1 | """ |
---|
2 | SAS data representations. |
---|
3 | |
---|
4 | Plotting functions for data sets: |
---|
5 | |
---|
6 | :func:`plot_data` plots the data file. |
---|
7 | |
---|
8 | :func:`plot_theory` plots a calculated result from the model. |
---|
9 | |
---|
10 | Wrappers for the sasview data loader and data manipulations: |
---|
11 | |
---|
12 | :func:`load_data` loads a sasview data file. |
---|
13 | |
---|
14 | :func:`set_beam_stop` masks the beam stop from the data. |
---|
15 | |
---|
16 | :func:`set_half` selects the right or left half of the data, which can |
---|
17 | be useful for shear measurements which have not been properly corrected |
---|
18 | for path length and reflections. |
---|
19 | |
---|
20 | :func:`set_top` cuts the top part off the data. |
---|
21 | |
---|
22 | |
---|
23 | Empty data sets for evaluating models without data: |
---|
24 | |
---|
25 | :func:`empty_data1D` creates an empty dataset, which is useful for plotting |
---|
26 | a theory function before the data is measured. |
---|
27 | |
---|
28 | :func:`empty_data2D` creates an empty 2D dataset. |
---|
29 | |
---|
30 | Note that the empty datasets use a minimal representation of the SasView |
---|
31 | objects so that models can be run without SasView on the path. You could |
---|
32 | also use these for your own data loader. |
---|
33 | |
---|
34 | """ |
---|
35 | import traceback |
---|
36 | |
---|
37 | import numpy as np # type: ignore |
---|
38 | from numpy import sqrt, sin, cos, pi |
---|
39 | |
---|
40 | # pylint: disable=unused-import |
---|
41 | try: |
---|
42 | from typing import Union, Dict, List, Optional |
---|
43 | except ImportError: |
---|
44 | pass |
---|
45 | else: |
---|
46 | Data = Union["Data1D", "Data2D", "SesansData"] |
---|
47 | # pylint: enable=unused-import |
---|
48 | |
---|
49 | def load_data(filename, index=0): |
---|
50 | # type: (str) -> Data |
---|
51 | """ |
---|
52 | Load data using a sasview loader. |
---|
53 | """ |
---|
54 | from sas.sascalc.dataloader.loader import Loader # type: ignore |
---|
55 | loader = Loader() |
---|
56 | # Allow for one part in multipart file |
---|
57 | if '[' in filename: |
---|
58 | filename, indexstr = filename[:-1].split('[') |
---|
59 | index = int(indexstr) |
---|
60 | datasets = loader.load(filename) |
---|
61 | if not datasets: # None or [] |
---|
62 | raise IOError("Data %r could not be loaded" % filename) |
---|
63 | if not isinstance(datasets, list): |
---|
64 | datasets = [datasets] |
---|
65 | for data in datasets: |
---|
66 | if hasattr(data, 'x'): |
---|
67 | data.qmin, data.qmax = data.x.min(), data.x.max() |
---|
68 | data.mask = (np.isnan(data.y) if data.y is not None |
---|
69 | else np.zeros_like(data.x, dtype='bool')) |
---|
70 | elif hasattr(data, 'qx_data'): |
---|
71 | data.mask = ~data.mask |
---|
72 | return datasets[index] if index != 'all' else datasets |
---|
73 | |
---|
74 | |
---|
75 | def set_beam_stop(data, radius, outer=None): |
---|
76 | # type: (Data, float, Optional[float]) -> None |
---|
77 | """ |
---|
78 | Add a beam stop of the given *radius*. If *outer*, make an annulus. |
---|
79 | """ |
---|
80 | from sas.sascalc.dataloader.manipulations import Ringcut |
---|
81 | if hasattr(data, 'qx_data'): |
---|
82 | data.mask = Ringcut(0, radius)(data) |
---|
83 | if outer is not None: |
---|
84 | data.mask += Ringcut(outer, np.inf)(data) |
---|
85 | else: |
---|
86 | data.mask = (data.x < radius) |
---|
87 | if outer is not None: |
---|
88 | data.mask |= (data.x >= outer) |
---|
89 | |
---|
90 | |
---|
91 | def set_half(data, half): |
---|
92 | # type: (Data, str) -> None |
---|
93 | """ |
---|
94 | Select half of the data, either "right" or "left". |
---|
95 | """ |
---|
96 | from sas.sascalc.dataloader.manipulations import Boxcut |
---|
97 | if half == 'right': |
---|
98 | data.mask += \ |
---|
99 | Boxcut(x_min=-np.inf, x_max=0.0, y_min=-np.inf, y_max=np.inf)(data) |
---|
100 | if half == 'left': |
---|
101 | data.mask += \ |
---|
102 | Boxcut(x_min=0.0, x_max=np.inf, y_min=-np.inf, y_max=np.inf)(data) |
---|
103 | |
---|
104 | |
---|
105 | def set_top(data, cutoff): |
---|
106 | # type: (Data, float) -> None |
---|
107 | """ |
---|
108 | Chop the top off the data, above *cutoff*. |
---|
109 | """ |
---|
110 | from sas.sascalc.dataloader.manipulations import Boxcut |
---|
111 | data.mask += \ |
---|
112 | Boxcut(x_min=-np.inf, x_max=np.inf, y_min=-np.inf, y_max=cutoff)(data) |
---|
113 | |
---|
114 | |
---|
115 | class Data1D(object): |
---|
116 | """ |
---|
117 | 1D data object. |
---|
118 | |
---|
119 | Note that this definition matches the attributes from sasview, with |
