| 1 |
/********************************************************************** |
| 2 |
atom2md.cpp - OpenBabel-based conversion program to OOPSE MD file, |
| 3 |
command-line handling. |
| 4 |
|
| 5 |
Copyright (C) 1998-2001 by OpenEye Scientific Software, Inc. |
| 6 |
Some portions Copyright (C) 2001-2006 by Geoffrey R. Hutchison |
| 7 |
Some portions Copyright (C) 2004-2006 by Chris Morley |
| 8 |
Some portions Copyright (C) 2008 by J. Daniel Gezelter |
| 9 |
|
| 10 |
This file is part of both the OOPSE and Open Babel projects. |
| 11 |
For more information, see <http://oopse.org> and <http://openbabel.sourceforge.net/> |
| 12 |
|
| 13 |
This program is free software; you can redistribute it and/or modify |
| 14 |
it under the terms of the GNU General Public License as published by |
| 15 |
the Free Software Foundation version 2 of the License. |
| 16 |
|
| 17 |
This program is distributed in the hope that it will be useful, |
| 18 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 19 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 20 |
GNU General Public License for more details. |
| 21 |
***********************************************************************/ |
| 22 |
|
| 23 |
#include "config.h" |
| 24 |
|
| 25 |
// used to set import/export for Cygwin DLLs |
| 26 |
#ifdef WIN32 |
| 27 |
#define USING_OBDLL |
| 28 |
#endif |
| 29 |
|
| 30 |
#include <openbabel/babelconfig.h> |
| 31 |
|
| 32 |
#include <iostream> |
| 33 |
#include <fstream> |
| 34 |
#include <sstream> |
| 35 |
|
| 36 |
#include <string> |
| 37 |
#include <map> |
| 38 |
#if HAVE_CONIO_H |
| 39 |
#include <conio.h> |
| 40 |
#endif |
| 41 |
#include <cstdlib> |
| 42 |
|
| 43 |
#if !HAVE_STRNCASECMP |
| 44 |
extern "C" int strncasecmp(const char *s1, const char *s2, size_t n); |
| 45 |
#endif |
| 46 |
|
| 47 |
#include <openbabel/obconversion.h> |
| 48 |
|
| 49 |
using namespace std; |
| 50 |
using namespace OpenBabel; |
| 51 |
|
| 52 |
void DoOption(const char* p, OBConversion& Conv, |
| 53 |
OBConversion::Option_type typ, int& arg, int argc, |
| 54 |
char *argv[]); |
| 55 |
void usage(); |
| 56 |
void help(); |
| 57 |
|
| 58 |
// There isn't a great way to do this -- we need to save argv[0] for usage() |
| 59 |
static char *program_name; |
| 60 |
|
| 61 |
int main(int argc,char *argv[]) |
| 62 |
{ |
| 63 |
OBConversion Conv(&cin, &cout); //default input and output are console |
| 64 |
|
| 65 |
OBFormat* pInFormat = NULL; |
| 66 |
OBFormat* pOutFormat = NULL; |
| 67 |
vector<string> FileList, OutputFileList; |
| 68 |
string OutputFileName; |
| 69 |
|
| 70 |
// Parse commandline |
| 71 |
bool gotInType = false, gotOutType = false; |
| 72 |
bool SplitOrBatch=false; |
| 73 |
|
| 74 |
char *oext; |
| 75 |
char *iext; |
| 76 |
string inputExt; |
| 77 |
string outputExt; |
| 78 |
|
| 79 |
//Save name of program without its path (and .exe) |
| 80 |
string pn(argv[0]); |
| 81 |
string::size_type pos; |
| 82 |
#ifdef _WIN32 |
| 83 |
pos = pn.find(".exe"); |
| 84 |
if(pos!=string::npos) |
| 85 |
argv[0][pos]='\0'; |
| 86 |
#endif |
| 87 |
pos = pn.find_last_of("/\\"); |
| 88 |
if(pos==string::npos) |
| 89 |
program_name=argv[0]; |
| 90 |
else |
| 91 |
program_name=argv[0]+pos+1; |
| 92 |
|
| 93 |
const char* p; |
