| 1 |
/********************************************************************** |
| 2 |
Copyright (C) 2000 by OpenEye Scientific Software, Inc. |
| 3 |
Some portions Copyright (C) 2001-2006 by Geoffrey R. Hutchison |
| 4 |
Some portions Copyright (C) 2004 by Chris Morley |
| 5 |
Some portions Copyright (C) 2008-2009 by J. Daniel Gezelter |
| 6 |
|
| 7 |
This program is free software; you can redistribute it and/or modify |
| 8 |
it under the terms of the GNU General Public License as published by |
| 9 |
the Free Software Foundation version 2 of the License. |
| 10 |
|
| 11 |
This program is distributed in the hope that it will be useful, |
| 12 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 |
GNU General Public License for more details. |
| 15 |
***********************************************************************/ |
| 16 |
|
| 17 |
#include <openbabel/babelconfig.h> |
| 18 |
#include <openbabel/obmolecformat.h> |
| 19 |
#include <openbabel/obiter.h> |
| 20 |
#include <openbabel/mol.h> |
| 21 |
#include <openbabel/chains.h> |
| 22 |
#include <openbabel/data.h> |
| 23 |
#include <fstream> |
| 24 |
|
| 25 |
#include "utils/StringUtils.hpp" |
| 26 |
|
| 27 |
using namespace std; |
| 28 |
namespace OpenBabel |
| 29 |
{ |
| 30 |
|
| 31 |
class OpenMDFormat : public OBMoleculeFormat |
| 32 |
{ |
| 33 |
public: |
| 34 |
//Register this format type ID |
| 35 |
OpenMDFormat() |
| 36 |
{ |
| 37 |
OBConversion::RegisterFormat("md",this); |
| 38 |
} |
| 39 |
|
| 40 |
virtual const char* Description() //required |
| 41 |
{ |
| 42 |
return |
| 43 |
"OpenMD combined meta-data / cartesian coordinates format\n\ |
| 44 |
No comments yet\n"; |
| 45 |
}; |
| 46 |
|
| 47 |
virtual const char* SpecificationURL() |
| 48 |
{return "http://www.openmd.net";}; //optional |
| 49 |
|
| 50 |
virtual const char* GetMIMEType() |
| 51 |
{return "chemical/x-md"; }; |
| 52 |
|
| 53 |
virtual unsigned int Flags() |
| 54 |
{ |
| 55 |
return NOTREADABLE | WRITEONEONLY; |
| 56 |
} |
| 57 |
|
| 58 |
virtual bool WriteMolecule(OBBase* pOb, OBConversion* pConv); |
| 59 |
|
| 60 |
private: |
| 61 |
bool AreSameFragments(OBMol& mol, vector<int>& frag1, vector<int>& frag2); |
| 62 |
OBMol* createMolFromFragment(OBMol& mol, vector<int>& fragment); |
| 63 |
void WriteMDFile(vector<OBMol*> mols, vector<int> numMols, ostream& os, |
| 64 |
OBMol& mol, vector<int>& indices); |
| 65 |
void CalcBoundingBox(OBMol &mol, |
| 66 |
double &min_x, double &max_x, |
| 67 |
double &min_y, double &max_y, |
| 68 |
double &min_z, double &max_z); |
| 69 |
|
| 70 |
}; |
| 71 |
|
| 72 |
//Make an instance of the format class |
| 73 |
OpenMDFormat theOpenMDFormat; |
| 74 |
|
| 75 |
bool OpenMDFormat::WriteMolecule(OBBase* pOb, OBConversion* pConv) { |
| 76 |
OBMol* pmol = dynamic_cast<OBMol*>(pOb); |
| 77 |
if(pmol==NULL) |
| 78 |
return false; |
| 79 |
|
| 80 |
vector<vector<int> > fragmentLists; |
| 81 |
pmol->ContigFragList(fragmentLists); |
| 82 |
OBBitVec unused; |
| 83 |
vector<bool> used(fragmentLists.size(), 0); |
| 84 |
vector<vector<int> > molecules; |
| 85 |
vector<int> indices; |
| 86 |
for(int i =0; i < used.size(); ++i) { |
| 87 |
if (used[i]) |
| 88 |
continue; |
| 89 |
|
| 90 |
used[i] = true; |