---|
120 | some generic 1D data vectors and some SAS specific definitions. Some |
---|
121 | refactoring to allow consistent naming conventions between 1D, 2D and |
---|
122 | SESANS data would be helpful. |
---|
123 | |
---|
124 | **Attributes** |
---|
125 | |
---|
126 | *x*, *dx*: $q$ vector and gaussian resolution |
---|
127 | |
---|
128 | *y*, *dy*: $I(q)$ vector and measurement uncertainty |
---|
129 | |
---|
130 | *mask*: values to include in plotting/analysis |
---|
131 | |
---|
132 | *dxl*: slit widths for slit smeared data, with *dx* ignored |
---|
133 | |
---|
134 | *qmin*, *qmax*: range of $q$ values in *x* |
---|
135 | |
---|
136 | *filename*: label for the data line |
---|
137 | |
---|
138 | *_xaxis*, *_xunit*: label and units for the *x* axis |
---|
139 | |
---|
140 | *_yaxis*, *_yunit*: label and units for the *y* axis |
---|
141 | """ |
---|
142 | def __init__(self, |
---|
143 | x=None, # type: Optional[np.ndarray] |
---|
144 | y=None, # type: Optional[np.ndarray] |
---|
145 | dx=None, # type: Optional[np.ndarray] |
---|
146 | dy=None # type: Optional[np.ndarray] |
---|
147 | ): |
---|
148 | # type: (...) -> None |
---|
149 | self.x, self.y, self.dx, self.dy = x, y, dx, dy |
---|
150 | self.dxl = None |
---|
151 | self.filename = None |
---|
152 | self.qmin = x.min() if x is not None else np.NaN |
---|
153 | self.qmax = x.max() if x is not None else np.NaN |
---|
154 | # TODO: why is 1D mask False and 2D mask True? |
---|
155 | self.mask = (np.isnan(y) if y is not None |
---|
156 | else np.zeros_like(x, 'b') if x is not None |
---|
157 | else None) |
---|
158 | self._xaxis, self._xunit = "x", "" |
---|
159 | self._yaxis, self._yunit = "y", "" |
---|
160 | |
---|
161 | def xaxis(self, label, unit): |
---|
162 | # type: (str, str) -> None |
---|
163 | """ |
---|
164 | set the x axis label and unit |
---|
165 | """ |
---|
166 | self._xaxis = label |
---|
167 | self._xunit = unit |
---|
168 | |
---|
169 | def yaxis(self, label, unit): |
---|
170 | # type: (str, str) -> None |
---|
171 | """ |
---|
172 | set the y axis label and unit |
---|
173 | """ |
---|
174 | self._yaxis = label |
---|
175 | self._yunit = unit |
---|
176 | |
---|
177 | class SesansData(Data1D): |
---|
178 | """ |
---|
179 | SESANS data object. |
---|
180 | |
---|
181 | This is just :class:`Data1D` with a wavelength parameter. |
---|
182 | |
---|
183 | *x* is spin echo length and *y* is polarization (P/P0). |
---|
184 | """ |
---|
185 | isSesans = True |
---|
186 | def __init__(self, **kw): |
---|
187 | Data1D.__init__(self, **kw) |
---|
188 | self.lam = None # type: Optional[np.ndarray] |
---|
189 | |
---|
190 | class Data2D(object): |
---|
191 | """ |
---|
192 | 2D data object. |
---|
193 | |
---|
194 | Note that this definition matches the attributes from sasview. Some |
---|
195 | refactoring to allow consistent naming conventions between 1D, 2D and |
---|
196 | SESANS data would be helpful. |
---|
197 | |
---|
198 | **Attributes** |
---|
199 | |
---|
200 | *qx_data*, *dqx_data*: $q_x$ matrix and gaussian resolution |
---|
201 | |
---|
202 | *qy_data*, *dqy_data*: $q_y$ matrix and gaussian resolution |
---|
203 | |
---|
204 | *data*, *err_data*: $I(q)$ matrix and measurement uncertainty |
---|
205 | |
---|
206 | *mask*: values to exclude from plotting/analysis |
---|
207 | |
---|
208 | *qmin*, *qmax*: range of $q$ values in *x* |
---|
209 | |
---|
210 | *filename*: label for the data line |
---|
211 | |
---|
212 | *_xaxis*, *_xunit*: label and units for the *x* axis |
---|
213 | |
---|
214 | *_yaxis*, *_yunit*: label and units for the *y* axis |
---|
215 | |
---|
216 | *_zaxis*, *_zunit*: label and units for the *y* axis |
---|
217 | |
---|
218 | *Q_unit*, *I_unit*: units for Q and intensity |
---|
219 | |
---|
220 | *x_bins*, *y_bins*: grid steps in *x* and *y* directions |
---|
221 | """ |
---|
222 | def __init__(self, |
---|
223 | x=None, # type: Optional[np.ndarray] |
---|
224 | y=None, # type: Optional[np.ndarray] |
---|
225 | z=None, # type: Optional[np.ndarray] |
---|
226 | dx=None, # type: Optional[np.ndarray] |
---|
227 | dy=None, # type: Optional[np.ndarray] |
---|
228 | dz=None # type: Optional[np.ndarray] |
---|
229 | ): |
---|
230 | # type: (...) -> None |
---|
231 | self.qx_data, self.dqx_data = x, dx |
---|
232 | self.qy_data, self.dqy_data = y, dy |
---|
233 | self.data, self.err_data = z, dz |
---|
234 | self.mask = (np.isnan(z) if z is not None |
---|
235 | else np.zeros_like(x, dtype='bool') if x is not None |
---|
236 | else None) |
---|
237 | self.q_data = np.sqrt(x**2 + y**2) |
---|
238 | self.qmin = 1e-16 |
---|
