[3240] | 1 | module math |
---|
| 2 | ! """ |
---|
| 3 | ! Contains useful math functions. |
---|
| 4 | ! """ |
---|
| 5 | implicit none |
---|
| 6 | |
---|
| 7 | integer, parameter :: & |
---|
| 8 | dp = selected_real_kind(15,300) |
---|
| 9 | |
---|
| 10 | real(kind=dp), parameter :: & |
---|
| 11 | pi = 3.141592653589793238460_dp |
---|
| 12 | |
---|
| 13 | doubleprecision, parameter :: & |
---|
| 14 | sqrtpi = 0.5641895835477563d0 ! = 1 / sqrt(pi) |
---|
| 15 | |
---|
| 16 | doubleprecision, parameter :: & |
---|
| 17 | prec_high = 10d0 ** (-precision(0d0)), prec_low = 10d0 ** (-precision(0.)) |
---|
| 18 | |
---|
| 19 | save |
---|
| 20 | |
---|
| 21 | contains |
---|
| 22 | doubleprecision function chi2(observed_data, calculated_data) |
---|
| 23 | ! """ |
---|
| 24 | ! Evaluate the goodeness of fit of calculated data in respect to observations. |
---|
| 25 | ! |
---|
| 26 | ! input: |
---|
| 27 | ! observed_data: array of observed data |
---|
| 28 | ! calculated_data: array of calculated data |
---|
| 29 | ! """ |
---|
| 30 | implicit None |
---|
| 31 | |
---|
| 32 | doubleprecision, dimension(:), intent(in) :: observed_data, calculated_data |
---|
| 33 | |
---|
| 34 | chi2 = sum((observed_data(:) - calculated_data(:))**2 / calculated_data(:)) |
---|
| 35 | end function chi2 |
---|
| 36 | |
---|
| 37 | |
---|
| 38 | doubleprecision function chi2_reduced(observed_data, calculated_data, input_deviation) |
---|
| 39 | ! """ |
---|
| 40 | ! Evaluate the goodeness of fit of calculated data in respect to observations and observational errors. |
---|
| 41 | ! |
---|
| 42 | ! input: |
---|
| 43 | ! observed_data: array of observed data |
---|
| 44 | ! calculated_data: array of calculated data |
---|
| 45 | ! input_deviation: array of observational error |
---|
| 46 | ! """ |
---|
| 47 | implicit None |
---|
| 48 | |
---|
| 49 | doubleprecision, dimension(:), intent(in) :: observed_data, calculated_data, input_deviation |
---|
| 50 | |
---|
| 51 | chi2_reduced = sum((observed_data(:) - calculated_data(:))**2 / input_deviation(:)**2) |
---|
| 52 | end function chi2_reduced |
---|
| 53 | |
---|
| 54 | |
---|
| 55 | doubleprecision function deg2rad(angle) |
---|
| 56 | ! """ |
---|
| 57 | ! Return an angle in radians from an angle in degrees. |
---|
| 58 | ! |
---|
| 59 | ! input: |
---|
| 60 | ! angle: (degree) an angle |
---|
| 61 | ! """ |
---|
| 62 | implicit none |
---|
| 63 | |
---|
| 64 | doubleprecision, intent(in) :: angle |
---|
| 65 | |
---|
| 66 | deg2rad = angle * (pi / 180.0_dp) |
---|
| 67 | end function deg2rad |
---|
| 68 | |
---|
| 69 | |
---|
| 70 | doubleprecision function ellipse_polar_form(semi_major_axis, semi_minor_axis, angle) |
---|
| 71 | ! """ |
---|
| 72 | ! Return the ellipse polar form of an ellipse relative to its center. |
---|
| 73 | ! |
---|
| 74 | ! inputs: |
---|
| 75 | ! angle: (deg) angle between the semi-major axis, the center, and the point of the ellipse |
---|
| 76 | ! semi_major_axis: semi-major axis of the ellipse |
---|
| 77 | ! semi_minor_axis: semi-minor axis of the ellipse |
---|
| 78 | ! """ |
---|
| 79 | implicit none |
---|
| 80 | |
---|
| 81 | doubleprecision, intent(in) :: angle, semi_major_axis, semi_minor_axis |
---|
| 82 | |
---|
| 83 | doubleprecision :: & |
---|
| 84 | theta |
---|
| 85 | |
---|
| 86 | theta = deg2rad(angle) |
---|
| 87 | |
---|
| 88 | ellipse_polar_form = semi_major_axis * semi_minor_axis / & |
---|
| 89 | sqrt((semi_minor_axis * cos(theta))**2 + (semi_major_axis * sin(theta))**2) |
---|
| 90 | |
---|
| 91 | return |
---|
| 92 | |
---|
| 93 | end function ellipse_polar_form |
---|
| 94 | |
---|
| 95 | |
---|
| 96 | doubleprecision function gaussian(x, fwhm) |
---|
| 97 | ! """ |
---|
| 98 | ! Return the gaussian of a value, for a given Full Width Half Maximum. |
---|
| 99 | ! |
---|
| 100 | ! inputs: |
---|
| 101 | ! x: value for which to calculate the gaussian |
---|
| 102 | ! fwhm: Full Width Half Maximum of the gaussian function |
---|
| 103 | ! """ |
---|
| 104 | implicit none |
---|
| 105 | |
---|
| 106 | doubleprecision, intent(in) :: x, fwhm |
---|
| 107 | |
---|
| 108 | doubleprecision :: & |
---|
| 109 | sigma |
---|
| 110 | |
---|
| 111 | sigma = fwhm / (2D0 * sqrt(2D0 * log(2D0))) |
---|
| 112 | gaussian = 1D0 / (sigma * sqrt(2D0 * pi)) * exp(-1D0/2D0 * (x / sigma)**2D0) |
---|
| 113 | end function gaussian |
---|
| 114 | |
---|
| 115 | |
---|
| 116 | doubleprecision function gaussian_noise() result(n) |
---|
| 117 | ! """ |
---|
| 118 | ! Return a random value following a standard normal distriubtion PDF. |
---|
| 119 | ! |
---|
| 120 | ! output: |
---|
| 121 | ! n: random gaussian noise |
---|
| 122 | ! """ |
---|
| 123 | implicit none |
---|
| 124 | |
---|
| 125 | doubleprecision ::& |
---|
| 126 | r ! random number between 0 and 1 |
---|
| 127 | |
---|
| 128 | call init_random_seed |
---|
| 129 | call random_number(r) |
---|
| 130 | |
---|
| 131 | n = sqrt(2D0) * erfinv(2 * r - 1) |
---|
| 132 | |
---|
| 133 | return |
---|
| 134 | |
---|
| 135 | end function gaussian_noise |
---|
| 136 | |
---|
| 137 | |
---|
| 138 | doubleprecision function sec(angle) |
---|
| 139 | ! """ |
---|
| 140 | ! Return the secant of an angle in degrees. |
---|
| 141 | ! |
---|
| 142 | ! input: |
---|
| 143 | ! angle: (degree) an angle |
---|
| 144 | ! """ |
---|
| 145 | implicit none |
---|
| 146 | |
---|
| 147 | doubleprecision, intent(in) :: angle |
---|
| 148 | |
---|
| 149 | sec = 1D0 / cos(deg2rad(angle)) |
---|
| 150 | end function sec |
---|
| 151 | |
---|
| 152 | |
---|
| 153 | doubleprecision function sgn(value) |
---|
| 154 | ! """ |
---|
| 155 | ! Return the sign of a value. |
---|
| 156 | ! Not to be confused with Fortran built-in function SIGN(A, B). |
---|
| 157 | ! """ |
---|
| 158 | implicit none |
---|
| 159 | |
---|
| 160 | doubleprecision, intent(in) :: value |
---|
| 161 | |
---|
| 162 | if (value >= 0D0) then |
---|
| 163 | sgn = 1D0 |
---|
| 164 | else |
---|
| 165 | sgn = -1D0 |
---|
| 166 | end if |
---|
| 167 | |
---|
| 168 | return |
---|
| 169 | end function sgn |
---|
| 170 | |
---|
| 171 | |
---|
| 172 | doubleprecision function sinc_fwhm(x, fwhm) |
---|
| 173 | ! """ |
---|
| 174 | ! Return the sine cardinal of a value, for a given Full Width Half Maximum. |
---|
| 175 | ! |
---|
| 176 | ! inputs: |
---|
| 177 | ! x: value for which to calculate sinc |
---|
| 178 | ! fwhm: Full Width Half Maximum of the sinc function |
---|
| 179 | ! """ |
---|
| 180 | implicit none |
---|
| 181 | |
---|
| 182 | doubleprecision,intent(in) :: x, fwhm |
---|
| 183 | |
---|
| 184 | doubleprecision, parameter :: & |
---|
| 185 | k = 1D0 / (2D0 * 1.89549D0) ! sinc fwhm constant |
---|
| 186 | |
---|
| 187 | if (x > 0D0 - tiny(0.) .and. x < 0D0 - tiny(0.)) then |
---|
| 188 | sinc_fwhm = 1D0 ! avoid NaN at x = 0 |
---|
| 189 | else |
---|
| 190 | sinc_fwhm = sin(x / (k * fwhm)) / (x / (k * fwhm)) |
---|
| 191 | end if |
---|
| 192 | end function sinc_fwhm |
---|
| 193 | |
---|
| 194 | |
---|
| 195 | function arange(start, stop, step) result(array) |
---|
| 196 | ! """ |
---|
| 197 | ! Return evenly spaced values within a given interval. |
---|
| 198 | ! |
---|
| 199 | ! inputs: |
---|
| 200 | ! start: start of interval; the array starts with this value |
---|
| 201 | ! stop: end of interval; the array does not inlude this value |
---|
| 202 | ! step: spacing between values |
---|
| 203 | ! |
---|
| 204 | ! outputs: |
---|
| 205 | ! array: array of evenly spaced values |
---|
| 206 | ! """ |
---|
| 207 | implicit none |
---|
| 208 | |
---|
| 209 | doubleprecision, intent(in) :: start, stop, step |
---|
| 210 | doubleprecision, dimension(:), allocatable :: array |
---|
| 211 | |
---|
| 212 | integer :: & |
---|
| 213 | i, & ! index |
---|
| 214 | n ! number of elements in array |