| 91 |
vector<int> sameMolTypes; |
| 92 |
sameMolTypes.push_back(i); |
| 93 |
indices.insert(indices.end(), fragmentLists[i].begin(), |
| 94 |
fragmentLists[i].end()); |
| 95 |
for (int j = i + 1;j < used.size(); ++j) { |
| 96 |
if (used[j]) |
| 97 |
continue; |
| 98 |
|
| 99 |
if (AreSameFragments(*pmol, fragmentLists[i], fragmentLists[j])) { |
| 100 |
sameMolTypes.push_back(j); |
| 101 |
indices.insert(indices.end(), fragmentLists[j].begin(), |
| 102 |
fragmentLists[j].end()); |
| 103 |
used[j]=true; |
| 104 |
} |
| 105 |
} |
| 106 |
molecules.push_back(sameMolTypes); |
| 107 |
} |
| 108 |
|
| 109 |
vector<OBMol*> mdMols; |
| 110 |
vector<int> numMols; |
| 111 |
for(vector<vector<int> >::iterator i = molecules.begin(); |
| 112 |
i != molecules.end(); ++i) { |
| 113 |
|
| 114 |
mdMols.push_back(createMolFromFragment(*pmol, |
| 115 |
fragmentLists[i->front()])); |
| 116 |
numMols.push_back((*i).size()); |
| 117 |
} |
| 118 |
|
| 119 |
string OutputFileName = pConv->GetInFilename(); |
| 120 |
size_t pos = OutputFileName.rfind("."); |
| 121 |
if(pos!=string::npos) |
| 122 |
OutputFileName = OutputFileName.substr(0, pos) + ".md"; |
| 123 |
else |
| 124 |
OutputFileName += ".md"; |
| 125 |
|
| 126 |
ofstream ofs(OutputFileName.c_str()); |
| 127 |
if(!ofs) { |
| 128 |
cerr << "Cannot write to " << OutputFileName <<endl; |
| 129 |
return false; |
| 130 |
} |
| 131 |
|
| 132 |
|
| 133 |
|
| 134 |
WriteMDFile(mdMols, numMols, ofs, *pmol, indices); |
| 135 |
|
| 136 |
for(vector<OBMol*>::iterator i = mdMols.begin(); i != mdMols.end(); ++i) { |
| 137 |
delete *i; |
| 138 |
} |
| 139 |
|
| 140 |
return(true); |
| 141 |
} |
| 142 |
|
| 143 |
bool OpenMDFormat::AreSameFragments(OBMol& mol, vector<int>& frag1, |
| 144 |
vector<int>& frag2) { |
| 145 |
if (frag1.size() != frag2.size()) |
| 146 |
return false; |
| 147 |
|
| 148 |
// Exact graph matching is an NP complete problem. |
| 149 |
// This just matches all of the atom atomic numbers and may falsely |
| 150 |
// detect identical fragments which aren't really identical. |
| 151 |
// @todo using sparse matrix to store the connectivities |
| 152 |
|
| 153 |
for (unsigned int i =0 ; i < frag1.size(); ++i) { |
| 154 |
OBAtom* atom1 = mol.GetAtom(frag1[i]); |
| 155 |
OBAtom* atom2 = mol.GetAtom(frag2[i]); |
| 156 |
|
| 157 |
if (atom1->GetAtomicNum() != atom2->GetAtomicNum()) |
| 158 |
return false; |
| 159 |
|
| 160 |
} |
| 161 |
return true; |
| 162 |
} |
| 163 |
|
| 164 |
struct SameAngle { |
| 165 |
bool operator()(const triple<OBAtom*,OBAtom*,OBAtom*> t1, |
| 166 |
const triple<OBAtom*,OBAtom*,OBAtom*> t2) const { |
| 167 |
return (t1.second == t2.second) && ( (t1.first == t2.first && t1.third == t2.third) || (t1.first == t2.third && t1.third == t2.first)); |
| 168 |
} |
| 169 |
}; |
| 170 |
|
| 171 |
|
| 172 |
OBMol* OpenMDFormat::createMolFromFragment(OBMol& mol, |
| 173 |
vector<int>& fragment) { |
| 174 |
|
| 175 |
OBMol* newMol = new OBMol(); |
| 176 |
|
| 177 |
newMol->ReserveAtoms(fragment.size()); |
| 178 |
newMol->BeginModify(); |
| 179 |
for(vector<int>::iterator i = fragment.begin(); i != fragment.end(); ++i) { |
| 180 |
OBAtom* newAtom = newMol->NewAtom(); |
| 181 |
*newAtom = *mol.GetAtom(*i); |
| 182 |
