[aa36f96] | 1 | |
---|
[89f3b66] | 2 | import copy |
---|
[c4d6900] | 3 | #import logging |
---|
[444c900e] | 4 | import sys |
---|
[89f3b66] | 5 | import numpy |
---|
| 6 | import math |
---|
| 7 | import park |
---|
[41517a6] | 8 | from sans.dataloader.data_info import Data1D |
---|
| 9 | from sans.dataloader.data_info import Data2D |
---|
[4b5bd73] | 10 | _SMALLVALUE = 1.0e-10 |
---|
[b2f25dc5] | 11 | |
---|
[48882d1] | 12 | class SansParameter(park.Parameter): |
---|
| 13 | """ |
---|
[aa36f96] | 14 | SANS model parameters for use in the PARK fitting service. |
---|
| 15 | The parameter attribute value is redirected to the underlying |
---|
| 16 | parameter value in the SANS model. |
---|
[48882d1] | 17 | """ |
---|
[1cff677] | 18 | def __init__(self, name, model, data): |
---|
[ca6d914] | 19 | """ |
---|
[1f9f3c8a] | 20 | :param name: the name of the model parameter |
---|
| 21 | :param model: the sans model to wrap as a park model |
---|
[ca6d914] | 22 | """ |
---|
[c4d6900] | 23 | park.Parameter.__init__(self, name) |
---|
[89f3b66] | 24 | self._model, self._name = model, name |
---|
[1cff677] | 25 | self.data = data |
---|
| 26 | self.model = model |
---|
[ca6d914] | 27 | #set the value for the parameter of the given name |
---|
| 28 | self.set(model.getParam(name)) |
---|
[48882d1] | 29 | |
---|
[ca6d914] | 30 | def _getvalue(self): |
---|
| 31 | """ |
---|
[aa36f96] | 32 | override the _getvalue of park parameter |
---|
| 33 | |
---|
| 34 | :return value the parameter associates with self.name |
---|
| 35 | |
---|
[ca6d914] | 36 | """ |
---|
| 37 | return self._model.getParam(self.name) |
---|
[48882d1] | 38 | |
---|
[89f3b66] | 39 | def _setvalue(self, value): |
---|
[ca6d914] | 40 | """ |
---|
[aa36f96] | 41 | override the _setvalue pf park parameter |
---|
| 42 | |
---|
| 43 | :param value: the value to set on a given parameter |
---|
| 44 | |
---|
[ca6d914] | 45 | """ |
---|
[48882d1] | 46 | self._model.setParam(self.name, value) |
---|
| 47 | |
---|
[c4d6900] | 48 | value = property(_getvalue, _setvalue) |
---|
[48882d1] | 49 | |
---|
| 50 | def _getrange(self): |
---|
[ca6d914] | 51 | """ |
---|
[aa36f96] | 52 | Override _getrange of park parameter |
---|
| 53 | return the range of parameter |
---|
[ca6d914] | 54 | """ |
---|
[920a6e5] | 55 | #if not self.name in self._model.getDispParamList(): |
---|
[89f3b66] | 56 | lo, hi = self._model.details[self.name][1:3] |
---|
[920a6e5] | 57 | if lo is None: lo = -numpy.inf |
---|
| 58 | if hi is None: hi = numpy.inf |
---|
[e0e22f2c] | 59 | if lo > hi: |
---|
[1f9f3c8a] | 60 | raise ValueError, "wrong fit range for parameters" |
---|
[05f14dd] | 61 | |
---|
[89f3b66] | 62 | return lo, hi |
---|
[48882d1] | 63 | |
---|
[b2f25dc5] | 64 | def get_name(self): |
---|
| 65 | """ |
---|
| 66 | """ |
---|
| 67 | return self._getname() |
---|
| 68 | |
---|
[89f3b66] | 69 | def _setrange(self, r): |
---|
[ca6d914] | 70 | """ |
---|
[aa36f96] | 71 | override _setrange of park parameter |
---|
| 72 | |
---|
| 73 | :param r: the value of the range to set |
---|
| 74 | |
---|
[ca6d914] | 75 | """ |
---|
[12b76cf] | 76 | self._model.details[self.name][1:3] = r |
---|
[89f3b66] | 77 | range = property(_getrange, _setrange) |
---|
[a9e04aa] | 78 | |
---|
[1f9f3c8a] | 79 | |
---|
[a9e04aa] | 80 | class Model(park.Model): |
---|
[48882d1] | 81 | """ |
---|
[aa36f96] | 82 | PARK wrapper for SANS models. |
---|
[48882d1] | 83 | """ |
---|
[1cff677] | 84 | def __init__(self, sans_model, sans_data=None, **kw): |
---|
[ca6d914] | 85 | """ |
---|
[aa36f96] | 86 | :param sans_model: the sans model to wrap using park interface |
---|
| 87 | |
---|
[ca6d914] | 88 | """ |
---|
[a9e04aa] | 89 | park.Model.__init__(self, **kw) |
---|
[48882d1] | 90 | self.model = sans_model |
---|
[ca6d914] | 91 | self.name = sans_model.name |
---|
[1cff677] | 92 | self.data = sans_data |
---|
[ca6d914] | 93 | #list of parameters names |
---|
[48882d1] | 94 | self.sansp = sans_model.getParamList() |
---|
[ca6d914] | 95 | #list of park parameter |
---|
[1cff677] | 96 | self.parkp = [SansParameter(p, sans_model, sans_data) for p in self.sansp] |
---|
[1f9f3c8a] | 97 | #list of parameter set |
---|
[89f3b66] | 98 | self.parameterset = park.ParameterSet(sans_model.name, pars=self.parkp) |
---|
| 99 | self.pars = [] |
---|
[ca6d914] | 100 | |
---|
[c4d6900] | 101 | def get_params(self, fitparams): |
---|
[ca6d914] | 102 | """ |
---|
[aa36f96] | 103 | return a list of value of paramter to fit |
---|
| 104 | |
---|
| 105 | :param fitparams: list of paramaters name to fit |
---|
| 106 | |
---|
[ca6d914] | 107 | """ |
---|
[c4d6900] | 108 | list_params = [] |
