| 1 | /********************************************************************** | 
| 2 | matrix.cpp - Operations on arbitrary-sized matrix. | 
| 3 |  | 
| 4 | Copyright (C) 1998-2001 by OpenEye Scientific Software, Inc. | 
| 5 | Some portions Copyright (C) 2001-2005 by Geoffrey R. Hutchison | 
| 6 |  | 
| 7 | This file is part of the Open Babel project. | 
| 8 | For more information, see <http://openbabel.sourceforge.net/> | 
| 9 |  | 
| 10 | This program is free software; you can redistribute it and/or modify | 
| 11 | it under the terms of the GNU General Public License as published by | 
| 12 | the Free Software Foundation version 2 of the License. | 
| 13 |  | 
| 14 | This program is distributed in the hope that it will be useful, | 
| 15 | but WITHOUT ANY WARRANTY; without even the implied warranty of | 
| 16 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the | 
| 17 | GNU General Public License for more details. | 
| 18 | ***********************************************************************/ | 
| 19 |  | 
| 20 | #include "matrix.hpp" | 
| 21 | #include "vector3.hpp" | 
| 22 |  | 
| 23 | using namespace std; | 
| 24 |  | 
| 25 | namespace OpenBabel | 
| 26 | { | 
| 27 |  | 
| 28 | void print_matrix(std::vector<std::vector<double> > &m) | 
| 29 | { | 
| 30 | unsigned int i,j; | 
| 31 |  | 
| 32 | for (i = 0; i < m.size(); i++) | 
| 33 | { | 
| 34 | for (j = 0; j < m[i].size(); j++) | 
| 35 | printf("%5.2f",m[i][j]); | 
| 36 | printf("\n"); | 
| 37 | } | 
| 38 | } | 
| 39 |  | 
| 40 | void print_matrix_f(double *m, int rows, int cols) | 
| 41 | { | 
| 42 | int i,j,idx; | 
| 43 |  | 
| 44 | for (i = 0; i < rows; i++) | 
| 45 | { | 
| 46 | idx = i * cols; | 
| 47 | for (j = 0; j < cols; j++) | 
| 48 | printf("%5.2f",m[idx+j]); | 
| 49 | printf("\n"); | 
| 50 | } | 
| 51 | } | 
| 52 |  | 
| 53 | void print_matrix_ff(double **m, int rows, int cols) | 
| 54 | { | 
| 55 | int i,j; | 
| 56 |  | 
| 57 | for (i = 0; i < rows; i++) | 
| 58 | { | 
| 59 | for (j = 0; j < cols; j++) | 
| 60 | printf("%5.2f",m[i][j]); | 
| 61 | printf("\n"); | 
| 62 | } | 
| 63 | } | 
| 64 |  | 
| 65 | bool mult_matrix(std::vector<std::vector<double> > &c, | 
| 66 | std::vector<std::vector<double> > &a, | 
| 67 | std::vector<std::vector<double> > &b) | 
| 68 | { | 
| 69 | unsigned int i,j,k; | 
| 70 |  | 
| 71 | if (a.size() != b.size()) | 
| 72 | return(false); | 
| 73 |  | 
| 74 | c.resize(a.size()); | 
| 75 |  | 
| 76 | for (i = 0; i < a.size(); i++) | 
| 77 | { | 
| 78 | c[i].resize(b[i].size()); | 
| 79 | for (j = 0; j < b[i].size(); j++) | 
| 80 | { | 
| 81 | c[i][j] = 0.0; | 
| 82 | for (k = 0; k < a[i].size(); k++) | 
| 83 | c[i][j] = c[i][j] + a[i][k] * b[k][j]; | 
| 84 | } | 
| 85 | } | 
| 86 |  | 