---|
| 215 | |
---|
| 216 | n = ceiling((stop - start) / step) |
---|
| 217 | |
---|
| 218 | allocate(array(n)) |
---|
| 219 | |
---|
| 220 | do i = 1, n |
---|
| 221 | array(i) = start + (i - 1) * step |
---|
| 222 | end do |
---|
| 223 | |
---|
| 224 | return |
---|
| 225 | end function arange |
---|
| 226 | |
---|
| 227 | |
---|
| 228 | function arange_include(start, stop, step) result(array) |
---|
| 229 | ! """ |
---|
| 230 | ! Return evenly spaced values including a given interval. |
---|
| 231 | ! |
---|
| 232 | ! inputs: |
---|
| 233 | ! start: start of interval; the array starts with this value |
---|
| 234 | ! stop: end of interval; the last element of array is always greater than this value by a max of 1 step |
---|
| 235 | ! step: spacing between values |
---|
| 236 | ! |
---|
| 237 | ! outputs: |
---|
| 238 | ! array: array of evenly spaced values |
---|
| 239 | ! """ |
---|
| 240 | implicit none |
---|
| 241 | |
---|
| 242 | doubleprecision, intent(in) :: start, stop, step |
---|
| 243 | doubleprecision, dimension(:), allocatable :: array |
---|
| 244 | |
---|
| 245 | integer :: & |
---|
| 246 | i, & ! index |
---|
| 247 | n ! number of elements in array |
---|
| 248 | |
---|
| 249 | n = ceiling((stop - start) / step) + 1 ! operation +1 ensures that stop is included in array |
---|
| 250 | |
---|
| 251 | allocate(array(n)) |
---|
| 252 | |
---|
| 253 | do i = 1, n |
---|
| 254 | array(i) = start + (i - 1) * step |
---|
| 255 | end do |
---|
| 256 | |
---|
| 257 | return |
---|
| 258 | end function arange_include |
---|
| 259 | |
---|
| 260 | |
---|
| 261 | function convolve(signal, filter) result(convolved_signal) |
---|
| 262 | ! """ |
---|
| 263 | ! Convolve the signal by the filter using classical convolution. |
---|
| 264 | ! |
---|
| 265 | ! inputs: |
---|
| 266 | ! signal: the signal array |
---|
| 267 | ! filter: the filter array |
---|
| 268 | ! |
---|
| 269 | ! output: |
---|
| 270 | ! convolved_signal: signal convolved by filter |
---|
| 271 | ! |
---|
| 272 | ! source: https://fortrandev.wordpress.com/2013/04/01/fortran-convolution-algorithm/ |
---|
| 273 | ! """ |
---|
| 274 | implicit none |
---|
| 275 | |
---|
| 276 | doubleprecision, dimension(:), intent(in) :: signal, filter |
---|
| 277 | doubleprecision, dimension(:), allocatable :: convolved_signal, convolution |
---|
| 278 | |
---|
| 279 | integer :: & |
---|
| 280 | i, & |
---|
| 281 | j, & |
---|
| 282 | k, & |
---|
| 283 | size_filter, & |
---|
| 284 | size_signal, & |
---|
| 285 | start_index |
---|
| 286 | |
---|
| 287 | size_signal = size(signal) |
---|
| 288 | size_filter = size(filter) |
---|
| 289 | |
---|
| 290 | start_index = floor(size_filter / 2d0) |
---|
| 291 | |
---|
| 292 | allocate(convolved_signal(size_signal), convolution(size_signal + size_filter)) |
---|
| 293 | |
---|
| 294 | ! Last part |
---|
| 295 | do i = size_signal, size_signal + size_filter |
---|
| 296 | convolution(i) = 0D0 |
---|
| 297 | j = size_signal |
---|
| 298 | |
---|
| 299 | do k = 1, size_filter |
---|
| 300 | convolution(i) = convolution(i) + signal(j) * filter(k) |
---|
| 301 | j = j - 1 |
---|
| 302 | end do |
---|
| 303 | end do |
---|
| 304 | |
---|
| 305 | ! Middle part |
---|
| 306 | do i = size_filter, size_signal |
---|
| 307 | convolution(i) = 0D0 |
---|
| 308 | j = i |
---|
| 309 | |
---|
| 310 | do k = 1, size_filter |
---|
| 311 | convolution(i) = convolution(i) + signal(j) * filter(k) |
---|
| 312 | j = j - 1 |
---|
| 313 | end do |
---|
| 314 | end do |
---|
| 315 | |
---|
| 316 | ! First part |
---|
| 317 | do i = 1, size_filter |
---|
| 318 | convolution(i) = 0D0 |
---|
| 319 | j = i |
---|
| 320 | k = 1 |
---|
| 321 | |
---|
| 322 | do while (j > 0) |
---|
| 323 | convolution(i) = convolution(i) + signal(j) * filter(k) |
---|
| 324 | j = j - 1 |
---|
| 325 | k = k + 1 |
---|
| 326 | end do |
---|
| 327 | end do |
---|
| 328 | |
---|
| 329 | convolved_signal(:) = convolution(start_index:start_index + size_signal - 1) |
---|
| 330 | end function convolve |
---|
| 331 | |
---|
| 332 | |
---|
| 333 | function slide_convolve(signal, filter) result(convolved_signal) |
---|
| 334 | ! """ |
---|
| 335 | ! Slide convolve the signal by a sliding filter using classical convolution. |
---|
| 336 | ! |
---|
| 337 | ! inputs: |
---|
| 338 | ! signal: the signal array |
---|
| 339 | ! filter: the filter 2D-array |
---|
| 340 | ! |
---|
| 341 | ! output: |
---|
| 342 | ! convolved_signal: signal convolved by filter |
---|
| 343 | ! """ |
---|
| 344 | implicit none |
---|
| 345 | |
---|
| 346 | doubleprecision, dimension(:), intent(in) :: signal |
---|
| 347 | doubleprecision, dimension(:, :), intent(in) :: filter |
---|
| 348 | doubleprecision, dimension(:), allocatable :: convolved_signal, convolution |
---|
| 349 | |
---|
| 350 | integer :: & |
---|
| 351 | i, & ! index |
---|
| 352 | j, & ! index |
---|
| 353 | k, & ! index |
---|
| 354 | size_filter, & ! size of the filter |
---|
| 355 | size_signal, & ! size of the signal |
---|
| 356 | start_index ! start index of the result |
---|
| 357 | |
---|
| 358 | size_signal = size(signal) |
---|
| 359 | size_filter = size(filter, 1) |
---|
| 360 | |
---|
| 361 | ! Check sizes |
---|
| 362 | ! TODO [low] remove the size constraint |
---|
| 363 | if (size_signal < size_filter) then |
---|
| 364 | print '("ERROR: slide_convolve: filter size must be lower than signal size, & |
---|
| 365 | &but sizes are ", I10, "and ", I10)', size_filter, size_signal |
---|
| 366 | print '("Catched error: ", ES15.8)', signal(size_filter) |
---|
| 367 | stop ! be sure the program stops here |
---|
| 368 | end if |
---|
| 369 | |
---|
| 370 | start_index = floor(size_filter / 2D0) + 1 |
---|
| 371 | |
---|
| 372 | allocate(convolved_signal(size_signal), convolution(size_signal + size_filter)) |
---|
| 373 | |
---|
| 374 | convolution(:) = 0D0 |
---|
| 375 | |
---|
| 376 | ! Last part |
---|
| 377 | do i = size_signal + 1, start_index + size_signal - 1 |
---|
| 378 | j = size_signal |
---|
| 379 | |
---|
| 380 | do k = 1, size_filter |
---|
| 381 | convolution(i) = convolution(i) + signal(j) * filter(k, j) |
---|
| 382 | j = j - 1 |
---|
| 383 | end do |
---|
| 384 | end do |
---|
| 385 | |
---|
| 386 | ! Middle part |
---|
| 387 | do i = size_filter, size_signal |
---|
| 388 | j = i |
---|
| 389 | |
---|
| 390 | do k = 1, size_filter |
---|
| 391 | convolution(i) = convolution(i) + signal(j) * filter(k, j) |
---|
| 392 | j = j - 1 |
---|
| 393 | end do |
---|
| 394 | end do |
---|
| 395 | |
---|
| 396 | ! First part |
---|
| 397 | do i = start_index + 1, size_filter - 1 |
---|
| 398 | j = i |
---|
| 399 | |
---|
| 400 | do k = 1, i |
---|
| 401 | convolution(i) = convolution(i) + signal(j) * filter(k, j) |
---|
| 402 | j = j - 1 |
---|
| 403 | end do |
---|
| 404 | end do |
---|
| 405 | |
---|
| 406 | ! First index |
---|
| 407 | i = start_index |
---|
| 408 | convolution(i) = 0D0 |
---|
| 409 | j = i |
---|
| 410 | |
---|
| 411 | do k = 1, i |
---|
| 412 | convolution(i) = convolution(i) + signal(j) * filter(k, j) |
---|
| 413 | j = j - 1 |
---|
| 414 | end do |
---|
| 415 | |
---|
| 416 | k = i + 1 |
---|
| 417 | convolution(i) = convolution(i) + signal(1) * filter(k, 1) |
---|
| 418 | |
---|
| 419 | convolved_signal(:) = convolution(start_index:start_index + size_signal - 1) |
---|
| 420 | end function slide_convolve |
---|
| 421 | |
---|
| 422 | |
---|
| 423 | function search_sorted(array, value) result(index) |
---|
| 424 | ! """ |
---|
| 425 | ! Find the index into a sorted array such that the corresponding value is the closest to 'value'. |
---|
| 426 | ! """ |
---|
| 427 | implicit none |
---|
| 428 | |
---|
| 429 | doubleprecision, intent(in) :: value |
---|
| 430 | doubleprecision, dimension(:), intent(in) :: array |
---|