239 | self.qmax = np.inf |
---|
240 | self.detector = [] |
---|
241 | self.source = Source() |
---|
242 | self.Q_unit = "1/A" |
---|
243 | self.I_unit = "1/cm" |
---|
244 | self.xaxis("Q_x", "1/A") |
---|
245 | self.yaxis("Q_y", "1/A") |
---|
246 | self.zaxis("Intensity", "1/cm") |
---|
247 | self._xaxis, self._xunit = "x", "" |
---|
248 | self._yaxis, self._yunit = "y", "" |
---|
249 | self._zaxis, self._zunit = "z", "" |
---|
250 | self.x_bins, self.y_bins = None, None |
---|
251 | self.filename = None |
---|
252 | |
---|
253 | def xaxis(self, label, unit): |
---|
254 | # type: (str, str) -> None |
---|
255 | """ |
---|
256 | set the x axis label and unit |
---|
257 | """ |
---|
258 | self._xaxis = label |
---|
259 | self._xunit = unit |
---|
260 | |
---|
261 | def yaxis(self, label, unit): |
---|
262 | # type: (str, str) -> None |
---|
263 | """ |
---|
264 | set the y axis label and unit |
---|
265 | """ |
---|
266 | self._yaxis = label |
---|
267 | self._yunit = unit |
---|
268 | |
---|
269 | def zaxis(self, label, unit): |
---|
270 | # type: (str, str) -> None |
---|
271 | """ |
---|
272 | set the y axis label and unit |
---|
273 | """ |
---|
274 | self._zaxis = label |
---|
275 | self._zunit = unit |
---|
276 | |
---|
277 | |
---|
278 | class Vector(object): |
---|
279 | """ |
---|
280 | 3-space vector of *x*, *y*, *z* |
---|
281 | """ |
---|
282 | def __init__(self, x=None, y=None, z=None): |
---|
283 | # type: (float, float, Optional[float]) -> None |
---|
284 | self.x, self.y, self.z = x, y, z |
---|
285 | |
---|
286 | class Detector(object): |
---|
287 | """ |
---|
288 | Detector attributes. |
---|
289 | """ |
---|
290 | def __init__(self, pixel_size=(None, None), distance=None): |
---|
291 | # type: (Tuple[float, float], float) -> None |
---|
292 | self.pixel_size = Vector(*pixel_size) |
---|
293 | self.distance = distance |
---|
294 | |
---|
295 | class Source(object): |
---|
296 | """ |
---|
297 | Beam attributes. |
---|
298 | """ |
---|
299 | def __init__(self): |
---|
300 | # type: () -> None |
---|
301 | self.wavelength = np.NaN |
---|
302 | self.wavelength_unit = "A" |
---|
303 | |
---|
304 | class Sample(object): |
---|
305 | """ |
---|
306 | Sample attributes. |
---|
307 | """ |
---|
308 | def __init__(self): |
---|
309 | # type: () -> None |
---|
310 | pass |
---|
311 | |
---|
312 | def empty_data1D(q, resolution=0.0, L=0., dL=0.): |
---|
313 | # type: (np.ndarray, float) -> Data1D |
---|
314 | r""" |
---|
315 | Create empty 1D data using the given *q* as the x value. |
---|
316 | |
---|
317 | rms *resolution* $\Delta q/q$ defaults to 0%. If wavelength *L* and rms |
---|
318 | wavelength divergence *dL* are defined, then *resolution* defines |
---|
319 | rms $\Delta \theta/\theta$ for the lowest *q*, with $\theta$ derived from |
---|
320 | $q = 4\pi/\lambda \sin(\theta)$. |
---|
321 | """ |
---|
322 | |
---|
323 | #Iq = 100 * np.ones_like(q) |
---|
324 | #dIq = np.sqrt(Iq) |
---|
325 | Iq, dIq = None, None |
---|
326 | q = np.asarray(q) |
---|
327 | if L != 0 and resolution != 0: |
---|
328 | theta = np.arcsin(q*L/(4*pi)) |
---|
329 | dtheta = theta[0]*resolution |
---|
330 | ## Solving Gaussian error propagation from |
---|
331 | ## Dq^2 = (dq/dL)^2 DL^2 + (dq/dtheta)^2 Dtheta^2 |
---|
332 | ## gives |
---|
333 | ## (Dq/q)^2 = (DL/L)**2 + (Dtheta/tan(theta))**2 |
---|
334 | ## Take the square root and multiply by q, giving |
---|
335 | ## Dq = (4*pi/L) * sqrt((sin(theta)*dL/L)**2 + (cos(theta)*dtheta)**2) |
---|
336 | dq = (4*pi/L) * sqrt((sin(theta)*dL/L)**2 + (cos(theta)*dtheta)**2) |
---|
337 | else: |
---|
338 | dq = resolution * q |
---|
339 | |
---|
340 | data = Data1D(q, Iq, dx=dq, dy=dIq) |
---|
341 | data.filename = "fake data" |
---|
342 | return data |
---|
343 | |
---|
344 | |
---|
345 | def empty_data2D(qx, qy=None, resolution=0.0): |
---|
346 | # type: (np.ndarray, Optional[np.ndarray], float) -> Data2D |
---|
347 | """ |
---|
348 | Create empty 2D data using the given mesh. |
---|
349 | |
---|
350 | If *qy* is missing, create a square mesh with *qy=qx*. |
---|
351 | |
---|
352 | *resolution* dq/q defaults to 5%. |
---|
353 | """ |
---|
354 | if qy is None: |
---|
355 | qy = qx |
---|
356 | qx, qy = np.asarray(qx), np.asarray(qy) |
---|
357 | # 5% dQ/Q resolution |
---|
358 | Qx, Qy = np.meshgrid(qx, qy) |
---|
359 | Qx, Qy = Qx.flatten(), Qy.flatten() |
---|
360 | Iq = 100 * np.ones_like(Qx) # type: np.ndarray |
---|
361 | dIq = np.sqrt(Iq) |
---|
362 | if resolution != 0: |
---|
363 | # https://www.ncnr.nist.gov/staff/hammouda/distance_learning/chapter_15.pdf |
---|
364 | # Should have an additional constant which depends on distances and |