| 87 | return(true); | 
| 88 | } | 
| 89 |  | 
| 90 | bool mult_matrix_f(double *c, double *a, double *b, int rows, int cols) | 
| 91 | { | 
| 92 | int i,j,k,idx; | 
| 93 |  | 
| 94 | for ( i = 0 ; i < rows ; i++ ) | 
| 95 | { | 
| 96 | idx = i * cols; | 
| 97 | for ( j = 0; j < cols ; j++ ) | 
| 98 | { | 
| 99 | c[idx+j] = 0.0; | 
| 100 | for ( k = 0; k < cols ; k++ ) | 
| 101 | c[idx+j] = c[idx+j] + a[idx+k] * b[(k*cols)+j]; | 
| 102 | } | 
| 103 | } | 
| 104 |  | 
| 105 | return(true); | 
| 106 | } | 
| 107 |  | 
| 108 | bool mult_matrix_ff(double **c, double **a, double **b, int rows, int cols) | 
| 109 | { | 
| 110 | int i,j,k; | 
| 111 |  | 
| 112 | for ( i = 0 ; i < rows ; i++ ) | 
| 113 | for ( j = 0; j < cols ; j++ ) | 
| 114 | { | 
| 115 | c[i][j] = 0.0; | 
| 116 | for ( k = 0; k < cols ; k++ ) | 
| 117 | c[i][j] = c[i][j] + a[i][k] * b[k][j]; | 
| 118 | } | 
| 119 |  | 
| 120 | return(true); | 
| 121 | } | 
| 122 |  | 
| 123 | bool invert_matrix(std::vector<std::vector<double> > &mat, double &det) | 
| 124 | { | 
| 125 | int  i, j, k, m, n, row = 0, col = 0; | 
| 126 | double tempo, big, pvt; | 
| 127 |  | 
| 128 | vector<int> pvt_ind; | 
| 129 | vector<vector<int> > index; | 
| 130 |  | 
| 131 | int cols = mat[0].size(); | 
| 132 | int rows = mat.size(); | 
| 133 |  | 
| 134 | pvt_ind.resize(mat[0].size()); | 
| 135 |  | 
| 136 | index.resize(mat.size()); | 
| 137 | for (i = 0; (unsigned)i < mat.size(); i++) | 
| 138 | index[i].resize(2); | 
| 139 |  | 
| 140 | // make sure we have a square matrix | 
| 141 | // #rows == #cols; | 
| 142 | if (cols != rows) | 
| 143 | { | 
| 144 | det = 0.0; | 
| 145 | return(false); | 
| 146 | } | 
| 147 |  | 
| 148 | det = 1.0; | 
| 149 |  | 
| 150 | for (i = 0; i < cols; i++) | 
| 151 | pvt_ind[i] = rows+1; | 
| 152 |  | 
| 153 | for (i = 0; i < cols; i++) | 
| 154 | { | 
| 155 | big = 0.0; | 
| 156 | for (j = 0; j < cols; j++) | 
| 157 | { | 
| 158 | if (pvt_ind[j] != 0) | 
| 159 | for (k = 0; k < cols; k++) | 
| 160 | { | 
| 161 | if (fabs(big) < fabs(mat[j][k])) | 
| 162 | { | 
| 163 | row = j; | 
| 164 | col = k; | 
| 165 | big = mat[j][k]; | 
| 166 | } | 
| 167 | } | 
| 168 | } | 
| 169 |  | 
| 170 | pvt_ind[col]++; | 
| 171 | if (row != col) | 
| 172 | { | 
| 173 | det = -det; | 
| 174 | for (m = 0; m < cols; m++) | 
| 175 | { | 
| 176 | tempo = mat[row][m]; | 
| 177 | mat[row][m] = mat[col][m]; | 
| 178 | mat[col][m] = tempo; | 
| 179 | } | 
| 180 | } | 
| 181 |  | 
| 182 | index[i][0] = row; | 
| 183 | index[i][1] = col; | 
| 184 | pvt = mat[col][col]; | 
| 185 | det *= pvt; | 
| 186 |  | 
| 187 | mat[col][col] = 1.0; | 
| 188 |  | 
| 189 | for (m = 0; m < cols; m++) | 
| 190 | mat[col][m] /= pvt; | 
| 191 |  | 
| 192 | for (n = 0; n < cols; n++) | 