| 431 | |
---|
| 432 | integer :: index |
---|
| 433 | |
---|
| 434 | integer :: & |
---|
| 435 | i_low, & |
---|
| 436 | i_high, & |
---|
| 437 | i_mid |
---|
| 438 | |
---|
| 439 | i_low = 1 |
---|
| 440 | i_mid = 0 |
---|
| 441 | i_high = size(array) |
---|
| 442 | |
---|
| 443 | if(value < array(1)) then |
---|
| 444 | index = 1 |
---|
| 445 | return |
---|
| 446 | elseif(value > array(i_high)) then |
---|
| 447 | index = i_high |
---|
| 448 | return |
---|
| 449 | end if |
---|
| 450 | |
---|
| 451 | do while(i_low <= i_high) |
---|
| 452 | i_mid = i_low + (i_high - i_low) / 2 |
---|
| 453 | |
---|
| 454 | if(value < array(i_mid)) then |
---|
| 455 | i_high = i_mid - 1 |
---|
| 456 | elseif(value > array(i_mid)) then |
---|
| 457 | i_low = i_mid + 1 |
---|
| 458 | else |
---|
| 459 | index = i_mid |
---|
| 460 | return |
---|
| 461 | end if |
---|
| 462 | end do |
---|
| 463 | |
---|
| 464 | if(array(i_low) - value < value - array(i_high)) then |
---|
| 465 | index = i_low |
---|
| 466 | else |
---|
| 467 | index = i_high |
---|
| 468 | end if |
---|
| 469 | |
---|
| 470 | return |
---|
| 471 | end function search_sorted |
---|
| 472 | |
---|
| 473 | |
---|
| 474 | function erfinv(x) result(r) |
---|
| 475 | ! """ |
---|
| 476 | ! Calculate the inverse of the erf function. |
---|
| 477 | ! |
---|
| 478 | ! input: |
---|
| 479 | ! x: a number between -1 and 1 |
---|
| 480 | ! |
---|
| 481 | ! output: |
---|
| 482 | ! r: x = erf(r) |
---|
| 483 | ! """ |
---|
| 484 | implicit none |
---|
| 485 | |
---|
| 486 | doubleprecision, parameter :: & ! erfinv-approximation factors |
---|
| 487 | a0 = 0.886226899, & |
---|
| 488 | a1 = -1.645349621, & |
---|
| 489 | a2 = 0.914624893, & |
---|
| 490 | a3 = -0.140543331, & |
---|
| 491 | b0 = 1, & |
---|
| 492 | b1 = -2.118377725, & |
---|
| 493 | b2 = 1.442710462, & |
---|
| 494 | b3 = -0.329097515, & |
---|
| 495 | b4 = 0.012229801, & |
---|
| 496 | c0 = -1.970840454, & |
---|
| 497 | c1 = -1.62490649, & |
---|
| 498 | c2 = 3.429567803, & |
---|
| 499 | c3 = 1.641345311, & |
---|
| 500 | d0 = 1, & |
---|
| 501 | d1 = 3.543889200, & |
---|
| 502 | d2 = 1.637067800 |
---|
| 503 | |
---|
| 504 | doubleprecision :: x, r |
---|
| 505 | |
---|
| 506 | integer ::& |
---|
| 507 | sign_x ! sign of elements of array x |
---|
| 508 | |
---|
| 509 | doubleprecision :: & |
---|
| 510 | x2, & ! square of x |
---|
| 511 | y ! intermediate value |
---|
| 512 | |
---|
| 513 | if(x < -1 .or. x > 1) then |
---|
| 514 | print '("ERROR: erfinv(x): x must be in [-1;1]")' |
---|
| 515 | stop |
---|
| 516 | end if |
---|
| 517 | |
---|
| 518 | if (x > -tiny(0.) .and. x < tiny(0.)) then |
---|
| 519 | r = 0 |
---|
| 520 | return |
---|
| 521 | end if |
---|
| 522 | |
---|
| 523 | if (x > 0) then |
---|
| 524 | sign_x = 1 |
---|
| 525 | else |
---|
| 526 | sign_x = -1 |
---|
| 527 | x = -x |
---|
| 528 | end if |
---|
| 529 | |
---|
| 530 | if (x <= 0.7) then |
---|
| 531 | x2 = x * x |
---|
| 532 | r = x * (((a3 * x2 + a2) * x2 + a1) * x2 + a0) |
---|
| 533 | r = r / (((b4 * x2 + b3) * x2 + b2) * x2 + b1) * x2 + b0 |
---|
| 534 | else |
---|
| 535 | y = sqrt(-log((1 - x) / 2)) |
---|
| 536 | r = (((c3 * y + c2) * y + c1) * y + c0) |
---|
| 537 | r = r / ((d2 * y + d1) * y + d0) |
---|
| 538 | end if |
---|
| 539 | |
---|
| 540 | r = r * sign_x |
---|
| 541 | x = x * sign_x |
---|
| 542 | |
---|
| 543 | r = r - (erf(r) - x) / (2 / sqrt(PI) * exp (-r * r)) |
---|
| 544 | r = r - (erf(r) - x) / (2 / sqrt(PI) * exp (-r * r)) |
---|
| 545 | |
---|
| 546 | return |
---|
| 547 | |
---|
| 548 | end function erfinv |
---|
| 549 | |
---|
| 550 | |
---|
| 551 | function interp(x_new, x, y) result(y_new) |
---|
| 552 | ! """ |
---|
| 553 | ! Interpolate array y of abscisse x to array y_new of abscisse x_new. |
---|
| 554 | ! |
---|
| 555 | ! inputs: |
---|
| 556 | ! x_new: abscisses on which y will be interpolated |
---|
| 557 | ! x: abscisses of y |
---|
| 558 | ! y: array to interpole |
---|
| 559 | ! |
---|
| 560 | ! output: |
---|
| 561 | ! y_new: interpolation of y(x) on abscisses x_new |
---|
| 562 | ! """ |
---|
| 563 | |
---|
| 564 | implicit none |
---|
| 565 | |
---|
| 566 | doubleprecision, dimension(:), intent(in) :: x, y, x_new |
---|
| 567 | doubleprecision, dimension(size(x_new)) :: y_new |
---|
| 568 | |
---|
| 569 | integer ::& |
---|
| 570 | i, & ! index |
---|
| 571 | j ! index |
---|
| 572 | |
---|
| 573 | i = 1 |
---|
| 574 | j = 1 |
---|
| 575 | y_new(:) = 0D0 |
---|
| 576 | |
---|
| 577 | if(size(x) /= size(y)) then |
---|
| 578 | print '("ERROR: interp: x and y must have the same size")' |
---|
| 579 | stop |
---|
| 580 | end if |
---|
| 581 | |
---|
| 582 | if(x(1) < x(size(x))) then ! ascending numerical order x array |
---|
| 583 | if(x_new(1) < x(1) .or. x_new(size(x_new)) > x(size(x))) then |
---|
| 584 | print '("ERROR: interp: x_new is outside x boundaries")' |
---|
| 585 | print *, x(1), ' < ' , x_new(1), '--', x_new(size(x_new)), ' < ', x(size(x)) |
---|
| 586 | stop |
---|
| 587 | end if |
---|
| 588 | |
---|
| 589 | do while (i <= size(x_new)) |
---|
| 590 | if(x_new(i) <= x(j + 1)) then |
---|
| 591 | y_new(i) = (x_new(i) - x(j)) * (y(j + 1) - y(j)) / (x(j + 1) - x(j)) + y(j) |
---|
| 592 | |
---|
| 593 | i = i + 1 |
---|
| 594 | else |
---|
| 595 | j = j + 1 |
---|
| 596 | end if |
---|
| 597 | end do |
---|
| 598 | else ! descending numerical order x array |
---|
| 599 | if(x_new(1) > x(1) .or. x_new(size(x_new)) < x(size(x))) then |
---|
| 600 | print '("ERROR: interp: x_new is outside x boundaries")' |
---|
| 601 | print *,x(1), ' > ', x_new(1), '--', x_new(size(x_new)), ' > ', x(size(x)) |
---|
| 602 | stop |
---|
| 603 | end if |
---|
| 604 | |
---|
| 605 | do while (i <= size(x_new)) |
---|
| 606 | if(x_new(i) >= x(j + 1)) then |
---|
| 607 | y_new(i) = (x_new(i) - x(j)) * (y(j + 1) - y(j)) / (x(j + 1) - x(j)) + y(j) |
---|
| 608 | |
---|
| 609 | i = i + 1 |
---|
| 610 | else |
---|
| 611 | j = j + 1 |
---|
| 612 | end if |
---|
| 613 | end do |
---|
| 614 | end if |
---|
| 615 | |
---|
| 616 | return |
---|
| 617 | end function interp |
---|
| 618 | |
---|
| 619 | |
---|
| 620 | function interp_fast(x_new, x, y) result(y_new) |
---|
| 621 | ! """ |
---|
| 622 | ! Interpolate array y of abscisse x to array y_new of abscisse x_new. |
---|
| 623 | ! Assumptions (no check performed): |
---|
| 624 | ! - x_new is within x boundaries |
---|
| 625 | ! - x and y have the same size |
---|
| 626 | ! |
---|
| 627 | ! inputs: |
---|
| 628 | ! x_new: abscisses on which y will be interpolated |
---|
| 629 | ! x: abscisses of y |
---|
| 630 | ! y: array to interpole |
---|
| 631 | ! |
---|
| 632 | ! output: |
---|
| 633 | ! y_new: interpolation of y(x) on abscisses x_new |
---|
| 634 | ! """ |
---|
| 635 | |
---|
| 636 | implicit none |
---|
| 637 | |
---|
| 638 | doubleprecision, dimension(:), intent(in) :: x, y, x_new |
---|
| 639 | doubleprecision, dimension(size(x_new)) :: y_new |
---|
| 640 | |
---|
| 641 | integer ::& |
---|
| 642 | i, & ! index |
---|
| 643 | j ! index |
---|
| 644 | |
---|
| 645 | i = 1 |
---|
| 646 | j = 1 |
---|
| 647 | y_new(:) = 0D0 |
---|
| 648 | |
---|
| 649 | if(x(1) < x(size(x))) then ! ascending numerical order x array |
---|
| 650 | do while(i <= size(x_new)) |
---|
| 651 | if(x_new(i) <= x(j + 1)) then |
---|
| 652 | y_new(i) = (x_new(i) - x(j)) * (y(j + 1) - y(j)) / (x(j + 1) - x(j)) + y(j) |
---|
| 653 | |
---|
| 654 | i = i + 1 |
---|
| 655 | else |
---|
| 656 | j = j + 1 |
---|
| 657 | end if |
---|
| 658 | end do |
---|
| 659 | else ! descending numerical order x array |
---|
| 660 | do while(i <= size(x_new)) |
---|
| 661 | if(x_new(i) >= x(j + 1)) then |
---|
| 662 | y_new(i) = (x_new(i) - x(j)) * (y(j + 1) - y(j)) / (x(j + 1) - x(j)) + y(j) |
---|
| 663 | |
---|