---|
[89f3b66] | 109 | self.pars = [] |
---|
| 110 | self.pars = fitparams |
---|
[48882d1] | 111 | for item in fitparams: |
---|
| 112 | for element in self.parkp: |
---|
[c4d6900] | 113 | if element.name == str(item): |
---|
| 114 | list_params.append(element.value) |
---|
| 115 | return list_params |
---|
[48882d1] | 116 | |
---|
[c4d6900] | 117 | def set_params(self, paramlist, params): |
---|
[ca6d914] | 118 | """ |
---|
[aa36f96] | 119 | Set value for parameters to fit |
---|
| 120 | |
---|
[1f9f3c8a] | 121 | :param params: list of value for parameters to fit |
---|
[aa36f96] | 122 | |
---|
[ca6d914] | 123 | """ |
---|
[e71440c] | 124 | try: |
---|
| 125 | for i in range(len(self.parkp)): |
---|
| 126 | for j in range(len(paramlist)): |
---|
[89f3b66] | 127 | if self.parkp[i].name == paramlist[j]: |
---|
[e71440c] | 128 | self.parkp[i].value = params[j] |
---|
[89f3b66] | 129 | self.model.setParam(self.parkp[i].name, params[j]) |
---|
[e71440c] | 130 | except: |
---|
| 131 | raise |
---|
[ca6d914] | 132 | |
---|
[89f3b66] | 133 | def eval(self, x): |
---|
[ca6d914] | 134 | """ |
---|
[1f9f3c8a] | 135 | Override eval method of park model. |
---|
[aa36f96] | 136 | |
---|
[1f9f3c8a] | 137 | :param x: the x value used to compute a function |
---|
[ca6d914] | 138 | """ |
---|
[d8a2e31] | 139 | try: |
---|
[393f0f3] | 140 | return self.model.evalDistribution(x) |
---|
[d8a2e31] | 141 | except: |
---|
[393f0f3] | 142 | raise |
---|
[c4d6900] | 143 | |
---|
| 144 | def eval_derivs(self, x, pars=[]): |
---|
| 145 | """ |
---|
| 146 | Evaluate the model and derivatives wrt pars at x. |
---|
| 147 | |
---|
| 148 | pars is a list of the names of the parameters for which derivatives |
---|
| 149 | are desired. |
---|
| 150 | |
---|
| 151 | This method needs to be specialized in the model to evaluate the |
---|
| 152 | model function. Alternatively, the model can implement is own |
---|
| 153 | version of residuals which calculates the residuals directly |
---|
| 154 | instead of calling eval. |
---|
| 155 | """ |
---|
| 156 | return [] |
---|
| 157 | |
---|
[b64fa56] | 158 | |
---|
[1e3169c] | 159 | class FitData1D(Data1D): |
---|
[1f9f3c8a] | 160 | """ |
---|
| 161 | Wrapper class for SANS data |
---|
| 162 | FitData1D inherits from DataLoader.data_info.Data1D. Implements |
---|
| 163 | a way to get residuals from data. |
---|
[1e3169c] | 164 | """ |
---|
[634ca14] | 165 | def __init__(self, x, y, dx=None, dy=None, smearer=None, data=None): |
---|
[7d0c1a8] | 166 | """ |
---|
[1f9f3c8a] | 167 | :param smearer: is an object of class QSmearer or SlitSmearer |
---|
| 168 | that will smear the theory data (slit smearing or resolution |
---|
| 169 | smearing) when set. |
---|
| 170 | |
---|
| 171 | The proper way to set the smearing object would be to |
---|
| 172 | do the following: :: |
---|
[109e60ab] | 173 | |
---|
[1f9f3c8a] | 174 | from DataLoader.qsmearing import smear_selection |
---|
| 175 | smearer = smear_selection(some_data) |
---|
| 176 | fitdata1d = FitData1D( x= [1,3,..,], |
---|
| 177 | y= [3,4,..,8], |
---|
| 178 | dx=None, |
---|
| 179 | dy=[1,2...], smearer= smearer) |
---|
| 180 | |
---|
| 181 | :Note: that some_data _HAS_ to be of |
---|
| 182 | class DataLoader.data_info.Data1D |
---|
| 183 | Setting it back to None will turn smearing off. |
---|
| 184 | |
---|
[7d0c1a8] | 185 | """ |
---|
[89f3b66] | 186 | Data1D.__init__(self, x=x, y=y, dx=dx, dy=dy) |
---|
[634ca14] | 187 | self.sans_data = data |
---|
[b461b6d7] | 188 | self.smearer = smearer |
---|
[c4d6900] | 189 | self._first_unsmeared_bin = None |
---|
| 190 | self._last_unsmeared_bin = None |
---|
[189be4e] | 191 | # Check error bar; if no error bar found, set it constant(=1) |
---|
[c4d6900] | 192 | # TODO: Should provide an option for users to set it like percent, |
---|
| 193 | # constant, or dy data |
---|
[89f3b66] | 194 | if dy == None or dy == [] or dy.all() == 0: |
---|
[1f9f3c8a] | 195 | self.dy = numpy.ones(len(y)) |
---|
[189be4e] | 196 | else: |
---|
[89f3b66] | 197 | self.dy = numpy.asarray(dy).copy() |
---|
[189be4e] | 198 | |
---|
[109e60ab] | 199 | ## Min Q-value |
---|
[4bd557d] | 200 | #Skip the Q=0 point, especially when y(q=0)=None at x[0]. |
---|
[1f9f3c8a] | 201 | if min(self.x) == 0.0 and self.x[0] == 0 and\ |
---|
[89f3b66] | 202 | not numpy.isfinite(self.y[0]): |
---|
[1f9f3c8a] | 203 | self.qmin = min(self.x[self.x != 0]) |
---|
| 204 | else: |
---|
[89f3b66] | 205 | self.qmin = min(self.x) |
---|
[109e60ab] | 206 | ## Max Q-value |
---|
[89f3b66] | 207 | self.qmax = max(self.x) |
---|
[058b2d7] | 208 | |
---|
