void testIsomorphism1()
{
OBMol mol;
OBConversion conv;
conv.SetInFormat("smi");
conv.ReadString(&mol, "CC1CCC(C)CC1");
OBQuery *query = CompileMoleculeQuery(&mol);
OBIsomorphismMapper *mapper = OBIsomorphismMapper::GetInstance(query);
OBIsomorphismMapper::Mappings maps;
mapper->MapAll(&mol, maps);
OB_ASSERT( maps.size() == 4 );
delete query;
delete mapper;
query = CompileSmilesQuery("C1(C)CCC(C)CC1");
mapper = OBIsomorphismMapper::GetInstance(query);
OBIsomorphismMapper::Mapping map;
mapper->MapFirst(&mol, map);
OB_ASSERT( map.size() == 8 );
mapper->MapUnique(&mol, maps);
OB_ASSERT( maps.size() == 1 );
mapper->MapAll(&mol, maps);
OB_ASSERT( maps.size() == 4 );
delete query;
delete mapper;
}
void testIsomorphismMask()
{
// read file: 3 6-rings
//
// /\ /\ /\
// | | | |
// \/ \/ \/
//
OBMol mol;
OBConversion conv;
conv.SetInFormat("cml");
std::ifstream ifs(OBTestUtil::GetFilename("isomorphism1.cml").c_str());
OB_REQUIRE( ifs );
conv.Read(&mol, &ifs);
OBQuery *query = CompileSmilesQuery("C1CCCCC1");
OBIsomorphismMapper *mapper = OBIsomorphismMapper::GetInstance(query);
// no mask
OBIsomorphismMapper::Mappings maps;
mapper->MapUnique(&mol, maps);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 3 );
// mask first ring
OBBitVec mask;
for (int i = 0; i < 6; ++i)
mask.SetBitOn(i+1);
mapper->MapUnique(&mol, maps, mask);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 1 );
// mask second ring also
for (int i = 6; i < 10; ++i)
mask.SetBitOn(i+1);
mapper->MapUnique(&mol, maps, mask);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 2 );
// just mask last ring (atomIds 7-8, 10-13)
mask.Clear();
for (int i = 10; i < 14; ++i)
mask.SetBitOn(i+1);
mask.SetBitOn(7 + 1); mask.SetBitOn(8 + 1);
mapper->MapUnique(&mol, maps, mask);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 1 ); // Should be same result as masking just the first ring
delete query;
delete mapper;
}
void testAutomorphismMask() {
// read file: 3 6-rings
//
// /\ /\ /\
// | | | |
// \/ \/ \/
//
cout << "testAutomorphismMask" << endl;
OBMol mol;
OBConversion conv;
conv.SetInFormat("cml");
std::ifstream ifs(OBTestUtil::GetFilename("isomorphism1.cml").c_str());
OB_REQUIRE( ifs );
conv.Read(&mol, &ifs);
OBIsomorphismMapper::Mappings maps;
// First of all, how many automorphisms are there without any mask?
// This takes about 20 seconds, so you may want to comment this out while debugging
FindAutomorphisms(&mol, maps);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 4 );
// Now, let's remove the bridge (atomId 6) of the central ring.
