##// END OF EJS Templates
schain
RSS
Fix pairsList in SpectraProc
jespinoza -
r1341:34c23d10b2ab
parent child
Show More
Side by Side Unified
Context file:
r1341:34c23d10b2ab Loading diff...:
Show file before | Show file after | Hide comments
@@ -1,874 +1,876
1 1 # Copyright (c) 2012-2020 Jicamarca Radio Observatory
2 2 # All rights reserved.
3 3 #
4 4 # Distributed under the terms of the BSD 3-clause license.
5 5 """Spectra processing Unit and operations
6 6
7 7 Here you will find the processing unit `SpectraProc` and several operations
8 8 to work with Spectra data type
9 9 """
10 10
11 11 import time
12 12 import itertools
13 13
14 14 import numpy
15 15
16 16 from schainpy.model.proc.jroproc_base import ProcessingUnit, MPDecorator, Operation
17 17 from schainpy.model.data.jrodata import Spectra
18 18 from schainpy.model.data.jrodata import hildebrand_sekhon
19 19 from schainpy.utils import log
20 20
21 21
22 22 class SpectraProc(ProcessingUnit):
23 23
24 24 def __init__(self):
25 25
26 26 ProcessingUnit.__init__(self)
27 27
28 28 self.buffer = None
29 29 self.firstdatatime = None
30 30 self.profIndex = 0
31 31 self.dataOut = Spectra()
32 32 self.id_min = None
33 33 self.id_max = None
34 34 self.setupReq = False #Agregar a todas las unidades de proc
35 35
36 36 def __updateSpecFromVoltage(self):
37 37
38 38 self.dataOut.timeZone = self.dataIn.timeZone
39 39 self.dataOut.dstFlag = self.dataIn.dstFlag
40 40 self.dataOut.errorCount = self.dataIn.errorCount
41 41 self.dataOut.useLocalTime = self.dataIn.useLocalTime
42 42 try:
43 43 self.dataOut.processingHeaderObj = self.dataIn.processingHeaderObj.copy()
44 44 except:
45 45 pass
46 46 self.dataOut.radarControllerHeaderObj = self.dataIn.radarControllerHeaderObj.copy()
47
48 47 self.dataOut.systemHeaderObj = self.dataIn.systemHeaderObj.copy()
49 48 self.dataOut.channelList = self.dataIn.channelList
50 49 self.dataOut.heightList = self.dataIn.heightList
51 50 self.dataOut.dtype = numpy.dtype([('real', '<f4'), ('imag', '<f4')])
52 51 self.dataOut.nProfiles = self.dataOut.nFFTPoints
53 52 self.dataOut.flagDiscontinuousBlock = self.dataIn.flagDiscontinuousBlock
54 53 self.dataOut.utctime = self.firstdatatime
55 54 self.dataOut.flagDecodeData = self.dataIn.flagDecodeData
56 55 self.dataOut.flagDeflipData = self.dataIn.flagDeflipData
57 56 self.dataOut.flagShiftFFT = False
58 57 self.dataOut.nCohInt = self.dataIn.nCohInt
59 58 self.dataOut.nIncohInt = 1
60 59 self.dataOut.windowOfFilter = self.dataIn.windowOfFilter
61 60 self.dataOut.frequency = self.dataIn.frequency
62 61 self.dataOut.realtime = self.dataIn.realtime
63 62 self.dataOut.azimuth = self.dataIn.azimuth
64 63 self.dataOut.zenith = self.dataIn.zenith
65 64 self.dataOut.beam.codeList = self.dataIn.beam.codeList
66 65 self.dataOut.beam.azimuthList = self.dataIn.beam.azimuthList
67 66 self.dataOut.beam.zenithList = self.dataIn.beam.zenithList
68 67
69 68 def __getFft(self):
70 69 """
71 70 Convierte valores de Voltaje a Spectra
72 71
73 72 Affected:
74 73 self.dataOut.data_spc
75 74 self.dataOut.data_cspc
76 75 self.dataOut.data_dc
77 76 self.dataOut.heightList
78 77 self.profIndex
79 78 self.buffer
80 79 self.dataOut.flagNoData
81 80 """
82 81 fft_volt = numpy.fft.fft(
83 82 self.buffer, n=self.dataOut.nFFTPoints, axis=1)
84 83 fft_volt = fft_volt.astype(numpy.dtype('complex'))
85 84 dc = fft_volt[:, 0, :]
86 85
87 86 # calculo de self-spectra
88 87 fft_volt = numpy.fft.fftshift(fft_volt, axes=(1,))
89 88 spc = fft_volt * numpy.conjugate(fft_volt)
90 89 spc = spc.real
91 90
92 91 blocksize = 0
93 92 blocksize += dc.size
94 93 blocksize += spc.size
95 94
96 95 cspc = None
97 96 pairIndex = 0
98 97 if self.dataOut.pairsList != None:
99 98 # calculo de cross-spectra
100 99 cspc = numpy.zeros(