| 664 | i = i + 1 |
---|
| 665 | else |
---|
| 666 | j = j + 1 |
---|
| 667 | end if |
---|
| 668 | end do |
---|
| 669 | end if |
---|
| 670 | |
---|
| 671 | return |
---|
| 672 | end function interp_fast |
---|
| 673 | |
---|
| 674 | |
---|
| 675 | function interp_ex(x_new, x, y) result(y_new) |
---|
| 676 | ! """ |
---|
| 677 | ! Interpolate array y of abscisse x to array y_new of abscisse x_new. |
---|
| 678 | ! The points outside x range are linearly extrapolated. |
---|
| 679 | ! Data must be in strict increasing order. |
---|
| 680 | ! |
---|
| 681 | ! inputs: |
---|
| 682 | ! x_new: abscisses on which y will be interpolated |
---|
| 683 | ! x: abscisses of y |
---|
| 684 | ! y: array to interpole |
---|
| 685 | ! |
---|
| 686 | ! output: |
---|
| 687 | ! y_new: interpolation of y(x) on abscisses x_new |
---|
| 688 | ! """ |
---|
| 689 | |
---|
| 690 | implicit none |
---|
| 691 | |
---|
| 692 | doubleprecision, dimension(:), intent(in) :: x, y, x_new |
---|
| 693 | doubleprecision, dimension(size(x_new)) :: y_new |
---|
| 694 | |
---|
| 695 | integer ::& |
---|
| 696 | i, & ! index |
---|
| 697 | i_max, & |
---|
| 698 | i_min, & |
---|
| 699 | n ! index |
---|
| 700 | |
---|
| 701 | n = size(x) |
---|
| 702 | i_min = 0 |
---|
| 703 | i_max = 0 |
---|
| 704 | |
---|
| 705 | ! Check order |
---|
| 706 | if(x(1) > x(n)) then |
---|
| 707 | if(x_new(1) < x_new(size(x_new))) then |
---|
| 708 | write(*, '("Error: interp_ex: x_new numerical order must be the same than x (descending)")') |
---|
| 709 | |
---|
| 710 | stop |
---|
| 711 | end if |
---|
| 712 | |
---|
| 713 | ! Find where x_new is in data range |
---|
| 714 | do i = 1, size(x_new) |
---|
| 715 | if(x_new(i) <= x(1) .and. i_min == 0) then |
---|
| 716 | i_min = i |
---|
| 717 | end if |
---|
| 718 | |
---|
| 719 | if(x_new(i) >= x(n)) then |
---|
| 720 | i_max = i |
---|
| 721 | end if |
---|
| 722 | |
---|
| 723 | if(x_new(i) < x(n)) then |
---|
| 724 | exit |
---|
| 725 | end if |
---|
| 726 | end do |
---|
| 727 | else |
---|
| 728 | if(x_new(1) > x_new(size(x_new))) then |
---|
| 729 | write(*, '("Error: interp_ex: x_new numerical order must be the same than x (ascending)")') |
---|
| 730 | |
---|
| 731 | stop |
---|
| 732 | end if |
---|
| 733 | |
---|
| 734 | ! Find where x_new is in data range |
---|
| 735 | do i = 1, size(x_new) |
---|
| 736 | if(x_new(i) >= x(1) .and. i_min == 0) then |
---|
| 737 | i_min = i |
---|
| 738 | end if |
---|
| 739 | |
---|
| 740 | if(x_new(i) <= x(n)) then |
---|
| 741 | i_max = i |
---|
| 742 | end if |
---|
| 743 | |
---|
| 744 | if(x_new(i) > x(n)) then |
---|
| 745 | exit |
---|
| 746 | end if |
---|
| 747 | end do |
---|
| 748 | end if |
---|
| 749 | |
---|
| 750 | ! Interpolation within data range |
---|
| 751 | if(i_min > 0 .and. i_max > 0) then |
---|
| 752 | y_new(i_min:i_max) = interp_fast(x_new(i_min:i_max), x, y) |
---|
| 753 | end if |
---|
| 754 | |
---|
| 755 | if(i_min == 0 .or. i_max == 0) then |
---|
| 756 | if(i_min > 0) then |
---|
| 757 | ! Higher bound extrapolation |
---|
| 758 | do i = 1, size(x_new) |
---|
| 759 | y_new(i) = (x_new(i) - x(n - 1)) * (y(n) - y(n - 1)) / (x(n) - x(n - 1)) + y(n - 1) |
---|
| 760 | end do |
---|
| 761 | else if (i_max > 0) then |
---|
| 762 | ! Lower bound extrapolation |
---|
| 763 | do i = 1, size(x_new) |
---|
| 764 | y_new(i) = (x_new(i) - x(1)) * (y(2) - y(1)) / (x(2) - x(1)) + y(1) |
---|
| 765 | end do |
---|
| 766 | else |
---|
| 767 | write(*, '("Error: interp_ex: unable to determine boundary")') |
---|
| 768 | |
---|
| 769 | stop |
---|
| 770 | end if |
---|
| 771 | else |
---|
| 772 | ! Lower bound extrapolation |
---|
| 773 | if(i_min > 1) then |
---|
| 774 | do i = 1, i_min - 1 |
---|
| 775 | y_new(i) = (x_new(i) - x(1)) * (y(2) - y(1)) / (x(2) - x(1)) + y(1) |
---|
| 776 | end do |
---|
| 777 | end if |
---|
| 778 | |
---|
| 779 | ! Higher bound extrapolation |
---|
| 780 | if(i_max < size(x_new)) then |
---|
| 781 | do i = i_max + 1, size(x_new) |
---|
| 782 | y_new(i) = (x_new(i) - x(n - 1)) * (y(n) - y(n - 1)) / (x(n) - x(n - 1)) + y(n - 1) |
---|
| 783 | end do |
---|
| 784 | end if |
---|
| 785 | end if |
---|
| 786 | end function interp_ex |
---|
| 787 | |
---|
| 788 | |
---|
| 789 | function interp_ex_0d(x_new, x, y) result(y_new) |
---|
| 790 | ! """ |
---|
| 791 | ! Interpolate array y of abscisse x to value y_new at x_new. |
---|
| 792 | ! This function only takes a double as x_new, not an array. Return a double, not an array. |
---|
| 793 | ! The points outside x range are linearly extrapolated. |
---|
| 794 | ! Data must be in strict increasing order. |
---|
| 795 | ! |
---|
| 796 | ! inputs: |
---|
| 797 | ! x_new: value at which y will be interpolated |
---|
| 798 | ! x: abscisses of y |
---|
| 799 | ! y: array to interpole |
---|
| 800 | ! |
---|
| 801 | ! output: |
---|
| 802 | ! y_new: interpolation of y(x) at x_new |
---|
| 803 | ! """ |
---|
| 804 | implicit none |
---|
| 805 | |
---|
| 806 | doubleprecision, dimension(:), intent(in) :: x, y |
---|
| 807 | doubleprecision :: x_new, y_new |
---|
| 808 | |
---|
| 809 | doubleprecision, dimension(1) :: arr_tmp |
---|
| 810 | |
---|
| 811 | arr_tmp = interp_ex([x_new], x, y) |
---|
| 812 | y_new = arr_tmp(1) |
---|
| 813 | end function interp_ex_0d |
---|
| 814 | |
---|
| 815 | |
---|
| 816 | function mean_restep(x_new, x, y) result(y_new) |
---|
| 817 | ! """ |
---|
| 818 | ! Change the values of y of abscisse x so that the new values y_new of abcisse x_new are the mean of y |
---|
| 819 | ! between a step of x_new centered on x_new. |
---|
| 820 | ! Both x and x_new must be regularly spaced. |
---|
| 821 | ! Example: |
---|
| 822 | ! >>> x_new = [0, 3, 6] |
---|
| 823 | ! >>> x = [0, 1, 2, 3, 4, 5, 6] |
---|
| 824 | ! >>> y = [1, 2, 4, 3, 6, 2, 0] |
---|
| 825 | ! >>> mean_restep(x_new, x, y) |
---|
| 826 | ! >>> [1.50, 4.33, 1.00] |
---|
| 827 | ! |
---|
| 828 | ! inputs: |
---|
| 829 | ! x_new: abscisses on which y will be interpolated |
---|
| 830 | ! x: abscisses of y |
---|
| 831 | ! y: array to interpole |
---|
| 832 | ! |
---|
| 833 | ! output: |
---|
| 834 | ! y_new: interpolation of y(x) on abscisses x_new |
---|
| 835 | ! """ |
---|
| 836 | implicit none |
---|
| 837 | |
---|
| 838 | doubleprecision, dimension(:), intent(in) :: x, y, x_new |
---|
| 839 | doubleprecision, dimension(size(x_new)) :: y_new |
---|
| 840 | |
---|
| 841 | integer :: & |
---|
| 842 | i, & ! index |
---|
| 843 | j, & ! index |
---|
| 844 | n ! number of elements |
---|
| 845 | |
---|
| 846 | doubleprecision :: & |
---|
| 847 | sum_x, & |
---|
| 848 | min_x, & |
---|
| 849 | max_x |
---|
| 850 | |
---|
| 851 | call check_inputs() |
---|
| 852 | |
---|
| 853 | j = 1 |
---|
| 854 | |
---|
| 855 | ! First value |
---|
| 856 | i = 1 |
---|
| 857 | n = 0 |
---|
| 858 | sum_x = 0D0 |
---|
| 859 | max_x = x_new(i) + (x_new(i + 1) - x_new(i)) / 2D0 |
---|
| 860 | min_x = x_new(i) |
---|
| 861 | |
---|
| 862 | do while (x(j) < max_x .and. j < size(x)) |
---|
| 863 | if (x(j) >= min_x) then |
---|
| 864 | sum_x = sum_x + y(j) |
---|
| 865 | n = n + 1 |
---|
| 866 | end if |
---|
| 867 | |
---|
| 868 | j = j + 1 |
---|
| 869 | end do |
---|
| 870 | |
---|
| 871 | y_new(i) = sum_x / n |
---|
| 872 | |
---|
| 873 | ! Intermediate values |
---|
| 874 | do i = 2, size(x_new) - 1 |
---|
| 875 | n = 0 |
---|
| 876 | sum_x = 0D0 |
---|
| 877 | max_x = x_new(i) + (x_new(i + 1) - x_new(i)) / 2D0 |
---|
| 878 | min_x = x_new(i) - (x_new(i) - x_new(i - 1)) / 2D0 |
---|
| 879 | |
---|
| 880 | do while (x(j) < max_x .and. j < size(x)) |
---|
| 881 | if (x(j) >= min_x) then |
---|
| 882 | sum_x = sum_x + y(j) |
---|
| 883 | n = n + 1 |
---|
| 884 | end if |
---|
| 885 | |
---|
| 886 | j = j + 1 |
---|
| 887 | end do |
---|
| 888 | |
---|
| 889 | y_new(i) = sum_x / n |