[72c7d31] | 209 | # Range used for input to smearing |
---|
| 210 | self._qmin_unsmeared = self.qmin |
---|
| 211 | self._qmax_unsmeared = self.qmax |
---|
[fd0d30fd] | 212 | # Identify the bin range for the unsmeared and smeared spaces |
---|
[89f3b66] | 213 | self.idx = (self.x >= self.qmin) & (self.x <= self.qmax) |
---|
| 214 | self.idx_unsmeared = (self.x >= self._qmin_unsmeared) \ |
---|
| 215 | & (self.x <= self._qmax_unsmeared) |
---|
[fd0d30fd] | 216 | |
---|
[c4d6900] | 217 | def set_fit_range(self, qmin=None, qmax=None): |
---|
[7d0c1a8] | 218 | """ to set the fit range""" |
---|
[09975cbb] | 219 | # Skip Q=0 point, (especially for y(q=0)=None at x[0]). |
---|
[189be4e] | 220 | # ToDo: Find better way to do it. |
---|
[89f3b66] | 221 | if qmin == 0.0 and not numpy.isfinite(self.y[qmin]): |
---|
| 222 | self.qmin = min(self.x[self.x != 0]) |
---|
[1f9f3c8a] | 223 | elif qmin != None: |
---|
| 224 | self.qmin = qmin |
---|
[89f3b66] | 225 | if qmax != None: |
---|
[eef2e0ed] | 226 | self.qmax = qmax |
---|
[4bb2917] | 227 | # Determine the range needed in unsmeared-Q to cover |
---|
| 228 | # the smeared Q range |
---|
[72c7d31] | 229 | self._qmin_unsmeared = self.qmin |
---|
[1f9f3c8a] | 230 | self._qmax_unsmeared = self.qmax |
---|
[72c7d31] | 231 | |
---|
[4bb2917] | 232 | self._first_unsmeared_bin = 0 |
---|
[1f9f3c8a] | 233 | self._last_unsmeared_bin = len(self.x) - 1 |
---|
[4bb2917] | 234 | |
---|
[c4d6900] | 235 | if self.smearer != None: |
---|
[89f3b66] | 236 | self._first_unsmeared_bin, self._last_unsmeared_bin = \ |
---|
| 237 | self.smearer.get_bin_range(self.qmin, self.qmax) |
---|
[1e3169c] | 238 | self._qmin_unsmeared = self.x[self._first_unsmeared_bin] |
---|
| 239 | self._qmax_unsmeared = self.x[self._last_unsmeared_bin] |
---|
[4bb2917] | 240 | |
---|
[fd0d30fd] | 241 | # Identify the bin range for the unsmeared and smeared spaces |
---|
[89f3b66] | 242 | self.idx = (self.x >= self.qmin) & (self.x <= self.qmax) |
---|
| 243 | ## zero error can not participate for fitting |
---|
[1f9f3c8a] | 244 | self.idx = self.idx & (self.dy != 0) |
---|
[89f3b66] | 245 | self.idx_unsmeared = (self.x >= self._qmin_unsmeared) \ |
---|
| 246 | & (self.x <= self._qmax_unsmeared) |
---|
[0766d6d] | 247 | |
---|
[c4d6900] | 248 | def get_fit_range(self): |
---|
[7d0c1a8] | 249 | """ |
---|
[1f9f3c8a] | 250 | Return the range of data.x to fit |
---|
[7d0c1a8] | 251 | """ |
---|
| 252 | return self.qmin, self.qmax |
---|
[72c7d31] | 253 | |
---|
[7d0c1a8] | 254 | def residuals(self, fn): |
---|
[1f9f3c8a] | 255 | """ |
---|
| 256 | Compute residuals. |
---|
| 257 | |
---|
| 258 | If self.smearer has been set, use if to smear |
---|
| 259 | the data before computing chi squared. |
---|
| 260 | |
---|
| 261 | :param fn: function that return model value |
---|
| 262 | |
---|
| 263 | :return: residuals |
---|
[109e60ab] | 264 | """ |
---|
| 265 | # Compute theory data f(x) |
---|
[89f3b66] | 266 | fx = numpy.zeros(len(self.x)) |
---|
[7e752fe] | 267 | fx[self.idx_unsmeared] = fn(self.x[self.idx_unsmeared]) |
---|
[fd0d30fd] | 268 | |
---|
[d5b488b] | 269 | ## Smear theory data |
---|
[109e60ab] | 270 | if self.smearer is not None: |
---|
[1f9f3c8a] | 271 | fx = self.smearer(fx, self._first_unsmeared_bin, |
---|
[89f3b66] | 272 | self._last_unsmeared_bin) |
---|
[d5b488b] | 273 | ## Sanity check |
---|
[89f3b66] | 274 | if numpy.size(self.dy) != numpy.size(fx): |
---|
| 275 | msg = "FitData1D: invalid error array " |
---|
[1f9f3c8a] | 276 | msg += "%d <> %d" % (numpy.shape(self.dy), numpy.size(fx)) |
---|
| 277 | raise RuntimeError, msg |
---|
[425e49ca] | 278 | return (self.y[self.idx] - fx[self.idx]) / self.dy[self.idx], fx[self.idx] |
---|
[444c900e] | 279 | |
---|
[7d0c1a8] | 280 | def residuals_deriv(self, model, pars=[]): |
---|
[1f9f3c8a] | 281 | """ |
---|
| 282 | :return: residuals derivatives . |
---|
| 283 | |
---|
| 284 | :note: in this case just return empty array |
---|
[7d0c1a8] | 285 | """ |
---|
| 286 | return [] |
---|
| 287 | |
---|
[1f9f3c8a] | 288 | |
---|
[1e3169c] | 289 | class FitData2D(Data2D): |
---|
[1f9f3c8a] | 290 | """ |
---|
| 291 | Wrapper class for SANS data |
---|
| 292 | """ |
---|
[89f3b66] | 293 | def __init__(self, sans_data2d, data=None, err_data=None): |
---|
[c4d6900] | 294 | Data2D.__init__(self, data=data, err_data=err_data) |
---|
[7d0c1a8] | 295 | """ |
---|
[1f9f3c8a] | 296 | Data can be initital with a data (sans plottable) |
---|
| 297 | or with vectors. |
---|
[7d0c1a8] | 298 | """ |
---|
[89f3b66] | 299 | self.res_err_image = [] |
---|
[444c900e] | 300 | self.idx = [] |
---|
[89f3b66] | 301 | self.qmin = None |
---|