| 193 | if (n != col) | 
| 194 | { | 
| 195 | tempo = mat[n][col]; | 
| 196 | mat[n][col] = 0.0; | 
| 197 | for (m = 0; m < cols; m++) | 
| 198 | mat[n][m] -= mat[col][m] * tempo; | 
| 199 | } | 
| 200 | } | 
| 201 |  | 
| 202 | for (i = 0; i < cols; i++) | 
| 203 | { | 
| 204 | m = cols - 1; | 
| 205 | if (index[m][0] != index[m][1]) | 
| 206 | { | 
| 207 | row = index[m][0]; | 
| 208 | col = index[m][1]; | 
| 209 | for (k = 0; k < cols; k++) | 
| 210 | { | 
| 211 | tempo = mat[k][row]; | 
| 212 | mat[k][row] = mat[k][col]; | 
| 213 | mat[k][col] = tempo; | 
| 214 | } | 
| 215 | } | 
| 216 | } | 
| 217 |  | 
| 218 | return(true); | 
| 219 | } | 
| 220 |  | 
| 221 | bool invert_matrix_f(double *mat, double &det, int rows, int cols) | 
| 222 | { | 
| 223 | int  i, j, k, m, n, row = 0, col = 0, idx1, idx2; | 
| 224 | double tempo, big, pvt; | 
| 225 |  | 
| 226 | vector<int> pvt_ind; | 
| 227 | vector<vector<int> > index; | 
| 228 |  | 
| 229 | pvt_ind.resize(cols); | 
| 230 | index.resize(rows); | 
| 231 |  | 
| 232 | for (i = 0; i < rows; i++) | 
| 233 | index[i].resize(2); | 
| 234 |  | 
| 235 | // make sure we have a square matrix | 
| 236 | // #rows == #cols; | 
| 237 | if (cols != rows) | 
| 238 | { | 
| 239 | det = 0.0; | 
| 240 | return(false); | 
| 241 | } | 
| 242 |  | 
| 243 | det = 1.0; | 
| 244 |  | 
| 245 | for (i = 0; i < cols; i++) | 
| 246 | pvt_ind[i] = rows+1; | 
| 247 |  | 
| 248 | for (i = 0; i < cols; i++) | 
| 249 | { | 
| 250 | big = 0.0; | 
| 251 | for (j = 0; j < cols; j++) | 
| 252 | { | 
| 253 | if (pvt_ind[j] != 0) | 
| 254 | { | 
| 255 | idx1 = (j * cols); | 
| 256 | for (k = 0; k < cols; k++) | 
| 257 | { | 
| 258 | idx2 = idx1 + k; | 
| 259 | if (fabs(big) < fabs(mat[idx2])) | 
| 260 | { | 
| 261 | row = j; | 
| 262 | col = k; | 
| 263 | big = mat[idx2]; | 
| 264 | } | 
| 265 | } | 
| 266 | } | 
| 267 | } | 
| 268 |  | 
| 269 | pvt_ind[col]++; | 
| 270 | if (row != col) | 
| 271 | { | 
| 272 | det  = -det; | 
| 273 | idx1 = row * cols; | 
| 274 | idx2 = col * cols; | 
| 275 | for (m = 0; m < cols; m++) | 
| 276 | { | 
| 277 | tempo = mat[idx1+m]; | 
| 278 | mat[idx1+m] = mat[idx2+m]; | 
| 279 | mat[idx2+m] = tempo; | 
| 280 | } | 
| 281 | } | 
| 282 |  | 
| 283 | index[i][0] = row; | 
| 284 | index[i][1] = col; | 
| 285 |  | 
| 286 | idx1 = (col*cols); | 
| 287 | pvt  = mat[idx1+col]; | 
| 288 | det *= pvt; | 
| 289 |  | 
| 290 | mat[idx1+col] = 1.0; | 
| 291 |  | 
| 292 | for (m = 0; m < cols; m++) | 
| 293 | mat[idx1+m] /= pvt; | 
| 294 |  | 
| 295 | for (n = 0; n < cols; n++) | 
| 296 | if (n != col) | 
| 297 | { | 
| 298 | idx1  = n * cols; | 
| 299 | tempo = mat[idx1 + col]; | 
| 300 | mat[idx1 + col] = 0.0; | 
| 301 |  | 
| 302 | idx2 = col * cols; | 
