1 | |
---|
2 | from sans.models.BaseComponent import BaseComponent |
---|
3 | from sans.models.ReflModel import ReflModel |
---|
4 | from copy import deepcopy |
---|
5 | from math import floor |
---|
6 | from scipy.special import erf |
---|
7 | func_list = {'Erf':0, 'Linear':1, 'RParabolic':2, \ |
---|
8 | 'LParabola':3, 'RCubic':4, 'LCubic':5} |
---|
9 | max_nshells = 10 |
---|
10 | class ReflectivityModel(BaseComponent): |
---|
11 | """ |
---|
12 | This multi-model is based on Parratt formalism and provides the capability |
---|
13 | of changing the number of layers between 0 and 10. |
---|
14 | """ |
---|
15 | def __init__(self, multfactor=1): |
---|
16 | BaseComponent.__init__(self) |
---|
17 | """ |
---|
18 | :param multfactor: number of layers in the model, assumes 0<= n_shells <=10. |
---|
19 | """ |
---|
20 | |
---|
21 | ## Setting model name model description |
---|
22 | self.description="" |
---|
23 | model = ReflModel() |
---|
24 | self.model = model |
---|
25 | self.name = "ReflectivityModel" |
---|
26 | self.description=model.description |
---|
27 | self.n_layers = multfactor |
---|
28 | ## Define parameters |
---|
29 | self.params = {} |
---|
30 | |
---|
31 | ## Parameter details [units, min, max] |
---|
32 | self.details = {} |
---|
33 | |
---|
34 | # non-fittable parameters |
---|
35 | self.non_fittable = model.non_fittable |
---|
36 | |
---|
37 | # list of function in order of the function number |
---|
38 | self.fun_list = self._get_func_list() |
---|
39 | ## dispersion |
---|
40 | self._set_dispersion() |
---|
41 | ## Define parameters |
---|
42 | self._set_params() |
---|
43 | |
---|
44 | ## Parameter details [units, min, max] |
---|
45 | self._set_details() |
---|
46 | |
---|
47 | #list of parameter that can be fitted |
---|
48 | self._set_fixed_params() |
---|
49 | self.model.params['n_layers'] = self.n_layers |
---|
50 | |
---|
51 | ## functional multiplicity info of the model |
---|
52 | # [int(maximum no. of functionality),"str(Titl), |
---|
53 | # [str(name of function0),...], [str(x-asix name of sld),...]] |
---|
54 | self.multiplicity_info = [max_nshells,"No. of Layers:",[],['Depth']] |
---|
55 | |
---|
56 | |
---|
57 | def _clone(self, obj): |
---|
58 | """ |
---|
59 | Internal utility function to copy the internal |
---|
60 | data members to a fresh copy. |
---|
61 | """ |
---|
62 | obj.params = deepcopy(self.params) |
---|
63 | obj.non_fittable = deepcopy(self.non_fittable) |
---|
64 | obj.description = deepcopy(self.description) |
---|
65 | obj.details = deepcopy(self.details) |
---|
66 | obj.dispersion = deepcopy(self.dispersion) |
---|
67 | obj.model = self.model.clone() |
---|
68 | |
---|
69 | return obj |
---|
70 | |
---|
71 | |
---|
72 | def _set_dispersion(self): |
---|
73 | """ |
---|
74 | model dispersions |
---|
75 | """ |
---|
76 | ##set dispersion from model |
---|
77 | self.dispersion = {} |
---|
78 | |
---|
79 | |
---|
80 | def _set_params(self): |
---|
81 | """ |
---|
82 | Concatenate the parameters of the model to create |
---|
83 | this model parameters |
---|
84 | """ |
---|
85 | # rearrange the parameters for the given # of shells |
---|
86 | for name , value in self.model.params.iteritems(): |
---|
87 | n = 0 |
---|
88 | pos = len(name.split('_'))-1 |
---|
89 | if name.split('_')[0] == 'sldIM': |
---|
90 | continue |