| 302 | self.qmax = None |
---|
[f72333f] | 303 | self.smearer = None |
---|
[c4d6900] | 304 | self.radius = 0 |
---|
| 305 | self.res_err_data = [] |
---|
[634ca14] | 306 | self.sans_data = sans_data2d |
---|
[89f3b66] | 307 | self.set_data(sans_data2d) |
---|
[f72333f] | 308 | |
---|
[89f3b66] | 309 | def set_data(self, sans_data2d, qmin=None, qmax=None): |
---|
[1e3169c] | 310 | """ |
---|
[1f9f3c8a] | 311 | Determine the correct qx_data and qy_data within range to fit |
---|
[1e3169c] | 312 | """ |
---|
[89f3b66] | 313 | self.data = sans_data2d.data |
---|
[83195f7] | 314 | self.err_data = sans_data2d.err_data |
---|
| 315 | self.qx_data = sans_data2d.qx_data |
---|
| 316 | self.qy_data = sans_data2d.qy_data |
---|
[89f3b66] | 317 | self.mask = sans_data2d.mask |
---|
[83195f7] | 318 | |
---|
| 319 | x_max = max(math.fabs(sans_data2d.xmin), math.fabs(sans_data2d.xmax)) |
---|
| 320 | y_max = max(math.fabs(sans_data2d.ymin), math.fabs(sans_data2d.ymax)) |
---|
[20d30e9] | 321 | |
---|
| 322 | ## fitting range |
---|
[027e8f2] | 323 | if qmin == None: |
---|
| 324 | self.qmin = 1e-16 |
---|
| 325 | if qmax == None: |
---|
[89f3b66] | 326 | self.qmax = math.sqrt(x_max * x_max + y_max * y_max) |
---|
[70bf68c] | 327 | ## new error image for fitting purpose |
---|
[89f3b66] | 328 | if self.err_data == None or self.err_data == []: |
---|
| 329 | self.res_err_data = numpy.ones(len(self.data)) |
---|
[70bf68c] | 330 | else: |
---|
[da58fcc] | 331 | self.res_err_data = copy.deepcopy(self.err_data) |
---|
[9e8c150] | 332 | #self.res_err_data[self.res_err_data==0]=1 |
---|
[d8a2e31] | 333 | |
---|
[89f3b66] | 334 | self.radius = numpy.sqrt(self.qx_data**2 + self.qy_data**2) |
---|
[83195f7] | 335 | |
---|
| 336 | # Note: mask = True: for MASK while mask = False for NOT to mask |
---|
[1f9f3c8a] | 337 | self.idx = ((self.qmin <= self.radius) &\ |
---|
[89f3b66] | 338 | (self.radius <= self.qmax)) |
---|
[444c900e] | 339 | self.idx = (self.idx) & (self.mask) |
---|
| 340 | self.idx = (self.idx) & (numpy.isfinite(self.data)) |
---|
[0766d6d] | 341 | |
---|
[1f9f3c8a] | 342 | def set_smearer(self, smearer): |
---|
[f72333f] | 343 | """ |
---|
[1f9f3c8a] | 344 | Set smearer |
---|
[f72333f] | 345 | """ |
---|
| 346 | if smearer == None: |
---|
| 347 | return |
---|
| 348 | self.smearer = smearer |
---|
[444c900e] | 349 | self.smearer.set_index(self.idx) |
---|
[f72333f] | 350 | self.smearer.get_data() |
---|
| 351 | |
---|
[c4d6900] | 352 | def set_fit_range(self, qmin=None, qmax=None): |
---|
[1f9f3c8a] | 353 | """ |
---|
| 354 | To set the fit range |
---|
| 355 | """ |
---|
[89f3b66] | 356 | if qmin == 0.0: |
---|
[773806e] | 357 | self.qmin = 1e-16 |
---|
[1f9f3c8a] | 358 | elif qmin != None: |
---|
| 359 | self.qmin = qmin |
---|
[89f3b66] | 360 | if qmax != None: |
---|
[1f9f3c8a] | 361 | self.qmax = qmax |
---|
[89f3b66] | 362 | self.radius = numpy.sqrt(self.qx_data**2 + self.qy_data**2) |
---|
[1f9f3c8a] | 363 | self.idx = ((self.qmin <= self.radius) &\ |
---|
[89f3b66] | 364 | (self.radius <= self.qmax)) |
---|
[1f9f3c8a] | 365 | self.idx = (self.idx) & (self.mask) |
---|
[444c900e] | 366 | self.idx = (self.idx) & (numpy.isfinite(self.data)) |
---|
| 367 | self.idx = (self.idx) & (self.res_err_data != 0) |
---|
[0766d6d] | 368 | |
---|
[c4d6900] | 369 | def get_fit_range(self): |
---|
[7d0c1a8] | 370 | """ |
---|
[aa36f96] | 371 | return the range of data.x to fit |
---|
[7d0c1a8] | 372 | """ |
---|
[20d30e9] | 373 | return self.qmin, self.qmax |
---|
[7d0c1a8] | 374 | |
---|
[1f9f3c8a] | 375 | def residuals(self, fn): |
---|
[83195f7] | 376 | """ |
---|
[aa36f96] | 377 | return the residuals |
---|
[1f9f3c8a] | 378 | """ |
---|
[f72333f] | 379 | if self.smearer != None: |
---|
[444c900e] | 380 | fn.set_index(self.idx) |
---|
[f72333f] | 381 | # Get necessary data from self.data and set the data for smearing |
---|
| 382 | fn.get_data() |
---|
| 383 | |
---|
[1f9f3c8a] | 384 | gn = fn.get_value() |
---|
[f72333f] | 385 | else: |
---|
[444c900e] | 386 | gn = fn([self.qx_data[self.idx], |
---|
| 387 | self.qy_data[self.idx]]) |
---|
[83195f7] | 388 | # use only the data point within ROI range |
---|
[1f9f3c8a] | 389 | res = (self.data[self.idx] - gn) / self.res_err_data[self.idx] |
---|
[0766d6d] | 390 | |
---|
[1f9f3c8a] | 391 | return res, gn |
---|
[0e51519] | 392 | |
---|
[7d0c1a8] | 393 | def residuals_deriv(self, model, pars=[]): |
---|
[1f9f3c8a] | 394 | """ |
---|
[aa36f96] | 395 | :return: residuals derivatives . |
---|
| 396 | |
---|
| 397 | :note: in this case just return empty array |
---|
| 398 | |
---|
[7d0c1a8] | 399 | """ |
---|
| 400 | return [] |