| 94 |
int arg; |
| 95 |
for (arg = 1; arg < argc; ++arg) |
| 96 |
{ |
| 97 |
if (argv[arg]) |
| 98 |
{ |
| 99 |
if (argv[arg][0] == '-') |
| 100 |
{ |
| 101 |
switch (argv[arg][1]) |
| 102 |
{ |
| 103 |
|
| 104 |
case 'V': |
| 105 |
{ |
| 106 |
cout << program_name << ": part of OOPSE " << |
| 107 |
OOPSE_VERSION_MAJOR << "." << OOPSE_VERSION_MINOR << |
| 108 |
"." << OOPSE_VERSION_TINY << |
| 109 |
" and Open Babel " << BABEL_VERSION << " -- " |
| 110 |
<< __DATE__ << " -- " << __TIME__ << endl; |
| 111 |
exit(0); |
| 112 |
} |
| 113 |
|
| 114 |
case 'i': |
| 115 |
gotInType = true; |
| 116 |
iext = argv[arg] + 2; |
| 117 |
if(!*iext) |
| 118 |
iext = argv[++arg]; // space left after -i: use next |
| 119 |
// argument |
| 120 |
|
| 121 |
if (strncasecmp(iext, "MIME", 4) == 0) |
| 122 |
{ |
| 123 |
// get the MIME type from the next argument |
| 124 |
iext = argv[++arg]; |
| 125 |
pInFormat = Conv.FormatFromMIME(iext); |
| 126 |
} |
| 127 |
else |
| 128 |
{ |
| 129 |
// The ID provided by the OBFormat class is used as the |
| 130 |
// identifying file extension |
| 131 |
pInFormat = Conv.FindFormat(iext); |
| 132 |
} |
| 133 |
if(pInFormat==NULL) |
| 134 |
{ |
| 135 |
cerr << program_name << ": cannot read input format!" |
| 136 |
<< endl; |
| 137 |
usage(); |
| 138 |
} |
| 139 |
break; |
| 140 |
|
| 141 |
case 'o': |
| 142 |
gotOutType = true; |
| 143 |
oext = argv[arg] + 2; |
| 144 |
if(!*oext) |
| 145 |
oext = argv[++arg]; // space left after -i: use next |
| 146 |
// argument |
| 147 |
|
| 148 |
if (strncasecmp(oext, "MIME", 4) == 0) |
| 149 |
{ |
| 150 |
// get the MIME type from the next argument |
| 151 |
oext = argv[++arg]; |
| 152 |
pOutFormat = Conv.FormatFromMIME(oext); |
| 153 |
} |
| 154 |
else |
| 155 |
pOutFormat = Conv.FindFormat(oext); |
| 156 |
|
| 157 |
if(pOutFormat==NULL) |
| 158 |
{ |
| 159 |
cerr << program_name << ": cannot write output format!" |
| 160 |
<< endl; |
| 161 |
usage(); |
| 162 |
} |
| 163 |
break; |
| 164 |
|
| 165 |
case 'F': |
| 166 |
if(!Conv.SetOutFormat("fpt")) |
| 167 |
cout << "FingerprintFormat needs to be loaded" << endl; |
| 168 |
else |
| 169 |
{ |
| 170 |
Conv.AddOption("F",OBConversion::OUTOPTIONS); |
| 171 |
Conv.Write(NULL); |
| 172 |
} |
| 173 |
return 0; |
| 174 |
|
| 175 |
case '?': |
| 176 |
case 'H': |
| 177 |
if(isalnum(argv[arg][2])) |
| 178 |
{ |
| 179 |
if(strncasecmp(argv[arg]+2,"all",3)) |
| 180 |
{ |
| 181 |
OBFormat* pFormat = Conv.FindFormat(argv[arg]+2); |
| 182 |
if(pFormat) |
| 183 |
{ |
| 184 |
cout << argv[arg]+2 << " " << pFormat->Description() << endl; |
| 185 |
if(strlen(pFormat->SpecificationURL())) |
| 186 |
cout << "Specification at: " << pFormat->SpecificationURL() << endl; |
| 187 |
} |
| 188 |
else |
| 189 |
cout << "Format type: " << argv[arg]+2 << " was not recognized" <<endl; |
| 190 |
} |
| 191 |
else |
| 192 |
{ |
| 193 |
Formatpos pos; |
| 194 |
OBFormat* pFormat; |
| 195 |
const char* str=NULL; |
| 196 |
while(OBConversion::GetNextFormat(pos,str,pFormat)) |
| 197 |
{ |
| 198 |
if((pFormat->Flags() & NOTWRITABLE) && (pFormat->Flags() & NOTREADABLE)) |
| 199 |
continue; |
| 200 |
cout << str << endl; |
| 201 |
const char* p = strchr(pFormat->Description(),'\n'); |
| 202 |