---|
365 | # radii of the aperture, pixel dimensions and wavelength spread |
---|
366 | # Instead, assume radial dQ/Q is constant, and perpendicular matches |
---|
367 | # radial (which instead it should be inverse). |
---|
368 | Q = np.sqrt(Qx**2 + Qy**2) |
---|
369 | dqx = resolution * Q |
---|
370 | dqy = resolution * Q |
---|
371 | else: |
---|
372 | dqx = dqy = None |
---|
373 | |
---|
374 | data = Data2D(x=Qx, y=Qy, z=Iq, dx=dqx, dy=dqy, dz=dIq) |
---|
375 | data.x_bins = qx |
---|
376 | data.y_bins = qy |
---|
377 | data.filename = "fake data" |
---|
378 | |
---|
379 | # pixel_size in mm, distance in m |
---|
380 | detector = Detector(pixel_size=(5, 5), distance=4) |
---|
381 | data.detector.append(detector) |
---|
382 | data.source.wavelength = 5 # angstroms |
---|
383 | data.source.wavelength_unit = "A" |
---|
384 | return data |
---|
385 | |
---|
386 | |
---|
387 | def plot_data(data, view='log', limits=None): |
---|
388 | # type: (Data, str, Optional[Tuple[float, float]]) -> None |
---|
389 | """ |
---|
390 | Plot data loaded by the sasview loader. |
---|
391 | |
---|
392 | *data* is a sasview data object, either 1D, 2D or SESANS. |
---|
393 | |
---|
394 | *view* is log or linear. |
---|
395 | |
---|
396 | *limits* sets the intensity limits on the plot; if None then the limits |
---|
397 | are inferred from the data. |
---|
398 | """ |
---|
399 | # Note: kind of weird using the plot result functions to plot just the |
---|
400 | # data, but they already handle the masking and graph markup already, so |
---|
401 | # do not repeat. |
---|
402 | if hasattr(data, 'isSesans') and data.isSesans: |
---|
403 | _plot_result_sesans(data, None, None, use_data=True, limits=limits) |
---|
404 | elif hasattr(data, 'qx_data') and not getattr(data, 'radial', False): |
---|
405 | _plot_result2D(data, None, None, view, use_data=True, limits=limits) |
---|
406 | else: |
---|
407 | _plot_result1D(data, None, None, view, use_data=True, limits=limits) |
---|
408 | |
---|
409 | |
---|
410 | def plot_theory(data, # type: Data |
---|
411 | theory, # type: Optional[np.ndarray] |
---|
412 | resid=None, # type: Optional[np.ndarray] |
---|
413 | view='log', # type: str |
---|
414 | use_data=True, # type: bool |
---|
415 | limits=None, # type: Optional[np.ndarray] |
---|
416 | Iq_calc=None # type: Optional[np.ndarray] |
---|
417 | ): |
---|
418 | # type: (...) -> None |
---|
419 | """ |
---|
420 | Plot theory calculation. |
---|
421 | |
---|
422 | *data* is needed to define the graph properties such as labels and |
---|
423 | units, and to define the data mask. |
---|
424 | |
---|
425 | *theory* is a matrix of the same shape as the data. |
---|
426 | |
---|
427 | *view* is log or linear |
---|
428 | |
---|
429 | *use_data* is True if the data should be plotted as well as the theory. |
---|
430 | |
---|
431 | *limits* sets the intensity limits on the plot; if None then the limits |
---|
432 | are inferred from the data. |
---|
433 | |
---|
434 | *Iq_calc* is the raw theory values without resolution smearing |
---|
435 | """ |
---|
436 | if hasattr(data, 'isSesans') and data.isSesans: |
---|
437 | _plot_result_sesans(data, theory, resid, use_data=True, limits=limits) |
---|
438 | elif hasattr(data, 'qx_data') and not getattr(data, 'radial', False): |
---|
439 | _plot_result2D(data, theory, resid, view, use_data, limits=limits) |
---|
440 | else: |
---|
441 | _plot_result1D(data, theory, resid, view, use_data, |
---|
442 | limits=limits, Iq_calc=Iq_calc) |
---|
443 | |
---|
444 | |
---|
445 | def protect(func): |
---|
446 | # type: (Callable) -> Callable |
---|
447 | """ |
---|
448 | Decorator to wrap calls in an exception trapper which prints the |
---|
449 | exception and continues. Keyboard interrupts are ignored. |
---|
450 | """ |
---|
451 | def wrapper(*args, **kw): |
---|
452 | """ |
---|
453 | Trap and print errors from function. |
---|
454 | """ |
---|
455 | try: |
---|
456 | return func(*args, **kw) |
---|
457 | except Exception: |
---|
458 | traceback.print_exc() |
---|
459 | |
---|
460 | return wrapper |
---|
461 | |
---|
462 | |
---|
463 | @protect |
---|
464 | def _plot_result1D(data, # type: Data1D |
---|
465 | theory, # type: Optional[np.ndarray] |
---|
466 | resid, # type: Optional[np.ndarray] |
---|
467 | view, # type: str |
---|
468 | use_data, # type: bool |
---|
469 | limits=None, # type: Optional[Tuple[float, float]] |
---|
470 | Iq_calc=None # type: Optional[np.ndarray] |
---|
471 | ): |
---|
472 | # type: (...) -> None |
---|
473 | """ |
---|
474 | Plot the data and residuals for 1D data. |
---|
475 | """ |
---|