---|
[48882d1] | 401 | |
---|
[1f9f3c8a] | 402 | |
---|
[4bd557d] | 403 | class FitAbort(Exception): |
---|
| 404 | """ |
---|
[aa36f96] | 405 | Exception raise to stop the fit |
---|
[4bd557d] | 406 | """ |
---|
[1ab9dc1] | 407 | #pass |
---|
| 408 | #print"Creating fit abort Exception" |
---|
[4bd557d] | 409 | |
---|
| 410 | |
---|
[70bf68c] | 411 | class SansAssembly: |
---|
[ca6d914] | 412 | """ |
---|
[aa36f96] | 413 | Sans Assembly class a class wrapper to be call in optimizer.leastsq method |
---|
[ca6d914] | 414 | """ |
---|
[1f9f3c8a] | 415 | def __init__(self, paramlist, model=None, data=None, fitresult=None, |
---|
[ba7dceb] | 416 | handler=None, curr_thread=None, msg_q=None): |
---|
[ca6d914] | 417 | """ |
---|
[aa36f96] | 418 | :param Model: the model wrapper fro sans -model |
---|
| 419 | :param Data: the data wrapper for sans data |
---|
| 420 | |
---|
[ca6d914] | 421 | """ |
---|
[e0072082] | 422 | self.model = model |
---|
[1f9f3c8a] | 423 | self.data = data |
---|
[e0072082] | 424 | self.paramlist = paramlist |
---|
[ba7dceb] | 425 | self.msg_q = msg_q |
---|
[e0072082] | 426 | self.curr_thread = curr_thread |
---|
| 427 | self.handler = handler |
---|
| 428 | self.fitresult = fitresult |
---|
| 429 | self.res = [] |
---|
[4b5bd73] | 430 | self.true_res = [] |
---|
[e0072082] | 431 | self.func_name = "Functor" |
---|
[425e49ca] | 432 | self.theory = None |
---|
[e0072082] | 433 | |
---|
[c4d6900] | 434 | def chisq(self): |
---|
[48882d1] | 435 | """ |
---|
[aa36f96] | 436 | Calculates chi^2 |
---|
| 437 | |
---|
| 438 | :param params: list of parameter values |
---|
| 439 | |
---|
| 440 | :return: chi^2 |
---|
| 441 | |
---|
[48882d1] | 442 | """ |
---|
[1f9f3c8a] | 443 | total = 0 |
---|
[4b5bd73] | 444 | for item in self.true_res: |
---|
[1f9f3c8a] | 445 | total += item * item |
---|
[4b5bd73] | 446 | if len(self.true_res) == 0: |
---|
[4bd557d] | 447 | return None |
---|
[1f9f3c8a] | 448 | return total / len(self.true_res) |
---|
[20d30e9] | 449 | |
---|
[c4d6900] | 450 | def __call__(self, params): |
---|
[ca6d914] | 451 | """ |
---|
[1f9f3c8a] | 452 | Compute residuals |
---|
| 453 | :param params: value of parameters to fit |
---|
| 454 | """ |
---|
[4b5bd73] | 455 | #import thread |
---|
| 456 | self.model.set_params(self.paramlist, params) |
---|
[0766d6d] | 457 | #print "params", params |
---|
[5722d66] | 458 | self.true_res, theory = self.data.residuals(self.model.eval) |
---|
| 459 | self.theory = copy.deepcopy(theory) |
---|
[4b5bd73] | 460 | # check parameters range |
---|
| 461 | if self.check_param_range(): |
---|
| 462 | # if the param value is outside of the bound |
---|
| 463 | # just silent return res = inf |
---|
| 464 | return self.res |
---|
[1f9f3c8a] | 465 | self.res = self.true_res |
---|
[ba7dceb] | 466 | |
---|
| 467 | if self.fitresult is not None: |
---|
[e0072082] | 468 | self.fitresult.set_model(model=self.model) |
---|
[444c900e] | 469 | self.fitresult.residuals = self.true_res |
---|
[bd7a426] | 470 | self.fitresult.iterations += 1 |
---|
[444c900e] | 471 | self.fitresult.theory = theory |
---|
[ba7dceb] | 472 | |
---|
[4b5bd73] | 473 | #fitness = self.chisq(params=params) |
---|
[c4d6900] | 474 | fitness = self.chisq() |
---|
[511c6810] | 475 | self.fitresult.pvec = params |
---|
[90c9cdf] | 476 | self.fitresult.set_fitness(fitness=fitness) |
---|
[ba7dceb] | 477 | if self.msg_q is not None: |
---|
| 478 | self.msg_q.put(self.fitresult) |
---|
| 479 | |
---|
| 480 | if self.handler is not None: |
---|
| 481 | self.handler.set_result(result=self.fitresult) |
---|
| 482 | self.handler.update_fit() |
---|
[4b5bd73] | 483 | |
---|
[1f9f3c8a] | 484 | if self.curr_thread != None: |
---|
[d5f0f5e3] | 485 | try: |
---|
[078f2f2] | 486 | self.curr_thread.isquit() |
---|
| 487 | except: |
---|
[1f9f3c8a] | 488 | msg = "Fitting: Terminated... Note: Forcing to stop " |
---|
[acfff8b] | 489 | msg += "fitting may cause a 'Functor error message' " |
---|
| 490 | msg += "being recorded in the log file....." |
---|
[1ab9dc1] | 491 | self.handler.error(msg) |
---|
| 492 | raise |
---|
[12cd4ec] | 493 | |
---|
[48882d1] | 494 | return self.res |
---|
| 495 | |
---|
[4b5bd73] | 496 | def check_param_range(self): |
---|
| 497 | """ |
---|
| 498 | Check the lower and upper bound of the parameter value |
---|
| 499 | and set res to the inf if the value is outside of the |
---|
| 500 | range |
---|
| 501 | :limitation: the initial values must be within range. |
---|
| 502 | """ |
---|
| 503 | |
---|
[bdc25e2] | 504 | #time.sleep(0.01) |
---|
[4b5bd73] | 505 | is_outofbound = False |
---|
| 506 | # loop through the fit parameters |
