bool MOL2Format::ReadMolecule(OBBase* pOb, OBConversion* pConv)
{
OBMol* pmol = pOb->CastAndClear<OBMol>();
if(pmol==NULL)
return false;
//Define some references so we can use the old parameter names
istream &ifs = *pConv->GetInStream();
OBMol &mol = *pmol;
//Old code follows...
bool foundAtomLine = false;
char buffer[BUFF_SIZE];
char *comment = NULL;
string str,str1;
vector<string> vstr;
int len;
mol.BeginModify();
for (;;)
{
if (!ifs.getline(buffer,BUFF_SIZE))
return(false);
if (EQn(buffer,"@<TRIPOS>MOLECULE",17))
break;
}
int lcount;
int natoms,nbonds;
for (lcount=0;;lcount++)
{
if (!ifs.getline(buffer,BUFF_SIZE))
return(false);
if (EQn(buffer,"@<TRIPOS>ATOM",13))
{
foundAtomLine = true;
break;
}
if (lcount == 0)
{
tokenize(vstr,buffer);
if (!vstr.empty())
mol.SetTitle(buffer);
}
else if (lcount == 1)
sscanf(buffer,"%d%d",&natoms,&nbonds);
else if (lcount == 4) //energy
{
tokenize(vstr,buffer);
if (!vstr.empty() && vstr.size() == 3)
if (vstr[0] == "Energy")
mol.SetEnergy(atof(vstr[2].c_str()));
}
else if (lcount == 5) //comment
{
if ( buffer[0] )
{
len = (int) strlen(buffer)+1;
// TODO allow better multi-line comments
// which don't allow ill-formed data to consume memory
// Thanks to Andrew Dalke for the pointer
if (comment != NULL)
delete [] comment;
comment = new char [len];
memcpy(comment,buffer,len);
}
}
}
if (!foundAtomLine)
{
mol.EndModify();
mol.Clear();
obErrorLog.ThrowError(__FUNCTION__, "Unable to read Mol2 format file. No atoms found.", obWarning);
return(false);
}
mol.ReserveAtoms(natoms);
int i;
vector3 v;
OBAtom atom;
bool hasPartialCharges=false;
double x,y,z,pcharge;
char temp_type[BUFF_SIZE], resname[BUFF_SIZE], atmid[BUFF_SIZE];
int elemno, resnum = -1;
ttab.SetFromType("SYB");
for (i = 0;i < natoms;i++)
{
if (!ifs.getline(buffer,BUFF_SIZE))
return(false);
sscanf(buffer," %*s %1024s %lf %lf %lf %1024s %d %1024s %lf",
atmid, &x,&y,&z, temp_type, &resnum, resname, &pcharge);
atom.SetVector(x, y, z);
// Handle "CL" and "BR" and other mis-typed atoms
str = temp_type;
if (strncmp(temp_type, "CL", 2) == 0) {
str = "Cl";
} else if (strncmp(temp_type,"BR",2) == 0) {
str = "Br";
} else if (strncmp(temp_type,"S.o2", 4) == 02) {
str = "S.O2";
} else if (strncmp(temp_type,"S.o", 3) == 0) {
str = "S.O";
} else if (strncmp(temp_type,"SI", 2) == 0) {
str = "Si";
// The following cases are entries which are not in openbabel/data/types.txt
// and should probably be added there
} else if (strncmp(temp_type,"S.1", 3) == 0) {
str = "S.2"; // no idea what the best type might be here
} else if (strncmp(temp_type,"P.", 2) == 0) {
str = "P.3";
} else if (strncasecmp(temp_type,"Ti.", 3) == 0) { // e.g. Ti.th
str = "Ti";
} else if (strncasecmp(temp_type,"Ru.", 3) == 0) { // e.g. Ru.oh
str = "Ru";
}
ttab.SetToType("ATN");
ttab.Translate(str1,str);
elemno = atoi(str1.c_str());
ttab.SetToType("IDX");
// We might have missed some SI or FE type things above, so here's
// another check
if( !elemno && isupper(temp_type[1]) )
{
temp_type[1] = (char)tolower(temp_type[1]);
str = temp_type;
ttab.Translate(str1,str);
elemno = atoi(str1.c_str());
}
// One last check if there isn't a period in the type,
// it's a malformed atom type, but it may be the element symbol
// GaussView does this (PR#1739905)
if ( !elemno ) {
obErrorLog.ThrowError(__FUNCTION__, "This Mol2 file is non-standard. Cannot interpret atom types correctly, instead attempting to interpret as elements instead.", obWarning);
string::size_type dotPos = str.find('.');
if (dotPos == string::npos) {
elemno = etab.GetAtomicNum(str.c_str());
}
}
atom.SetAtomicNum(elemno);
ttab.SetToType("INT");
ttab.Translate(str1,str);
atom.SetType(str1);
atom.SetPartialCharge(pcharge);
if (!mol.AddAtom(atom))
return(false);
if (!IsNearZero(pcharge))
hasPartialCharges = true;
// Add residue information if it exists
if (resnum != -1 && resnum != 0 &&
strlen(resname) != 0 && strncmp(resname,"<1>", 3) != 0)
{
OBResidue *res = (mol.NumResidues() > 0) ?
