1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
|
subroutine spline(energy, xmu, mxmu, xkstd, chistd, mstd,
$ e0_in, rbkgin, r1stin, toler, nknots,
$ qmin, qmax, xkw, dk1, dk2, winnam,
$ stfind, fnorm, enor1, enor2, pre1, pre2, estep,
$ lclmp1, xclmp1, lclmp2, xclmp2, nclmp,
$ cnorm, fixstd, usestd, varye0, spstep,
$ dofit, bkg, mxk, xk, chi)
c
c//////////////////////////////////////////////////////////////////////
c Copyright (c) 1997--2000 Matthew Newville, The University of Chicago
c Copyright (c) 1992--1996 Matthew Newville, University of Washington
c
c Permission to use and redistribute the source code or binary forms of
c this software and its documentation, with or without modification is
c hereby granted provided that the above notice of copyright, these
c terms of use, and the disclaimer of warranty below appear in the
c source code and documentation, and that none of the names of The
c University of Chicago, The University of Washington, or the authors
c appear in advertising or endorsement of works derived from this
c software without specific prior written permission from all parties.
c
c THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
c EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
c MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
c IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
c CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
c TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
c SOFTWARE OR THE USE OR OTHER DEALINGS IN THIS SOFTWARE.
c//////////////////////////////////////////////////////////////////////
c
c purpose: automated xafs background spline (the autobk algorithm)
c
c arguments:
c energy array of energy values [in]
c xmu array of xmu values [in]
c mxmu maximum array length of energy/xmu [in]
c xkstd array of k values for standard [in]
c chistd array of chi values for standard [in]
c mstd maximum array length of xkstd/chistd [in]
c fixstd l=flag for fixing amp of standard [in]
c usestd l=flag for using standard at all [in]
c rbkgin rbkg value (nyquist value) [in/out]
c toler fitting tolerance [in/out]
c nknots number of knots in spline [in/out]
c qmin low-k value for fft window [in/out]
c qmax high-k value for fft window [in/out]
c xkw k-weight for fft [in]
c dk1 low-k value for fft window [in]
c dk2 low-k value for fft window [in]
c winnam fft window type [in]
c varye0 l=flag for varying e0 in fit [in]
c e0_in energy origin [in/out]
c stfind l=flag for redefining estep [in]
c fnorm l=flag for functional normalization [in]
c enor1 low-energy range for post-step [in]
c enor2 high-energy range for post-step [in]
c pre1 low-energy range for pre-edge [in]
c pre2 high-energy range for pre-edge [in]
c estep edge-step, aka delta_xmu [in/out]
c dofit l=flag for doing fit at all [in]
c bkg array of mu0 values (mxmu points) [out]
c mxk maximum array length of xk/chi [out]
c xk array of k values [out]
c chi array of chi values [out]
c
c notes:
c 1. for use in ifeffit: uses ifeffit include files
c 2. based on/related to suroutine autnls from autobk
c 3. not well tested with standard chi(k) data
c
c see documentation and code comments for more details
c
c requires: include files consts.h, spline.h, fft.h
c splfun, lmdif1, cffti, qintrp, lintrp, polyft,
c fitfft, window, sumsqr, nofx, echo, chrdmp
c
c
implicit none
double precision energy(*), xmu(*), bkg(*), xkstd(*), chistd(*)
double precision xk(*), chi(*), enor1, enor2, e0_in, estep
double precision rbkgin, toler,qmin,qmax
double precision xkw,dk1,dk2, pre1, pre2, cnorm(*)
integer mxmu, mxk, mstd, nknots, nx1
logical fnorm, fixstd, usestd, varye0, stfind, dofit
logical lclmp1, lclmp2
double precision xclmp1, xclmp2
c
include 'consts.h'
include 'spline.h'
include 'fft.h'
c
integer lenwrk, loop, lminfo, iup, ilo, nnorm , iterp
integer iwork(mtknot), nemin, nemax, nofx, nrfit, nclmp
integer i, nr, iend, ntest, ipos, nqmax, j
parameter(lenwrk = 2*maxpts*mtknot + 20*mtknot + 2*maxpts )
character*80 messg, winnam*32
double precision work(lenwrk), fvect(2*maxpts), varys(mtknot)
double precision esplin(mtknot), bscoef(mtknot), thiq(maxpts)
double precision small, toldef, spstep, r1stin
double precision sumsqr, qmn, qmx, qknot, slope, offset
double precision e0tst, qmntst, qmxtst, xtmp, qtmp, e0t, tmp
parameter (small = 1.d-10, toldef = 1.d-5)
external splfun, nofx, sumsqr
save
c
nnorm = 3
if (toler .le. zero) toler = toldef
spstep = min(20.d0, max(1.d-6, spstep))
loop = 0
if (.not.wftset) then
call cffti(maxfft, wfftc)
wftset = .true.