} |
| 183 |
|
| 184 |
newMol->EndModify(); |
| 185 |
newMol->ConnectTheDots(); |
| 186 |
newMol->PerceiveBondOrders(); |
| 187 |
|
| 188 |
return newMol; |
| 189 |
} |
| 190 |
|
| 191 |
void OpenMDFormat::WriteMDFile(vector<OBMol*> mols, vector<int> numMols, |
| 192 |
ostream& os, OBMol& mol, |
| 193 |
vector<int>& indices) { |
| 194 |
|
| 195 |
std::string molPrefix("MolName"); |
| 196 |
std::string resName; |
| 197 |
unsigned int i; |
| 198 |
const int BUFFLEN = 1024; |
| 199 |
char buffer[BUFFLEN]; |
| 200 |
string str, str1, str2, str3; |
| 201 |
bool molIsWater = false; |
| 202 |
OBResidue *r; |
| 203 |
double min_x, max_x, min_y, max_y, min_z, max_z; /* Edges of bounding box */ |
| 204 |
|
| 205 |
os << "<OpenMD version=2>" << endl; |
| 206 |
os << " <MetaData>" << endl << endl; |
| 207 |
|
| 208 |
for(i = 0; i < mols.size(); ++i) { |
| 209 |
OBMol* pmol = mols[i]; |
| 210 |
map<OBAtom*, int> atomMap; |
| 211 |
|
| 212 |
molIsWater = false; |
| 213 |
FOR_RESIDUES_OF_MOL(residue, *pmol) { |
| 214 |
if (residue->GetName().compare("HOH") == 0) { |
| 215 |
molIsWater = true; |
| 216 |
} |
| 217 |
} |
| 218 |
|
| 219 |
if (molIsWater) { |
| 220 |
// water include files define all of the known water types |
| 221 |
os << "#include \"water.md\";\n"; |
| 222 |
pmol->SetTitle("HOH"); |
| 223 |
} else { |
| 224 |
|
| 225 |
os << "molecule {\n"; |
| 226 |
sprintf(buffer, "%u", i); |
| 227 |
os << " name = \"" << molPrefix << buffer << "\";\n"; |
| 228 |
|
| 229 |
int ai = 0; |
| 230 |
FOR_ATOMS_OF_MOL(atom, *pmol ) { |
| 231 |
str = atom->GetType(); |
| 232 |
r = atom->GetResidue(); |
| 233 |
|
| 234 |
if (r == NULL) |
| 235 |
resName = "NULL"; |
| 236 |
else |
| 237 |
resName = r->GetName(); |
| 238 |
|
| 239 |
if (resName.compare("NULL") ==0 || |
| 240 |
resName.compare("LIG") == 0 || |
| 241 |
resName.compare("UNK") == 0) { |
| 242 |
// Either couldn't find a residue at all or couldn't find a |
| 243 |
// reasonable residue name to use. We'll punt and use |
| 244 |
// OpenBabel's internal atom typing: |
| 245 |
ttab.SetFromType("INT"); |
| 246 |
ttab.SetToType("INT"); |
| 247 |
ttab.Translate(str1, str); |
| 248 |
} else { |
| 249 |
|
| 250 |
// If we know what residue we've got, the specific atom name can |
| 251 |
// be used to help specify partial charges. |
| 252 |
|
| 253 |
//resdat.SetResName(resName); |
| 254 |
|
| 255 |
// atom type from residue: |
| 256 |
str = r->GetAtomID(&*atom); |
| 257 |
|
| 258 |
// arginine has separate indices for chemically-identical |
| 259 |
// nitrogen atoms: |
| 260 |
if (resName.compare("ARG") == 0) { |
| 261 |
if (str.compare("NH1") == 0 || str.compare("NH2") == 0) { |
| 262 |
str = "NH"; |
| 263 |
} |
| 264 |
} |
| 265 |
if (resName.compare("VAL") == 0) { |
| 266 |
if (str.compare("CG1") == 0 || str.compare("CG2") == 0) { |
| 267 |
str = "CG"; |
| 268 |
} |
| 269 |
} |
| 270 |
if (resName.compare("LEU") == 0) { |
| 271 |
if (str.compare("CD1") == 0 || str.compare("CD2") == 0) { |
| 272 |
str = "CD"; |
| 273 |
} |
| 274 |
} |
| 275 |
if (resName.compare("ASP") == 0) { |
| 276 |
if (str.compare("OD1") == 0 || str.compare("OD2") == 0) { |
| 277 |
str = "OD"; |
| 278 |
} |
| 279 |
} |
| 280 |
if (resName.compare("GLU") == 0) { |