101 100 (self.dataOut.nPairs, self.dataOut.nFFTPoints, self.dataOut.nHeights), dtype='complex')
102 101 for pair in self.dataOut.pairsList:
103 102 if pair[0] not in self.dataOut.channelList:
104 103 raise ValueError("Error getting CrossSpectra: pair 0 of %s is not in channelList = %s" % (
105 104 str(pair), str(self.dataOut.channelList)))
106 105 if pair[1] not in self.dataOut.channelList:
107 106 raise ValueError("Error getting CrossSpectra: pair 1 of %s is not in channelList = %s" % (
108 107 str(pair), str(self.dataOut.channelList)))
109 108
110 109 cspc[pairIndex, :, :] = fft_volt[pair[0], :, :] * \
111 110 numpy.conjugate(fft_volt[pair[1], :, :])
112 111 pairIndex += 1
113 112 blocksize += cspc.size
114 113
115 114 self.dataOut.data_spc = spc
116 115 self.dataOut.data_cspc = cspc
117 116 self.dataOut.data_dc = dc
118 117 self.dataOut.blockSize = blocksize
119 118 self.dataOut.flagShiftFFT = False
120 119
121 120 def run(self, nProfiles=None, nFFTPoints=None, pairsList=None, ippFactor=None, shift_fft=False):
122 121
123 122 if self.dataIn.type == "Spectra":
124 123 self.dataOut.copy(self.dataIn)
125 124 if shift_fft:
126 125 #desplaza a la derecha en el eje 2 determinadas posiciones
127 126 shift = int(self.dataOut.nFFTPoints/2)
128 127 self.dataOut.data_spc = numpy.roll(self.dataOut.data_spc, shift , axis=1)
129 128
130 129 if self.dataOut.data_cspc is not None:
131 130 #desplaza a la derecha en el eje 2 determinadas posiciones
132 131 self.dataOut.data_cspc = numpy.roll(self.dataOut.data_cspc, shift, axis=1)
132 if pairsList:
133 self.__selectPairs(pairsList)
133 134
134 135 elif self.dataIn.type == "Voltage":
135 136
136 137 self.dataOut.flagNoData = True
137 138
138 139 if nFFTPoints == None:
139 140 raise ValueError("This SpectraProc.run() need nFFTPoints input variable")
140 141
141 142 if nProfiles == None:
142 143 nProfiles = nFFTPoints
143 144
144 145 if ippFactor == None:
145 146 self.dataOut.ippFactor = 1
146 147
147 148 self.dataOut.nFFTPoints = nFFTPoints
148 149
149 150 if self.buffer is None:
150 151 self.buffer = numpy.zeros((self.dataIn.nChannels,
151 152 nProfiles,
152 153 self.dataIn.nHeights),
153 154 dtype='complex')
154 155
155 156 if self.dataIn.flagDataAsBlock:
156 157 nVoltProfiles = self.dataIn.data.shape[1]
157 158
158 159 if nVoltProfiles == nProfiles:
159 160 self.buffer = self.dataIn.data.copy()
160 161 self.profIndex = nVoltProfiles
161 162
162 163 elif nVoltProfiles < nProfiles:
163 164
164 165 if self.profIndex == 0:
165 166 self.id_min = 0
166 167 self.id_max = nVoltProfiles
167 168
168 169 self.buffer[:, self.id_min:self.id_max,
169 170 :] = self.dataIn.data
170 171 self.profIndex += nVoltProfiles
171 172 self.id_min += nVoltProfiles
172 173 self.id_max += nVoltProfiles
173 174 else:
174 175 raise ValueError("The type object %s has %d profiles, it should just has %d profiles" % (
175 176 self.dataIn.type, self.dataIn.data.shape[1], nProfiles))
176 177 self.dataOut.flagNoData = True
177 178 else:
178 179 self.buffer[:, self.profIndex, :] = self.dataIn.data.copy()
179 180 self.profIndex += 1
180 181
181 182 if self.firstdatatime == None:
182 183 self.firstdatatime = self.dataIn.utctime
183 184
184 185 if self.profIndex == nProfiles:
185 186 self.__updateSpecFromVoltage()
186 187 if pairsList == None:
187 188 self.dataOut.pairsList = [pair for pair in itertools.combinations(self.dataOut.channelList, 2)]
189 else:
190 self.dataOut.pairsList = pairsList
188 191 self.__getFft()
189
190 192 self.dataOut.flagNoData = False
191 193 self.firstdatatime = None
192 194 self.profIndex = 0
193 195 else:
194 196 raise ValueError("The type of input object '%s' is not valid".format(
195 197 self.dataIn.type))
196 198
197 199 def __selectPairs(self, pairsList):