//
// /\ /\ /\
// | | | |
// \/ \/
// both rings can be flipped around exocyclic bond, the whole molecule can be mirrored
// horizontally, this results in 2 x 2 x 2 = 8 automorphisms
OBBitVec mask;
mask.SetRangeOn(1, mol.NumAtoms());
mask.SetBitOff(6+1);
FindAutomorphisms(&mol, maps, mask);
cout << maps.size() << endl;
for (unsigned int i = 0; i < maps.size(); ++i) {
OBIsomorphismMapper::Mapping::const_iterator j;
for (j = maps[i].begin(); j != maps[i].end(); ++j)
cout << j->second << " ";
cout << endl;
}
OB_ASSERT( maps.size() == 8 );
// Verify that atom Id 6 does not occur anywhere in the mappings
OBIsomorphismMapper::Mappings::const_iterator a;
OBIsomorphismMapper::Mapping::const_iterator b;
for (a = maps.begin(); a != maps.end(); ++a)
for (b = a->begin(); b!= a->end(); ++b) {
OB_ASSERT( b->first != 6 );
OB_ASSERT( b->second != 6 );
}
}
void testIsomorphism4()
{
cout << "testIsomorphism4" << endl;
OBMol mol;
OBConversion conv;
conv.SetInFormat("smi");
conv.ReadString(&mol, "C12C(C2)C1");
OBQuery *query = CompileSmilesQuery("C1CC1");
OBIsomorphismMapper *mapper = OBIsomorphismMapper::GetInstance(query);
OBIsomorphismMapper::Mappings maps;
mapper->MapUnique(&mol, maps);
cout << maps.size() << endl;
OB_ASSERT( maps.size() == 2 );
delete query;
delete mapper;
}
bool OpNewS::Do(OBBase* pOb, const char* OptionText, OpMap* pmap, OBConversion* pConv)
{
OBMol* pmol = dynamic_cast<OBMol*>(pOb);
if(!pmol)
return false;
// The SMARTS and any other parameters are extracted on the first molecule
// and stored in the member variables. The parameter is cleared so that
// the original -s option in transform.cpp is inactive
//string txt(pmap->find(GetID())->second); // ID can be "s" or "v"
vector<OBQuery*>::iterator qiter;
if(OptionText && *OptionText)//(!pConv || pConv->IsFirstInput())
{
//Set up on first call
queries.clear();
query=NULL;
nPatternAtoms=0;
inv=false;
tokenize(vec, OptionText);
inv = GetID()[0]=='v';
if(vec[0][0]=='~')
{
inv = true;
vec[0].erase(0,1);
}
//Do not filter out any molecules if there is a parameter "showall";
//allows -s option to be used for highlighting substructures (--highlight also does this)
vector<string>::iterator it = std::remove(vec.begin(), vec.end(),"showall");
showAll = it != vec.end();
if(showAll)
vec.erase(it);
//Store the number of matches required, if as a number in the second parameter, else 0.
nmatches = 0;
comparechar = '\0';
if(vec.size()>1)
{
comparechar = vec[1][0];
if(comparechar=='>' || comparechar=='<')
vec[1].erase(0,1);
else
comparechar = '\0';
nmatches = atoi(vec[1].c_str());
if(nmatches) //remove this parameter to still allow coloring
vec.erase(vec.begin()+1);
}
//Interpret as a filename if possible
MakeQueriesFromMolInFile(queries, vec[0], &nPatternAtoms, strstr(OptionText,"noH"));
vec.erase(remove(vec.begin(),vec.end(),"noH"),vec.end());//to prevent "noH2" being seen as a color
if(queries.empty())
{
//SMARTS supplied
// Explicit H in SMARTS requires explicit H in the molecule.
// Calling AddHydrogens() on a copy of the molecule is done in parsmart.cpp
// only when SMARTS contains [H]. Doing more has complications with atom typing,
// so AddHydrogens here on the molecule (not a copy) when #1 detected.
addHydrogens = (vec[0].find("#1]")!=string::npos);
// If extra target mols have been supplied, make a composite SMARTS
// to test for any of the targets.
if(ExtraMols.size()>0)
{
for(unsigned i=0;i<ExtraMols.size();++i)
{
OBConversion extraConv;
extraConv.AddOption("h");
if(!extraConv.SetOutFormat("smi"))
return false;
// Add option which avoids implicit H being added to the SMARTS.
// The parameter must be present but can be anything.
extraConv.AddOption("h",OBConversion::OUTOPTIONS, "X");
xsmarts += ",$(" + extraConv.WriteString(ExtraMols[i], true) + ")";
}
}
string ysmarts = xsmarts.empty() ? vec[0] : "[$(" + vec[0] + ")" + xsmarts +"]";
xsmarts.clear();
if(!sp.Init(ysmarts))
{
string msg = ysmarts + " cannot be interpreted as either valid SMARTS "
"or the name of a file with an extension known to OpenBabel "
"that contains one or more pattern molecules.";
obErrorLog.ThrowError(__FUNCTION__, msg, obError, onceOnly);
delete pmol;
pmol = NULL;
pConv->SetOneObjectOnly(); //stop conversion
return false;
}
}
else
{
// Target is in a file. Add extra targets if any supplied
for(unsigned i=0;i<ExtraMols.size();++i)
queries.push_back(CompileMoleculeQuery(static_cast<OBMol*>(ExtraMols[i])));
ExtraMols.clear();
}
if(vec.size()>1 && vec[1]=="exact")
{
if(queries.empty())
{
//Convert SMARTS to SMILES to count number of atoms
OBConversion conv;
OBMol patmol;
if(!conv.SetInFormat("smi") || !conv.ReadString(&patmol, vec[0]))
{
obErrorLog.ThrowError(__FUNCTION__, "Cannot read the parameter of -s option, "
"which has to be valid SMILES when the exact option is used.", obError, onceOnly);
delete pmol;
if(pConv)
pConv->SetOneObjectOnly(); //stop conversion
return false;
}
nPatternAtoms = patmol.NumHvyAtoms();
}
}
else
nPatternAtoms = 0;
//disable old versions
if(pConv)
pConv->AddOption(GetID(), OBConversion::GENOPTIONS, "");
}
bool match = false;
//These are a vector of each mapping, each containing atom indxs.