476 | import matplotlib.pyplot as plt # type: ignore |
---|
477 | from numpy.ma import masked_array, masked # type: ignore |
---|
478 | |
---|
479 | if getattr(data, 'radial', False): |
---|
480 | data.x = data.q_data |
---|
481 | data.y = data.data |
---|
482 | |
---|
483 | use_data = use_data and data.y is not None |
---|
484 | use_theory = theory is not None |
---|
485 | use_resid = resid is not None |
---|
486 | use_calc = use_theory and Iq_calc is not None |
---|
487 | num_plots = (use_data or use_theory) + use_calc + use_resid |
---|
488 | non_positive_x = (data.x <= 0.0).any() |
---|
489 | |
---|
490 | scale = data.x**4 if view == 'q4' else 1.0 |
---|
491 | xscale = yscale = 'linear' if view == 'linear' else 'log' |
---|
492 | |
---|
493 | if use_data or use_theory: |
---|
494 | if num_plots > 1: |
---|
495 | plt.subplot(1, num_plots, 1) |
---|
496 | |
---|
497 | #print(vmin, vmax) |
---|
498 | all_positive = True |
---|
499 | some_present = False |
---|
500 | if use_data: |
---|
501 | mdata = masked_array(data.y, data.mask.copy()) |
---|
502 | mdata[~np.isfinite(mdata)] = masked |
---|
503 | if view is 'log': |
---|
504 | mdata[mdata <= 0] = masked |
---|
505 | plt.errorbar(data.x, scale*mdata, yerr=data.dy, fmt='.') |
---|
506 | all_positive = all_positive and (mdata > 0).all() |
---|
507 | some_present = some_present or (mdata.count() > 0) |
---|
508 | |
---|
509 | |
---|
510 | if use_theory: |
---|
511 | # Note: masks merge, so any masked theory points will stay masked, |
---|
512 | # and the data mask will be added to it. |
---|
513 | #mtheory = masked_array(theory, data.mask.copy()) |
---|
514 | theory_x = data.x[data.mask == 0] |
---|
515 | mtheory = masked_array(theory) |
---|
516 | mtheory[~np.isfinite(mtheory)] = masked |
---|
517 | if view is 'log': |
---|
518 | mtheory[mtheory <= 0] = masked |
---|
519 | plt.plot(theory_x, scale*mtheory, '-') |
---|
520 | all_positive = all_positive and (mtheory > 0).all() |
---|
521 | some_present = some_present or (mtheory.count() > 0) |
---|
522 | |
---|
523 | if limits is not None: |
---|
524 | plt.ylim(*limits) |
---|
525 | |
---|
526 | |
---|
527 | xscale = ('linear' if not some_present or non_positive_x |
---|
528 | else view if view is not None |
---|
529 | else 'log') |
---|
530 | yscale = ('linear' |
---|
531 | if view == 'q4' or not some_present or not all_positive |
---|
532 | else view if view is not None |
---|
533 | else 'log') |
---|
534 | plt.xscale(xscale) |
---|
535 | plt.xlabel("$q$/A$^{-1}$") |
---|
536 | plt.yscale(yscale) |
---|
537 | plt.ylabel('$I(q)$') |
---|
538 | title = ("data and model" if use_theory and use_data |
---|
539 | else "data" if use_data |
---|
540 | else "model") |
---|
541 | plt.title(title) |
---|
542 | |
---|
543 | if use_calc: |
---|
544 | # Only have use_calc if have use_theory |
---|
545 | plt.subplot(1, num_plots, 2) |
---|
546 | qx, qy, Iqxy = Iq_calc |
---|
547 | plt.pcolormesh(qx, qy[qy > 0], np.log10(Iqxy[qy > 0, :])) |
---|
548 | plt.xlabel("$q_x$/A$^{-1}$") |
---|
549 | plt.xlabel("$q_y$/A$^{-1}$") |
---|
550 | plt.xscale('log') |
---|
551 | plt.yscale('log') |
---|
552 | #plt.axis('equal') |
---|
553 | |
---|
554 | if use_resid: |
---|
555 | theory_x = data.x[data.mask == 0] |
---|
556 | mresid = masked_array(resid) |
---|
557 | mresid[~np.isfinite(mresid)] = masked |
---|
558 | some_present = (mresid.count() > 0) |
---|
559 | |
---|
560 | if num_plots > 1: |
---|
561 | plt.subplot(1, num_plots, use_calc + 2) |
---|
562 | plt.plot(theory_x, mresid, '.') |
---|
563 | plt.xlabel("$q$/A$^{-1}$") |
---|
564 | plt.ylabel('residuals') |
---|
565 | plt.title('(model - Iq)/dIq') |
---|
566 | plt.xscale(xscale) |
---|
567 | plt.yscale('linear') |
---|
568 | |
---|
569 | |
---|
570 | @protect |
---|
571 | def _plot_result_sesans(data, # type: SesansData |
---|
572 | theory, # type: Optional[np.ndarray] |
---|
573 | resid, # type: Optional[np.ndarray] |
---|
574 | use_data, # type: bool |
---|
575 | limits=None # type: Optional[Tuple[float, float]] |
---|
576 | ): |
---|
577 | # type: (...) -> None |
---|
578 | """ |
---|
579 | Plot SESANS results. |
---|
580 | """ |
---|
581 | import matplotlib.pyplot as plt # type: ignore |
---|
582 | use_data = use_data and data.y is not None |
---|
583 | use_theory = theory is not None |
---|
584 | use_resid = resid is not None |
---|
585 | num_plots = (use_data or use_theory) + use_resid |
---|
586 | |
---|
587 | if use_data or use_theory: |
---|
588 | is_tof = data.lam is not None and (data.lam != data.lam[0]).any() |
---|
589 | if num_plots > 1: |
---|
590 | plt.subplot(1, num_plots, 1) |
---|
591 | if use_data: |