mol.GetResidue(mol.NumResidues()-1) : NULL;
if (res == NULL || res->GetName() != resname ||
static_cast<int>(res->GetNum()) != resnum)
{
vector<OBResidue*>::iterator ri;
for (res = mol.BeginResidue(ri) ; res ; res = mol.NextResidue(ri))
if (res->GetName() == resname &&
static_cast<int>(res->GetNum())
== resnum)
break;
if (res == NULL)
{
res = mol.NewResidue();
res->SetName(resname);
res->SetNum(resnum);
}
}
OBAtom *atomPtr = mol.GetAtom(mol.NumAtoms());
res->AddAtom(atomPtr);
res->SetAtomID(atomPtr, atmid);
} // end adding residue info
}
for (;;)
{
if (!ifs.getline(buffer,BUFF_SIZE))
return(false);
str = buffer;
if (!strncmp(buffer,"@<TRIPOS>BOND",13))
break;
}
int start,end,order;
for (i = 0; i < nbonds; i++)
{
if (!ifs.getline(buffer,BUFF_SIZE))
return(false);
sscanf(buffer,"%*d %d %d %1024s",&start,&end,temp_type);
str = temp_type;
order = 1;
if (str == "ar" || str == "AR" || str == "Ar")
order = 5;
else if (str == "AM" || str == "am" || str == "Am")
order = 1;
else
order = atoi(str.c_str());
mol.AddBond(start,end,order);
}
// update neighbour bonds information for each atom.
vector<OBAtom*>::iterator apos;
vector<OBBond*>::iterator bpos;
OBAtom* patom;
OBBond* pbond;
for (patom = mol.BeginAtom(apos); patom; patom = mol.NextAtom(apos))
{
patom->ClearBond();
for (pbond = mol.BeginBond(bpos); pbond; pbond = mol.NextBond(bpos))
{
if (patom == pbond->GetBeginAtom() || patom == pbond->GetEndAtom())
{
patom->AddBond(pbond);
}
}
}
// Suggestion by Liu Zhiguo 2008-01-26
// Mol2 files define atom types -- there is no need to re-perceive
mol.SetAtomTypesPerceived();
mol.EndModify();
//must add generic data after end modify - otherwise it will be blown away
if (comment)
{
OBCommentData *cd = new OBCommentData;
cd->SetData(comment);
cd->SetOrigin(fileformatInput);
mol.SetData(cd);
delete [] comment;
comment = NULL;
}
if (hasPartialCharges)
mol.SetPartialChargesPerceived();
// continue untill EOF or untill next molecule record
streampos pos;
for(;;)
{
pos = ifs.tellg();
if (!ifs.getline(buffer,BUFF_SIZE))
break;
if (EQn(buffer,"@<TRIPOS>MOLECULE",17))
break;
}
ifs.seekg(pos); // go back to the end of the molecule
return(true);
}
static bool parseAtomRecord(char *buffer, OBMol &mol,int /*chainNum*/)
/* ATOMFORMAT "(i5,1x,a4,a1,a3,1x,a1,i4,a1,3x,3f8.3,2f6.2,a2,a2)" */
{
string sbuf = &buffer[6];
if (sbuf.size() < 48)
return(false);
bool hetatm = (EQn(buffer,"HETATM",6)) ? true : false;
bool elementFound = false; // true if correct element found in col 77-78
/* serial number */