---|
91 | elif name.split('_')[0] == 'func': |
---|
92 | n= -1 |
---|
93 | while n<self.n_layers: |
---|
94 | n += 1 |
---|
95 | if name.split('_')[pos] == 'inter%s' % str(n): |
---|
96 | self.params[name]=value |
---|
97 | continue |
---|
98 | #continue |
---|
99 | elif name.split('_')[pos][0:5] == 'inter': |
---|
100 | n= -1 |
---|
101 | while n<self.n_layers: |
---|
102 | n += 1 |
---|
103 | if name.split('_')[pos] == 'inter%s' % str(n): |
---|
104 | self.params[name]= value |
---|
105 | continue |
---|
106 | elif name.split('_')[pos][0:4] == 'flat': |
---|
107 | while n<self.n_layers: |
---|
108 | n += 1 |
---|
109 | if name.split('_')[pos] == 'flat%s' % str(n): |
---|
110 | self.params[name]= value |
---|
111 | continue |
---|
112 | elif name == 'n_layers': |
---|
113 | continue |
---|
114 | else: |
---|
115 | self.params[name]= value |
---|
116 | |
---|
117 | self.model.params['n_layers'] = self.n_layers |
---|
118 | |
---|
119 | # set constrained values for the original model params |
---|
120 | self._set_xtra_model_param() |
---|
121 | |
---|
122 | def _set_details(self): |
---|
123 | """ |
---|
124 | Concatenate details of the original model to create |
---|
125 | this model details |
---|
126 | """ |
---|
127 | for name ,detail in self.model.details.iteritems(): |
---|
128 | if name in self.params.iterkeys(): |
---|
129 | self.details[name]= detail |
---|
130 | |
---|
131 | |
---|
132 | def _set_xtra_model_param(self): |
---|
133 | """ |
---|
134 | Set params of original model that are hidden from this model |
---|
135 | """ |
---|
136 | # look for the model parameters that are not in param list |
---|
137 | for key in self.model.params.iterkeys(): |
---|
138 | if key not in self.params.keys(): |
---|
139 | if key.split('_')[0] == 'thick': |
---|
140 | self.model.setParam(key, 0) |
---|
141 | continue |
---|
142 | if key.split('_')[0] == 'func': |
---|
143 | self.model.setParam(key, 0) |
---|
144 | continue |
---|
145 | for nshell in range(self.n_layers,max_nshells): |
---|
146 | if key.split('_')[1] == 'flat%s' % str(nshell+1): |
---|
147 | try: |
---|
148 | if key.split('_')[0] == 'sld': |
---|
149 | value = self.model.params['sld_medium'] |
---|
150 | elif key.split('_')[0] == 'sldIM': |
---|
151 | value = self.model.params['sldIM_medium'] |
---|
152 | self.model.setParam(key, value) |
---|
153 | except: pass |
---|
154 | |
---|
155 | def _get_func_list(self): |
---|
156 | """ |
---|
157 | Get the list of functions in each layer (shell) |
---|
158 | """ |
---|
159 | #func_list = {} |
---|
160 | return func_list |
---|
161 | |
---|
162 | def getProfile(self): |
---|
163 | """ |
---|
164 | Get SLD profile |
---|
165 | |
---|
166 | : return: (z, beta) where z is a list of depth of the transition points |
---|
167 | beta is a list of the corresponding SLD values |
---|
168 | """ |
---|
169 | # max_pts for each layers |
---|
170 | n_sub = 21 |
---|
171 | z = [] |
---|
172 | beta = [] |
---|
173 | sub_range = floor(n_sub/2.0) |
---|
174 | z.append(0) |
---|
175 | beta.append(self.params['sld_sub0']) |
---|
176 | |
---|
177 | z0 = 0 |
---|
178 | # for layers from the top |
---|
179 | for n in range(1,self.n_layers+2): |
---|
180 | i = n |
---|
181 | |
---|
182 | for j in range(0,2): |
---|
183 | for n_s in range(-sub_range,sub_range+1): |
---|
184 | if j==1: |
---|