| 281 |
if (str.compare("OE1") == 0 || str.compare("OE2") == 0) { |
| 282 |
str = "OE"; |
| 283 |
} |
| 284 |
} |
| 285 |
if (resName.compare("TYR") == 0) { |
| 286 |
if (str.compare("CD1") == 0 || str.compare("CD2") == 0) { |
| 287 |
str = "CD"; |
| 288 |
} |
| 289 |
if (str.compare("CE1") == 0 || str.compare("CE2") == 0) { |
| 290 |
str = "CE"; |
| 291 |
} |
| 292 |
} |
| 293 |
|
| 294 |
|
| 295 |
if ((&*atom)->IsHydrogen()) { |
| 296 |
FOR_NBORS_OF_ATOM(nbr, *atom) { |
| 297 |
str2 = r->GetAtomID(&*nbr); |
| 298 |
size_t startpos = str2.find_first_not_of(" "); |
| 299 |
size_t endpos = str2.find_last_not_of(" "); |
| 300 |
if ((endpos - startpos) < 1) { |
| 301 |
// if the bonded atom type has only one character (i.e. N) |
| 302 |
// then the hydrogen will be labeled "HN" to show what |
| 303 |
// kind of proton it is: |
| 304 |
str3 = str2; |
| 305 |
} else { |
| 306 |
if (str2.compare("OH") == 0) { |
| 307 |
str3 = "O"; |
| 308 |
} else { |
| 309 |
// When the bonded atom type is more specific, we drop |
| 310 |
// the first character: i.e. H bonded to OG1 is HG1 type: |
| 311 |
str3 = str2.substr(startpos+1, endpos-startpos); |
| 312 |
} |
| 313 |
} |
| 314 |
str = "H" + str3; |
| 315 |
} |
| 316 |
// same problem with arginine NH atoms, but now for connected hydrogens |
| 317 |
if (resName.compare("ARG") == 0) { |
| 318 |
if (str.compare("HH1") == 0 || str.compare("HH2") == 0) { |
| 319 |
str = "HH"; |
| 320 |
} |
| 321 |
} |
| 322 |
if (resName.compare("VAL") == 0) { |
| 323 |
if (str.compare("HG1") == 0 || str.compare("HG2") == 0) { |
| 324 |
str = "HG"; |
| 325 |
} |
| 326 |
} |
| 327 |
if (resName.compare("LEU") == 0) { |
| 328 |
if (str.compare("HD1") == 0 || str.compare("HD2") == 0) { |
| 329 |
str = "HD"; |
| 330 |
} |
| 331 |
} |
| 332 |
if (resName.compare("TYR") == 0) { |
| 333 |
if (str.compare("HD1") == 0 || str.compare("HD2") == 0) { |
| 334 |
str = "HD"; |
| 335 |
} |
| 336 |
if (str.compare("HE1") == 0 || str.compare("HE2") == 0) { |
| 337 |
str = "HE"; |
| 338 |
} |
| 339 |
} |
| 340 |
|
| 341 |
} |
| 342 |
|
| 343 |
// atom type from residue table: |
| 344 |
//resdat.LookupType(str, str2, hyb); |
| 345 |
size_t startpos = str.find_first_not_of(" "); |
| 346 |
size_t endpos = str.find_last_not_of(" "); |
| 347 |
str = str.substr( startpos, endpos-startpos+1 ); |
| 348 |
str1 = resName + "-" + str; |
| 349 |
} |
| 350 |
os << " atom[" << ai << "] { "; |
| 351 |
os << "type = " << "\"" << str1 << "\"" << "; "; |
| 352 |
os << "position( " << (&*atom)->GetX() << ", " << (&*atom)->GetY() << ", " << (&*atom)->GetZ() << ");"; |
| 353 |
os << "}\n"; |
| 354 |
atomMap[&(*atom)] = ai++; |
| 355 |
} |
| 356 |
os << "\n"; |
| 357 |
|
| 358 |
//bond |
| 359 |
|
| 360 |
int b1, b2; |
| 361 |
FOR_BONDS_OF_MOL(bond, *pmol ) { |
| 362 |
b1 = atomMap[bond->GetBeginAtom()]; |
| 363 |
b2 = atomMap[bond->GetEndAtom()]; |
| 364 |
|
| 365 |
os << " bond { "; |
| 366 |
|
| 367 |
if (b1 < b2) |
| 368 |
os << "members(" << b1 << ", " << b2 << "); "; |
| 369 |
else |
| 370 |
os << "members(" << b2 << ", " << b1 << "); "; |
| 371 |
|
| 372 |
os << "}" << endl; |
| 373 |
} |
| 374 |
|
| 375 |
os << endl; |
| 376 |
|
| 377 |
os << "}" << endl; |
| 378 |