string serno = sbuf.substr(0,5);
/* atom name */
string atmid = sbuf.substr(6,4);
/* chain */
char chain = sbuf.substr(15,1)[0];
/* element */
string element = " ";
if (sbuf.size() > 71)
{
element = sbuf.substr(70,2);
if (isalpha(element[1]))
{
if (element[0] == ' ')
{
element.erase(0, 1);
elementFound = true;
}
else if (isalpha(element[0]))
{
elementFound = true;
}
}
}
if (!elementFound)
{
stringstream errorMsg;
errorMsg << "WARNING: Problems reading a PDB file\n"
<< " Problems reading a HETATM or ATOM record.\n"
<< " According to the PDB specification,\n"
<< " columns 77-78 should contain the element symbol of an atom.\n"
<< " but OpenBabel found '" << element << "' (atom " << mol.NumAtoms()+1 << ")";
obErrorLog.ThrowError(__FUNCTION__, errorMsg.str(), obWarning);
}
// charge - optional
string scharge;
if (sbuf.size() > 73)
{
scharge = sbuf.substr(72,2);
}
//trim spaces on the right and left sides
while (!atmid.empty() && atmid[0] == ' ')
atmid = atmid.erase(0, 1);
while (!atmid.empty() && atmid[atmid.size()-1] == ' ')
atmid = atmid.substr(0,atmid.size()-1);
/* residue name */
string resname = sbuf.substr(11,3);
if (resname == " ")
resname = "UNK";
else
{
while (!resname.empty() && resname[0] == ' ')
resname = resname.substr(1,resname.size()-1);
while (!resname.empty() && resname[resname.size()-1] == ' ')
resname = resname.substr(0,resname.size()-1);
}
string type;
if (!elementFound) {
// OK, we have to fall back to determining the element from the atom type
// This is unreliable, but there's no other choice
if (EQn(buffer,"ATOM",4)) {
type = atmid.substr(0,2);
if (isdigit(type[0])) {
// sometimes non-standard files have, e.g 11HH
if (!isdigit(type[1])) type = atmid.substr(1,1);
else type = atmid.substr(2,1);
} else if ((sbuf[6] == ' ' &&
strncasecmp(type.c_str(), "Zn", 2) != 0 &&
strncasecmp(type.c_str(), "Fe", 2) != 0) ||
isdigit(type[1])) //type[1] is digit in Platon
type = atmid.substr(0,1); // one-character element
if (resname.substr(0,2) == "AS" || resname[0] == 'N') {
if (atmid == "AD1")
type = "O";
if (atmid == "AD2")
type = "N";
}
if (resname.substr(0,3) == "HIS" || resname[0] == 'H') {
if (atmid == "AD1" || atmid == "AE2")
type = "N";
if (atmid == "AE1" || atmid == "AD2")
type = "C";
}
if (resname.substr(0,2) == "GL" || resname[0] == 'Q') {
if (atmid == "AE1")
type = "O";
if (atmid == "AE2")
type = "N";
}
// fix: #2002557
if (atmid[0] == 'H' &&
(atmid[1] == 'D' || atmid[1] == 'E' ||
atmid[1] == 'G' || atmid[1] == 'H')) // HD, HE, HG, HH, ..
