bool FastSearchFormat::ObtainTarget(OBConversion* pConv, OBMol& patternMol, const string& indexname)
{
//Obtains an OBMol
// either from the SMARTS string in the -s option
// or by converting the file in the -S option
//or, if neither option is provided, displays information on the index file.
stringstream smiles(stringstream::out);
ifstream patternstream;
OBConversion PatternConv(&patternstream,&smiles);
const char* p = pConv->IsOption("s",OBConversion::GENOPTIONS);
string txt;
if(p)
{
// Use the -s option
txt=p;
stringstream smarts(txt, stringstream::in);
OBConversion Convsm(&smarts);
if(!Convsm.SetInFormat("smi")) return false;
Convsm.Read(&patternMol);
//erase -s option in GeneralOptions since it will be rewritten
pConv->RemoveOption("s",OBConversion::GENOPTIONS);
if(patternMol.Empty())
{
obErrorLog.ThrowError(__FUNCTION__,
"Could not make a molecule from " + smarts.str()
+ "\nThis needs to be valid SMILES when using fastsearch."
"You can use the more versatile SMARTS in a normal substructure search." , obError);
return false;
}
}
else
{
// or Make OBMol from file in -S option or -aS option
p = pConv->IsOption("S",OBConversion::GENOPTIONS);
if(!p)
p = pConv->IsOption("S",OBConversion::INOPTIONS);//for GUI mainly
}
if(!p)
{
//neither -s or -S options provided. Output info rather than doing search
const FptIndexHeader& header = fs.GetIndexHeader();
string id(header.fpid);
if(id.empty())
id = "default";
clog << indexname << " is an index of\n " << header.datafilename
<< ".\n It contains " << header.nEntries
<< " molecules. The fingerprint type is " << id << " with "
<< OBFingerprint::Getbitsperint() * header.words << " bits.\n"
<< "Typical usage for a substructure search:\n"
<< "babel indexfile.fs -osmi -sSMILES" << endl;
return false;
}
if(p && patternMol.Empty())
{
txt=p;
string::size_type pos = txt.find_last_of('.');
if(pos==string::npos)
{
obErrorLog.ThrowError(__FUNCTION__, "Filename of pattern molecule in -S option must have an extension", obError);
return false;
}
patternstream.open(txt.c_str());
if(!patternstream)
{
stringstream errorMsg;
errorMsg << "Cannot open " << txt << endl;
obErrorLog.ThrowError(__FUNCTION__, errorMsg.str(), obError);
return false;
}
PatternConv.SetOneObjectOnly();
if(PatternConv.SetInFormat(txt.substr(pos+1).c_str()))
PatternConv.Read(&patternMol);
}
if(patternMol.Empty())
{
obErrorLog.ThrowError(__FUNCTION__, "Cannot derive a molecule from the -s or -S options", obWarning);
return false;
}
patternMol.ConvertDativeBonds();//use standard form for dative bonds
//Convert to SMILES and generate a -s option for use in the final filtering
if(!PatternConv.SetOutFormat("smi"))
return false;
PatternConv.Write(&patternMol);
//remove name to leave smiles string
string smilesstr(smiles.str());
string::size_type pos = smilesstr.find_first_of(" \t\r\n");
if(pos!=string::npos)
smilesstr = smilesstr.substr(0,pos);
pConv->AddOption("s", OBConversion::GENOPTIONS, smilesstr.c_str());
return true;
}
bool fragmentMol(const ROMol& mol,
std::vector<std::pair<ROMOL_SPTR, ROMOL_SPTR> >& res,
unsigned int minCuts,
unsigned int maxCuts,
unsigned int maxCutBonds,
const std::string& pattern) {
#ifdef _DEBUG
for (size_t i = 0; i < mol.getNumAtoms(); i++) {
std::string symbol = mol.getAtomWithIdx(i)->getSymbol();
int label = 0;
mol.getAtomWithIdx(i)->getPropIfPresent(common_properties::molAtomMapNumber,
label);
char a1[32];
if (0 == label)
sprintf(a1, "\'%s\'", symbol.c_str(), label);
else
sprintf(a1, "\'%s:%u\'", symbol.c_str(), label);
std::cout << "Atom " << i << ": " << a1; //<<" Bonds:";
std::cout << "\n";
}
#endif
res.clear();
std::auto_ptr<const ROMol> smarts((const ROMol*)SmartsToMol(pattern));
std::vector<MatchVectType>
matching_atoms; // one bond per match ! with default pattern
unsigned int total = SubstructMatch(mol, *smarts, matching_atoms);
#ifdef _DEBUG
std::cout << "total substructs =" << total
<< "\nmatching bonds (atom1, atom2):\n";
#endif
if (0 == total) // Not found. Return empty set of molecules
return false;
#ifdef _DEBUG
for (size_t i = 0; i < matching_atoms.size(); i++) {
std::string symbol =
mol.getAtomWithIdx(matching_atoms[i][0].second)->getSymbol();
int label = 0;
mol.getAtomWithIdx(matching_atoms[i][0].second)
->getPropIfPresent(common_properties::molAtomMapNumber, label);
char a1[32];
if (0 == label)
sprintf(a1, "\'%s\'", symbol.c_str(), label);
else
sprintf(a1, "\'%s:%u\'", symbol.c_str(), label);
symbol = mol.getAtomWithIdx(matching_atoms[i][1].second)->getSymbol();
label = 0;
mol.getAtomWithIdx(matching_atoms[i][1].second)
->getPropIfPresent(common_properties::molAtomMapNumber, label);
char a2[32];
if (0 == label)
sprintf(a2, "\'%s\'", symbol.c_str(), label);
else
sprintf(a2, "\'%s:%u\'", symbol.c_str(), label);
std::cout << i << ": (" << matching_atoms[i][0].second << a1 << ","
<< matching_atoms[i][1].second << a2 << ") \n";
}
#endif
std::vector<BondVector_t> matching_bonds; // List of matched query's bonds
convertMatchingToBondVect(matching_bonds, matching_atoms, mol);
if (matching_bonds.size() > maxCutBonds) return false;
#ifdef _DEBUG
std::cout << "total matching_bonds = " << matching_bonds.size() << "\n";
#endif
// loop to generate every cut in the molecule
BondVector_t bonds_selected;
processCuts(0, minCuts, maxCuts, bonds_selected, matching_bonds, mol, res);
return true;
}