198 200
199 201 if not pairsList:
200 202 return
201 203
202 204 pairs = []
203 205 pairsIndex = []
204 206
205 207 for pair in pairsList:
206 208 if pair[0] not in self.dataOut.channelList or pair[1] not in self.dataOut.channelList:
207 209 continue
208 210 pairs.append(pair)
209 211 pairsIndex.append(pairs.index(pair))
210 212
211 213 self.dataOut.data_cspc = self.dataOut.data_cspc[pairsIndex]
212 214 self.dataOut.pairsList = pairs
213 215
214 216 return
215 217
216 218 def selectFFTs(self, minFFT, maxFFT ):
217 219 """
218 220 Selecciona un bloque de datos en base a un grupo de valores de puntos FFTs segun el rango
219 221 minFFT<= FFT <= maxFFT
220 222 """
221 223
222 224 if (minFFT > maxFFT):
223 225 raise ValueError("Error selecting heights: Height range (%d,%d) is not valid" % (minFFT, maxFFT))
224 226
225 227 if (minFFT < self.dataOut.getFreqRange()[0]):
226 228 minFFT = self.dataOut.getFreqRange()[0]
227 229
228 230 if (maxFFT > self.dataOut.getFreqRange()[-1]):
229 231 maxFFT = self.dataOut.getFreqRange()[-1]
230 232
231 233 minIndex = 0
232 234 maxIndex = 0
233 235 FFTs = self.dataOut.getFreqRange()
234 236
235 237 inda = numpy.where(FFTs >= minFFT)
236 238 indb = numpy.where(FFTs <= maxFFT)
237 239
238 240 try:
239 241 minIndex = inda[0][0]
240 242 except:
241 243 minIndex = 0
242 244
243 245 try:
244 246 maxIndex = indb[0][-1]
245 247 except:
246 248 maxIndex = len(FFTs)
247 249
248 250 self.selectFFTsByIndex(minIndex, maxIndex)
249 251
250 252 return 1
251 253
252 254 def getBeaconSignal(self, tauindex=0, channelindex=0, hei_ref=None):
253 255 newheis = numpy.where(
254 256 self.dataOut.heightList > self.dataOut.radarControllerHeaderObj.Taus[tauindex])
255 257
256 258 if hei_ref != None:
257 259 newheis = numpy.where(self.dataOut.heightList > hei_ref)
258 260
259 261 minIndex = min(newheis[0])
260 262 maxIndex = max(newheis[0])
261 263 data_spc = self.dataOut.data_spc[:, :, minIndex:maxIndex + 1]
262 264 heightList = self.dataOut.heightList[minIndex:maxIndex + 1]
263 265
264 266 # determina indices
265 267 nheis = int(self.dataOut.radarControllerHeaderObj.txB /
266 268 (self.dataOut.heightList[1] - self.dataOut.heightList[0]))
267 269 avg_dB = 10 * \
268 270 numpy.log10(numpy.sum(data_spc[channelindex, :, :], axis=0))
269 271 beacon_dB = numpy.sort(avg_dB)[-nheis:]
270 272 beacon_heiIndexList = []
271 273 for val in avg_dB.tolist():
272 274 if val >= beacon_dB[0]:
273 275 beacon_heiIndexList.append(avg_dB.tolist().index(val))
274 276
275 277 #data_spc = data_spc[:,:,beacon_heiIndexList]
276 278 data_cspc = None
277 279 if self.dataOut.data_cspc is not None:
278 280 data_cspc = self.dataOut.data_cspc[:, :, minIndex:maxIndex + 1]
279 281 #data_cspc = data_cspc[:,:,beacon_heiIndexList]
280 282
281 283 data_dc = None
282 284 if self.dataOut.data_dc is not None:
283 285 data_dc = self.dataOut.data_dc[:, minIndex:maxIndex + 1]
284 286 #data_dc = data_dc[:,beacon_heiIndexList]
285 287
286 288 self.dataOut.data_spc = data_spc
287 289 self.dataOut.data_cspc = data_cspc
288 290 self.dataOut.data_dc = data_dc
289 291 self.dataOut.heightList = heightList
290 292 self.dataOut.beacon_heiIndexList = beacon_heiIndexList
291 293
292 294 return 1
293 295
294 296 def selectFFTsByIndex(self, minIndex, maxIndex):
295 297 """
296 298
297 299 """
298 300
299 301 if (minIndex < 0) or (minIndex > maxIndex):
300 302 raise ValueError("Error selecting heights: Index range (%d,%d) is not valid" % (minIndex, maxIndex))
301 303
302 304 if (maxIndex >= self.dataOut.nProfiles):
303 305 maxIndex = self.dataOut.nProfiles-1
304 306
305 307 #Spectra
306 308 data_spc = self.dataOut.data_spc[:,minIndex:maxIndex+1,:]
307 309
308 310 data_cspc = None
309 311 if self.dataOut.data_cspc is not None:
310 312 data_cspc = self.dataOut.data_cspc[:,minIndex:maxIndex+1,:]
311 313
312 314 data_dc = None
313 315 if self.dataOut.data_dc is not None:
314 316 data_dc = self.dataOut.data_dc[minIndex:maxIndex+1,:]
315 317
316 318 self.dataOut.data_spc = data_spc
317 319 self.dataOut.data_cspc = data_cspc
318 320 self.dataOut.data_dc = data_dc
319 321
320 322 self.dataOut.ippSeconds = self.dataOut.ippSeconds*(self.dataOut.nFFTPoints / numpy.shape(data_cspc)[1])
321 323 self.dataOut.nFFTPoints = numpy.shape(data_cspc)[1]
322 324 self.dataOut.profilesPerBlock = numpy.shape(data_cspc)[1]
323 325
324 326 return 1
325 327
326 328 def getNoise(self, minHei=None, maxHei=None, minVel=None, maxVel=None):
327 329 # validacion de rango
328 330 if minHei == None:
329 331 minHei = self.dataOut.heightList[0]
330 332
331 333 if maxHei == None:
332 334 maxHei = self.dataOut.heightList[-1]
333 335
334 336 if (minHei < self.dataOut.heightList[0]) or (minHei > maxHei):
335 337 print('minHei: %.2f is out of the heights range' % (minHei))
336 338 print('minHei is setting to %.2f' % (self.dataOut.heightList[0]))
337 339 minHei = self.dataOut.heightList[0]
338 340
339 341 if (maxHei > self.dataOut.heightList[-1]) or (maxHei < minHei):
340 342 print('maxHei: %.2f is out of the heights range' % (maxHei))
341 343 print('maxHei is setting to %.2f' % (self.dataOut.heightList[-1]))
342 344 maxHei = self.dataOut.heightList[-1]
343 345
344 346 # validacion de velocidades
345 347 velrange = self.dataOut.getVelRange(1)
346 348
347 349 if minVel == None:
348 350 minVel = velrange[0]
349 351
350 352 if maxVel == None:
351 353 maxVel = velrange[-1]
352 354
353 355 if (minVel < velrange[0]) or (minVel > maxVel):
354 356 print('minVel: %.2f is out of the velocity range' % (minVel))
355 357 print('minVel is setting to %.2f' % (velrange[0]))
356 358 minVel = velrange[0]
357 359
358 360 if (maxVel > velrange[-1]) or (maxVel < minVel):
359 361 print('maxVel: %.2f is out of the velocity range' % (maxVel))
360 362 print('maxVel is setting to %.2f' % (velrange[-1]))
361 363 maxVel = velrange[-1]
362 364
363 365 # seleccion de indices para rango
364 366 minIndex = 0
365 367 maxIndex = 0
366 368 heights = self.dataOut.heightList
367 369
368 370 inda = numpy.where(heights >= minHei)
369 371 indb = numpy.where(heights <= maxHei)
370 372
371 373 try:
372 374 minIndex = inda[0][0]
373 375 except:
374 376 minIndex = 0
375 377
376 378 try:
377 379 maxIndex = indb[0][-1]
378 380 except:
379 381 maxIndex = len(heights)
380 382
381 383 if (minIndex < 0) or (minIndex > maxIndex):
382 384 raise ValueError("some value in (%d,%d) is not valid" % (
383 385 minIndex, maxIndex))
384 386
385 387 if (maxIndex >= self.dataOut.nHeights):
386 388 maxIndex = self.dataOut.nHeights - 1
387 389
388 390 # seleccion de indices para velocidades
389 391 indminvel = numpy.where(velrange >= minVel)
390 392 indmaxvel = numpy.where(velrange <= maxVel)
391 393 try:
392 394 minIndexVel = indminvel[0][0]
393 395 except:
394 396 minIndexVel = 0
395 397
396 398 try:
397 399 maxIndexVel = indmaxvel[0][-1]
398 400 except:
399 401 maxIndexVel = len(velrange)
400 402
401 403 # seleccion del espectro
402 404 data_spc = self.dataOut.data_spc[:,
403 405 minIndexVel:maxIndexVel + 1, minIndex:maxIndex + 1]
404 406 # estimacion de ruido
405 407 noise = numpy.zeros(self.dataOut.nChannels)
406 408
407 409 for channel in range(self.dataOut.nChannels):
408 410 daux = data_spc[channel, :, :]
409 411 sortdata = numpy.sort(daux, axis=None)
410 412 noise[channel] = hildebrand_sekhon(sortdata, self.dataOut.nIncohInt)
411 413
412 414 self.dataOut.noise_estimation = noise.copy()
413 415
414 416 return 1
415 417
416 418 class removeDC(Operation):
417 419