---|
| 890 | end do |
---|
| 891 | |
---|
| 892 | ! Last value |
---|
| 893 | i = size(x_new) |
---|
| 894 | n = 0 |
---|
| 895 | sum_x = 0D0 |
---|
| 896 | max_x = x_new(i) |
---|
| 897 | min_x = x_new(i) - (x_new(i) - x_new(i - 1)) / 2D0 |
---|
| 898 | |
---|
| 899 | do while (x(j) < max_x .and. j < size(x)) |
---|
| 900 | if (x(j) >= min_x) then |
---|
| 901 | sum_x = sum_x + y(j) |
---|
| 902 | n = n + 1 |
---|
| 903 | end if |
---|
| 904 | |
---|
| 905 | j = j + 1 |
---|
| 906 | end do |
---|
| 907 | |
---|
| 908 | if(n == 0) then |
---|
| 909 | y_new(i) = y(j) |
---|
| 910 | else |
---|
| 911 | y_new(i) = sum_x / n |
---|
| 912 | end if |
---|
| 913 | |
---|
| 914 | return |
---|
| 915 | |
---|
| 916 | contains |
---|
| 917 | subroutine check_inputs() |
---|
| 918 | implicit none |
---|
| 919 | |
---|
| 920 | if (size(x) /= size(y)) then |
---|
| 921 | print '("ERROR: mean_restep: x and y must have the same size")' |
---|
| 922 | stop |
---|
| 923 | end if |
---|
| 924 | |
---|
| 925 | if (x(1) > x(size(x))) then |
---|
| 926 | print '("ERROR: mean_restep: x must be in increasing order")' |
---|
| 927 | stop |
---|
| 928 | end if |
---|
| 929 | |
---|
| 930 | if(x_new(1) < x(1) .or. x_new(size(x_new)) > x(size(x))) then |
---|
| 931 | print '("ERROR: interp: x_new is outside x boundaries")' |
---|
| 932 | print *,x_new(1),x(1),x_new(size(x_new)),x(size(x)) |
---|
| 933 | stop |
---|
| 934 | end if |
---|
| 935 | end subroutine check_inputs |
---|
| 936 | end function mean_restep |
---|
| 937 | |
---|
| 938 | |
---|
| 939 | function reverse_array(array) result(reversed_array) |
---|
| 940 | ! """ |
---|
| 941 | ! Reverse a 1-D double precision array. |
---|
| 942 | ! """ |
---|
| 943 | implicit none |
---|
| 944 | |
---|
| 945 | doubleprecision, dimension(:), intent(in) :: array |
---|
| 946 | integer :: head, tail |
---|
| 947 | |
---|
| 948 | doubleprecision, dimension(size(array)) :: reversed_array |
---|
| 949 | |
---|
| 950 | head = 1 |
---|
| 951 | tail = size(array) |
---|
| 952 | reversed_array(:) = 0d0 |
---|
| 953 | |
---|
| 954 | do while(head < tail) |
---|
| 955 | reversed_array(tail) = array(head) |
---|
| 956 | reversed_array(head) = array(tail) |
---|
| 957 | head = head + 1 |
---|
| 958 | tail = tail - 1 |
---|
| 959 | end do |
---|
| 960 | |
---|
| 961 | return |
---|
| 962 | end function reverse_array |
---|
| 963 | |
---|
| 964 | |
---|
| 965 | function voigt(x, y) |
---|
| 966 | ! """ |
---|
| 967 | ! Calculate the voigt function using an algorithm written by Humlicek JQSRT, 27, 437 (1982). |
---|
| 968 | ! Calculate the complex probability function W(z) = exp(-z^2) * erfc(-z^2) in the superior complex plan |
---|
| 969 | ! (i.e. for y >= 0). The real part of this function is the Voigt function. |
---|
| 970 | ! |
---|
| 971 | ! The article shows that the rational function W can be written as (Eq. 11): |
---|
| 972 | ! W_n(z) = (-iz) * sum_{k=1}^{n/2}(c_k((-iz)^2)^{k-1}) / ((-iz)^n + sum_{k=1}^{n/2}(d_k((-iz)^2)^{k-1})) |
---|
| 973 | ! With z = x + iy, (-iz = t). The coefficent c and d are given in Table 3 of the article. |
---|
| 974 | ! The maximal relative error on the imaginary and real parts is < 1e-4. |
---|
| 975 | ! |
---|
| 976 | ! inputs: |
---|
| 977 | ! x: real part coordinate |
---|
| 978 | ! y: imaginary part coordinate |
---|
| 979 | ! |
---|
| 980 | ! output: |
---|
| 981 | ! voigt: value of the voigt function at z = x + iy |
---|
| 982 | ! """ |
---|
| 983 | implicit none |
---|
| 984 | |
---|
| 985 | doubleprecision, dimension(:), intent(in) :: x |
---|
| 986 | doubleprecision, intent(in) :: y |
---|
| 987 | doubleprecision, dimension(size(x)) :: voigt |
---|
| 988 | |
---|
| 989 | integer :: & |
---|
| 990 | i ! index |
---|
| 991 | |
---|
| 992 | double precision, dimension(size(x)) :: & |
---|
| 993 | s ! intermediate value |
---|
| 994 | |
---|
| 995 | complex (kind=8), dimension(size(x)) :: & |
---|
| 996 | t, & ! intermediate value |
---|
| 997 | u, & ! intermediate value |
---|
| 998 | w ! intermediate value |
---|
| 999 | |
---|
| 1000 | do i=1,size(x) |
---|
| 1001 | t(i) = cmplx(y, -x(i), kind=8) |
---|
| 1002 | end do |
---|
| 1003 | |
---|
| 1004 | s(:) = abs(x(:)) + y |
---|
| 1005 | |
---|
| 1006 | where (s >= 15D0) ! n = 2 (region i) |
---|
| 1007 | w = t * sqrtpi / (0.5D0 + t**2) |
---|
| 1008 | voigt = dble(w * sqrtpi) |
---|
| 1009 | elsewhere (s >= 5.5D0) ! n = 4 (region II) |
---|
| 1010 | u = t**2 |
---|
| 1011 | w = t * (1.410474D0 + u * sqrtpi) / (0.75D0 + u * (3D0 + u)) |
---|
| 1012 | voigt = dble(w * sqrtpi) |
---|
| 1013 | elsewhere (0.195D0 * abs(x) - 0.176D0 <= y) ! n = 6 (region III) |
---|
| 1014 | w = (16.4955D0 + t * (20.20933D0 + t * (11.96482D0 + t * (3.778987D0 + t * 0.5642236D0)))) / & |
---|
| 1015 | (16.4955D0 + t * (38.82363D0 + t * (39.27121D0 + t * (21.69274D0 + t * (6.699398D0 + t))))) |
---|
| 1016 | voigt = dble(w * sqrtpi) |
---|
| 1017 | elsewhere ! n = 8 (region IV) |
---|
| 1018 | u = t**2 |
---|
| 1019 | w = exp(u) - & |
---|
| 1020 | t * (36183.31D0 - & |
---|
| 1021 | u * (3321.9905D0 - & |
---|
| 1022 | u * (1540.787D0 - & |
---|
| 1023 | u * (219.0313D0 - u *(35.76683D0 - u * (1.320522D0 - u * sqrtpi)))))) / & |
---|
| 1024 | (32066.6D0 - & |
---|
| 1025 | u * (24322.84D0 - & |
---|
| 1026 | u * (9022.228D0 - & |
---|
| 1027 | u * (2186.181D0 - u * (364.2191D0 - u * (61.57037D0 - u * (1.841439D0 - u))))))) |
---|
| 1028 | voigt = dble(w * sqrtpi) |
---|
| 1029 | end where |
---|
| 1030 | |
---|
| 1031 | return |
---|
| 1032 | end function voigt |
---|
| 1033 | |
---|
| 1034 | |
---|
| 1035 | doubleprecision function voigt_from_data(x, y, v) result(voigt) |
---|
| 1036 | ! """ |
---|
| 1037 | ! """ |
---|
| 1038 | implicit none |
---|
| 1039 | |
---|
| 1040 | doubleprecision, intent(in) :: & |
---|
| 1041 | x, & |
---|
| 1042 | y, & |
---|
| 1043 | v(400, 100) |
---|
| 1044 | |
---|
| 1045 | doubleprecision, parameter :: & |
---|
| 1046 | dp = 0.0025d0, & |
---|
| 1047 | ds = 0.01d0, & |
---|
| 1048 | yl = 1d0 / 99d0, & |
---|
| 1049 | xl = 399d0 |
---|
| 1050 | |
---|
| 1051 | integer :: & |
---|
| 1052 | ip, & |
---|
| 1053 | is |
---|
| 1054 | |
---|
| 1055 | doubleprecision :: & |
---|
| 1056 | a, & |
---|
| 1057 | b, & |
---|
| 1058 | pressure_space, & |
---|
| 1059 | x2, & |
---|
| 1060 | x2y2, & |
---|
| 1061 | s, & |
---|
| 1062 | y2, & |
---|
| 1063 | f |
---|
| 1064 | |
---|
| 1065 | if(y > 5.4d0) then |
---|
| 1066 | x2 = x * x |
---|
| 1067 | y2 = y * y |
---|
| 1068 | x2y2 = x2 + y2 |
---|
| 1069 | voigt = sqrtpi * y * (1d0 + (3d0 * x2 - y2) / (2d0 * x2y2 * x2y2) + & |
---|
| 1070 | 0.75d0 * (5d0 * x2 * x2 + y2 * y2 - 10d0 * x2 * y2) / (x2y2 * x2y2 * x2y2 * x2y2)) / x2y2 |
---|
| 1071 | else |
---|
| 1072 | if(y <= yl) then |
---|
| 1073 | b = exp(-x * x) |
---|
| 1074 | pressure_space = 1d0 / (1d0 + x) |
---|
| 1075 | ip = int(pressure_space / dp) |
---|
| 1076 | f = pressure_space / dp - ip |
---|
| 1077 | if(ip >= 1) then |
---|
| 1078 | if(ip < 400) then |
---|
| 1079 | a = (1d0 - f) * v(ip, 1) + f * v(ip + 1, 1) |
---|
| 1080 | else |
---|
| 1081 | a = v(ip, 1) |
---|
| 1082 | end if |
---|
| 1083 | else |
---|
| 1084 | a = f * v(ip + 1, 1) |
---|
| 1085 | end if |
---|
| 1086 | a = a * a |
---|
| 1087 | f = y / yl |
---|
| 1088 | voigt = (1d0 - f) * b + f * a |
---|
| 1089 | else |
---|
| 1090 | if(x > xl) then |
---|
| 1091 | s = y / (1d0 + y) |
---|
| 1092 | is = int(s / ds) |
---|
| 1093 | b = v(1, is) * xl / x |
---|
| 1094 | b = b * b |
---|
| 1095 | a = v(1, is + 1) * xl / x |
---|
| 1096 | a = a * a |
---|
| 1097 | f = s / ds - is |
---|
| 1098 | voigt = (1d0 - f) * b + f * a |
---|
| 1099 | else |
---|
| 1100 | s = y / (1d0 + y) |
---|
| 1101 | pressure_space = 1d0 / (1d0 + x) |
---|
| 1102 | is = int(s / ds) |