vector<vector<int> > vecatomvec;
vector<vector<int> >* pMappedAtoms = NULL;
if(nPatternAtoms)
if(pmol->NumHvyAtoms() != nPatternAtoms)
return false;
unsigned int imol=0; //index of mol in pattern file
if(!queries.empty()) //filename supplied
{
//match is set true if any of the structures match - OR behaviour
for(qiter=queries.begin();qiter!=queries.end();++qiter, ++imol)
{
OBIsomorphismMapper* mapper = OBIsomorphismMapper::GetInstance(*qiter);
OBIsomorphismMapper::Mappings mappings;
mapper->MapUnique(pmol, mappings);
if( (match = !mappings.empty()) ) // extra parens to indicate truth value
{
OBIsomorphismMapper::Mappings::iterator ita;
OBIsomorphismMapper::Mapping::iterator itb;
for(ita=mappings.begin(); ita!=mappings.end();++ita)//each mapping
{
vector<int> atomvec;
for(itb=ita->begin(); itb!=ita->end();++itb)//each atom index
atomvec.push_back(itb->second+1);
vecatomvec.push_back(atomvec);
atomvec.clear();
}
pMappedAtoms = &vecatomvec;
break;
}
}
}
else //SMARTS supplied
{
if(addHydrogens)
pmol->AddHydrogens(false,false);
if( (match = sp.Match(*pmol)) ) // extra parens to indicate truth value
{
pMappedAtoms = &sp.GetMapList();
if(nmatches!=0)
{
int n = sp.GetUMapList().size();
if(comparechar=='>') match = (n > nmatches);
else if(comparechar=='<') match = (n < nmatches);
else match = (n == nmatches);
}
}
}
if((!showAll && (!match && !inv)) || (match && inv))
{
//delete a non-matching mol
delete pmol;
pmol = NULL;
return false;
}
if(match)
//Copy the idxes of the first match to a member variable so that it can be retrieved from outside
firstmatch.assign(pMappedAtoms->begin()->begin(), pMappedAtoms->begin()->end());
else
firstmatch.clear();
if(match && !inv && vec.size()>=2 && !vec[1].empty() && !nPatternAtoms)
{
vector<vector<int> >::iterator iter;
if (vec[1]=="extract" || (vec.size()>3 && vec[2]=="extract"))
{
//Delete all unmatched atoms. Use only the first match
ExtractSubstruct(pmol, *pMappedAtoms->begin());
return true;
}
// color the substructure if there is a second parameter which is not "exact" or "extract" or "noH"
// with multiple color parameters use the one corresponding to the query molecule, or the last
if(imol>vec.size()-2)
imol = vec.size()-2;
for(iter=pMappedAtoms->begin();iter!=pMappedAtoms->end();++iter)//each match
AddDataToSubstruct(pmol, *iter, "color", vec[imol+1]);
return true;
}
if(pConv && pConv->IsLast())
{
for(qiter=queries.begin();qiter!=queries.end();++qiter)
delete *qiter;
queries.clear();
}
return true;
}
bool OpNewS::Do(OBBase* pOb, const char* OptionText, OpMap* pmap, OBConversion* pConv)
{
OBMol* pmol = dynamic_cast<OBMol*>(pOb);
if(!pmol)
return false;
// The SMARTS and any other parameters are extracted on the first molecule
// and stored in the static variables vec, inv. The parameter is cleared so that:
// (a) the original -s option in transform.cpp is inactive, and
// (b) the parsing does not have to be done again for multi-molecule files
string txt(pmap->find(GetID())->second); // ID can be "s" or "v"
static vector<string> vec;
static bool inv;
static int nPatternAtoms; //non-zero for exact matches
static OBQuery* query;
static vector<OBQuery*> queries;
vector<OBQuery*>::iterator qiter;
if(!txt.empty())
{