418 420 def run(self, dataOut, mode=2):
419 421 self.dataOut = dataOut
420 422 jspectra = self.dataOut.data_spc
421 423 jcspectra = self.dataOut.data_cspc
422 424
423 425 num_chan = jspectra.shape[0]
424 426 num_hei = jspectra.shape[2]
425 427
426 428 if jcspectra is not None:
427 429 jcspectraExist = True
428 430 num_pairs = jcspectra.shape[0]
429 431 else:
430 432 jcspectraExist = False
431 433
432 434 freq_dc = int(jspectra.shape[1] / 2)
433 435 ind_vel = numpy.array([-2, -1, 1, 2]) + freq_dc
434 436 ind_vel = ind_vel.astype(int)
435 437
436 438 if ind_vel[0] < 0:
437 439 ind_vel[list(range(0, 1))] = ind_vel[list(range(0, 1))] + self.num_prof
438 440
439 441 if mode == 1:
440 442 jspectra[:, freq_dc, :] = (
441 443 jspectra[:, ind_vel[1], :] + jspectra[:, ind_vel[2], :]) / 2 # CORRECCION
442 444
443 445 if jcspectraExist:
444 446 jcspectra[:, freq_dc, :] = (
445 447 jcspectra[:, ind_vel[1], :] + jcspectra[:, ind_vel[2], :]) / 2
446 448
447 449 if mode == 2:
448 450
449 451 vel = numpy.array([-2, -1, 1, 2])
450 452 xx = numpy.zeros([4, 4])
451 453
452 454 for fil in range(4):
453 455 xx[fil, :] = vel[fil]**numpy.asarray(list(range(4)))
454 456
455 457 xx_inv = numpy.linalg.inv(xx)
456 458 xx_aux = xx_inv[0, :]
457 459
458 460 for ich in range(num_chan):
459 461 yy = jspectra[ich, ind_vel, :]
460 462 jspectra[ich, freq_dc, :] = numpy.dot(xx_aux, yy)
461 463
462 464 junkid = jspectra[ich, freq_dc, :] <= 0
463 465 cjunkid = sum(junkid)
464 466
465 467 if cjunkid.any():
466 468 jspectra[ich, freq_dc, junkid.nonzero()] = (
467 469 jspectra[ich, ind_vel[1], junkid] + jspectra[ich, ind_vel[2], junkid]) / 2
468 470
469 471 if jcspectraExist:
470 472 for ip in range(num_pairs):
471 473 yy = jcspectra[ip, ind_vel, :]
472 474 jcspectra[ip, freq_dc, :] = numpy.dot(xx_aux, yy)
473 475
474 476 self.dataOut.data_spc = jspectra
475 477 self.dataOut.data_cspc = jcspectra
476 478
477 479 return self.dataOut
478 480
479 481 class removeInterference(Operation):
480 482
481 483 def removeInterference2(self):
482 484
483 485 cspc = self.dataOut.data_cspc
484 486 spc = self.dataOut.data_spc
485 487 Heights = numpy.arange(cspc.shape[2])
486 488 realCspc = numpy.abs(cspc)
487 489
488 490 for i in range(cspc.shape[0]):
489 491 LinePower= numpy.sum(realCspc[i], axis=0)
490 492 Threshold = numpy.amax(LinePower)-numpy.sort(LinePower)[len(Heights)-int(len(Heights)*0.1)]
491 493 SelectedHeights = Heights[ numpy.where( LinePower < Threshold ) ]
492 494 InterferenceSum = numpy.sum( realCspc[i,:,SelectedHeights], axis=0 )
493 495 InterferenceThresholdMin = numpy.sort(InterferenceSum)[int(len(InterferenceSum)*0.98)]
494 496 InterferenceThresholdMax = numpy.sort(InterferenceSum)[int(len(InterferenceSum)*0.99)]
495 497
496 498
497 499 InterferenceRange = numpy.where( ([InterferenceSum > InterferenceThresholdMin]))# , InterferenceSum < InterferenceThresholdMax]) )
498 500 #InterferenceRange = numpy.where( ([InterferenceRange < InterferenceThresholdMax]))
499 501 if len(InterferenceRange)<int(cspc.shape[1]*0.3):
500 502 cspc[i,InterferenceRange,:] = numpy.NaN
501 503
502 504 self.dataOut.data_cspc = cspc
503 505
504 506 def removeInterference(self, interf = 2, hei_interf = None, nhei_interf = None, offhei_interf = None):
505 507
506 508 jspectra = self.dataOut.data_spc
507 509 jcspectra = self.dataOut.data_cspc
508 510 jnoise = self.dataOut.getNoise()
509 511 num_incoh = self.dataOut.nIncohInt
510 512
511 513 num_channel = jspectra.shape[0]
512 514 num_prof = jspectra.shape[1]
513 515 num_hei = jspectra.shape[2]
514 516
515 517 # hei_interf
516 518 if hei_interf is None:
517 519 count_hei = int(num_hei / 2)
518 520 hei_interf = numpy.asmatrix(list(range(count_hei))) + num_hei - count_hei
519 521 hei_interf = numpy.asarray(hei_interf)[0]