---|
| 1103 | ip = int(pressure_space / dp) |
---|
| 1104 | f = pressure_space / dp - ip |
---|
| 1105 | if(ip < 400) then |
---|
| 1106 | b = (1d0 - f) * v(ip, is) + f * v(ip + 1, is) |
---|
| 1107 | a = (1d0 - f) * v(ip, is + 1) + f * v(ip + 1, is + 1) |
---|
| 1108 | else |
---|
| 1109 | b = v(ip, is) |
---|
| 1110 | a = v(ip, is + 1) |
---|
| 1111 | end if |
---|
| 1112 | b = b * b |
---|
| 1113 | a = a * a |
---|
| 1114 | f = s / ds - is |
---|
| 1115 | voigt = (1d0 - f) * b + f * a |
---|
| 1116 | end if |
---|
| 1117 | end if |
---|
| 1118 | end if |
---|
| 1119 | |
---|
| 1120 | return |
---|
| 1121 | end function voigt_from_data |
---|
| 1122 | |
---|
| 1123 | |
---|
| 1124 | recursive subroutine fft(x) |
---|
| 1125 | ! """ |
---|
| 1126 | ! Calculate the Cooley-Tukey FFT functions. |
---|
| 1127 | ! |
---|
| 1128 | ! input: |
---|
| 1129 | ! x: double complex vector of size 2^n |
---|
| 1130 | ! |
---|
| 1131 | ! notes: |
---|
| 1132 | ! Source: https://rosettacode.org/wiki/Fast_Fourier_transform#Fortran |
---|
| 1133 | ! """ |
---|
| 1134 | implicit none |
---|
| 1135 | |
---|
| 1136 | complex(kind=dp), dimension(:), intent(inout) :: x |
---|
| 1137 | complex(kind=dp) ::& |
---|
| 1138 | t |
---|
| 1139 | |
---|
| 1140 | integer ::& |
---|
| 1141 | i, & ! index |
---|
| 1142 | n ! size of array x |
---|
| 1143 | |
---|
| 1144 | complex(kind=dp), dimension(:), allocatable ::& |
---|
| 1145 | even, & ! even-number indexed values of x |
---|
| 1146 | odd ! odd-number indexed values of x |
---|
| 1147 | |
---|
| 1148 | n = size(x) |
---|
| 1149 | |
---|
| 1150 | if(n <= 1) return |
---|
| 1151 | |
---|
| 1152 | allocate(odd((n+1)/2)) |
---|
| 1153 | allocate(even(n/2)) |
---|
| 1154 | |
---|
| 1155 | ! divide |
---|
| 1156 | odd(:) = x(1:n:2) |
---|
| 1157 | even(:) = x(2:n:2) |
---|
| 1158 | |
---|
| 1159 | ! conquer |
---|
| 1160 | call fft(odd) |
---|
| 1161 | call fft(even) |
---|
| 1162 | |
---|
| 1163 | ! combine |
---|
| 1164 | do i = 1, n/2 |
---|
| 1165 | t = exp(cmplx(0.0_dp, -2.0_dp * pi * real(i-1, dp) / real(n, dp), kind=dp)) * even(i) |
---|
| 1166 | x(i) = odd(i) + t |
---|
| 1167 | x(i+n/2) = odd(i) - t |
---|
| 1168 | end do |
---|
| 1169 | |
---|
| 1170 | deallocate(odd) |
---|
| 1171 | deallocate(even) |
---|
| 1172 | |
---|
| 1173 | end subroutine fft |
---|
| 1174 | |
---|
| 1175 | |
---|
| 1176 | recursive subroutine quicksort(array) |
---|
| 1177 | ! """ |
---|
| 1178 | ! Sort double precision numbers into ascending numerical order. |
---|
| 1179 | ! |
---|
| 1180 | ! input: |
---|
| 1181 | ! array: double precision array to sort |
---|
| 1182 | ! |
---|
| 1183 | ! output: |
---|
| 1184 | ! array: sorted double precision array |
---|
| 1185 | ! |
---|
| 1186 | ! notes: |
---|
| 1187 | ! Author: t-nissie, some tweaks by 1AdAstra1 |
---|
| 1188 | ! Source: https://gist.github.com/t-nissie/479f0f16966925fa29ea |
---|
| 1189 | ! """ |
---|
| 1190 | implicit none |
---|
| 1191 | |
---|
| 1192 | double precision, intent(inout), dimension(:) :: array |
---|
| 1193 | |
---|
| 1194 | double precision :: & |
---|
| 1195 | x, & ! pivot point |
---|
| 1196 | tmp ! temporary value |
---|
| 1197 | |
---|
| 1198 | integer :: & |
---|
| 1199 | first = 1, & ! index of the beginning of the array |
---|
| 1200 | last, & ! index of the end of the array |
---|
| 1201 | insertion_size_threshold = 32 |
---|
| 1202 | |
---|
| 1203 | integer :: & |
---|
| 1204 | i, & |
---|
| 1205 | j |
---|
| 1206 | |
---|
| 1207 | last = size(array, 1) |
---|
| 1208 | |
---|
| 1209 | if(last < insertion_size_threshold) then ! use insertion sort on small arrays |
---|
| 1210 | do i = 2, last |
---|
| 1211 | tmp = array(i) |
---|
| 1212 | |
---|
| 1213 | do j = i - 1, 1, -1 |
---|
| 1214 | if(array(j) < tmp) then |
---|
| 1215 | exit |
---|
| 1216 | else |
---|
| 1217 | array(j + 1) = array(j) |
---|
| 1218 | end if |
---|
| 1219 | end do |
---|
| 1220 | |
---|
| 1221 | array(j + 1) = tmp |
---|
| 1222 | end do |
---|
| 1223 | else |
---|
| 1224 | x = array((first+last)/2) |
---|
| 1225 | i = first |
---|
| 1226 | j = last |
---|
| 1227 | |
---|
| 1228 | do |
---|
| 1229 | do while (array(i) < x) |
---|
| 1230 | i = i + 1 |
---|
| 1231 | end do |
---|
| 1232 | |
---|
| 1233 | do while (x < array(j)) |
---|
| 1234 | j = j - 1 |
---|
| 1235 | end do |
---|
| 1236 | |
---|
| 1237 | if (i >= j) then |
---|
| 1238 | exit |
---|
| 1239 | end if |
---|
| 1240 | |
---|
| 1241 | tmp = array(i) |
---|
| 1242 | array(i) = array(j) |
---|
| 1243 | array(j) = tmp |
---|
| 1244 | |
---|
| 1245 | i = i + 1 |
---|
| 1246 | j = j - 1 |
---|
| 1247 | end do |
---|
| 1248 | |
---|
| 1249 | if (first < i - 1) call quicksort(array(first:i-1)) |
---|
| 1250 | if (j + 1 < last) call quicksort(array(j+1:last)) |
---|
| 1251 | end if |
---|
| 1252 | end subroutine quicksort |
---|
| 1253 | |
---|
| 1254 | |
---|
| 1255 | recursive subroutine quicksort_index(array, sorted_index) |
---|
| 1256 | ! """ |
---|
| 1257 | ! Sort double precision numbers into ascending numerical order. |
---|
| 1258 | ! |
---|
| 1259 | ! input: |
---|
| 1260 | ! array: double precision array to sort |
---|
| 1261 | ! |
---|
| 1262 | ! output: |
---|
| 1263 | ! sorted_index: sorted index of the array |
---|
| 1264 | ! """ |
---|
| 1265 | implicit none |
---|
| 1266 | |
---|
| 1267 | double precision, intent(inout), dimension(:) :: array |
---|
| 1268 | integer, intent(out), dimension(:) :: sorted_index |
---|
| 1269 | |
---|
| 1270 | integer :: & |
---|
| 1271 | i |
---|
| 1272 | |
---|
| 1273 | do i = 1, size(array) |
---|
| 1274 | sorted_index(i) = i |
---|
| 1275 | end do |
---|
| 1276 | |
---|
| 1277 | call partition(array, sorted_index) |
---|
| 1278 | |
---|
| 1279 | contains |
---|
| 1280 | recursive subroutine partition(array, sorted_index) |
---|
| 1281 | implicit none |
---|
| 1282 | |
---|
| 1283 | integer, intent(inout), dimension(:) :: sorted_index |
---|
| 1284 | double precision, intent(inout), dimension(:) :: array |
---|
| 1285 | |
---|
| 1286 | integer :: & |
---|
| 1287 | tmp_i |
---|
| 1288 | |
---|
| 1289 | double precision :: & |
---|
| 1290 | x, & ! pivot point |
---|
| 1291 | tmp ! temporary value |
---|
| 1292 | |
---|
| 1293 | integer :: & |
---|
| 1294 | first = 1, & ! index of the beginning of the array |
---|
| 1295 | last, & ! index of the end of the array |
---|
| 1296 | insertion_size_threshold = 128 |
---|
| 1297 | |
---|
| 1298 | integer :: & |
---|
| 1299 | i, & |
---|
| 1300 | j |
---|
| 1301 | |
---|
| 1302 | last = size(array, 1) |
---|
| 1303 | |
---|
| 1304 | if(last < insertion_size_threshold) then ! use insertion sort on small arrays |
---|
| 1305 | do i = 2, last |
---|
| 1306 | tmp = array(i) |
---|
| 1307 | tmp_i = sorted_index(i) |
---|
| 1308 | |
---|
| 1309 | do j = i - 1, 1, -1 |
---|
| 1310 | if(array(j) < tmp) then |
---|
| 1311 | exit |
---|
| 1312 | else |
---|
| 1313 | array(j + 1) = array(j) |
---|
| 1314 | sorted_index(j + 1) = sorted_index(j) |
---|
| 1315 | end if |
---|
| 1316 | end do |
---|
| 1317 | |
---|
| 1318 | array(j + 1) = tmp |
---|
| 1319 | sorted_index(j + 1) = tmp_i |
---|
| 1320 | end do |
---|
| 1321 | else |
---|
| 1322 | x = array((first + last) / 2) |
---|
| 1323 | i = first |
---|
| 1324 | j = last |
---|
| 1325 | |
---|
| 1326 | do |
---|
| 1327 | do while (array(i) < x) |
---|
| 1328 | i = i + 1 |
---|
| 1329 | end do |
---|
| 1330 | |
---|
| 1331 | do while (x < array(j)) |
---|
| 1332 | j = j - 1 |
---|
| 1333 | end do |
---|
| 1334 | |
---|
| 1335 | if (i >= j) then |
---|
| 1336 | exit |
---|
| 1337 | end if |
---|
| 1338 | |
---|
| 1339 | tmp = array(i) |
---|
| 1340 | array(i) = array(j) |
---|
| 1341 | array(j) = tmp |
---|
| 1342 | |
---|
| 1343 | tmp_i = sorted_index(i) |
---|
| 1344 | sorted_index(i) = sorted_index(j) |