| 303 | for (m = 0; m < cols; m++) | 
| 304 | mat[idx1 + m] -= mat[idx2 + m] * tempo; | 
| 305 | } | 
| 306 | } | 
| 307 |  | 
| 308 | for (i = 0; i < cols; i++) | 
| 309 | { | 
| 310 | m = cols - 1; | 
| 311 | if (index[m][0] != index[m][1]) | 
| 312 | { | 
| 313 | row = index[m][0]; | 
| 314 | col = index[m][1]; | 
| 315 | for (k = 0; k < cols; k++) | 
| 316 | { | 
| 317 | idx1  = (k * cols); | 
| 318 | idx2  = idx1 + col; | 
| 319 | idx1 += row; | 
| 320 |  | 
| 321 | tempo = mat[idx1]; | 
| 322 | mat[idx1] = mat[idx2]; | 
| 323 | mat[idx2] = tempo; | 
| 324 | } | 
| 325 | } | 
| 326 | } | 
| 327 |  | 
| 328 | return(true); | 
| 329 | } | 
| 330 |  | 
| 331 | bool invert_matrix_ff(double **mat, double &det, int rows, int cols) | 
| 332 | { | 
| 333 | int  i, j, k, m, n, row = 0, col = 0; | 
| 334 | double tempo, big, pvt; | 
| 335 |  | 
| 336 | vector<int> pvt_ind; | 
| 337 | vector<vector<int> > index; | 
| 338 |  | 
| 339 | pvt_ind.resize(cols); | 
| 340 | index.resize(rows); | 
| 341 |  | 
| 342 | for (i = 0; i < rows; i++) | 
| 343 | index[i].resize(2); | 
| 344 |  | 
| 345 | // make sure we have a square matrix | 
| 346 | // #rows == #cols; | 
| 347 | if (cols != rows) | 
| 348 | { | 
| 349 | det = 0.0; | 
| 350 | return(false); | 
| 351 | } | 
| 352 |  | 
| 353 | det = 1.0; | 
| 354 |  | 
| 355 | for (i = 0; i < cols; i++) | 
| 356 | pvt_ind[i] = rows+1; | 
| 357 |  | 
| 358 | for (i = 0; i < cols; i++) | 
| 359 | { | 
| 360 | big = 0.0; | 
| 361 | for (j = 0; j < cols; j++) | 
| 362 | { | 
| 363 | if (pvt_ind[j] != 0) | 
| 364 | for (k = 0; k < cols; k++) | 
| 365 | { | 
| 366 | if (fabs(big) < fabs(mat[j][k])) | 
| 367 | { | 
| 368 | row = j; | 
| 369 | col = k; | 
| 370 | big = mat[j][k]; | 
| 371 | } | 
| 372 | } | 
| 373 | } | 
| 374 |  | 
| 375 | pvt_ind[col]++; | 
| 376 | if (row != col) | 
| 377 | { | 
| 378 | det = -det; | 
| 379 | for (m = 0; m < cols; m++) | 
| 380 | { | 
| 381 | tempo = mat[row][m]; | 
| 382 | mat[row][m] = mat[col][m]; | 
| 383 | mat[col][m] = tempo; | 
| 384 | } | 
| 385 | } | 
| 386 |  | 
| 387 | index[i][0] = row; | 
| 388 | index[i][1] = col; | 
| 389 | pvt = mat[col][col]; | 
| 390 | det *= pvt; | 
| 391 |  | 
| 392 | mat[col][col] = 1.0; | 
| 393 |  | 
| 394 | for (m = 0; m < cols; m++) | 
| 395 | mat[col][m] /= pvt; | 
| 396 |  | 
| 397 | for (n = 0; n < cols; n++) | 
| 398 | if (n != col) | 
| 399 | { | 
| 400 | tempo = mat[n][col]; | 
| 401 | mat[n][col] = 0.0; | 
| 402 | for (m = 0; m < cols; m++) | 
| 403 | mat[n][m] -= mat[col][m] * tempo; | 
| 404 | } | 
| 405 | } | 
| 406 |  | 
| 407 | for (i = 0; i < cols; i++) | 
| 408 | { | 
| 409 | m = cols - 1; | 
| 410 | if (index[m][0] != index[m][1]) | 
| 411 | { | 
| 412 | row = index[m][0]; | 