520 522 # nhei_interf
521 523 if (nhei_interf == None):
522 524 nhei_interf = 5
523 525 if (nhei_interf < 1):
524 526 nhei_interf = 1
525 527 if (nhei_interf > count_hei):
526 528 nhei_interf = count_hei
527 529 if (offhei_interf == None):
528 530 offhei_interf = 0
529 531
530 532 ind_hei = list(range(num_hei))
531 533 # mask_prof = numpy.asarray(range(num_prof - 2)) + 1
532 534 # mask_prof[range(num_prof/2 - 1,len(mask_prof))] += 1
533 535 mask_prof = numpy.asarray(list(range(num_prof)))
534 536 num_mask_prof = mask_prof.size
535 537 comp_mask_prof = [0, num_prof / 2]
536 538
537 539 # noise_exist: Determina si la variable jnoise ha sido definida y contiene la informacion del ruido de cada canal
538 540 if (jnoise.size < num_channel or numpy.isnan(jnoise).any()):
539 541 jnoise = numpy.nan
540 542 noise_exist = jnoise[0] < numpy.Inf
541 543
542 544 # Subrutina de Remocion de la Interferencia
543 545 for ich in range(num_channel):
544 546 # Se ordena los espectros segun su potencia (menor a mayor)
545 547 power = jspectra[ich, mask_prof, :]
546 548 power = power[:, hei_interf]
547 549 power = power.sum(axis=0)
548 550 psort = power.ravel().argsort()
549 551
550 552 # Se estima la interferencia promedio en los Espectros de Potencia empleando
551 553 junkspc_interf = jspectra[ich, :, hei_interf[psort[list(range(
552 554 offhei_interf, nhei_interf + offhei_interf))]]]
553 555
554 556 if noise_exist:
555 557 # tmp_noise = jnoise[ich] / num_prof
556 558 tmp_noise = jnoise[ich]
557 559 junkspc_interf = junkspc_interf - tmp_noise
558 560 #junkspc_interf[:,comp_mask_prof] = 0
559 561
560 562 jspc_interf = junkspc_interf.sum(axis=0) / nhei_interf
561 563 jspc_interf = jspc_interf.transpose()
562 564 # Calculando el espectro de interferencia promedio
563 565 noiseid = numpy.where(
564 566 jspc_interf <= tmp_noise / numpy.sqrt(num_incoh))
565 567 noiseid = noiseid[0]
566 568 cnoiseid = noiseid.size
567 569 interfid = numpy.where(
568 570 jspc_interf > tmp_noise / numpy.sqrt(num_incoh))
569 571 interfid = interfid[0]
570 572 cinterfid = interfid.size
571 573
572 574 if (cnoiseid > 0):
573 575 jspc_interf[noiseid] = 0
574 576
575 577 # Expandiendo los perfiles a limpiar
576 578 if (cinterfid > 0):
577 579 new_interfid = (
578 580 numpy.r_[interfid - 1, interfid, interfid + 1] + num_prof) % num_prof
579 581 new_interfid = numpy.asarray(new_interfid)
580 582 new_interfid = {x for x in new_interfid}
581 583 new_interfid = numpy.array(list(new_interfid))
582 584 new_cinterfid = new_interfid.size
583 585 else:
584 586 new_cinterfid = 0
585 587
586 588 for ip in range(new_cinterfid):
587 589 ind = junkspc_interf[:, new_interfid[ip]].ravel().argsort()
588 590 jspc_interf[new_interfid[ip]
589 591 ] = junkspc_interf[ind[nhei_interf // 2], new_interfid[ip]]
590 592
591 593 jspectra[ich, :, ind_hei] = jspectra[ich, :,
592 594 ind_hei] - jspc_interf # Corregir indices
593 595
594 596 # Removiendo la interferencia del punto de mayor interferencia
595 597 ListAux = jspc_interf[mask_prof].tolist()
596 598 maxid = ListAux.index(max(ListAux))
597 599
598 600 if cinterfid > 0:
599 601 for ip in range(cinterfid * (interf == 2) - 1):
600 602 ind = (jspectra[ich, interfid[ip], :] < tmp_noise *
601 603 (1 + 1 / numpy.sqrt(num_incoh))).nonzero()
602 604 cind = len(ind)
603 605
604 606 if (cind > 0):
605 607 jspectra[ich, interfid[ip], ind] = tmp_noise * \
606 608 (1 + (numpy.random.uniform(cind) - 0.5) /
607 609 numpy.sqrt(num_incoh))
608 610
609 611 ind = numpy.array([-2, -1, 1, 2])
610 612 xx = numpy.zeros([4, 4])
611 613
612 614 for id1 in range(4):
613 615 xx[:, id1] = ind[id1]**numpy.asarray(list(range(4)))
614 616
615 617 xx_inv = numpy.linalg.inv(xx)
616 618 xx = xx_inv[:, 0]