---|
| 507 | for p in self.model.parameterset: |
---|
| 508 | param_name = p.get_name() |
---|
| 509 | if param_name in self.paramlist: |
---|
| 510 | |
---|
| 511 | # if the range was defined, check the range |
---|
| 512 | if numpy.isfinite(p.range[0]): |
---|
| 513 | if p.value == 0: |
---|
| 514 | # This value works on Scipy |
---|
| 515 | # Do not change numbers below |
---|
[1f9f3c8a] | 516 | value = _SMALLVALUE |
---|
[4b5bd73] | 517 | else: |
---|
| 518 | value = p.value |
---|
| 519 | # For leastsq, it needs a bit step back from the boundary |
---|
[1f9f3c8a] | 520 | val = p.range[0] - value * _SMALLVALUE |
---|
| 521 | if p.value < val: |
---|
[4b5bd73] | 522 | self.res *= 1e+6 |
---|
| 523 | |
---|
| 524 | is_outofbound = True |
---|
| 525 | break |
---|
| 526 | if numpy.isfinite(p.range[1]): |
---|
| 527 | # This value works on Scipy |
---|
| 528 | # Do not change numbers below |
---|
| 529 | if p.value == 0: |
---|
[1f9f3c8a] | 530 | value = _SMALLVALUE |
---|
[4b5bd73] | 531 | else: |
---|
| 532 | value = p.value |
---|
| 533 | # For leastsq, it needs a bit step back from the boundary |
---|
[1f9f3c8a] | 534 | val = p.range[1] + value * _SMALLVALUE |
---|
[4b5bd73] | 535 | if p.value > val: |
---|
| 536 | self.res *= 1e+6 |
---|
| 537 | is_outofbound = True |
---|
| 538 | break |
---|
| 539 | |
---|
| 540 | return is_outofbound |
---|
| 541 | |
---|
| 542 | |
---|
[4c718654] | 543 | class FitEngine: |
---|
[ee5b04c] | 544 | def __init__(self): |
---|
[ca6d914] | 545 | """ |
---|
[aa36f96] | 546 | Base class for scipy and park fit engine |
---|
[ca6d914] | 547 | """ |
---|
| 548 | #List of parameter names to fit |
---|
[b2f25dc5] | 549 | self.param_list = [] |
---|
[ca6d914] | 550 | #Dictionnary of fitArrange element (fit problems) |
---|
[b2f25dc5] | 551 | self.fit_arrange_dict = {} |
---|
[06e7c26] | 552 | self.fitter_id = None |
---|
[7db52f1] | 553 | |
---|
[1f9f3c8a] | 554 | def set_model(self, model, id, pars=[], constraints=[], data=None): |
---|
[4c718654] | 555 | """ |
---|
[c4d6900] | 556 | set a model on a given in the fit engine. |
---|
[aa36f96] | 557 | |
---|
| 558 | :param model: sans.models type |
---|
[c4d6900] | 559 | :param : is the key of the fitArrange dictionary where model is |
---|
[aa36f96] | 560 | saved as a value |
---|
| 561 | :param pars: the list of parameters to fit |
---|
| 562 | :param constraints: list of |
---|
| 563 | tuple (name of parameter, value of parameters) |
---|
| 564 | the value of parameter must be a string to constraint 2 different |
---|
| 565 | parameters. |
---|
| 566 | Example: |
---|
| 567 | we want to fit 2 model M1 and M2 both have parameters A and B. |
---|
| 568 | constraints can be: |
---|
| 569 | constraints = [(M1.A, M2.B+2), (M1.B= M2.A *5),...,] |
---|
| 570 | |
---|
| 571 | |
---|
| 572 | :note: pars must contains only name of existing model's parameters |
---|
| 573 | |
---|
[ca6d914] | 574 | """ |
---|
[fd6b789] | 575 | if model == None: |
---|
| 576 | raise ValueError, "AbstractFitEngine: Need to set model to fit" |
---|
[393f0f3] | 577 | |
---|
[89f3b66] | 578 | new_model = model |
---|
[393f0f3] | 579 | if not issubclass(model.__class__, Model): |
---|
[1cff677] | 580 | new_model = Model(model, data) |
---|
[fd6b789] | 581 | |
---|
[89f3b66] | 582 | if len(constraints) > 0: |
---|
[fd6b789] | 583 | for constraint in constraints: |
---|
| 584 | name, value = constraint |
---|
| 585 | try: |
---|
[89f3b66] | 586 | new_model.parameterset[str(name)].set(str(value)) |
---|
[fd6b789] | 587 | except: |
---|
[89f3b66] | 588 | msg = "Fit Engine: Error occurs when setting the constraint" |
---|
[c4d6900] | 589 | msg += " %s for parameter %s " % (value, name) |
---|
[fd6b789] | 590 | raise ValueError, msg |
---|
| 591 | |
---|
[89f3b66] | 592 | if len(pars) > 0: |
---|
| 593 | temp = [] |
---|
[fd6b789] | 594 | for item in pars: |
---|
| 595 | if item in new_model.model.getParamList(): |
---|
| 596 | temp.append(item) |
---|
[b2f25dc5] | 597 | self.param_list.append(item) |
---|
[fd6b789] | 598 | else: |
---|
| 599 | |
---|
[89f3b66] | 600 | msg = "wrong parameter %s used" % str(item) |
---|
| 601 | msg += "to set model %s. Choose" % str(new_model.model.name) |
---|
| 602 | msg += "parameter name within %s" % \ |
---|
| 603 | str(new_model.model.getParamList()) |
---|
| 604 | raise ValueError, msg |
---|
[fd6b789] | 605 | |
---|
[c4d6900] | 606 | #A fitArrange is already created but contains data_list only at id |
---|
| 607 | if self.fit_arrange_dict.has_key(id): |
---|
| 608 | self.fit_arrange_dict[id].set_model(new_model) |
---|
| 609 | self.fit_arrange_dict[id].pars = pars |