---|
592 | if is_tof: |
---|
593 | plt.errorbar(data.x, np.log(data.y)/(data.lam*data.lam), |
---|
594 | yerr=data.dy/data.y/(data.lam*data.lam)) |
---|
595 | else: |
---|
596 | plt.errorbar(data.x, data.y, yerr=data.dy) |
---|
597 | if theory is not None: |
---|
598 | if is_tof: |
---|
599 | plt.plot(data.x, np.log(theory)/(data.lam*data.lam), '-') |
---|
600 | else: |
---|
601 | plt.plot(data.x, theory, '-') |
---|
602 | if limits is not None: |
---|
603 | plt.ylim(*limits) |
---|
604 | |
---|
605 | plt.xlabel('spin echo length ({})'.format(data._xunit)) |
---|
606 | if is_tof: |
---|
607 | plt.ylabel(r'(Log (P/P$_0$))/$\lambda^2$') |
---|
608 | else: |
---|
609 | plt.ylabel('polarization (P/P0)') |
---|
610 | |
---|
611 | |
---|
612 | if resid is not None: |
---|
613 | if num_plots > 1: |
---|
614 | plt.subplot(1, num_plots, (use_data or use_theory) + 1) |
---|
615 | plt.plot(data.x, resid, 'x') |
---|
616 | plt.xlabel('spin echo length ({})'.format(data._xunit)) |
---|
617 | plt.ylabel('residuals (P/P0)') |
---|
618 | |
---|
619 | |
---|
620 | @protect |
---|
621 | def _plot_result2D(data, # type: Data2D |
---|
622 | theory, # type: Optional[np.ndarray] |
---|
623 | resid, # type: Optional[np.ndarray] |
---|
624 | view, # type: str |
---|
625 | use_data, # type: bool |
---|
626 | limits=None # type: Optional[Tuple[float, float]] |
---|
627 | ): |
---|
628 | # type: (...) -> None |
---|
629 | """ |
---|
630 | Plot the data and residuals for 2D data. |
---|
631 | """ |
---|
632 | import matplotlib.pyplot as plt # type: ignore |
---|
633 | use_data = use_data and data.data is not None |
---|
634 | use_theory = theory is not None |
---|
635 | use_resid = resid is not None |
---|
636 | num_plots = use_data + use_theory + use_resid |
---|
637 | |
---|
638 | # Put theory and data on a common colormap scale |
---|
639 | vmin, vmax = np.inf, -np.inf |
---|
640 | target = None # type: Optional[np.ndarray] |
---|
641 | if use_data: |
---|
642 | target = data.data[~data.mask] |
---|
643 | datamin = target[target > 0].min() if view == 'log' else target.min() |
---|
644 | datamax = target.max() |
---|
645 | vmin = min(vmin, datamin) |
---|
646 | vmax = max(vmax, datamax) |
---|
647 | if use_theory: |
---|
648 | theorymin = theory[theory > 0].min() if view == 'log' else theory.min() |
---|
649 | theorymax = theory.max() |
---|
650 | vmin = min(vmin, theorymin) |
---|
651 | vmax = max(vmax, theorymax) |
---|
652 | |
---|
653 | # Override data limits from the caller |
---|
654 | if limits is not None: |
---|
655 | vmin, vmax = limits |
---|
656 | |
---|
657 | # Plot data |
---|
658 | if use_data: |
---|
659 | if num_plots > 1: |
---|
660 | plt.subplot(1, num_plots, 1) |
---|
661 | _plot_2d_signal(data, target, view=view, vmin=vmin, vmax=vmax) |
---|
662 | plt.title('data') |
---|
663 | h = plt.colorbar() |
---|
664 | h.set_label('$I(q)$') |
---|
665 | |
---|
666 | # plot theory |
---|
667 | if use_theory: |
---|
668 | if num_plots > 1: |
---|
669 | plt.subplot(1, num_plots, use_data+1) |
---|
670 | _plot_2d_signal(data, theory, view=view, vmin=vmin, vmax=vmax) |
---|
671 | plt.title('theory') |
---|
672 | h = plt.colorbar() |
---|
673 | h.set_label(r'$\log_{10}I(q)$' if view == 'log' |
---|
674 | else r'$q^4 I(q)$' if view == 'q4' |
---|
675 | else '$I(q)$') |
---|
676 | |
---|
677 | # plot resid |
---|
678 | if use_resid: |
---|
679 | if num_plots > 1: |
---|
680 | plt.subplot(1, num_plots, use_data+use_theory+1) |
---|
681 | _plot_2d_signal(data, resid, view='linear') |
---|
682 | plt.title('residuals') |
---|
683 | h = plt.colorbar() |
---|
684 | h.set_label(r'$\Delta I(q)$') |
---|
685 | |
---|
686 | |
---|
687 | @protect |
---|
688 | def _plot_2d_signal(data, # type: Data2D |
---|
689 | signal, # type: np.ndarray |
---|
690 | vmin=None, # type: Optional[float] |
---|
691 | vmax=None, # type: Optional[float] |
---|
692 | view='log' # type: str |
---|
693 | ): |
---|
694 | # type: (...) -> Tuple[float, float] |
---|
695 | """ |
---|
696 | Plot the target value for the data. This could be the data itself, |
---|
697 | the theory calculation, or the residuals. |
---|
698 | |
---|
699 | *scale* can be 'log' for log scale data, or 'linear'. |
---|
700 | """ |
---|
701 | import matplotlib.pyplot as plt # type: ignore |
---|
702 | from numpy.ma import masked_array # type: ignore |
---|
703 | |
---|
704 | image = np.zeros_like(data.qx_data) |
---|
705 | image[~data.mask] = signal |
---|
706 | valid = np.isfinite(image) |
---|
707 | if view == 'log': |
---|
708 | valid[valid] = (image[valid] > 0) |
---|
709 | if vmin is None: |
---|
710 | vmin = image[valid & ~data.mask].min() |