---|
| 1345 | sorted_index(j) = tmp_i |
---|
| 1346 | |
---|
| 1347 | i = i + 1 |
---|
| 1348 | j = j - 1 |
---|
| 1349 | end do |
---|
| 1350 | |
---|
| 1351 | if (first < i - 1) call partition(array(first:i-1), sorted_index(first:i-1)) |
---|
| 1352 | if (j + 1 < last) call partition(array(j+1:last), sorted_index(j+1:last)) |
---|
| 1353 | end if |
---|
| 1354 | end subroutine partition |
---|
| 1355 | end subroutine quicksort_index |
---|
| 1356 | |
---|
| 1357 | |
---|
| 1358 | subroutine reallocate_1Ddouble(double_array, new_size) |
---|
| 1359 | ! """ |
---|
| 1360 | ! Change the size of a 1D double precision array. |
---|
| 1361 | ! :param double_array: array to change the size |
---|
| 1362 | ! :param new_size: new size of the array |
---|
| 1363 | ! :return double_array: resized array |
---|
| 1364 | ! """ |
---|
| 1365 | implicit none |
---|
| 1366 | |
---|
| 1367 | integer, intent(in) :: new_size |
---|
| 1368 | doubleprecision, intent(inout), dimension(:), allocatable :: double_array |
---|
| 1369 | |
---|
| 1370 | integer :: & |
---|
| 1371 | size_array ! size of the array |
---|
| 1372 | doubleprecision, dimension(:), allocatable :: & |
---|
| 1373 | tmp ! temporary array |
---|
| 1374 | |
---|
| 1375 | size_array = size(double_array) |
---|
| 1376 | |
---|
| 1377 | call move_alloc(double_array, tmp) |
---|
| 1378 | allocate(double_array(new_size)) |
---|
| 1379 | |
---|
| 1380 | double_array(:) = 0d0 |
---|
| 1381 | |
---|
| 1382 | size_array = min(new_size, size_array) |
---|
| 1383 | |
---|
| 1384 | double_array(:size_array) = tmp(:size_array) |
---|
| 1385 | end subroutine reallocate_1Ddouble |
---|
| 1386 | |
---|
| 1387 | |
---|
| 1388 | subroutine reallocate_1Dinteger(integer_array, new_size) |
---|
| 1389 | ! """ |
---|
| 1390 | ! Change the size of a 1D integer array. |
---|
| 1391 | ! :param integer_array: array to change the size |
---|
| 1392 | ! :param new_size: new size of the array |
---|
| 1393 | ! :return integer_array: resized array |
---|
| 1394 | ! """ |
---|
| 1395 | implicit none |
---|
| 1396 | |
---|
| 1397 | integer, intent(in) :: new_size |
---|
| 1398 | integer, intent(inout), dimension(:), allocatable :: integer_array |
---|
| 1399 | |
---|
| 1400 | integer :: & |
---|
| 1401 | size_array ! size of the array |
---|
| 1402 | integer, dimension(:), allocatable :: & |
---|
| 1403 | tmp ! temporary array |
---|
| 1404 | |
---|
| 1405 | size_array = size(integer_array) |
---|
| 1406 | |
---|
| 1407 | call move_alloc(integer_array, tmp) |
---|
| 1408 | allocate(integer_array(new_size)) |
---|
| 1409 | |
---|
| 1410 | integer_array(:) = 0 |
---|
| 1411 | |
---|
| 1412 | size_array = min(new_size, size_array) |
---|
| 1413 | |
---|
| 1414 | integer_array(:size_array) = tmp(:size_array) |
---|
| 1415 | end subroutine reallocate_1Dinteger |
---|
| 1416 | |
---|
| 1417 | |
---|
| 1418 | subroutine reallocate_2Ddouble(double_array, new_size_1, new_size_2) |
---|
| 1419 | ! """ |
---|
| 1420 | ! Change the size of a 2D double precision array. |
---|
| 1421 | ! :param double_array: array to change the size |
---|
| 1422 | ! :param new_size_1: new size of dimension 1 of the array |
---|
| 1423 | ! :param new_size_2: new size of dimension 2 the array |
---|
| 1424 | ! :return double_array: resized array |
---|
| 1425 | ! """ |
---|
| 1426 | implicit none |
---|
| 1427 | |
---|
| 1428 | integer, intent(in) :: new_size_1, new_size_2 |
---|
| 1429 | doubleprecision, intent(inout), dimension(:, :), allocatable :: double_array |
---|
| 1430 | |
---|
| 1431 | integer :: & |
---|
| 1432 | shape_array(2) ! shape of the array |
---|
| 1433 | |
---|
| 1434 | doubleprecision, dimension(:, :), allocatable :: & |
---|
| 1435 | tmp ! temporary array |
---|
| 1436 | |
---|
| 1437 | shape_array = shape(double_array) |
---|
| 1438 | |
---|
| 1439 | call move_alloc(double_array, tmp) |
---|
| 1440 | allocate(double_array(new_size_1, new_size_2)) |
---|
| 1441 | |
---|
| 1442 | double_array(:, :) = 0d0 |
---|
| 1443 | |
---|
| 1444 | shape_array(1) = min(new_size_1, shape_array(1)) |
---|
| 1445 | shape_array(2) = min(new_size_2, shape_array(2)) |
---|
| 1446 | |
---|
| 1447 | double_array(:shape_array(1), :shape_array(2)) = tmp(:shape_array(1), :shape_array(2)) |
---|
| 1448 | end subroutine reallocate_2Ddouble |
---|
| 1449 | |
---|
| 1450 | |
---|
| 1451 | subroutine reallocate_2Dinteger(integer_array, new_size_1, new_size_2) |
---|
| 1452 | ! """ |
---|
| 1453 | ! Change the size of a 2D integer array. |
---|
| 1454 | ! :param integer_array: array to change the size |
---|
| 1455 | ! :param new_size_1: new size of dimension 1 of the array |
---|
| 1456 | ! :param new_size_2: new size of dimension 2 the array |
---|
| 1457 | ! :return integer_array: resized array |
---|
| 1458 | ! """ |
---|
| 1459 | implicit none |
---|
| 1460 | |
---|
| 1461 | integer, intent(in) :: new_size_1, new_size_2 |
---|
| 1462 | integer, intent(inout), dimension(:, :), allocatable :: integer_array |
---|
| 1463 | |
---|
| 1464 | integer :: & |
---|
| 1465 | shape_array(2) ! shape of the array |
---|
| 1466 | |
---|
| 1467 | integer, dimension(:, :), allocatable :: & |
---|
| 1468 | tmp ! temporary array |
---|
| 1469 | |
---|
| 1470 | shape_array = shape(integer_array) |
---|
| 1471 | |
---|
| 1472 | call move_alloc(integer_array, tmp) |
---|
| 1473 | allocate(integer_array(new_size_1, new_size_2)) |
---|
| 1474 | |
---|
| 1475 | integer_array(:, :) = 0 |
---|
| 1476 | |
---|
| 1477 | shape_array(1) = min(new_size_1, shape_array(1)) |
---|
| 1478 | shape_array(2) = min(new_size_2, shape_array(2)) |
---|
| 1479 | |
---|
| 1480 | integer_array(:shape_array(1), :shape_array(2)) = tmp(:shape_array(1), :shape_array(2)) |
---|
| 1481 | end subroutine reallocate_2Dinteger |
---|
| 1482 | |
---|
| 1483 | |
---|
| 1484 | subroutine reallocate_3Ddouble(double_array, new_size_1, new_size_2, new_size_3) |
---|
| 1485 | ! """ |
---|
| 1486 | ! Change the size of a 2D double precision array. |
---|
| 1487 | ! :param double_array: array to change the size |
---|
| 1488 | ! :param new_size_1: new size of dimension 1 of the array |
---|
| 1489 | ! :param new_size_2: new size of dimension 2 of the array |
---|
| 1490 | ! :param new_size_3: new size of dimension 3 of the array |
---|
| 1491 | ! :return double_array: resized array |
---|
| 1492 | ! """ |
---|
| 1493 | implicit none |
---|
| 1494 | |
---|
| 1495 | integer, intent(in) :: new_size_1, new_size_2, new_size_3 |
---|
| 1496 | doubleprecision, intent(inout), dimension(:, :, :), allocatable :: double_array |
---|
| 1497 | |
---|
| 1498 | integer :: & |
---|
| 1499 | shape_array(3) ! shape of the array |
---|
| 1500 | |
---|
| 1501 | doubleprecision, dimension(:, :, :), allocatable :: & |
---|
| 1502 | tmp ! temporary array |
---|
| 1503 | |
---|
| 1504 | shape_array = shape(double_array) |
---|
| 1505 | |
---|
| 1506 | call move_alloc(double_array, tmp) |
---|
| 1507 | allocate(double_array(new_size_1, new_size_2, new_size_3)) |
---|
| 1508 | |
---|
| 1509 | double_array(:, :, :) = 0d0 |
---|
| 1510 | |
---|
| 1511 | shape_array(1) = min(new_size_1, shape_array(1)) |
---|
| 1512 | shape_array(2) = min(new_size_2, shape_array(2)) |
---|
| 1513 | shape_array(3) = min(new_size_3, shape_array(3)) |
---|
| 1514 | |
---|
| 1515 | double_array(:shape_array(1), :shape_array(2), :shape_array(3)) = & |
---|
| 1516 | tmp(:shape_array(1), :shape_array(2), :shape_array(3)) |
---|
| 1517 | end subroutine reallocate_3Ddouble |
---|
| 1518 | |
---|
| 1519 | |
---|
| 1520 | subroutine ifft(x) |
---|
| 1521 | ! """ |
---|
| 1522 | ! Calculate the Cooley-Tukey IFFT functions. |
---|
| 1523 | ! |
---|
| 1524 | ! input: |
---|
| 1525 | ! x: double complex vector of size 2^n |
---|
| 1526 | ! """ |
---|
| 1527 | implicit none |
---|