//Set up on first call
tokenize(vec, txt);
inv = GetID()[0]=='v';
if(vec[0][0]=='~')
{
inv = true;
vec[0].erase(0,1);
}
//Interpret as a filename if possible
MakeQueriesFromMolInFile(queries, vec[0], &nPatternAtoms);
if(vec.size()>1 && vec[1]=="exact")
{
if(queries.empty())
{
//Convert SMARTS to SMILES to count number of atoms
OBConversion conv;
OBMol patmol;
if(!conv.SetInFormat("smi") || !conv.ReadString(&patmol, vec[0]))
{
obErrorLog.ThrowError(__FUNCTION__, "Cannot read the parameter of -s option, "
"which has to be valid SMILES when the exact option is used.", obError, onceOnly);
delete pmol;
pConv->SetOneObjectOnly(); //stop conversion
return false;
}
nPatternAtoms = patmol.NumHvyAtoms();
}
}
else
nPatternAtoms = 0;
//disable old versions
pConv->AddOption(GetID(), OBConversion::GENOPTIONS, "");
}
bool match;
//These are a vector of each mapping, each containing atom indxs.
vector<vector<int> > vecatomvec;
vector<vector<int> >* pMappedAtoms = NULL;
OBSmartsPattern sp;
if(nPatternAtoms)
if(pmol->NumHvyAtoms() != nPatternAtoms)
return false;
int imol=0; //index of mol in pattern file
if(!queries.empty()) //filename supplied
{
//match is set true if any of the structures match - OR behaviour
for(qiter=queries.begin();qiter!=queries.end();++qiter, ++imol)
{
OBIsomorphismMapper* mapper = OBIsomorphismMapper::GetInstance(*qiter);
OBIsomorphismMapper::Mappings mappings;
mapper->MapUnique(pmol, mappings);
if( (match = !mappings.empty()) ) // extra parens to indicate truth value
{
OBIsomorphismMapper::Mappings::iterator ita;
OBIsomorphismMapper::Mapping::iterator itb;
for(ita=mappings.begin(); ita!=mappings.end();++ita)//each mapping
{
vector<int> atomvec;
for(itb=ita->begin(); itb!=ita->end();++itb)//each atom index
atomvec.push_back(itb->second+1);
vecatomvec.push_back(atomvec);
atomvec.clear();
}
pMappedAtoms = &vecatomvec;
break;
}
}
}
else //SMARTS supplied
{
if(!sp.Init(vec[0]))
{
string msg = vec[0] + " cannot be interpreted as either valid SMARTS "
"or the name of a file with an extension known to OpenBabel "
"that contains one or more pattern molecules.";
obErrorLog.ThrowError(__FUNCTION__, msg, obError, onceOnly);
delete pmol;
pmol = NULL;
pConv->SetOneObjectOnly(); //stop conversion
return false;
}
if( (match = sp.Match(*pmol)) ) // extra parens to indicate truth value
pMappedAtoms = &sp.GetMapList();
}
if((!match && !inv) || (match && inv))
{
//delete a non-matching mol
delete pmol;
pmol = NULL;
return false;
}
if(!inv && vec.size()>=2 && !vec[1].empty() && !nPatternAtoms)
{
vector<vector<int> >::iterator iter;
if(vec[1]=="extract")
{
//Delete all unmatched atoms. Use only the first match
ExtractSubstruct(pmol, *pMappedAtoms->begin());
return true;
}
// color the substructure if there is a second parameter which is not "exact" or "extract"
// with multiple color parameters use the one corresponding to the query molecule, or the last
if(imol>vec.size()-2)
imol = vec.size()-2;
for(iter=pMappedAtoms->begin();iter!=pMappedAtoms->end();++iter)//each match
AddDataToSubstruct(pmol, *iter, "color", vec[imol+1]);
return true;
}
if(pConv && pConv->IsLast())
{
for(qiter=queries.begin();qiter!=queries.end();++qiter)
delete *qiter;
queries.clear();
}
return true;
}