---|
711 | if vmax is None: |
---|
712 | vmax = image[valid & ~data.mask].max() |
---|
713 | image[valid] = np.log10(image[valid]) |
---|
714 | elif view == 'q4': |
---|
715 | image[valid] *= (data.qx_data[valid]**2+data.qy_data[valid]**2)**2 |
---|
716 | if vmin is None: |
---|
717 | vmin = image[valid & ~data.mask].min() |
---|
718 | if vmax is None: |
---|
719 | vmax = image[valid & ~data.mask].max() |
---|
720 | else: |
---|
721 | if vmin is None: |
---|
722 | vmin = image[valid & ~data.mask].min() |
---|
723 | if vmax is None: |
---|
724 | vmax = image[valid & ~data.mask].max() |
---|
725 | |
---|
726 | image[~valid | data.mask] = 0 |
---|
727 | #plottable = Iq |
---|
728 | plottable = masked_array(image, ~valid | data.mask) |
---|
729 | # Divide range by 10 to convert from angstroms to nanometers |
---|
730 | xmin, xmax = min(data.qx_data), max(data.qx_data) |
---|
731 | ymin, ymax = min(data.qy_data), max(data.qy_data) |
---|
732 | if view == 'log': |
---|
733 | vmin_scaled, vmax_scaled = np.log10(vmin), np.log10(vmax) |
---|
734 | else: |
---|
735 | vmin_scaled, vmax_scaled = vmin, vmax |
---|
736 | #nx, ny = len(data.x_bins), len(data.y_bins) |
---|
737 | x_bins, y_bins, image = _build_matrix(data, plottable) |
---|
738 | plt.imshow(image, |
---|
739 | interpolation='nearest', aspect=1, origin='lower', |
---|
740 | extent=[xmin, xmax, ymin, ymax], |
---|
741 | vmin=vmin_scaled, vmax=vmax_scaled) |
---|
742 | plt.xlabel("$q_x$/A$^{-1}$") |
---|
743 | plt.ylabel("$q_y$/A$^{-1}$") |
---|
744 | return vmin, vmax |
---|
745 | |
---|
746 | |
---|
747 | # === The following is modified from sas.sasgui.plottools.PlotPanel |
---|
748 | def _build_matrix(self, plottable): |
---|
749 | """ |
---|
750 | Build a matrix for 2d plot from a vector |
---|
751 | Returns a matrix (image) with ~ square binning |
---|
752 | Requirement: need 1d array formats of |
---|
753 | self.data, self.qx_data, and self.qy_data |
---|
754 | where each one corresponds to z, x, or y axis values |
---|
755 | |
---|
756 | """ |
---|
757 | # No qx or qy given in a vector format |
---|
758 | if self.qx_data is None or self.qy_data is None \ |
---|
759 | or self.qx_data.ndim != 1 or self.qy_data.ndim != 1: |
---|
760 | return self.x_bins, self.y_bins, plottable |
---|
761 | |
---|
762 | # maximum # of loops to fillup_pixels |
---|
763 | # otherwise, loop could never stop depending on data |
---|
764 | max_loop = 1 |
---|
765 | # get the x and y_bin arrays. |
---|
766 | x_bins, y_bins = _get_bins(self) |
---|
767 | # set zero to None |
---|
768 | |
---|
769 | #Note: Can not use scipy.interpolate.Rbf: |
---|
770 | # 'cause too many data points (>10000)<=JHC. |
---|
771 | # 1d array to use for weighting the data point averaging |
---|
772 | #when they fall into a same bin. |
---|
773 | weights_data = np.ones([self.data.size]) |
---|
774 | # get histogram of ones w/len(data); this will provide |
---|
775 | #the weights of data on each bins |
---|
776 | weights, xedges, yedges = np.histogram2d(x=self.qy_data, |
---|
777 | y=self.qx_data, |
---|
778 | bins=[y_bins, x_bins], |
---|
779 | weights=weights_data) |
---|
780 | # get histogram of data, all points into a bin in a way of summing |
---|
781 | image, xedges, yedges = np.histogram2d(x=self.qy_data, |
---|
782 | y=self.qx_data, |
---|
783 | bins=[y_bins, x_bins], |
---|
784 | weights=plottable) |
---|
785 | # Now, normalize the image by weights only for weights>1: |
---|
786 | # If weight == 1, there is only one data point in the bin so |
---|
787 | # that no normalization is required. |
---|
788 | image[weights > 1] = image[weights > 1] / weights[weights > 1] |
---|
789 | # Set image bins w/o a data point (weight==0) as None (was set to zero |
---|
790 | # by histogram2d.) |
---|
791 | image[weights == 0] = None |
---|
792 | |
---|
793 | # Fill empty bins with 8 nearest neighbors only when at least |
---|
794 | #one None point exists |
---|
795 | loop = 0 |
---|
796 | |
---|
797 | # do while loop until all vacant bins are filled up up |
---|
798 | #to loop = max_loop |
---|
799 | while (weights == 0).any(): |
---|
800 | if loop >= max_loop: # this protects never-ending loop |
---|
801 | break |
---|
802 | image = _fillup_pixels(image=image, weights=weights) |
---|
803 | loop += 1 |
---|
804 | |
---|
805 | return x_bins, y_bins, image |
---|
806 | |
---|
807 | def _get_bins(self): |
---|
808 | """ |
---|
809 | get bins |
---|
810 | set x_bins and y_bins into self, 1d arrays of the index with |
---|
811 | ~ square binning |
---|
812 | Requirement: need 1d array formats of |
---|
813 | self.qx_data, and self.qy_data |