| 1528 | |
---|
| 1529 | complex(kind=dp), dimension(:), intent(inout) :: x |
---|
| 1530 | |
---|
| 1531 | call ifft_r(x) |
---|
| 1532 | |
---|
| 1533 | x = x / real(size(x), dp) |
---|
| 1534 | |
---|
| 1535 | contains |
---|
| 1536 | recursive subroutine ifft_r(x) |
---|
| 1537 | implicit none |
---|
| 1538 | |
---|
| 1539 | complex(kind=dp), dimension(:), intent(inout) :: x |
---|
| 1540 | |
---|
| 1541 | integer ::& |
---|
| 1542 | i, & ! index |
---|
| 1543 | n ! size of array x |
---|
| 1544 | |
---|
| 1545 | complex(kind=dp) ::& |
---|
| 1546 | t ! intermediate value |
---|
| 1547 | |
---|
| 1548 | complex(kind=dp), dimension(:), allocatable ::& |
---|
| 1549 | even, & ! even-number indexed values of x |
---|
| 1550 | odd ! odd-number indexed values of x |
---|
| 1551 | |
---|
| 1552 | n=size(x) |
---|
| 1553 | |
---|
| 1554 | if(n <= 1) return |
---|
| 1555 | |
---|
| 1556 | allocate(odd((n + 1) / 2)) |
---|
| 1557 | allocate(even(n / 2)) |
---|
| 1558 | |
---|
| 1559 | ! divide |
---|
| 1560 | odd(:) = x(1:n:2) |
---|
| 1561 | even(:) = x(2:n:2) |
---|
| 1562 | |
---|
| 1563 | ! conquer |
---|
| 1564 | call ifft_r(odd) |
---|
| 1565 | call ifft_r(even) |
---|
| 1566 | |
---|
| 1567 | ! combine |
---|
| 1568 | do i = 1, n/2 |
---|
| 1569 | t = exp(cmplx(0.0_dp, 2.0_dp * pi * real(i - 1, dp) / real(n, dp), kind=dp)) * even(i) |
---|
| 1570 | x(i) = odd(i) + t |
---|
| 1571 | x(i+n/2) = odd(i) - t |
---|
| 1572 | end do |
---|
| 1573 | |
---|
| 1574 | deallocate(odd) |
---|
| 1575 | deallocate(even) |
---|
| 1576 | |
---|
| 1577 | end subroutine ifft_r |
---|
| 1578 | |
---|
| 1579 | end subroutine ifft |
---|
| 1580 | |
---|
| 1581 | |
---|
| 1582 | subroutine init_random_seed() |
---|
| 1583 | ! """ |
---|
| 1584 | ! Initialize the random seed with a varying seed in order to ensure a different random number sequence for |
---|
| 1585 | ! each invocation of the program. |
---|
| 1586 | ! |
---|
| 1587 | ! notes: |
---|
| 1588 | ! Using the random number generator file takes far too long to be useful. |
---|
| 1589 | ! Source: https://gcc.gnu.org/onlinedocs/gfortran/RANDOM_005fSEED.html |
---|
| 1590 | ! """ |
---|
| 1591 | use iso_fortran_env, only: int64 |
---|
| 1592 | |
---|
| 1593 | implicit none |
---|
| 1594 | |
---|
| 1595 | integer :: & |
---|
| 1596 | i, & ! index |
---|
| 1597 | n, & ! size of random seed |
---|
| 1598 | dt(8), & ! date and time |
---|
| 1599 | pid ! pid |
---|
| 1600 | |
---|
| 1601 | integer, allocatable :: & |
---|
| 1602 | seed(:) ! random generato seed |
---|
| 1603 | |
---|
| 1604 | integer(int64) :: & |
---|
| 1605 | t ! (s) time |
---|
| 1606 | |
---|
| 1607 | call random_seed(size = n) |
---|
| 1608 | |
---|
| 1609 | allocate(seed(n)) |
---|
| 1610 | |
---|
| 1611 | ! XOR:ing the current time and pid. The PID is useful in case one launches multiple instances of the same |
---|
| 1612 | ! program in parallel. |
---|
| 1613 | call system_clock(t) |
---|
| 1614 | |
---|
| 1615 | if (t == 0) then |
---|
| 1616 | call date_and_time(values=dt) |
---|
| 1617 | |
---|
| 1618 | t = (dt(1) - 1970) * 365_int64 * 24 * 60 * 60 * 1000 & |
---|
| 1619 | + dt(2) * 31_int64 * 24 * 60 * 60 * 1000 & |
---|
| 1620 | + dt(3) * 24_int64 * 60 * 60 * 1000 & |
---|
| 1621 | + dt(5) * 60 * 60 * 1000 & |
---|
| 1622 | + dt(6) * 60 * 1000 + dt(7) * 1000 & |
---|
| 1623 | + dt(8) |
---|
| 1624 | end if |
---|
| 1625 | |
---|
| 1626 | pid = getpid() |
---|
| 1627 | t = ieor(t, int(pid, kind(t))) |
---|
| 1628 | |
---|
| 1629 | do i = 1, n |
---|
| 1630 | seed(i) = lcg(t) |
---|
| 1631 | end do |
---|
| 1632 | |
---|
| 1633 | call random_seed(put=seed) |
---|
| 1634 | |
---|
| 1635 | contains |
---|
| 1636 | function lcg(s) |
---|
| 1637 | integer :: lcg |
---|
| 1638 | integer(int64) :: s |
---|
| 1639 | |
---|
| 1640 | if (s == 0) then |
---|
| 1641 | s = 104729 |
---|
| 1642 | else |
---|
| 1643 | s = mod(s, 4294967296_int64) |
---|
| 1644 | end if |
---|
| 1645 | s = mod(s * 279470273_int64, 4294967291_int64) |
---|
| 1646 | lcg = int(mod(s, int(huge(0), int64)), kind(0)) |
---|
| 1647 | end function lcg |
---|
| 1648 | end subroutine init_random_seed |
---|
| 1649 | |
---|
| 1650 | |
---|
| 1651 | subroutine matinv(a, n) |
---|
| 1652 | ! """ |
---|
| 1653 | ! Inverse a 2D matrix of dimension (n,n) |
---|
| 1654 | ! The original matrix a(n,n) will be destroyed during the calculation |
---|
| 1655 | ! Method: Based on Doolittle LU factorization for Ax=b |
---|
| 1656 | ! Source: http://ww2.odu.edu/~agodunov/computing/programs/book2/Ch06/Inverse.f90 |
---|
| 1657 | ! by Alex G. December 2009 |
---|
| 1658 | ! |
---|
| 1659 | ! inputs: |
---|
| 1660 | ! a(n,n): array of coefficients for matrix A |
---|
| 1661 | ! n: dimension |
---|
| 1662 | ! |
---|
| 1663 | ! output: |
---|
| 1664 | ! c(n,n): inverse matrix of A |
---|
| 1665 | ! """ |
---|
| 1666 | implicit none |
---|
| 1667 | |
---|
| 1668 | integer, intent(in) :: n |
---|
| 1669 | double precision, intent(inout) :: a(n,n) |
---|
| 1670 | |
---|
| 1671 | double precision :: & |
---|
| 1672 | l(n,n), & ! lower triangular matrix |
---|
| 1673 | U(n,n), & ! upper triangular matrix |
---|
| 1674 | b(n), & ! intermediate value |
---|
| 1675 | c(n,n), & ! temporary matrix |
---|
| 1676 | d(n), & ! intermediate value |
---|
| 1677 | x(n) ! intermediate value |
---|
| 1678 | |
---|
| 1679 | double precision :: & |
---|
| 1680 | coeff ! intermediate value |
---|
| 1681 | |
---|
| 1682 | integer :: & |
---|
| 1683 | i, & ! index |
---|
| 1684 | j, & ! index |
---|
| 1685 | k ! index |
---|
| 1686 | |
---|
| 1687 | ! Step 0: initialization for matrices l and U and b |
---|
| 1688 | l(:, :) = 0d0 |
---|
| 1689 | U(:, :) = 0d0 |
---|
| 1690 | b = 0d0 |
---|
| 1691 | c(:, :) = 0d0 |
---|
| 1692 | |
---|
| 1693 | ! Step 1: forward elimination |
---|
| 1694 | do k = 1, n - 1 |
---|
| 1695 | do i = k + 1, n |
---|
| 1696 | coeff = a(i, k) / a(k, k) |
---|
| 1697 | |
---|
| 1698 | l(i, k) = coeff |
---|
| 1699 | |
---|
| 1700 | do j = k + 1, n |
---|
| 1701 | a(i, j) = a(i, j) - coeff * a(k, j) |
---|
| 1702 | end do |
---|
| 1703 | end do |
---|
| 1704 | end do |
---|
| 1705 | |
---|
| 1706 | ! Step 2: prepare l and U matrices |
---|
| 1707 | ! l matrix is a matrix of the elimination coefficient + the diagonal elements are 1.0 |
---|
| 1708 | do i = 1, n |
---|
| 1709 | l(i, i) = 1.0 |
---|
| 1710 | end do |
---|
| 1711 | |
---|
| 1712 | ! U matrix is the upper triangular part of A |
---|
| 1713 | do j = 1, n |
---|
| 1714 | do i = 1, j |
---|
| 1715 | U(i, j) = a(i, j) |
---|
| 1716 | end do |
---|
| 1717 | end do |
---|
| 1718 | |
---|
| 1719 | ! Step 3: compute columns of the inverse matrix C |
---|
| 1720 | do k = 1, n |
---|
| 1721 | b(k) = 1.0 |
---|
| 1722 | d(1) = b(1) |
---|
| 1723 | |
---|
| 1724 | ! Step 3a: Solve Ld=b using the forward substitution |
---|
| 1725 | do i = 2, n |
---|
| 1726 | d(i) = b(i) |
---|
| 1727 | |
---|
| 1728 | do j=1, i - 1 |
---|
| 1729 | d(i) = d(i) - l(i, j) * d(j) |
---|
| 1730 | end do |
---|
| 1731 | end do |
---|
| 1732 | |
---|
| 1733 | ! Step 3b: Solve Ux=d using the back substitution |
---|
| 1734 | x(n) = d(n) / U(n,n) |
---|
| 1735 | |
---|
| 1736 | do i = n - 1, 1, -1 |
---|
| 1737 | x(i) = d(i) |
---|
| 1738 | |
---|
| 1739 | do j = n, i + 1, -1 |
---|
| 1740 | x(i) = x(i) - U(i, j) * x(j) |
---|
| 1741 | end do |
---|
| 1742 | |
---|
| 1743 | x(i) = x(i) / u(i, i) |
---|
| 1744 | end do |
---|
| 1745 | |
---|
| 1746 | ! Step 3c: fill the solutions x(n) into column k of C |
---|
| 1747 | do i = 1, n |
---|
| 1748 | c(i, k) = x(i) |
---|
| 1749 | end do |
---|
| 1750 | |
---|
| 1751 | b(k) = 0d0 |
---|
| 1752 | end do |
---|
| 1753 | |
---|
| 1754 | a(:, :) = c(:, :) |
---|
| 1755 | end subroutine matinv |
---|
| 1756 | end module math |
---|