end if
rbkg = rbkgin
r1st = rbkg + r1stin
if (rbkg .le. small) rbkg = one
nrbkg = 2 * int(1d-2 + (rbkg /rgrid))+ 2
nr1st = 2 * int(1d-2 + (r1st /rgrid))+ 2
if (qmax .le. small) qmax = 0
c
c initialize bkgdat common block
splqw = xkw
nsplin = nknots
nautbk = 0
thefix = fixstd
theory = usestd
eevary = varye0 .and. theory
cc print*, ' spline: eevary, = ', eevary, theory
step = estep
e0 = e0_in
emin = e0 + qmin**2 / etok
emax = e0 + qmax**2 / etok
e0t = e0
de0 = 0
thessq = one
thebkg = one
nxmu = mxmu
do 10 i = 1, maxpts
endat(i) = zero
xmudat(i) = zero
spldat(i) = zero
chie(i) = zero
splfit(i) = zero
splwin(i) = zero
10 continue
do 20 i = 1, nxmu
endat(i) = energy(i)
xmudat(i) = xmu(i)
spldat(i) = xmu(i)
20 continue
c do xafsft window
c put standard chi(k) on an absolute q-grid, and do fft
if (mstd.gt.0) then
nqmax = int( xkstd(mstd) / qgrid)
ipos = 1
do 30 i = 1, nqmax
qtmp = qgrid * (i-1)
call qintrp(xkstd, chistd, mstd, qtmp, ipos, thiq(i))
30 continue
end if
c
100 continue
lminfo = 0
loop = loop + 1
c
c----start fitting:
do 120 i = 1, nxmu
bkg(i) = spldat(i)
120 continue
c
clamp(1) = lclmp1
clamp(2) = lclmp2
sclamp(1) = xclmp1
sclamp(2) = xclmp2
nclamp = nclmp
cc print*,' clamps: ', clamp, sclamp, nclamp
c
c initialize fvect and work arrays for lmdif1
nrfit = nrbkg
if (theory) nrfit = nr1st
if (clamp(1)) nrfit = nrfit + nclamp
if (clamp(2)) nrfit = nrfit + nclamp
do 210 i =1, nrfit
fvect(i) = zero
210 continue
do 220 i =1, lenwrk
work(i) = zero
220 continue
do 240 i =1, mtknot
iwork(i) = 0
varys(i) = zero
esplin(i) = zero
bscoef(i) = zero
240 continue
c
c e0-shift: e0, emin, and emax
if (abs(de0).gt.1d-3) then
e0t = e0t + de0
emin = emin + de0
emax = emax + de0
end if
c
if (emax.le.emin) emax = energy(nxmu)
nemin = nofx(emin,energy,nxmu)
nemax = nofx(emax,energy,nxmu)
emin = max(e0t, energy(nemin))
emax = max(emin, energy(nemax))
c
c evaluate energies for the spline variables:
c - emin and emax are not relative to the edge
c - energy contains the input energy values, not relative to the edge.
qmin = sqrt(etok* abs(emin - e0t) )
qmax = sqrt(etok* (emax - e0t) )
c
c calculate number of independent points in r-space
if (nsplin.le.1)
$ nsplin = 2 * int(rbkg * (qmax - qmin)/ pi) + 1
nsplin = min(mtknot-5, max(5,nsplin))
nknots = nsplin
cc print*, ' SPLINE ', rbkg, nsplin
c initialize energy values through which the first guess for the
c spline must go (evenly spaced in q), and get the initial value
c for the spline value at this point. we'll also get the initial
c guesses for the variables (the b-spline coefficients) from this
c initial spline.
do 300 i = 1, nsplin
qtmp = qmin + (i-1)*(qmax-qmin)/(nsplin - 1)
xtmp = e0t + ( qtmp**2 / etok )
j = nofx(xtmp,energy,nxmu)
esplin(i) = energy(j)
iup = min(nxmu, j + 5)
ilo = max(1, j - 5)
bscoef(i) = (2*spldat(j)+ spldat(iup)+spldat(ilo))/4
300 continue
esplin(nsplin) = one + esplin(nsplin)
c the first and last korder knots in the b-spline are nearly degenerate at
c the endpoints. spstep sets the spacing between these points. the
c default is one -- this may help eliminate "spikes" at the endpoints.
c since each knot represents a place where a derivative can break,
c having all four of these at one place allows a complete break from
c at this point. by moving a few of the knots just off the ends, the
c spline is a little bit stiffer at the endpoints.