---|
[6831a99] | 610 | else: |
---|
[c4d6900] | 611 | #no fitArrange object has been create with this id |
---|
[48882d1] | 612 | fitproblem = FitArrange() |
---|
[fd6b789] | 613 | fitproblem.set_model(new_model) |
---|
[89f3b66] | 614 | fitproblem.pars = pars |
---|
[c4d6900] | 615 | self.fit_arrange_dict[id] = fitproblem |
---|
[7db52f1] | 616 | vals = [] |
---|
| 617 | for name in pars: |
---|
| 618 | vals.append(new_model.model.getParam(name)) |
---|
| 619 | self.fit_arrange_dict[id].vals = vals |
---|
[d4b0687] | 620 | else: |
---|
[6831a99] | 621 | raise ValueError, "park_integration:missing parameters" |
---|
[48882d1] | 622 | |
---|
[c4d6900] | 623 | def set_data(self, data, id, smearer=None, qmin=None, qmax=None): |
---|
[1f9f3c8a] | 624 | """ |
---|
[aa36f96] | 625 | Receives plottable, creates a list of data to fit,set data |
---|
[1f9f3c8a] | 626 | in a FitArrange object and adds that object in a dictionary |
---|
[c4d6900] | 627 | with key id. |
---|
[aa36f96] | 628 | |
---|
| 629 | :param data: data added |
---|
[c4d6900] | 630 | :param id: unique key corresponding to a fitArrange object with data |
---|
[ca6d914] | 631 | """ |
---|
[89f3b66] | 632 | if data.__class__.__name__ == 'Data2D': |
---|
| 633 | fitdata = FitData2D(sans_data2d=data, data=data.data, |
---|
| 634 | err_data=data.err_data) |
---|
[f8ce013] | 635 | else: |
---|
[1f9f3c8a] | 636 | fitdata = FitData1D(x=data.x, y=data.y, |
---|
[89f3b66] | 637 | dx=data.dx, dy=data.dy, smearer=smearer) |
---|
[634ca14] | 638 | fitdata.sans_data = data |
---|
[393f0f3] | 639 | |
---|
[c4d6900] | 640 | fitdata.set_fit_range(qmin=qmin, qmax=qmax) |
---|
| 641 | #A fitArrange is already created but contains model only at id |
---|
[1f9f3c8a] | 642 | if id in self.fit_arrange_dict: |
---|
[c4d6900] | 643 | self.fit_arrange_dict[id].add_data(fitdata) |
---|
[d4b0687] | 644 | else: |
---|
[c4d6900] | 645 | #no fitArrange object has been create with this id |
---|
[89f3b66] | 646 | fitproblem = FitArrange() |
---|
[f8ce013] | 647 | fitproblem.add_data(fitdata) |
---|
[1f9f3c8a] | 648 | self.fit_arrange_dict[id] = fitproblem |
---|
[20d30e9] | 649 | |
---|
[c4d6900] | 650 | def get_model(self, id): |
---|
[1f9f3c8a] | 651 | """ |
---|
[c4d6900] | 652 | :param id: id is key in the dictionary containing the model to return |
---|
[aa36f96] | 653 | |
---|
[1f9f3c8a] | 654 | :return: a model at this id or None if no FitArrange element was |
---|
[c4d6900] | 655 | created with this id |
---|
[d4b0687] | 656 | """ |
---|
[1f9f3c8a] | 657 | if id in self.fit_arrange_dict: |
---|
[c4d6900] | 658 | return self.fit_arrange_dict[id].get_model() |
---|
[d4b0687] | 659 | else: |
---|
| 660 | return None |
---|
| 661 | |
---|
[c4d6900] | 662 | def remove_fit_problem(self, id): |
---|
| 663 | """remove fitarrange in id""" |
---|
[1f9f3c8a] | 664 | if id in self.fit_arrange_dict: |
---|
[c4d6900] | 665 | del self.fit_arrange_dict[id] |
---|
[a9e04aa] | 666 | |
---|
[c4d6900] | 667 | def select_problem_for_fit(self, id, value): |
---|
[a9e04aa] | 668 | """ |
---|
[c4d6900] | 669 | select a couple of model and data at the id position in dictionary |
---|
[aa36f96] | 670 | and set in self.selected value to value |
---|
| 671 | |
---|
[1f9f3c8a] | 672 | :param value: the value to allow fitting. |
---|
[aa36f96] | 673 | can only have the value one or zero |
---|
[a9e04aa] | 674 | """ |
---|
[1f9f3c8a] | 675 | if id in self.fit_arrange_dict: |
---|
[c4d6900] | 676 | self.fit_arrange_dict[id].set_to_fit(value) |
---|
[eef2e0ed] | 677 | |
---|
[c4d6900] | 678 | def get_problem_to_fit(self, id): |
---|
[a9e04aa] | 679 | """ |
---|
[c4d6900] | 680 | return the self.selected value of the fit problem of id |
---|
[aa36f96] | 681 | |
---|
[c4d6900] | 682 | :param id: the id of the problem |
---|
[a9e04aa] | 683 | """ |
---|
[1f9f3c8a] | 684 | if id in self.fit_arrange_dict: |
---|
[c4d6900] | 685 | self.fit_arrange_dict[id].get_to_fit() |
---|
[4c718654] | 686 | |
---|
[1f9f3c8a] | 687 | |
---|
[d4b0687] | 688 | class FitArrange: |
---|
| 689 | def __init__(self): |
---|
| 690 | """ |
---|
[aa36f96] | 691 | Class FitArrange contains a set of data for a given model |
---|
| 692 | to perform the Fit.FitArrange must contain exactly one model |
---|
| 693 | and at least one data for the fit to be performed. |
---|
| 694 | |
---|
| 695 | model: the model selected by the user |
---|
| 696 | Ldata: a list of data what the user wants to fit |
---|
[d4b0687] | 697 | |
---|
| 698 | """ |
---|
| 699 | self.model = None |
---|
[c4d6900] | 700 | self.data_list = [] |
---|
[89f3b66] | 701 | self.pars = [] |
---|