185 | if i==self.n_layers+1: |
---|
186 | break |
---|
187 | # shift half sub thickness for the first point |
---|
188 | z0 += dz/2.0 |
---|
189 | z.append(z0) |
---|
190 | #z0 -= dz/2.0 |
---|
191 | z0 += self.params['thick_flat%s'% str(i)] |
---|
192 | |
---|
193 | sld_i = self.params['sld_flat%s'% str(i)] |
---|
194 | beta.append(self.params['sld_flat%s'% str(i)]) |
---|
195 | else: |
---|
196 | |
---|
197 | dz = self.params['thick_inter%s'% str(i-1)]/n_sub |
---|
198 | |
---|
199 | if n_s == -sub_range: |
---|
200 | # shift half sub thickness for the first point |
---|
201 | z0 -= dz/2.0 |
---|
202 | #exec "dz = self.params['thick_inter[%s-1]'% str(i)]/9" |
---|
203 | #print "%d = %g \n"% (i,self.params['thick_inter3']) |
---|
204 | z0 += dz |
---|
205 | |
---|
206 | if i == 1: |
---|
207 | sld_l = self.params['sld_sub0'] |
---|
208 | else: |
---|
209 | sld_l = self.params['sld_flat%s'% str(i-1)] |
---|
210 | if i == self.n_layers+1: |
---|
211 | sld_r = self.params['sld_medium'] |
---|
212 | else: |
---|
213 | sld_r = self.params['sld_flat%s'% str(i)] |
---|
214 | func_idx = self.params['func_inter%s'% str(i-1)] |
---|
215 | func = self._get_func(n_s, n_sub, func_idx) |
---|
216 | if sld_r>sld_l: |
---|
217 | sld_i = (sld_r-sld_l)*func+sld_l |
---|
218 | elif sld_r<sld_l: |
---|
219 | sld_i = (sld_l-sld_r)*(1-func)+sld_r |
---|
220 | else: |
---|
221 | sld_i = sld_r |
---|
222 | z.append(z0) |
---|
223 | beta.append(sld_i) |
---|
224 | if j==1: break |
---|
225 | # put substrate and superstrate profile |
---|
226 | # shift half sub thickness for the first point |
---|
227 | z0 += dz/2.0 |
---|
228 | z.append(z0) |
---|
229 | beta.append(self.params['sld_medium']) |
---|
230 | z_ext = z0/6.0 |
---|
231 | |
---|
232 | # put the extra points for the substrate |
---|
233 | # and superstrate |
---|
234 | z.append(z0+z_ext) |
---|
235 | beta.append(self.params['sld_medium']) |
---|
236 | z.insert(0,-z_ext) |
---|
237 | beta.insert(0,self.params['sld_sub0']) |
---|
238 | z = [z0 - x for x in z] |
---|
239 | z.reverse() |
---|
240 | beta.reverse() |
---|
241 | return z, beta |
---|
242 | |
---|
243 | def _get_func(self, index, n_sub, func_idx): |
---|
244 | """ |
---|
245 | Get the function asked to buil sld profile |
---|
246 | : param index: index of sub_layer |
---|
247 | : param n_sub: total number of sub_layer |
---|
248 | : param func_idx: an integer to identify a function |
---|
249 | |
---|
250 | : return out: the output from the function, float |
---|
251 | """ |
---|
252 | # cal bin_size |
---|
253 | bin_size = 1.0/n_sub |
---|
254 | # erf |
---|
255 | if func_idx == 0: |
---|
256 | out = erf(index/(n_sub/5.0))/2.0 + 0.5 |
---|
257 | return out |
---|
258 | else: |
---|
259 | index += 0.5 |
---|
260 | # linear |
---|
261 | if func_idx == 1: |
---|
262 | out = ((index + floor(n_sub/2.0))*bin_size) |
---|
263 | # r_parabolic |
---|
264 | elif func_idx == 2: |
---|
265 | out = ((index + floor(n_sub/2.0))*bin_size)* \ |
---|
266 | ((index + floor(n_sub/2.0))*bin_size) |
---|
267 | # l_parabolic |
---|
268 | elif func_idx == 3: |
---|
269 | out = 1.0-(((index + floor(n_sub/2.0))*bin_size) - 1.0) *\ |
---|
270 | (((index + floor(n_sub/2.0))*bin_size) - 1.0) |
---|
271 | # r_cubic |
---|
272 | elif func_idx == 4: |
---|
273 | out = ((index + floor(n_sub/2.0))*bin_size)* \ |