617 619 ind = (ind + maxid + num_mask_prof) % num_mask_prof
618 620 yy = jspectra[ich, mask_prof[ind], :]
619 621 jspectra[ich, mask_prof[maxid], :] = numpy.dot(
620 622 yy.transpose(), xx)
621 623
622 624 indAux = (jspectra[ich, :, :] < tmp_noise *
623 625 (1 - 1 / numpy.sqrt(num_incoh))).nonzero()
624 626 jspectra[ich, indAux[0], indAux[1]] = tmp_noise * \
625 627 (1 - 1 / numpy.sqrt(num_incoh))
626 628
627 629 # Remocion de Interferencia en el Cross Spectra
628 630 if jcspectra is None:
629 631 return jspectra, jcspectra
630 632 num_pairs = int(jcspectra.size / (num_prof * num_hei))
631 633 jcspectra = jcspectra.reshape(num_pairs, num_prof, num_hei)
632 634
633 635 for ip in range(num_pairs):
634 636
635 637 #-------------------------------------------
636 638
637 639 cspower = numpy.abs(jcspectra[ip, mask_prof, :])
638 640 cspower = cspower[:, hei_interf]
639 641 cspower = cspower.sum(axis=0)
640 642
641 643 cspsort = cspower.ravel().argsort()
642 644 junkcspc_interf = jcspectra[ip, :, hei_interf[cspsort[list(range(
643 645 offhei_interf, nhei_interf + offhei_interf))]]]
644 646 junkcspc_interf = junkcspc_interf.transpose()
645 647 jcspc_interf = junkcspc_interf.sum(axis=1) / nhei_interf
646 648
647 649 ind = numpy.abs(jcspc_interf[mask_prof]).ravel().argsort()
648 650
649 651 median_real = int(numpy.median(numpy.real(
650 652 junkcspc_interf[mask_prof[ind[list(range(3 * num_prof // 4))]], :])))
651 653 median_imag = int(numpy.median(numpy.imag(
652 654 junkcspc_interf[mask_prof[ind[list(range(3 * num_prof // 4))]], :])))
653 655 comp_mask_prof = [int(e) for e in comp_mask_prof]
654 656 junkcspc_interf[comp_mask_prof, :] = numpy.complex(
655 657 median_real, median_imag)
656 658
657 659 for iprof in range(num_prof):
658 660 ind = numpy.abs(junkcspc_interf[iprof, :]).ravel().argsort()
659 661 jcspc_interf[iprof] = junkcspc_interf[iprof, ind[nhei_interf // 2]]
660 662
661 663 # Removiendo la Interferencia
662 664 jcspectra[ip, :, ind_hei] = jcspectra[ip,
663 665 :, ind_hei] - jcspc_interf
664 666
665 667 ListAux = numpy.abs(jcspc_interf[mask_prof]).tolist()
666 668 maxid = ListAux.index(max(ListAux))
667 669
668 670 ind = numpy.array([-2, -1, 1, 2])
669 671 xx = numpy.zeros([4, 4])
670 672
671 673 for id1 in range(4):
672 674 xx[:, id1] = ind[id1]**numpy.asarray(list(range(4)))
673 675
674 676 xx_inv = numpy.linalg.inv(xx)
675 677 xx = xx_inv[:, 0]
676 678
677 679 ind = (ind + maxid + num_mask_prof) % num_mask_prof
678 680 yy = jcspectra[ip, mask_prof[ind], :]
679 681 jcspectra[ip, mask_prof[maxid], :] = numpy.dot(yy.transpose(), xx)
680 682
681 683 # Guardar Resultados
682 684 self.dataOut.data_spc = jspectra
683 685 self.dataOut.data_cspc = jcspectra
684 686
685 687 return 1
686 688
687 689 def run(self, dataOut, interf = 2,hei_interf = None, nhei_interf = None, offhei_interf = None, mode=1):
688 690
689 691 self.dataOut = dataOut
690 692
691 693 if mode == 1:
692 694 self.removeInterference(interf = 2,hei_interf = None, nhei_interf = None, offhei_interf = None)
693 695 elif mode == 2:
694 696 self.removeInterference2()
695 697
696 698 return self.dataOut
697 699
698 700
699 701 class IncohInt(Operation):
700 702
701 703 __profIndex = 0
702 704 __withOverapping = False
703 705
704 706 __byTime = False
705 707 __initime = None
706 708 __lastdatatime = None
707 709 __integrationtime = None
708 710
709 711 __buffer_spc = None
710 712 __buffer_cspc = None
711 713 __buffer_dc = None
712 714
713 715 __dataReady = False
714 716
715 717 __timeInterval = None
716 718
717 719 n = None
718 720
719 721 def __init__(self):
720 722
721 723 Operation.__init__(self)
722 724
723 725 def setup(self, n=None, timeInterval=None, overlapping=False):
724 726 """
725 727 Set the parameters of the integration class.
