(mol.getAtomWithIdx(aid4)->getHybridization() == Atom::SP2)) {
doneBonds[bid2] = 1;
std::vector<int> atoms(4);
atoms[0] = aid1;
atoms[1] = aid2;
atoms[2] = aid3;
atoms[3] = aid4;
details.expTorsionAtoms.push_back(atoms);
std::vector<int> signs(6, 1);
signs[1] = -1; // MMFF sign for m = 2
std::vector<double> fconsts(6, 0.0);
fconsts[1] = 100.0; // 7.0 is MMFF force constants for aromatic rings
details.expTorsionAngles.push_back(std::make_pair(signs, fconsts));
/*if (verbose) {
std::cout << "SP2 ring: " << aid1 << " " << aid2 << " " << aid3 << "
" << aid4 << std::endl;
}*/
}
} // loop over atoms in ring
} // loop over rings
} // if useBasicKnowledge
} // end function
void testMORSE() {
std::cout << "=>start test MORSE\n";
std::string pathName = getenv("RDBASE");
std::string sdfName =
pathName + "/Code/GraphMol/Descriptors/test_data/PBF_egfr.sdf";
RDKit::SDMolSupplier reader(sdfName, true, false);
std::string fName =
pathName + "/Code/GraphMol/Descriptors/test_data/MORSE.out";
std::ifstream instrm(fName.c_str());
std::string line;
std::vector<std::vector<std::string> > data;
while (std::getline(instrm, line)) {
std::string phrase;
std::vector<std::string> row;
std::stringstream ss(line);
while (std::getline(ss, phrase, '\t')) {
row.push_back(phrase);
}
data.push_back(row);
}
std::cout << "=>read file\n";
int nDone = 0;
while (!reader.atEnd()) {
RDKit::ROMol *m = reader.next();
TEST_ASSERT(m);
std::string nm;
m->getProp("_Name", nm);
std::vector<double> dmorse;
RDKit::Descriptors::MORSE(*m, dmorse, -1);
std::vector<std::string> myrow = data[nDone];
std::string inm = myrow[0];
TEST_ASSERT(inm == nm);
for (int i = 0; i < dmorse.size(); i++) {
double ref = atof(myrow[i + 1].c_str());
if (fabs(ref) > 0.01) {
if (fabs((ref - dmorse[i]) / ref) > 1) {
std::cout << "value mismatch: pos" << i << " " << inm << " " << ref
<< " " << dmorse[i] << std::endl;
}
}
if (fabs(ref) < 0.01) {
if (fabs(ref - dmorse[i]) > 0.02) {
std::cout << "value mismatch: pos" << i << " " << inm << " " << ref
<< " " << dmorse[i] << std::endl;
}
}
if (ref > 1 && fabs(ref - dmorse[i]) / ref > 0.02) {
std::cout << "value mismatch: pos" << i << " " << inm << " " << ref
<< " " << dmorse[i] << std::endl;
}
// we're testing reasonably sized values and want to be sure that we're
// within 2% of the reference.
TEST_ASSERT(ref < 1 || fabs(ref - dmorse[i]) / ref < 0.02);
}
delete m;
++nDone;
}
BOOST_LOG(rdErrorLog) << "test on : " << nDone << " molecules done"
<< std::endl;
}
void canonicalizeMol(RDKit::ROMol &mol, bool normalizeCovar, bool ignoreHs) {
ROMol::ConformerIterator ci;
for (ci = mol.beginConformers(); ci != mol.endConformers(); ci++) {
canonicalizeConformer(*(*ci), 0, normalizeCovar, ignoreHs);
}
}
void testGETAWAY() {
std::cout << "=>start test GETAWAY\n";
std::string pathName = getenv("RDBASE");
std::string sdfName =
pathName + "/Code/GraphMol/Descriptors/test_data/PBF_egfr.sdf";
RDKit::SDMolSupplier reader(sdfName, true, false);
std::string fName =
pathName + "/Code/GraphMol/Descriptors/test_data/GETAWAY.new.out";
std::ifstream instrm(fName.c_str());
// std::string ofName =
// pathName + "/Code/GraphMol/Descriptors/test_data/GETAWAY.new.out";
// std::ofstream outstrm(ofName.c_str());
std::string line;
std::vector<std::vector<std::string> > data;
while (std::getline(instrm, line)) {
std::string phrase;
std::vector<std::string> row;
std::stringstream ss(line);
while (std::getline(ss, phrase, '\t')) {
row.push_back(phrase);
}
data.push_back(row);
}
int nDone = 0;
while (!reader.atEnd()) {
// if (nDone > 10) {
// break;
// }
RDKit::ROMol *m = reader.next();
TEST_ASSERT(m);
std::string nm;
m->getProp("_Name", nm);
std::vector<double> dgetaway;
RDKit::Descriptors::GETAWAY(*m, dgetaway);
std::vector<std::string> myrow = data[nDone];
std::string inm = myrow[0];
TEST_ASSERT(inm == nm);
// std::cout << "\n";
// int numAtoms = m->getNumAtoms();
// std::cout << "number of Atoms : " << numAtoms << "\n";
// outstrm << nm << "\t";
for (int i = 0; i < 273; i++) {
double ref = atof(myrow[i + 1].c_str());
if (fabs(ref) > 1) {
if (fabs((ref - dgetaway[i]) / ref) > 0.01) {
std::cerr << "value mismatch: pos" << i << " " << inm
<< " dragon: " << ref << " rdkit: " << dgetaway[i]
<< std::endl;
}
}
if (fabs(ref) <= 1) {
if (fabs(ref - dgetaway[i]) > 0.02) {
std::cerr << "value mismatch: pos" << i << " " << inm
<< " dragon: " << ref << " rdkit: " << dgetaway[i]
<< std::endl;
}
}
// if (i != 0) outstrm << "\t";
// outstrm << dgetaway[i];
TEST_ASSERT(fabs(ref - dgetaway[i]) < 0.05);
}
// outstrm << "\n";
delete m;
++nDone;
// if (nDone > 50) break;
}
BOOST_LOG(rdErrorLog) << "test on : " << nDone << " molecules done"
<< std::endl;
}