cc print*, ' spline : ', spstep, korder
do 310 i = 1, korder
eknot(i) = esplin(1) - spstep * (korder-i-1)
eknot(nsplin+i) = esplin(nsplin) + spstep * i
310 continue
qmn = sqrt( etok* abs(esplin(1) - e0t) )
if (e0.lt.esplin(1)) qmn = zero
qmx = sqrt( etok* (esplin(nsplin) - e0t) )
do 320 i = korder+1, nsplin
qknot = (i-korder)*(qmx - qmn)/(nsplin-korder+1)
eknot(i) = esplin(1) + qknot**2/etok
320 continue
c
c determine the knots for the spline:
c knots are points at which the spline has extra freedom.
if ( (korder.lt.3).or.(nsplin.lt.korder) ) then
call warn(2,
$ ' spline error: not enough data to create spline.')
return
end if
c the b-spline coefficients will be the variables in the fit,
c the above estimates for the elements of bscoef are good enough
c especially when the spline values are smoothed for the initial
c guesses in the fit, as below.
varys(1) = (3*bscoef(1) + bscoef(2))/ 4
varys(nsplin) = (3*bscoef(nsplin) + bscoef(nsplin-1))/ 4
do 380 i = 2, nsplin-1
varys(i) = (bscoef(i-1) + 2*bscoef(i) + bscoef(i+1) )/4
380 continue
nautbk = nsplin
c
c set up window function
call window(winnam,dk1,dk2,qmin,qmax,qgrid,maxpts,splwin)
c if a theory file is used, do its fft now
c and add another variable if e0 is to be shifted
if (theory) then
if (eevary) then
nautbk = nautbk + 1
varys(nautbk) = de0
end if
nr = nrbkg
call fitfft(thiq, maxpts, maxfft, wfftc, qgrid,
$ splwin, splqw, bkg, zero, 1, 0, zero, r1st,
$ nr, splfit)
thessq = max(small, sumsqr(splfit(nrbkg),nr1st))
thebkg = thessq
cc print*, ' spline.f: thebkg = ', thebkg
cc print*, ' nrbkg, nr1st = ', nrbkg, nr1st, nr
nautbk = nautbk + 1
varys(nautbk) = thessq
cc
if (nr.ne.nr1st) then
call warn(3,' spline error: fitfft is broken' )
return
end if
end if
c
if (eevary) then
call chrdmp(' spline: fitting background and e0 ... ')
else
call chrdmp(' spline: fitting background ... ')
end if
c
lminfo = 1
if (dofit) then
call lmdif1 (splfun, nrfit, nautbk, varys, fvect,
$ toler, lminfo, iwork, work, lenwrk)
call echo(' done.')
call lm_err(lminfo,toler)
end if
c
c if an energy shift was done, the q values may be slightly off, so that
c the number of independent points changes. if so, re-do the fit until
c e0 is stable to 0.5 eV, up to 5 times
e0tst = e0 + de0
qmntst = sqrt(etok* abs(emin - e0tst ) )
qmxtst = sqrt(etok* abs(emax - e0tst ) )
if (e0tst.le.emin) qmntst = zero
ntest = 2 * int ( rbkg * abs(qmxtst - qmntst) / pi) + 1
ntest = min(mtknot-5, max(5,ntest))
c
if ((loop.le.5).and.eevary.and.(ntest.ne.nsplin) ) then
call echo(' spline warning: e0 was shifted enough that')
call warn(1, ' the # of knots in the spline changed'//
$ ' and the fit should be re-done.')
go to 100
end if
c
c now that we have a good spline, we may want to improve the estimate
c of the edge-step. so do a parabolic extrapolation of the background
c spline to the edge energy value, which is probably better than
c getting the edge step from a linear extrapolation of the xmu data.
c redo function evaluation with the final values for everything.
call splfun(nrfit, nautbk, varys, fvect, iend)
if (stfind.and.(.not.fnorm)) then
call echo(' spline: finding edge_step from bkg(E)')
call preedg(.false.,.true.,nxmu, endat, bkg, e0,
$ pre1, pre2, enor1, enor2, nnorm,
$ step, slope, offset,cnorm)
estep = step
end if
c redo function evaluation with the final values for everything.
call splfun(nrfit, nautbk, varys, fvect, iend)
if (eevary) then
e0 = e0_in + de0
e0_in = e0
end if
do 830 i = 1, nxmu
bkg(i) = spldat(i)
830 continue
c set chi array
call chie2k(endat,chie,nxmu,e0, mxk,xk,chi)
c print*, 'MM end of spline ', endat(1), endat(2),
c $ chie(1), chie(2), xk(1), xk(2),
c $ chi(1), chi(2), nxmu, e0, mxk
c end subroutine spline
return
end
|