| 413 | col = index[m][1]; | 
| 414 | for (k = 0; k < cols; k++) | 
| 415 | { | 
| 416 | tempo = mat[k][row]; | 
| 417 | mat[k][row] = mat[k][col]; | 
| 418 | mat[k][col] = tempo; | 
| 419 | } | 
| 420 | } | 
| 421 | } | 
| 422 |  | 
| 423 | return(true); | 
| 424 | } | 
| 425 |  | 
| 426 | bool convert_matrix_f(std::vector<std::vector<double> > &src, double *dst) | 
| 427 | { | 
| 428 | unsigned int i, j, idx = 0; | 
| 429 |  | 
| 430 | for ( i = 0 ; i < src.size() ; i++ ) | 
| 431 | for ( j = 0 ; j < src[i].size() ; j++ ) | 
| 432 | dst[idx++] = src[i][j]; | 
| 433 |  | 
| 434 | return true; | 
| 435 | } | 
| 436 |  | 
| 437 | bool convert_matrix_ff(std::vector<std::vector<double> > &src, double **dst) | 
| 438 | { | 
| 439 | unsigned int i, j; | 
| 440 |  | 
| 441 | for ( i = 0 ; i < src.size() ; i++ ) | 
| 442 | for ( j = 0 ; j < src[i].size() ; j++ ) | 
| 443 | dst[i][j] = src[i][j]; | 
| 444 |  | 
| 445 | return true; | 
| 446 | } | 
| 447 |  | 
| 448 | bool convert_matrix_f(double *src, std::vector<std::vector<double> > &dst, | 
| 449 | int rows, int cols) | 
| 450 | { | 
| 451 | int i, j, idx; | 
| 452 |  | 
| 453 | dst.resize(rows); | 
| 454 | for ( i = 0 ; i < rows ; i++ ) | 
| 455 | { | 
| 456 | idx = i * cols; | 
| 457 | dst[i].resize(cols); | 
| 458 | for ( j = 0 ; j < cols ; j++ ) | 
| 459 | dst[i][j] = src[idx+j]; | 
| 460 | } | 
| 461 |  | 
| 462 | return true; | 
| 463 | } | 
| 464 |  | 
| 465 | bool convert_matrix_ff(double **src, std::vector<std::vector<double> > &dst, | 
| 466 | int rows, int cols) | 
| 467 | { | 
| 468 | int i, j; | 
| 469 |  | 
| 470 | dst.resize(rows); | 
| 471 | for ( i = 0 ; i < rows ; i++ ) | 
| 472 | { | 
| 473 | dst[i].resize(cols); | 
| 474 | for ( j = 0 ; j < cols ; j++ ) | 
| 475 | dst[i][j] = src[i][j]; | 
| 476 | } | 
| 477 |  | 
| 478 | return true; | 
| 479 | } | 
| 480 |  | 
| 481 | bool convert_matrix_f_ff(double *src, double **dst, int rows, int cols) | 
| 482 | { | 
| 483 | int i, j, idx; | 
| 484 |  | 
| 485 | for ( i = 0 ; i < rows ; i++ ) | 
| 486 | { | 
| 487 | idx = i * cols; | 
| 488 | for ( j = 0 ; j < cols ; j++ ) | 
| 489 | dst[i][j] = src[idx+j]; | 
| 490 | } | 
| 491 |  | 
| 492 | return true; | 
| 493 | } | 
| 494 |  | 
| 495 | bool convert_matrix_ff_f(double **src, double *dst, int rows, int cols) | 
| 496 | { | 
| 497 | int i, j, idx; | 
| 498 |  | 
| 499 | for ( i = 0 ; i < rows ; i++ ) | 
| 500 | { | 
| 501 | idx = i * cols; | 
| 502 | for ( j = 0 ; j < cols ; j++ ) | 
| 503 | dst[idx+j] = src[i][j]; | 
| 504 | } | 
| 505 |  | 
| 506 | return true; | 
| 507 | } | 
| 508 |  | 
| 509 | } // end namespace OpenBabel | 
| 510 |  | 
| 511 | //! \file matrix.cpp | 
| 512 | //! \brief Operations on arbitrary-sized matrix. | 
| 513 |  |