cout << p+1; //second line of description |
| 203 |
if(strlen(pFormat->SpecificationURL())) |
| 204 |
cout << "Specification at: " << pFormat->SpecificationURL(); |
| 205 |
cout << endl << endl; |
| 206 |
} |
| 207 |
} |
| 208 |
} |
| 209 |
else |
| 210 |
help(); |
| 211 |
exit(0); |
| 212 |
|
| 213 |
case '-': //long option --name text |
| 214 |
{ |
| 215 |
//Do nothing if name is empty |
| 216 |
//Option's text is the next arg provided it doesn't start with - |
| 217 |
char* nam = argv[arg]+2; |
| 218 |
if(*nam != '\0') |
| 219 |
{ |
| 220 |
string txt; |
| 221 |
int i; |
| 222 |
for(i=0; i<Conv.GetOptionParams(nam, OBConversion::GENOPTIONS) |
| 223 |
&& arg<argc-1 && argv[arg+1];++i) //removed && *argv[arg+1]!='-' |
| 224 |
{ |
| 225 |
if(!txt.empty()) txt+=' '; |
| 226 |
txt += argv[++arg]; |
| 227 |
} |
| 228 |
if(*nam=='-') |
| 229 |
{ |
| 230 |
// Is a API directive, e.g.---errorlevel |
| 231 |
//Send to the pseudoformat "obapi" (without any leading -) |
| 232 |
OBConversion apiConv; |
| 233 |
OBFormat* pAPI= OBConversion::FindFormat("obapi"); |
| 234 |
if(pAPI) |
| 235 |
{ |
| 236 |
apiConv.SetOutFormat(pAPI); |
| 237 |
apiConv.AddOption(nam+1, OBConversion::GENOPTIONS, txt.c_str()); |
| 238 |
apiConv.Write(NULL, &std::cout); |
| 239 |
} |
| 240 |
} |
| 241 |
else |
| 242 |
// Is a long option name, e.g --addtotitle |
| 243 |
Conv.AddOption(nam,OBConversion::GENOPTIONS,txt.c_str()); |
| 244 |
} |
| 245 |
} |
| 246 |
break; |
| 247 |
|
| 248 |
case 'm': //multiple output files |
| 249 |
SplitOrBatch=true; |
| 250 |
break; |
| 251 |
|
| 252 |
case 'a': //single character input option |
| 253 |
p = argv[arg]+2; |
| 254 |
DoOption(p,Conv,OBConversion::INOPTIONS,arg,argc,argv); |
| 255 |
break; |
| 256 |
|
| 257 |
case 'x': //single character output option |
| 258 |
p = argv[arg]+2; |
| 259 |
DoOption(p,Conv,OBConversion::OUTOPTIONS,arg,argc,argv); |
| 260 |
break; |
| 261 |
|
| 262 |
default: //single character general option |
| 263 |
p = argv[arg]+1; |
| 264 |
DoOption(p,Conv,OBConversion::GENOPTIONS,arg,argc,argv); |
| 265 |
break; |
| 266 |
} |
| 267 |
} |
| 268 |
else |
| 269 |
{ |
| 270 |
//filenames |
| 271 |
if(!gotOutType) |
| 272 |
FileList.push_back(argv[arg]); |
| 273 |
else |
| 274 |
OutputFileName = argv[arg]; |
| 275 |
} |
| 276 |
} |
| 277 |
} |
| 278 |
|
| 279 |
// user didn't specify input and output format in commandline option |
| 280 |
// try to parse it from program name (pdb2mdin means input format is pdb, |
| 281 |
// output format is mdin) |
| 282 |
|
| 283 |
string formatName(program_name); |
| 284 |
pos = formatName.find_first_of("2"); |
| 285 |
if(pos!=string::npos) { |
| 286 |
if (!gotInType) |
| 287 |
{ |
| 288 |
string tmpExt = formatName.substr(0, pos); |
| 289 |
pInFormat = Conv.FindFormat(tmpExt.c_str()); |
| 290 |
if(pInFormat==NULL) |
| 291 |
{ |
| 292 |
cerr << program_name << ": cannot read input format!" << endl; |
| 293 |
usage(); |
| 294 |
} else |
| 295 |
{ |
| 296 |
gotInType = true; |
| 297 |
inputExt = tmpExt; |
| 298 |
} |
| 299 |
} |
| 300 |
|
| 301 |
if (!gotOutType) |
| 302 |
{ |
| 303 |
string tmpExt = formatName.substr(pos+1, string::npos); |
| 304 |
pOutFormat = Conv.FindFormat(tmpExt.c_str()); |
| 305 |
if(pOutFormat==NULL) |
| 306 |
{ |