type = "H";
} else { //must be hetatm record
if (isalpha(element[1]) && (isalpha(element[0]) || (element[0] == ' '))) {
if (isalpha(element[0]))
type = element.substr(0,2);
else
type = element.substr(1,1);
if (type.size() == 2)
type[1] = tolower(type[1]);
} else { // no element column to use
if (isalpha(atmid[0])) {
if (atmid.size() > 2 && (atmid[2] == '\0' || atmid[2] == ' '))
type = atmid.substr(0,2);
else if (atmid[0] == 'A') // alpha prefix
type = atmid.substr(1, atmid.size() - 1);
else
type = atmid.substr(0,1);
} else if (atmid[0] == ' ')
type = atmid.substr(1,1); // one char element
else
type = atmid.substr(1,2);
// Some cleanup steps
if (atmid == resname) {
type = atmid;
if (type.size() == 2)
type[1] = tolower(type[1]);
} else
if (resname == "ADR" || resname == "COA" || resname == "FAD" ||
resname == "GPG" || resname == "NAD" || resname == "NAL" ||
resname == "NDP" || resname == "ABA") {
if (type.size() > 1)
type = type.substr(0,1);
//type.erase(1,type.size()-1);
} else // other residues
if (isdigit(type[0])){
type = type.substr(1,1);
}
else
if (type.size() > 1 && isdigit(type[1]))
type = type.substr(0,1);
else
if (type.size() > 1 && isalpha(type[1])) {
if (type[0] == 'O' && type[1] == 'H')
type = type.substr(0,1); // no "Oh" element (e.g. 1MBN)
else if(isupper(type[1])) {
type[1] = tolower(type[1]);
}
}
}
} // HETATM records
} // no element column to use
OBAtom atom;
/* X, Y, Z */
string xstr = sbuf.substr(24,8);
string ystr = sbuf.substr(32,8);
string zstr = sbuf.substr(40,8);
vector3 v(atof(xstr.c_str()),atof(ystr.c_str()),atof(zstr.c_str()));
atom.SetVector(v);
atom.ForceImplH();
// useful for debugging unknown atom types (e.g., PR#1577238)
// cout << mol.NumAtoms() + 1 << " : '" << element << "'" << " " << etab.GetAtomicNum(element.c_str()) << endl;
if (elementFound)
atom.SetAtomicNum(etab.GetAtomicNum(element.c_str()));
else // use our old-style guess from athe atom type
atom.SetAtomicNum(etab.GetAtomicNum(type.c_str()));
if ( (! scharge.empty()) && " " != scharge )
{
if ( isdigit(scharge[0]) && ('+' == scharge[1] || '-' == scharge[1]) )
{
const char reorderCharge[3] = { scharge[1], scharge[0], '\0' };
const int charge = atoi(reorderCharge);
atom.SetFormalCharge(charge);
}
else
{
stringstream errorMsg;
errorMsg << "WARNING: Problems reading a PDB file\n"
<< " Problems reading a HETATM or ATOM record.\n"
<< " According to the PDB specification,\n"
<< " columns 79-80 should contain charge of the atom\n"
<< " but OpenBabel found '" << scharge << "' (atom " << mol.NumAtoms()+1 << ").";
obErrorLog.ThrowError(__FUNCTION__, errorMsg.str(), obWarning);
}
}
else {
atom.SetFormalCharge(0);
}
/* residue sequence number */
string resnum = sbuf.substr(16,4);
OBResidue *res = (mol.NumResidues() > 0) ? mol.GetResidue(mol.NumResidues()-1) : NULL;
if (res == NULL
|| res->GetName() != resname
|| res->GetNumString() != resnum
|| res->GetChain() != chain)
{
vector<OBResidue*>::iterator ri;
for (res = mol.BeginResidue(ri) ; res ; res = mol.NextResidue(ri))
if (res->GetName() == resname
&& res->GetNumString() == resnum
&& static_cast<int>(res->GetChain()) == chain)
break;
if (res == NULL) {
res = mol.NewResidue();
res->SetChain(chain);
res->SetName(resname);
res->SetNum(resnum);
}
}
if (!mol.AddAtom(atom))
return(false);
else {
OBAtom *atom = mol.GetAtom(mol.NumAtoms());
res->AddAtom(atom);
res->SetSerialNum(atom, atoi(serno.c_str()));
res->SetAtomID(atom, sbuf.substr(6,4));
res->SetHetAtom(atom, hetatm);
return(true);
}
} // end reading atom records