---|
814 | where each one corresponds to x, or y axis values |
---|
815 | """ |
---|
816 | # find max and min values of qx and qy |
---|
817 | xmax = self.qx_data.max() |
---|
818 | xmin = self.qx_data.min() |
---|
819 | ymax = self.qy_data.max() |
---|
820 | ymin = self.qy_data.min() |
---|
821 | |
---|
822 | # calculate the range of qx and qy: this way, it is a little |
---|
823 | # more independent |
---|
824 | x_size = xmax - xmin |
---|
825 | y_size = ymax - ymin |
---|
826 | |
---|
827 | # estimate the # of pixels on each axes |
---|
828 | npix_y = int(np.floor(np.sqrt(len(self.qy_data)))) |
---|
829 | npix_x = int(np.floor(len(self.qy_data) / npix_y)) |
---|
830 | |
---|
831 | # bin size: x- & y-directions |
---|
832 | xstep = x_size / (npix_x - 1) |
---|
833 | ystep = y_size / (npix_y - 1) |
---|
834 | |
---|
835 | # max and min taking account of the bin sizes |
---|
836 | xmax = xmax + xstep / 2.0 |
---|
837 | xmin = xmin - xstep / 2.0 |
---|
838 | ymax = ymax + ystep / 2.0 |
---|
839 | ymin = ymin - ystep / 2.0 |
---|
840 | |
---|
841 | # store x and y bin centers in q space |
---|
842 | x_bins = np.linspace(xmin, xmax, npix_x) |
---|
843 | y_bins = np.linspace(ymin, ymax, npix_y) |
---|
844 | |
---|
845 | return x_bins, y_bins |
---|
846 | |
---|
847 | def _fillup_pixels(image=None, weights=None): |
---|
848 | """ |
---|
849 | Fill z values of the empty cells of 2d image matrix |
---|
850 | with the average over up-to next nearest neighbor points |
---|
851 | |
---|
852 | :param image: (2d matrix with some zi = None) |
---|
853 | |
---|
854 | :return: image (2d array ) |
---|
855 | |
---|
856 | :TODO: Find better way to do for-loop below |
---|
857 | |
---|
858 | """ |
---|
859 | # No image matrix given |
---|
860 | if image is None or np.ndim(image) != 2 \ |
---|
861 | or np.isfinite(image).all() \ |
---|
862 | or weights is None: |
---|
863 | return image |
---|
864 | # Get bin size in y and x directions |
---|
865 | len_y = len(image) |
---|
866 | len_x = len(image[1]) |
---|
867 | temp_image = np.zeros([len_y, len_x]) |
---|
868 | weit = np.zeros([len_y, len_x]) |
---|
869 | # do for-loop for all pixels |
---|
870 | for n_y in range(len(image)): |
---|
871 | for n_x in range(len(image[1])): |
---|
872 | # find only null pixels |
---|
873 | if weights[n_y][n_x] > 0 or np.isfinite(image[n_y][n_x]): |
---|
874 | continue |
---|
875 | else: |
---|
876 | # find 4 nearest neighbors |
---|
877 | # check where or not it is at the corner |
---|
878 | if n_y != 0 and np.isfinite(image[n_y - 1][n_x]): |
---|
879 | temp_image[n_y][n_x] += image[n_y - 1][n_x] |
---|
880 | weit[n_y][n_x] += 1 |
---|
881 | if n_x != 0 and np.isfinite(image[n_y][n_x - 1]): |
---|
882 | temp_image[n_y][n_x] += image[n_y][n_x - 1] |
---|
883 | weit[n_y][n_x] += 1 |
---|
884 | if n_y != len_y - 1 and np.isfinite(image[n_y + 1][n_x]): |
---|
885 | temp_image[n_y][n_x] += image[n_y + 1][n_x] |
---|
886 | weit[n_y][n_x] += 1 |
---|
887 | if n_x != len_x - 1 and np.isfinite(image[n_y][n_x + 1]): |
---|
888 | temp_image[n_y][n_x] += image[n_y][n_x + 1] |
---|
889 | weit[n_y][n_x] += 1 |
---|
890 | # go 4 next nearest neighbors when no non-zero |
---|
891 | # neighbor exists |
---|
892 | if n_y != 0 and n_x != 0 and \ |
---|
893 | np.isfinite(image[n_y - 1][n_x - 1]): |
---|
894 | temp_image[n_y][n_x] += image[n_y - 1][n_x - 1] |
---|
895 | weit[n_y][n_x] += 1 |
---|
896 | if n_y != len_y - 1 and n_x != 0 and \ |
---|
897 | np.isfinite(image[n_y + 1][n_x - 1]): |
---|
898 | temp_image[n_y][n_x] += image[n_y + 1][n_x - 1] |
---|
899 | weit[n_y][n_x] += 1 |
---|
900 | if n_y != len_y and n_x != len_x - 1 and \ |
---|
901 | np.isfinite(image[n_y - 1][n_x + 1]): |
---|
902 | temp_image[n_y][n_x] += image[n_y - 1][n_x + 1] |
---|
903 | weit[n_y][n_x] += 1 |
---|
904 | if n_y != len_y - 1 and n_x != len_x - 1 and \ |
---|
905 | np.isfinite(image[n_y + 1][n_x + 1]): |
---|
906 | temp_image[n_y][n_x] += image[n_y + 1][n_x + 1] |
---|
907 | weit[n_y][n_x] += 1 |
---|
908 | |
---|
909 | # get it normalized |
---|
910 | ind = (weit > 0) |
---|
911 | image[ind] = temp_image[ind] / weit[ind] |
---|
912 | |
---|
913 | return image |
---|
914 | |
---|
915 | |
---|
916 | def demo(): |
---|
917 | # type: () -> None |
---|
918 | """ |
---|
919 | Load and plot a SAS dataset. |
---|
920 | """ |
---|
921 | data = load_data('DEC07086.DAT') |
---|
922 | set_beam_stop(data, 0.004) |
---|
923 | plot_data(data) |
---|
924 | import matplotlib.pyplot as plt # type: ignore |
---|
925 | plt.show() |
---|
926 | |
---|
927 | |
---|
928 | if __name__ == "__main__": |
---|
929 | demo() |
---|