| 307 |
cerr << program_name << ": cannot write output format!" << endl; |
| 308 |
usage(); |
| 309 |
}else { |
| 310 |
gotOutType = true; |
| 311 |
outputExt = tmpExt; |
| 312 |
} |
| 313 |
} |
| 314 |
} |
| 315 |
|
| 316 |
if (!gotInType) |
| 317 |
{ |
| 318 |
if(FileList.empty()) |
| 319 |
{ |
| 320 |
cerr << "No input file or format spec!" <<endl; |
| 321 |
usage(); |
| 322 |
} |
| 323 |
} |
| 324 |
|
| 325 |
if (!gotOutType) |
| 326 |
{ |
| 327 |
pOutFormat = Conv.FormatFromExt(OutputFileName.c_str()); |
| 328 |
if(pOutFormat==NULL) |
| 329 |
{ |
| 330 |
cerr << program_name << ": cannot write output format!" << endl; |
| 331 |
usage(); |
| 332 |
} |
| 333 |
} |
| 334 |
|
| 335 |
Conv.SetInAndOutFormats(pInFormat, pOutFormat); |
| 336 |
|
| 337 |
if(SplitOrBatch) |
| 338 |
{ |
| 339 |
//Put * into output file name before extension (or ext.gz) |
| 340 |
if(OutputFileName.empty()) |
| 341 |
{ |
| 342 |
OutputFileName = "*."; |
| 343 |
OutputFileName += oext; |
| 344 |
} |
| 345 |
else |
| 346 |
{ |
| 347 |
string::size_type pos = OutputFileName.rfind(".gz"); |
| 348 |
if(pos==string::npos) |
| 349 |
pos = OutputFileName.rfind('.'); |
| 350 |
else |
| 351 |
pos = OutputFileName.rfind('.',pos-1); |
| 352 |
if(pos==string::npos) |
| 353 |
OutputFileName += '*'; |
| 354 |
else |
| 355 |
OutputFileName.insert(pos,"*"); |
| 356 |
} |
| 357 |
} |
| 358 |
|
| 359 |
int count = Conv.FullConvert(FileList, OutputFileName, OutputFileList); |
| 360 |
|
| 361 |
// send info message to clog -- don't mess up cerr or cout for user programs |
| 362 |
//Get the last word on the first line of the description which should |
| 363 |
//be "molecules", "reactions", etc and remove the s if only one object converted |
| 364 |
std::string objectname(pOutFormat->TargetClassDescription()); |
| 365 |
pos = objectname.find('\n'); |
| 366 |
if(count==1) --pos; |
| 367 |
objectname.erase(pos); |
| 368 |
pos = objectname.rfind(' '); |
| 369 |
if(pos==std::string::npos) |
| 370 |
pos=0; |
| 371 |
std::clog << count << objectname.substr(pos) << " converted" << endl; |
| 372 |
if(OutputFileList.size()>1) |
| 373 |
{ |
| 374 |
clog << OutputFileList.size() << " files output. The first is " << OutputFileList[0] <<endl; |
| 375 |
} |
| 376 |
|
| 377 |
std::string messageSummary = obErrorLog.GetMessageSummary(); |
| 378 |
if (messageSummary.size()) |
| 379 |
{ |
| 380 |
clog << messageSummary << endl; |
| 381 |
} |
| 382 |
|
| 383 |
#ifdef _DEBUG |
| 384 |
//CM keep window open |
| 385 |
cout << "Press any key to finish" <<endl; |
| 386 |
getch(); |
| 387 |
#endif |
| 388 |
|
| 389 |
return 0; |
| 390 |
} |
| 391 |
|
| 392 |
void DoOption(const char* p, OBConversion& Conv, |
| 393 |
OBConversion::Option_type typ, int& arg, int argc, char *argv[]) |
| 394 |
{ |
| 395 |
while(p && *p) //can have multiple single char options |
| 396 |
{ |
| 397 |
char ch[2]="?"; |
| 398 |
*ch = *p++; |
| 399 |
const char* txt=NULL; |
| 400 |
//Get the option text if needed |
| 401 |
int nParams = Conv.GetOptionParams(ch, typ); |
| 402 |
if(nParams) |
| 403 |
{ |
| 404 |
if(*p) |
| 405 |
{ |
| 406 |
txt = p; //use text immediately following the option letter |
| 407 |
p=NULL; //no more single char options |
| 408 |
} |
| 409 |
else if(arg<argc-1) |