os << endl; |
| 379 |
} |
| 380 |
} |
| 381 |
|
| 382 |
os << endl; |
| 383 |
|
| 384 |
for(i=0; i < mols.size(); ++i) { |
| 385 |
OBMol* pmol = mols[i]; |
| 386 |
os << "component{" << endl; |
| 387 |
if (std::string(pmol->GetTitle()).compare("HOH") == 0) { |
| 388 |
os << " type = " << "\"HOH\"" << "; // change to appropriate water model" << endl; |
| 389 |
} else { |
| 390 |
sprintf(buffer, "%u", i); |
| 391 |
os << " type = " << molPrefix << buffer << ";" << endl; |
| 392 |
} |
| 393 |
os << " nMol = " << numMols[i]<< ";" << endl; |
| 394 |
os << "}" << endl; |
| 395 |
} |
| 396 |
|
| 397 |
os << " </MetaData>" << endl; |
| 398 |
os << " <Snapshot>" << endl; |
| 399 |
os << " <FrameData>" << endl; |
| 400 |
|
| 401 |
sprintf(buffer, " Time: %.10g", 0.0); |
| 402 |
|
| 403 |
os << buffer << endl; |
| 404 |
|
| 405 |
CalcBoundingBox(mol, min_x, max_x, min_y, max_y, min_z, max_z); |
| 406 |
|
| 407 |
// still to do: should compute a bounding box here |
| 408 |
sprintf(buffer, " Hmat: {{ %.10g, %.10g, %.10g }, { %.10g, %.10g, %.10g }, { %.10g, %.10g, %.10g }}", |
| 409 |
max_x - min_x, 0.0, 0.0, 0.0, max_y - min_y, 0.0, 0.0, 0.0, max_z - min_z); |
| 410 |
|
| 411 |
os << buffer << endl; |
| 412 |
os << " </FrameData>" << endl; |
| 413 |
os << " <StuntDoubles>" << endl; |
| 414 |
|
| 415 |
OBAtom *atom; |
| 416 |
|
| 417 |
for(vector<int>::iterator i = indices.begin();i != indices.end(); ++i) { |
| 418 |
|
| 419 |
atom = mol.GetAtom(*i); |
| 420 |
sprintf(buffer, "%10d %7s %18.10g %18.10g %18.10g %13e %13e %13e", *i - 1, |
| 421 |
"pv", atom->GetX(), atom->GetY(), atom->GetZ(), 0.0, 0.0, 0.0); |
| 422 |
os << buffer << endl; |
| 423 |
} |
| 424 |
os << " </StuntDoubles>" << endl; |
| 425 |
os << " </Snapshot>" << endl; |
| 426 |
os << "</OpenMD>" << endl; |
| 427 |
} |
| 428 |
|
| 429 |
void OpenMDFormat::CalcBoundingBox(OBMol &mol, |
| 430 |
double &min_x, double &max_x, |
| 431 |
double &min_y, double &max_y, |
| 432 |
double &min_z, double &max_z |
| 433 |
) |
| 434 |
{ |
| 435 |
/* ---- Init bounding-box variables ---- */ |
| 436 |
min_x = (double) 0.0; |
| 437 |
max_x = (double) 0.0; |
| 438 |
min_y = (double) 0.0; |
| 439 |
max_y = (double) 0.0; |
| 440 |
min_z = (double) 0.0; |
| 441 |
max_z = (double) 0.0; |
| 442 |
|
| 443 |
/* ---- Check all atoms ---- */ |
| 444 |
for(unsigned int i = 1; i <= mol.NumAtoms(); ++i) |
| 445 |
{ |
| 446 |
|
| 447 |
/* ---- Get a pointer to ith atom ---- */ |
| 448 |
OBAtom *atom = mol.GetAtom(i); |
| 449 |
|
| 450 |
/* ---- Check for minimal/maximal x-position ---- */ |
| 451 |
if (atom -> GetX() < min_x) |
| 452 |
min_x = atom -> GetX(); |
| 453 |
if (atom -> GetX() > max_x) |
| 454 |
max_x = atom -> GetX(); |
| 455 |
|
| 456 |
/* ---- Check for minimal/maximal y-position ---- */ |
| 457 |
if (atom -> GetY() < min_y) |
| 458 |
min_y = atom -> GetY(); |
| 459 |
if (atom -> GetY() > max_y) |
| 460 |
max_y = atom -> GetY(); |
| 461 |
|
| 462 |
/* ---- Check for minimal/maximal z-position ---- */ |
| 463 |
if (atom -> GetZ() < min_z) |
| 464 |
min_z = atom -> GetZ(); |
| 465 |
if (atom -> GetZ() > max_z) |
| 466 |
max_z = atom -> GetZ(); |
| 467 |
|
| 468 |
} |
| 469 |
} |
| 470 |
} //namespace OpenBabel |
| 471 |
|