726 728
727 729 Inputs:
728 730
729 731 n : Number of coherent integrations
730 732 timeInterval : Time of integration. If the parameter "n" is selected this one does not work
731 733 overlapping :
732 734
733 735 """
734 736
735 737 self.__initime = None
736 738 self.__lastdatatime = 0
737 739
738 740 self.__buffer_spc = 0
739 741 self.__buffer_cspc = 0
740 742 self.__buffer_dc = 0
741 743
742 744 self.__profIndex = 0
743 745 self.__dataReady = False
744 746 self.__byTime = False
745 747
746 748 if n is None and timeInterval is None:
747 749 raise ValueError("n or timeInterval should be specified ...")
748 750
749 751 if n is not None:
750 752 self.n = int(n)
751 753 else:
752 754
753 755 self.__integrationtime = int(timeInterval)
754 756 self.n = None
755 757 self.__byTime = True
756 758
757 759 def putData(self, data_spc, data_cspc, data_dc):
758 760 """
759 761 Add a profile to the __buffer_spc and increase in one the __profileIndex
760 762
761 763 """
762 764
763 765 self.__buffer_spc += data_spc
764 766
765 767 if data_cspc is None:
766 768 self.__buffer_cspc = None
767 769 else:
768 770 self.__buffer_cspc += data_cspc
769 771
770 772 if data_dc is None:
771 773 self.__buffer_dc = None
772 774 else:
773 775 self.__buffer_dc += data_dc
774 776
775 777 self.__profIndex += 1
776 778
777 779 return
778 780
779 781 def pushData(self):
780 782 """
781 783 Return the sum of the last profiles and the profiles used in the sum.
782 784
783 785 Affected:
784 786
785 787 self.__profileIndex
786 788
787 789 """
788 790
789 791 data_spc = self.__buffer_spc
790 792 data_cspc = self.__buffer_cspc
791 793 data_dc = self.__buffer_dc
792 794 n = self.__profIndex
793 795
794 796 self.__buffer_spc = 0
795 797 self.__buffer_cspc = 0
796 798 self.__buffer_dc = 0
797 799 self.__profIndex = 0
798 800
799 801 return data_spc, data_cspc, data_dc, n
800 802
801 803 def byProfiles(self, *args):
802 804
803 805 self.__dataReady = False
804 806 avgdata_spc = None
805 807 avgdata_cspc = None
806 808 avgdata_dc = None
807 809
808 810 self.putData(*args)
809 811
810 812 if self.__profIndex == self.n:
811 813
812 814 avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData()
813 815 self.n = n
814 816 self.__dataReady = True
815 817
816 818 return avgdata_spc, avgdata_cspc, avgdata_dc
817 819
818 820 def byTime(self, datatime, *args):
819 821
820 822 self.__dataReady = False
821 823 avgdata_spc = None
822 824 avgdata_cspc = None
823 825 avgdata_dc = None
824 826
825 827 self.putData(*args)
826 828
827 829 if (datatime - self.__initime) >= self.__integrationtime:
828 830 avgdata_spc, avgdata_cspc, avgdata_dc, n = self.pushData()
829 831 self.n = n
830 832 self.__dataReady = True
831 833
832 834 return avgdata_spc, avgdata_cspc, avgdata_dc
833 835
834 836 def integrate(self, datatime, *args):
835 837
836 838 if self.__profIndex == 0:
837 839 self.__initime = datatime
838 840
839 841 if self.__byTime:
840 842 avgdata_spc, avgdata_cspc, avgdata_dc = self.byTime(
841 843 datatime, *args)
842 844 else:
843 845 avgdata_spc, avgdata_cspc, avgdata_dc = self.byProfiles(*args)
844 846
845 847 if not self.__dataReady:
846 848 return None, None, None, None
847 849
848 850 return self.__initime, avgdata_spc, avgdata_cspc, avgdata_dc
849 851
850 852 def run(self, dataOut, n=None, timeInterval=None, overlapping=False):
851 853 if n == 1:
852 854 return dataOut
853 855
854 856 dataOut.flagNoData = True
855 857
856 858 if not self.isConfig:
857 859 self.setup(n, timeInterval, overlapping)
858 860 self.isConfig = True
859 861
860 862 avgdatatime, avgdata_spc, avgdata_cspc, avgdata_dc = self.integrate(dataOut.utctime,
861 863 dataOut.data_spc,
862 864 dataOut.data_cspc,
863 865 dataOut.data_dc)
864 866
865 867 if self.__dataReady:
866 868
867 869 dataOut.data_spc = avgdata_spc
868 870 dataOut.data_cspc = avgdata_cspc
869 871 dataOut.data_dc = avgdata_dc
870 872 dataOut.nIncohInt *= self.n
871 873 dataOut.utctime = avgdatatime
872 874 dataOut.flagNoData = False
873 875
874 876 return dataOut No newline at end of file
General Comments 0
You need to be logged in to leave comments. Login now