[7db52f1] | 702 | self.vals = [] |
---|
[a9e04aa] | 703 | self.selected = 0 |
---|
[d4b0687] | 704 | |
---|
[89f3b66] | 705 | def set_model(self, model): |
---|
[1f9f3c8a] | 706 | """ |
---|
[aa36f96] | 707 | set_model save a copy of the model |
---|
| 708 | |
---|
| 709 | :param model: the model being set |
---|
[d4b0687] | 710 | """ |
---|
| 711 | self.model = model |
---|
| 712 | |
---|
[89f3b66] | 713 | def add_data(self, data): |
---|
[1f9f3c8a] | 714 | """ |
---|
[c4d6900] | 715 | add_data fill a self.data_list with data to fit |
---|
[aa36f96] | 716 | |
---|
[1f9f3c8a] | 717 | :param data: Data to add in the list |
---|
[d4b0687] | 718 | """ |
---|
[c4d6900] | 719 | if not data in self.data_list: |
---|
| 720 | self.data_list.append(data) |
---|
[d4b0687] | 721 | |
---|
| 722 | def get_model(self): |
---|
[aa36f96] | 723 | """ |
---|
[1f9f3c8a] | 724 | :return: saved model |
---|
| 725 | """ |
---|
| 726 | return self.model |
---|
[d4b0687] | 727 | |
---|
| 728 | def get_data(self): |
---|
[1f9f3c8a] | 729 | """ |
---|
[c4d6900] | 730 | :return: list of data data_list |
---|
[aa36f96] | 731 | """ |
---|
[1f9f3c8a] | 732 | return self.data_list[0] |
---|
[d4b0687] | 733 | |
---|
[89f3b66] | 734 | def remove_data(self, data): |
---|
[d4b0687] | 735 | """ |
---|
[aa36f96] | 736 | Remove one element from the list |
---|
| 737 | |
---|
[c4d6900] | 738 | :param data: Data to remove from data_list |
---|
[d4b0687] | 739 | """ |
---|
[c4d6900] | 740 | if data in self.data_list: |
---|
| 741 | self.data_list.remove(data) |
---|
[aa36f96] | 742 | |
---|
[1f9f3c8a] | 743 | def set_to_fit(self, value=0): |
---|
[a9e04aa] | 744 | """ |
---|
[aa36f96] | 745 | set self.selected to 0 or 1 for other values raise an exception |
---|
| 746 | |
---|
| 747 | :param value: integer between 0 or 1 |
---|
[a9e04aa] | 748 | """ |
---|
[89f3b66] | 749 | self.selected = value |
---|
[a9e04aa] | 750 | |
---|
| 751 | def get_to_fit(self): |
---|
| 752 | """ |
---|
[aa36f96] | 753 | return self.selected value |
---|
[a9e04aa] | 754 | """ |
---|
| 755 | return self.selected |
---|
[444c900e] | 756 | |
---|
| 757 | |
---|
| 758 | IS_MAC = True |
---|
| 759 | if sys.platform.count("win32") > 0: |
---|
| 760 | IS_MAC = False |
---|
[1f9f3c8a] | 761 | |
---|
| 762 | |
---|
[444c900e] | 763 | class FResult(object): |
---|
| 764 | """ |
---|
| 765 | Storing fit result |
---|
| 766 | """ |
---|
| 767 | def __init__(self, model=None, param_list=None, data=None): |
---|
| 768 | self.calls = None |
---|
[06e7c26] | 769 | self.pars = [] |
---|
[444c900e] | 770 | self.fitness = None |
---|
| 771 | self.chisqr = None |
---|
| 772 | self.pvec = [] |
---|
| 773 | self.cov = [] |
---|
| 774 | self.info = None |
---|
| 775 | self.mesg = None |
---|
| 776 | self.success = None |
---|
| 777 | self.stderr = None |
---|
| 778 | self.residuals = [] |
---|
| 779 | self.index = [] |
---|
| 780 | self.parameters = None |
---|
| 781 | self.is_mac = IS_MAC |
---|
| 782 | self.model = model |
---|
| 783 | self.data = data |
---|
| 784 | self.theory = [] |
---|
| 785 | self.param_list = param_list |
---|
| 786 | self.iterations = 0 |
---|
| 787 | self.inputs = [] |
---|
[06e7c26] | 788 | self.fitter_id = None |
---|
[444c900e] | 789 | if self.model is not None and self.data is not None: |
---|
| 790 | self.inputs = [(self.model, self.data)] |
---|
| 791 | |
---|
| 792 | def set_model(self, model): |
---|
| 793 | """ |
---|
| 794 | """ |
---|
| 795 | self.model = model |
---|
| 796 | |
---|
| 797 | def set_fitness(self, fitness): |
---|
| 798 | """ |
---|
| 799 | """ |
---|
| 800 | self.fitness = fitness |
---|
| 801 | |
---|
| 802 | def __str__(self): |
---|
| 803 | """ |
---|
| 804 | """ |
---|
| 805 | if self.pvec == None and self.model is None and self.param_list is None: |
---|
| 806 | return "No results" |
---|
| 807 | n = len(self.model.parameterset) |
---|
[bd7a426] | 808 | |
---|
[444c900e] | 809 | result_param = zip(xrange(n), self.model.parameterset) |
---|
| 810 | msg1 = ["[Iteration #: %s ]" % self.iterations] |
---|
| 811 | msg3 = ["=== goodness of fit: %s ===" % (str(self.fitness))] |
---|
| 812 | if not self.is_mac: |
---|
| 813 | msg2 = ["P%-3d %s......|.....%s" % \ |
---|
| 814 | (p[0], p[1], p[1].value)\ |
---|
| 815 | for p in result_param if p[1].name in self.param_list] |
---|
[1f9f3c8a] | 816 | msg = msg1 + msg3 + msg2 |
---|
[444c900e] | 817 | else: |
---|
| 818 | msg = msg1 + msg3 |
---|
| 819 | msg = "\n".join(msg) |
---|
| 820 | return msg |
---|
| 821 | |
---|
| 822 | def print_summary(self): |
---|
| 823 | """ |
---|
| 824 | """ |
---|
[1f9f3c8a] | 825 | print self |
---|