| 410 |
{ |
| 411 |
txt = argv[++arg]; //use text from next arg |
| 412 |
if(*txt=='-') |
| 413 |
{ |
| 414 |
//...unless it is another option |
| 415 |
cerr << "Option -" << ch << " takes a parameter" << endl; |
| 416 |
exit(0); |
| 417 |
} |
| 418 |
} |
| 419 |
} |
| 420 |
Conv.AddOption(ch, typ, txt); |
| 421 |
} |
| 422 |
} |
| 423 |
|
| 424 |
void usage() |
| 425 |
{ |
| 426 |
cout << program_name << ": part of OOPSE " << |
| 427 |
OOPSE_VERSION_MAJOR << "." << OOPSE_VERSION_MINOR << "." << |
| 428 |
OOPSE_VERSION_TINY << " and OpenBabel " << BABEL_VERSION << " -- " |
| 429 |
<< __DATE__ << " -- " << __TIME__ << endl; |
| 430 |
cout << "Usage: " << program_name |
| 431 |
<< " [-i<input-type>] <name> [-o<output-type>] <name>" << endl; |
| 432 |
cout << "Try -H option for more information." << endl; |
| 433 |
/* |
| 434 |
#ifdef _DEBUG |
| 435 |
//CM keep window open |
| 436 |
cout << "Press any key to finish" <<endl; |
| 437 |
getch(); |
| 438 |
#endif |
| 439 |
*/ |
| 440 |
exit (0); |
| 441 |
} |
| 442 |
|
| 443 |
void help() |
| 444 |
{ |
| 445 |
cout << program_name << " converts chemical structures from one file format to another"<< endl << endl; |
| 446 |
cout << "Usage: " << program_name << " <input spec> <output spec> [Options]" << endl << endl; |
| 447 |
cout << "Each spec can be a file whose extension decides the format." << endl; |
| 448 |
cout << "Optionally the format can be specified by preceding the file by" << endl; |
| 449 |
cout << "-i<format-type> e.g. -ipdb, for input and -o<format-type> for output" << endl << endl; |
| 450 |
cout << "See below for available format-types, which are the same as the " << endl; |
| 451 |
cout << "file extensions and are case independent." << endl; |
| 452 |
cout << "If no input or output file is given stdin or stdout are used instead." << endl << endl; |
| 453 |
cout << "More than one input file can be specified and their names can contain" <<endl; |
| 454 |
cout << "wildcard chars (* and ?).The molecules are aggregated in the output file.\n" << endl; |
| 455 |
cout << OBConversion::Description(); // Conversion options |
| 456 |
cout << " -H Outputs this help text" << endl; |
| 457 |
cout << " -Hxxx (xxx is file format ID e.g. -Hpdb) gives format info" <<endl; |
| 458 |
cout << " -Hall Outputs details of all formats" <<endl; |
| 459 |
cout << " -V Outputs version number" <<endl; |
| 460 |
|
| 461 |
|
| 462 |
OBFormat* pDefault = OBConversion::GetDefaultFormat(); |
| 463 |
if(pDefault) |
| 464 |
cout << pDefault->TargetClassDescription();// some more options probably for OBMol |
| 465 |
|
| 466 |
OBFormat* pAPI= OBConversion::FindFormat("obapi"); |
| 467 |
if(pAPI) |
| 468 |
cout << pAPI->Description(); |
| 469 |
|
| 470 |
cout << "The following file formats are recognized:" << endl; |
| 471 |
Formatpos pos; |
| 472 |
OBFormat* pFormat; |
| 473 |
const char* str=NULL; |
| 474 |
while(OBConversion::GetNextFormat(pos,str,pFormat)) |
| 475 |
{ |
| 476 |
if((pFormat->Flags() & NOTWRITABLE) && (pFormat->Flags() & NOTREADABLE)) |
| 477 |
continue; |
| 478 |
cout << " " << str << endl; |
| 479 |
} |
| 480 |
cout << "\nSee further specific info and options using -H<format-type>, e.g. -Hpdb" << endl; |
| 481 |
} |
| 482 |
|