---|
274 | ((index + floor(n_sub/2.0))*bin_size)* \ |
---|
275 | ((index + floor(n_sub/2.0))*bin_size) |
---|
276 | # l_cubic |
---|
277 | elif func_idx == 5: |
---|
278 | out = 1.0+(((index + floor(n_sub/2.0)))*bin_size - 1.0) *\ |
---|
279 | (((index + floor(n_sub/2.0)))*bin_size - 1.0) *\ |
---|
280 | (((index + floor(n_sub/2.0)))*bin_size - 1.0) |
---|
281 | # return output |
---|
282 | return out |
---|
283 | |
---|
284 | def setParam(self, name, value): |
---|
285 | """ |
---|
286 | Set the value of a model parameter |
---|
287 | |
---|
288 | : param name: name of the parameter |
---|
289 | : param value: value of the parameter |
---|
290 | """ |
---|
291 | # set param to new model |
---|
292 | self._setParamHelper( name, value) |
---|
293 | |
---|
294 | ## setParam to model |
---|
295 | if name=='sld_medium': |
---|
296 | # the sld_*** model.params not in params must set to value of sld_solv |
---|
297 | for key in self.model.params.iterkeys(): |
---|
298 | if key not in self.params.keys()and key.split('_')[0] == 'sld': |
---|
299 | self.model.setParam(key, value) |
---|
300 | |
---|
301 | self.model.setParam( name, value) |
---|
302 | |
---|
303 | def _setParamHelper(self, name, value): |
---|
304 | """ |
---|
305 | Helper function to setParam |
---|
306 | """ |
---|
307 | |
---|
308 | # Look for standard parameter |
---|
309 | for item in self.params.keys(): |
---|
310 | if item.lower()==name.lower(): |
---|
311 | self.params[item] = value |
---|
312 | return |
---|
313 | |
---|
314 | raise ValueError, "Model does not contain parameter %s" % name |
---|
315 | |
---|
316 | |
---|
317 | def _set_fixed_params(self): |
---|
318 | """ |
---|
319 | Fill the self.fixed list with the model fixed list |
---|
320 | """ |
---|
321 | pass |
---|
322 | |
---|
323 | def run(self, x = 0.0): |
---|
324 | """ |
---|
325 | Evaluate the model |
---|
326 | |
---|
327 | :param x: input q, or [q,phi] |
---|
328 | |
---|
329 | :return: scattering function P(q) |
---|
330 | |
---|
331 | """ |
---|
332 | |
---|
333 | return self.model.run(x) |
---|
334 | |
---|
335 | def runXY(self, x = 0.0): |
---|
336 | """ |
---|
337 | Evaluate the model |
---|
338 | |
---|
339 | : param x: input q-value (float or [float, float] as [qx, qy]) |
---|
340 | : return: scattering function value |
---|
341 | """ |
---|
342 | |
---|
343 | return self.model.runXY(x) |
---|
344 | |
---|
345 | ## Now (May27,10) directly uses the model eval function |
---|
346 | ## instead of the for-loop in Base Component. |
---|
347 | def evalDistribution(self, x = []): |
---|
348 | """ |
---|
349 | Evaluate the model in cartesian coordinates |
---|
350 | |
---|
351 | : param x: input q[], or [qx[], qy[]] |
---|
352 | : return: scattering function P(q[]) |
---|
353 | """ |
---|
354 | # set effective radius and scaling factor before run |
---|
355 | return self.model.evalDistribution(x) |
---|
356 | def calculate_ER(self): |
---|
357 | """ |
---|
358 | """ |
---|
359 | return self.model.calculate_ER() |
---|
360 | def set_dispersion(self, parameter, dispersion): |
---|
361 | """ |
---|
362 | Set the dispersion object for a model parameter |
---|
363 | |
---|
364 | : param parameter: name of the parameter [string] |
---|
365 | :dispersion: dispersion object of type DispersionModel |
---|
366 | """ |
---|
367 | pass |
---|